diff --git a/.gitattributes b/.gitattributes index 914e929dcbe36c9b31249d04998034ca18e96910..3d1145c08d3b29a8e6fb71867c85203c7fae9894 100644 --- a/.gitattributes +++ b/.gitattributes @@ -499,3 +499,8 @@ checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/chec checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-500/tokenizer.json filter=lfs diff=lfs merge=lfs -text checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-200/tokenizer.json filter=lfs diff=lfs merge=lfs -text checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2736/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/model-00001-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9741fee6fadae4b25e2073a819044fe9b6d8da96 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-100/model-00001-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19c0c55e50c348bea1271afcf6d6c1f094b1a012e35aabe4faa889fa631b88d6 +size 4965798912 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/rng_state_0.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..67aa9b2fd4b943dfc7e6e73e6531cd5db98b9183 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1bf4b9590839a4e72249adde49bcbc20d8c1ca92be37d8477499c964fe0d9b89 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/rng_state_1.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..cbd8e850b95f5894ae4f678168478faa8a1e99d3 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c7cf60e175c5b99b395877d16092be0125dd5f69461b994b5cfe682f70ba2770 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/rng_state_2.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..1761012d042e5002043e1528b26e0400c5466f6d --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:acc33d63883e9d3a5ab5ae28d5908acb0928d0ecd1bc49240c4da7f8b6098927 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/scheduler.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..4354a5168876a519c315ffd886af65b9c68360b7 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d445261cc43c842ba2e5eb5e36f9e9b74b3c267d1b67e5a1f4bdb87bf8651668 +size 563 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/tokenizer.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb +size 17210374 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/training_args.bin b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..c2a348e56d106930f8adddd749139fb888853251 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1000/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b62eb270908475de51cd2bd0793440ff3c7234746943dad6610947db320ee11a +size 6011 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/model-00001-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..439b8188b3b27f7f57927bd14369ce515aa82b3a --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1100/model-00001-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c83e66ea788d530317daf72e55325a8aa8bc466392c79735a27829b9d6e5976 +size 4965798912 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/model-00001-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..561dfbcf4c81c4dada08534587f42968de3ae7b9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1200/model-00001-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:395a55f3ca94e2677f29022f31d0def79fc4ef262ff02ed987c6d60be0bb56bb +size 4965798912 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/model-00001-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..46f3293950a850748384f17c93dfb9ef06f3739d --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1600/model-00001-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ea9ea151354f7ef8dcabb0b3a0946e76348ab9cc0b3617aa6b3748cadfcd3c79 +size 4965798912 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/model-00001-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..79785a8f7e9ec88ca000cb40087327f257f43aac --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1700/model-00001-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:45689982aa3ce31ed851fc8be2a3bde8b151264ccedee9fd8b06613eeb567d9c +size 4965798912 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1800/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1800/model-00002-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8cc66edf8121484d876565f2228480dc96a837c8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1800/model-00002-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3a3e3222ea7e0646125afa444449af955d9cbf7c4ad0867353a62dc3649a87fb +size 2247734920 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1900/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1900/model-00002-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..87fcda4d549c5a40c29a623c74f54951854b4d01 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-1900/model-00002-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b10ff570b7341cff87827d55405b89413b1df24710e659f4b95a1d5d4048ef18 +size 2247734920 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/rng_state_0.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..9829253a45a30486360d23ef44a31e49a0fbb37c --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e1200c6256bb3c5211e1fa1cfcea6721e1b8871fd3c8953a59bf49a59f67404c +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/rng_state_1.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..ed05b8eb12ac744b29eb407adca7ab9af789a745 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4b59dfa224b09069bb1265a6a5cb86c22c34f53669b17ebd9335621f959bf7a4 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/rng_state_2.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..a26252516294710c0550a7751d8c4bc072b73762 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:316a81119e1ab6e1b3e5b9d890ce1d0ca9ea9223094844fdcf8b7d9f0f4bb6c2 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/scheduler.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..2df566bc576474c2fececbde3b12e8d11009a9de --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:63abdfcc6823ba777b6b899591d13beec16a9595774c1ecb042d2b482ae9f624 +size 563 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/tokenizer.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb +size 17210374 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/training_args.bin b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..c2a348e56d106930f8adddd749139fb888853251 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2100/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b62eb270908475de51cd2bd0793440ff3c7234746943dad6610947db320ee11a +size 6011 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/rng_state_0.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..cf9f7cc0f22e38cdfea7c36a09f25ba146158728 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b93ade60213134037a6e2a03060d84a67c5041245aff2ad374a5c97f337ee4c2 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/rng_state_1.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..22044561af392bd37b8a514ffc971741f1a025c2 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e17eca23ec632d21ab0115d8db1a9360104fa7b08d928954eee2eed9adf46b8a +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/rng_state_2.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..24228040a3cbfaaf0d0e62b9e12dbef736514695 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6614221a631cfb519fe3c961fa724d0b1c62628b10e8afa55fbb30d5380c88e3 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/scheduler.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..34d75e5931afa3a9d55ab6f7db32572cbe390352 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7544037d90feb060d4553c8b3a7779154e12763c2ee280b6042f565812048250 +size 563 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/tokenizer.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2300/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb +size 17210374 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/rng_state_0.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..7cebe108f7c295f538afca780f8bff98aa6232f2 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c7105b68cc7706a616e5184c049404c7551d0ad422b5b1a292aa19facaa3d0f +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/rng_state_1.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..618e221236c3debbaad7ada8aebf8a3335d91f6e --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9a994ddb08d50eccc51680080b7c8b587407a3900fb66b09b933d12402a0da4b +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/rng_state_2.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..f378850538621142c5cd15d9d6f50bc9a0f23b02 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0fd0c5d9fbbbbb845dde3ee7d26db4ac79f6c6261f597d751fb6b57476387a4b +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/scheduler.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..17204c467abd18e6895fb76e02140caca51478f3 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:22a539e3f96f62dcbf670ac680470aa77ce125c47f5cccb212c99d53ae20651d +size 563 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/tokenizer.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb +size 17210374 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/training_args.bin b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..c2a348e56d106930f8adddd749139fb888853251 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2400/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b62eb270908475de51cd2bd0793440ff3c7234746943dad6610947db320ee11a +size 6011 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2736/training_args.bin b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2736/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..c2a348e56d106930f8adddd749139fb888853251 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-2736/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b62eb270908475de51cd2bd0793440ff3c7234746943dad6610947db320ee11a +size 6011 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/rng_state_0.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..9015a2b291768ac1cbe0fb417354be7e82849ff8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c47a2aec04adfb2b774bed1d02a5453693feb66e4636bdf7cd908de748732842 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/rng_state_1.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..a17e6802aa777eb35577f53d24a38a0381fecbb7 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e375d2646839f0fc8836323cd9f6c6fec1f1316a3f21b7366755c68ab3c469c +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/rng_state_2.pth b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..8aab09805997622aec2edae7b565ed82010cfd53 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6802d0e782c931221303c74054cdedbab70d850d6310dc553b393a4f3f848807 +size 16567 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/scheduler.pt b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..7e7e519916848b630beefaeefaf5fba757d0c28e --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5be0ff57f0c443e5a68c38a2288c9b31e45320676ebfa6b2b08fe9b73072b75b +size 563 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/tokenizer.json b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb +size 17210374 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/training_args.bin b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..c2a348e56d106930f8adddd749139fb888853251 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-300/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b62eb270908475de51cd2bd0793440ff3c7234746943dad6610947db320ee11a +size 6011 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/model-00001-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..30230be34496e3b3f57e0ca82e00366cd5f84368 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-600/model-00001-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:70ac00818057f336a1c967632ccf0d9d255e2a2235364970b730d28a842bc45b +size 4965798912 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-700/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-700/model-00002-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d56e8c1a2131906a2a85fd3557f50a15efbc2b95 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-700/model-00002-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0c280e40d17570f0e7f7b4dc79750c00d416ff81061d2ab4121b570a0fb3948c +size 2247734920 diff --git a/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-800/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-800/model-00002-of-00002.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..968d619ee77ab93fcfd20060eac1c7cafdc9df03 --- /dev/null +++ b/checkpoints/Llama-3.2-3B-FTP/babylm_shuffle_nondeterministic_10M_seed0/runs/checkpoint-800/model-00002-of-00002.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:35ebb22f2d000cf3ceac4a18166e22486e52eb42bcf1cac29150d3002ecc88ad +size 2247734920 diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/latest b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/latest new file mode 100644 index 0000000000000000000000000000000000000000..744ae7dbad571b6f37ec6c7066549494261bb59e --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/latest @@ -0,0 +1 @@ +global_step100 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..8991b49e9c2a43fc527dab9e09ad8171f0cc5943 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/tokenizer_config.json @@ -0,0 +1,2086 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128257": { + "content": "🅂", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + }, + "128258": { + "content": "🄿", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3719b1967eb071b3f06936f1ef05e6206a126337 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/trainer_state.json @@ -0,0 +1,813 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.12626262626262627, + "eval_steps": 10, + "global_step": 100, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0012626262626262627, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8571, + "step": 1 + }, + { + "epoch": 0.0025252525252525255, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7758, + "step": 2 + }, + { + "epoch": 0.003787878787878788, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8383, + "step": 3 + }, + { + "epoch": 0.005050505050505051, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.9088, + "step": 4 + }, + { + "epoch": 0.006313131313131313, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8748, + "step": 5 + }, + { + "epoch": 0.007575757575757576, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8575, + "step": 6 + }, + { + "epoch": 0.008838383838383838, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8271, + "step": 7 + }, + { + "epoch": 0.010101010101010102, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8336, + "step": 8 + }, + { + "epoch": 0.011363636363636364, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8037, + "step": 9 + }, + { + "epoch": 0.012626262626262626, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8323, + "step": 10 + }, + { + "epoch": 0.012626262626262626, + "eval_loss": 2.8736937046051025, + "eval_runtime": 37.7802, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 10 + }, + { + "epoch": 0.013888888888888888, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8653, + "step": 11 + }, + { + "epoch": 0.015151515151515152, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8104, + "step": 12 + }, + { + "epoch": 0.016414141414141416, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8489, + "step": 13 + }, + { + "epoch": 0.017676767676767676, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8442, + "step": 14 + }, + { + "epoch": 0.01893939393939394, + "grad_norm": 5.356770992279053, + "learning_rate": 2.1008403361344538e-08, + "loss": 2.8207, + "step": 15 + }, + { + "epoch": 0.020202020202020204, + "grad_norm": 4.954462051391602, + "learning_rate": 4.2016806722689076e-08, + "loss": 2.8104, + "step": 16 + }, + { + "epoch": 0.021464646464646464, + "grad_norm": 5.248059272766113, + "learning_rate": 6.302521008403361e-08, + "loss": 2.8178, + "step": 17 + }, + { + "epoch": 0.022727272727272728, + "grad_norm": 5.277204513549805, + "learning_rate": 8.403361344537815e-08, + "loss": 2.8255, + "step": 18 + }, + { + "epoch": 0.023989898989898988, + "grad_norm": 5.344228267669678, + "learning_rate": 1.050420168067227e-07, + "loss": 2.8018, + "step": 19 + }, + { + "epoch": 0.025252525252525252, + "grad_norm": 5.004310131072998, + "learning_rate": 1.2605042016806723e-07, + "loss": 2.8337, + "step": 20 + }, + { + "epoch": 0.025252525252525252, + "eval_loss": 2.872413158416748, + "eval_runtime": 37.7426, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 20 + }, + { + "epoch": 0.026515151515151516, + "grad_norm": 5.209287166595459, + "learning_rate": 1.4705882352941178e-07, + "loss": 2.8246, + "step": 21 + }, + { + "epoch": 0.027777777777777776, + "grad_norm": 5.369327545166016, + "learning_rate": 1.680672268907563e-07, + "loss": 2.8791, + "step": 22 + }, + { + "epoch": 0.02904040404040404, + "grad_norm": 4.646446228027344, + "learning_rate": 1.8907563025210085e-07, + "loss": 2.8425, + "step": 23 + }, + { + "epoch": 0.030303030303030304, + "grad_norm": 4.949256420135498, + "learning_rate": 2.100840336134454e-07, + "loss": 2.8629, + "step": 24 + }, + { + "epoch": 0.03156565656565657, + "grad_norm": 4.709571838378906, + "learning_rate": 2.3109243697478996e-07, + "loss": 2.8264, + "step": 25 + }, + { + "epoch": 0.03282828282828283, + "grad_norm": 4.519557952880859, + "learning_rate": 2.5210084033613445e-07, + "loss": 2.8143, + "step": 26 + }, + { + "epoch": 0.03409090909090909, + "grad_norm": 4.445594310760498, + "learning_rate": 2.73109243697479e-07, + "loss": 2.8031, + "step": 27 + }, + { + "epoch": 0.03535353535353535, + "grad_norm": 4.26489782333374, + "learning_rate": 2.9411764705882356e-07, + "loss": 2.7719, + "step": 28 + }, + { + "epoch": 0.036616161616161616, + "grad_norm": 4.153639793395996, + "learning_rate": 3.151260504201681e-07, + "loss": 2.8251, + "step": 29 + }, + { + "epoch": 0.03787878787878788, + "grad_norm": 3.40531325340271, + "learning_rate": 3.361344537815126e-07, + "loss": 2.849, + "step": 30 + }, + { + "epoch": 0.03787878787878788, + "eval_loss": 2.825382709503174, + "eval_runtime": 37.0715, + "eval_samples_per_second": 26.975, + "eval_steps_per_second": 1.133, + "step": 30 + }, + { + "epoch": 0.039141414141414144, + "grad_norm": 3.4759507179260254, + "learning_rate": 3.5714285714285716e-07, + "loss": 2.7653, + "step": 31 + }, + { + "epoch": 0.04040404040404041, + "grad_norm": 3.4147374629974365, + "learning_rate": 3.781512605042017e-07, + "loss": 2.7916, + "step": 32 + }, + { + "epoch": 0.041666666666666664, + "grad_norm": 3.2775769233703613, + "learning_rate": 3.991596638655462e-07, + "loss": 2.8072, + "step": 33 + }, + { + "epoch": 0.04292929292929293, + "grad_norm": 3.3541715145111084, + "learning_rate": 4.201680672268908e-07, + "loss": 2.7944, + "step": 34 + }, + { + "epoch": 0.04419191919191919, + "grad_norm": 2.991971969604492, + "learning_rate": 4.4117647058823536e-07, + "loss": 2.7827, + "step": 35 + }, + { + "epoch": 0.045454545454545456, + "grad_norm": 2.9127862453460693, + "learning_rate": 4.621848739495799e-07, + "loss": 2.7332, + "step": 36 + }, + { + "epoch": 0.04671717171717172, + "grad_norm": 2.839707136154175, + "learning_rate": 4.831932773109245e-07, + "loss": 2.7568, + "step": 37 + }, + { + "epoch": 0.047979797979797977, + "grad_norm": 2.9395432472229004, + "learning_rate": 5.042016806722689e-07, + "loss": 2.7445, + "step": 38 + }, + { + "epoch": 0.04924242424242424, + "grad_norm": 2.869896650314331, + "learning_rate": 5.252100840336135e-07, + "loss": 2.7157, + "step": 39 + }, + { + "epoch": 0.050505050505050504, + "grad_norm": 2.9064571857452393, + "learning_rate": 5.46218487394958e-07, + "loss": 2.7049, + "step": 40 + }, + { + "epoch": 0.050505050505050504, + "eval_loss": 2.7477543354034424, + "eval_runtime": 37.0669, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 40 + }, + { + "epoch": 0.05176767676767677, + "grad_norm": 2.6264421939849854, + "learning_rate": 5.672268907563026e-07, + "loss": 2.7651, + "step": 41 + }, + { + "epoch": 0.05303030303030303, + "grad_norm": 2.518627405166626, + "learning_rate": 5.882352941176471e-07, + "loss": 2.7241, + "step": 42 + }, + { + "epoch": 0.054292929292929296, + "grad_norm": 2.521034002304077, + "learning_rate": 6.092436974789917e-07, + "loss": 2.6691, + "step": 43 + }, + { + "epoch": 0.05555555555555555, + "grad_norm": 2.4679362773895264, + "learning_rate": 6.302521008403362e-07, + "loss": 2.743, + "step": 44 + }, + { + "epoch": 0.056818181818181816, + "grad_norm": 2.5817418098449707, + "learning_rate": 6.512605042016808e-07, + "loss": 2.7067, + "step": 45 + }, + { + "epoch": 0.05808080808080808, + "grad_norm": 2.3704075813293457, + "learning_rate": 6.722689075630252e-07, + "loss": 2.6495, + "step": 46 + }, + { + "epoch": 0.059343434343434344, + "grad_norm": 2.2321906089782715, + "learning_rate": 6.932773109243699e-07, + "loss": 2.6606, + "step": 47 + }, + { + "epoch": 0.06060606060606061, + "grad_norm": 1.985407829284668, + "learning_rate": 7.142857142857143e-07, + "loss": 2.662, + "step": 48 + }, + { + "epoch": 0.06186868686868687, + "grad_norm": 1.945688247680664, + "learning_rate": 7.352941176470589e-07, + "loss": 2.6319, + "step": 49 + }, + { + "epoch": 0.06313131313131314, + "grad_norm": 1.9392870664596558, + "learning_rate": 7.563025210084034e-07, + "loss": 2.6094, + "step": 50 + }, + { + "epoch": 0.06313131313131314, + "eval_loss": 2.6554172039031982, + "eval_runtime": 37.0569, + "eval_samples_per_second": 26.986, + "eval_steps_per_second": 1.133, + "step": 50 + }, + { + "epoch": 0.06439393939393939, + "grad_norm": 1.9975577592849731, + "learning_rate": 7.77310924369748e-07, + "loss": 2.6045, + "step": 51 + }, + { + "epoch": 0.06565656565656566, + "grad_norm": 1.9733693599700928, + "learning_rate": 7.983193277310924e-07, + "loss": 2.649, + "step": 52 + }, + { + "epoch": 0.06691919191919192, + "grad_norm": 1.7324917316436768, + "learning_rate": 8.193277310924371e-07, + "loss": 2.5689, + "step": 53 + }, + { + "epoch": 0.06818181818181818, + "grad_norm": 1.7553645372390747, + "learning_rate": 8.403361344537816e-07, + "loss": 2.644, + "step": 54 + }, + { + "epoch": 0.06944444444444445, + "grad_norm": 1.8163801431655884, + "learning_rate": 8.613445378151261e-07, + "loss": 2.4943, + "step": 55 + }, + { + "epoch": 0.0707070707070707, + "grad_norm": 1.7717015743255615, + "learning_rate": 8.823529411764707e-07, + "loss": 2.6093, + "step": 56 + }, + { + "epoch": 0.07196969696969698, + "grad_norm": 1.7013663053512573, + "learning_rate": 9.033613445378152e-07, + "loss": 2.6376, + "step": 57 + }, + { + "epoch": 0.07323232323232323, + "grad_norm": 1.752943992614746, + "learning_rate": 9.243697478991598e-07, + "loss": 2.5172, + "step": 58 + }, + { + "epoch": 0.07449494949494949, + "grad_norm": 1.5362415313720703, + "learning_rate": 9.453781512605043e-07, + "loss": 2.6129, + "step": 59 + }, + { + "epoch": 0.07575757575757576, + "grad_norm": 1.514885425567627, + "learning_rate": 9.66386554621849e-07, + "loss": 2.6064, + "step": 60 + }, + { + "epoch": 0.07575757575757576, + "eval_loss": 2.5934340953826904, + "eval_runtime": 37.0782, + "eval_samples_per_second": 26.97, + "eval_steps_per_second": 1.133, + "step": 60 + }, + { + "epoch": 0.07702020202020202, + "grad_norm": 1.5915579795837402, + "learning_rate": 9.873949579831934e-07, + "loss": 2.5491, + "step": 61 + }, + { + "epoch": 0.07828282828282829, + "grad_norm": 1.601773977279663, + "learning_rate": 1.0084033613445378e-06, + "loss": 2.5685, + "step": 62 + }, + { + "epoch": 0.07954545454545454, + "grad_norm": 1.8327804803848267, + "learning_rate": 1.0294117647058825e-06, + "loss": 2.5763, + "step": 63 + }, + { + "epoch": 0.08080808080808081, + "grad_norm": 1.5276741981506348, + "learning_rate": 1.050420168067227e-06, + "loss": 2.5494, + "step": 64 + }, + { + "epoch": 0.08207070707070707, + "grad_norm": 1.7065422534942627, + "learning_rate": 1.0714285714285714e-06, + "loss": 2.449, + "step": 65 + }, + { + "epoch": 0.08333333333333333, + "grad_norm": 2.0551910400390625, + "learning_rate": 1.092436974789916e-06, + "loss": 2.4371, + "step": 66 + }, + { + "epoch": 0.0845959595959596, + "grad_norm": 2.154855728149414, + "learning_rate": 1.1134453781512607e-06, + "loss": 2.5417, + "step": 67 + }, + { + "epoch": 0.08585858585858586, + "grad_norm": 1.7922390699386597, + "learning_rate": 1.1344537815126051e-06, + "loss": 2.4754, + "step": 68 + }, + { + "epoch": 0.08712121212121213, + "grad_norm": 1.745632290840149, + "learning_rate": 1.1554621848739498e-06, + "loss": 2.4669, + "step": 69 + }, + { + "epoch": 0.08838383838383838, + "grad_norm": 1.8907713890075684, + "learning_rate": 1.1764705882352942e-06, + "loss": 2.4943, + "step": 70 + }, + { + "epoch": 0.08838383838383838, + "eval_loss": 2.518237829208374, + "eval_runtime": 37.074, + "eval_samples_per_second": 26.973, + "eval_steps_per_second": 1.133, + "step": 70 + }, + { + "epoch": 0.08964646464646464, + "grad_norm": 1.8178164958953857, + "learning_rate": 1.1974789915966389e-06, + "loss": 2.4891, + "step": 71 + }, + { + "epoch": 0.09090909090909091, + "grad_norm": 1.970969796180725, + "learning_rate": 1.2184873949579833e-06, + "loss": 2.4797, + "step": 72 + }, + { + "epoch": 0.09217171717171717, + "grad_norm": 2.0211431980133057, + "learning_rate": 1.2394957983193278e-06, + "loss": 2.4716, + "step": 73 + }, + { + "epoch": 0.09343434343434344, + "grad_norm": 2.650787591934204, + "learning_rate": 1.2605042016806724e-06, + "loss": 2.4915, + "step": 74 + }, + { + "epoch": 0.0946969696969697, + "grad_norm": 2.235499143600464, + "learning_rate": 1.2815126050420169e-06, + "loss": 2.3917, + "step": 75 + }, + { + "epoch": 0.09595959595959595, + "grad_norm": 2.245999813079834, + "learning_rate": 1.3025210084033615e-06, + "loss": 2.4942, + "step": 76 + }, + { + "epoch": 0.09722222222222222, + "grad_norm": 2.4414420127868652, + "learning_rate": 1.323529411764706e-06, + "loss": 2.4004, + "step": 77 + }, + { + "epoch": 0.09848484848484848, + "grad_norm": 2.279937982559204, + "learning_rate": 1.3445378151260504e-06, + "loss": 2.4043, + "step": 78 + }, + { + "epoch": 0.09974747474747475, + "grad_norm": 2.3830068111419678, + "learning_rate": 1.3655462184873949e-06, + "loss": 2.4096, + "step": 79 + }, + { + "epoch": 0.10101010101010101, + "grad_norm": 2.0119733810424805, + "learning_rate": 1.3865546218487397e-06, + "loss": 2.4227, + "step": 80 + }, + { + "epoch": 0.10101010101010101, + "eval_loss": 2.4316186904907227, + "eval_runtime": 37.0837, + "eval_samples_per_second": 26.966, + "eval_steps_per_second": 1.133, + "step": 80 + }, + { + "epoch": 0.10227272727272728, + "grad_norm": 2.1273374557495117, + "learning_rate": 1.4075630252100842e-06, + "loss": 2.3541, + "step": 81 + }, + { + "epoch": 0.10353535353535354, + "grad_norm": 2.6852898597717285, + "learning_rate": 1.4285714285714286e-06, + "loss": 2.3321, + "step": 82 + }, + { + "epoch": 0.10479797979797979, + "grad_norm": 2.2622721195220947, + "learning_rate": 1.4495798319327733e-06, + "loss": 2.3166, + "step": 83 + }, + { + "epoch": 0.10606060606060606, + "grad_norm": 2.98258900642395, + "learning_rate": 1.4705882352941177e-06, + "loss": 2.3733, + "step": 84 + }, + { + "epoch": 0.10732323232323232, + "grad_norm": 2.5140459537506104, + "learning_rate": 1.4915966386554624e-06, + "loss": 2.314, + "step": 85 + }, + { + "epoch": 0.10858585858585859, + "grad_norm": 2.637608051300049, + "learning_rate": 1.5126050420168068e-06, + "loss": 2.3479, + "step": 86 + }, + { + "epoch": 0.10984848484848485, + "grad_norm": 2.399508476257324, + "learning_rate": 1.5336134453781515e-06, + "loss": 2.2932, + "step": 87 + }, + { + "epoch": 0.1111111111111111, + "grad_norm": 3.0835793018341064, + "learning_rate": 1.554621848739496e-06, + "loss": 2.2886, + "step": 88 + }, + { + "epoch": 0.11237373737373738, + "grad_norm": 3.288769483566284, + "learning_rate": 1.5756302521008404e-06, + "loss": 2.2914, + "step": 89 + }, + { + "epoch": 0.11363636363636363, + "grad_norm": 2.538361072540283, + "learning_rate": 1.5966386554621848e-06, + "loss": 2.3247, + "step": 90 + }, + { + "epoch": 0.11363636363636363, + "eval_loss": 2.3510072231292725, + "eval_runtime": 37.0847, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 90 + }, + { + "epoch": 0.1148989898989899, + "grad_norm": 2.8016724586486816, + "learning_rate": 1.6176470588235297e-06, + "loss": 2.281, + "step": 91 + }, + { + "epoch": 0.11616161616161616, + "grad_norm": 2.9251859188079834, + "learning_rate": 1.6386554621848741e-06, + "loss": 2.2715, + "step": 92 + }, + { + "epoch": 0.11742424242424243, + "grad_norm": 2.595750570297241, + "learning_rate": 1.6596638655462186e-06, + "loss": 2.2127, + "step": 93 + }, + { + "epoch": 0.11868686868686869, + "grad_norm": 2.6989378929138184, + "learning_rate": 1.6806722689075632e-06, + "loss": 2.2623, + "step": 94 + }, + { + "epoch": 0.11994949494949494, + "grad_norm": 2.554863929748535, + "learning_rate": 1.7016806722689077e-06, + "loss": 2.2649, + "step": 95 + }, + { + "epoch": 0.12121212121212122, + "grad_norm": 2.7013020515441895, + "learning_rate": 1.7226890756302521e-06, + "loss": 2.2428, + "step": 96 + }, + { + "epoch": 0.12247474747474747, + "grad_norm": 2.7415151596069336, + "learning_rate": 1.7436974789915968e-06, + "loss": 2.1882, + "step": 97 + }, + { + "epoch": 0.12373737373737374, + "grad_norm": 2.672806739807129, + "learning_rate": 1.7647058823529414e-06, + "loss": 2.2302, + "step": 98 + }, + { + "epoch": 0.125, + "grad_norm": 2.9508326053619385, + "learning_rate": 1.7857142857142859e-06, + "loss": 2.1894, + "step": 99 + }, + { + "epoch": 0.12626262626262627, + "grad_norm": 2.4668166637420654, + "learning_rate": 1.8067226890756303e-06, + "loss": 2.2043, + "step": 100 + }, + { + "epoch": 0.12626262626262627, + "eval_loss": 2.2788307666778564, + "eval_runtime": 37.1156, + "eval_samples_per_second": 26.943, + "eval_steps_per_second": 1.132, + "step": 100 + } + ], + "logging_steps": 1, + "max_steps": 2376, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 3.11730940936192e+16, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-100/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/latest b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/latest new file mode 100644 index 0000000000000000000000000000000000000000..22cd5c3402316b70299aed2025d7943595f5d495 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/latest @@ -0,0 +1 @@ +global_step1100 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..8991b49e9c2a43fc527dab9e09ad8171f0cc5943 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/tokenizer_config.json @@ -0,0 +1,2086 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128257": { + "content": "🅂", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + }, + "128258": { + "content": "🄿", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3135ffc5e070d534f8035b6aee716ceae15b5c09 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/trainer_state.json @@ -0,0 +1,8613 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.3888888888888888, + "eval_steps": 10, + "global_step": 1100, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0012626262626262627, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8571, + "step": 1 + }, + { + "epoch": 0.0025252525252525255, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7758, + "step": 2 + }, + { + "epoch": 0.003787878787878788, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8383, + "step": 3 + }, + { + "epoch": 0.005050505050505051, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.9088, + "step": 4 + }, + { + "epoch": 0.006313131313131313, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8748, + "step": 5 + }, + { + "epoch": 0.007575757575757576, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8575, + "step": 6 + }, + { + "epoch": 0.008838383838383838, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8271, + "step": 7 + }, + { + "epoch": 0.010101010101010102, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8336, + "step": 8 + }, + { + "epoch": 0.011363636363636364, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8037, + "step": 9 + }, + { + "epoch": 0.012626262626262626, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8323, + "step": 10 + }, + { + "epoch": 0.012626262626262626, + "eval_loss": 2.8736937046051025, + "eval_runtime": 37.7802, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 10 + }, + { + "epoch": 0.013888888888888888, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8653, + "step": 11 + }, + { + "epoch": 0.015151515151515152, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8104, + "step": 12 + }, + { + "epoch": 0.016414141414141416, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8489, + "step": 13 + }, + { + "epoch": 0.017676767676767676, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8442, + "step": 14 + }, + { + "epoch": 0.01893939393939394, + "grad_norm": 5.356770992279053, + "learning_rate": 2.1008403361344538e-08, + "loss": 2.8207, + "step": 15 + }, + { + "epoch": 0.020202020202020204, + "grad_norm": 4.954462051391602, + "learning_rate": 4.2016806722689076e-08, + "loss": 2.8104, + "step": 16 + }, + { + "epoch": 0.021464646464646464, + "grad_norm": 5.248059272766113, + "learning_rate": 6.302521008403361e-08, + "loss": 2.8178, + "step": 17 + }, + { + "epoch": 0.022727272727272728, + "grad_norm": 5.277204513549805, + "learning_rate": 8.403361344537815e-08, + "loss": 2.8255, + "step": 18 + }, + { + "epoch": 0.023989898989898988, + "grad_norm": 5.344228267669678, + "learning_rate": 1.050420168067227e-07, + "loss": 2.8018, + "step": 19 + }, + { + "epoch": 0.025252525252525252, + "grad_norm": 5.004310131072998, + "learning_rate": 1.2605042016806723e-07, + "loss": 2.8337, + "step": 20 + }, + { + "epoch": 0.025252525252525252, + "eval_loss": 2.872413158416748, + "eval_runtime": 37.7426, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 20 + }, + { + "epoch": 0.026515151515151516, + "grad_norm": 5.209287166595459, + "learning_rate": 1.4705882352941178e-07, + "loss": 2.8246, + "step": 21 + }, + { + "epoch": 0.027777777777777776, + "grad_norm": 5.369327545166016, + "learning_rate": 1.680672268907563e-07, + "loss": 2.8791, + "step": 22 + }, + { + "epoch": 0.02904040404040404, + "grad_norm": 4.646446228027344, + "learning_rate": 1.8907563025210085e-07, + "loss": 2.8425, + "step": 23 + }, + { + "epoch": 0.030303030303030304, + "grad_norm": 4.949256420135498, + "learning_rate": 2.100840336134454e-07, + "loss": 2.8629, + "step": 24 + }, + { + "epoch": 0.03156565656565657, + "grad_norm": 4.709571838378906, + "learning_rate": 2.3109243697478996e-07, + "loss": 2.8264, + "step": 25 + }, + { + "epoch": 0.03282828282828283, + "grad_norm": 4.519557952880859, + "learning_rate": 2.5210084033613445e-07, + "loss": 2.8143, + "step": 26 + }, + { + "epoch": 0.03409090909090909, + "grad_norm": 4.445594310760498, + "learning_rate": 2.73109243697479e-07, + "loss": 2.8031, + "step": 27 + }, + { + "epoch": 0.03535353535353535, + "grad_norm": 4.26489782333374, + "learning_rate": 2.9411764705882356e-07, + "loss": 2.7719, + "step": 28 + }, + { + "epoch": 0.036616161616161616, + "grad_norm": 4.153639793395996, + "learning_rate": 3.151260504201681e-07, + "loss": 2.8251, + "step": 29 + }, + { + "epoch": 0.03787878787878788, + "grad_norm": 3.40531325340271, + "learning_rate": 3.361344537815126e-07, + "loss": 2.849, + "step": 30 + }, + { + "epoch": 0.03787878787878788, + "eval_loss": 2.825382709503174, + "eval_runtime": 37.0715, + "eval_samples_per_second": 26.975, + "eval_steps_per_second": 1.133, + "step": 30 + }, + { + "epoch": 0.039141414141414144, + "grad_norm": 3.4759507179260254, + "learning_rate": 3.5714285714285716e-07, + "loss": 2.7653, + "step": 31 + }, + { + "epoch": 0.04040404040404041, + "grad_norm": 3.4147374629974365, + "learning_rate": 3.781512605042017e-07, + "loss": 2.7916, + "step": 32 + }, + { + "epoch": 0.041666666666666664, + "grad_norm": 3.2775769233703613, + "learning_rate": 3.991596638655462e-07, + "loss": 2.8072, + "step": 33 + }, + { + "epoch": 0.04292929292929293, + "grad_norm": 3.3541715145111084, + "learning_rate": 4.201680672268908e-07, + "loss": 2.7944, + "step": 34 + }, + { + "epoch": 0.04419191919191919, + "grad_norm": 2.991971969604492, + "learning_rate": 4.4117647058823536e-07, + "loss": 2.7827, + "step": 35 + }, + { + "epoch": 0.045454545454545456, + "grad_norm": 2.9127862453460693, + "learning_rate": 4.621848739495799e-07, + "loss": 2.7332, + "step": 36 + }, + { + "epoch": 0.04671717171717172, + "grad_norm": 2.839707136154175, + "learning_rate": 4.831932773109245e-07, + "loss": 2.7568, + "step": 37 + }, + { + "epoch": 0.047979797979797977, + "grad_norm": 2.9395432472229004, + "learning_rate": 5.042016806722689e-07, + "loss": 2.7445, + "step": 38 + }, + { + "epoch": 0.04924242424242424, + "grad_norm": 2.869896650314331, + "learning_rate": 5.252100840336135e-07, + "loss": 2.7157, + "step": 39 + }, + { + "epoch": 0.050505050505050504, + "grad_norm": 2.9064571857452393, + "learning_rate": 5.46218487394958e-07, + "loss": 2.7049, + "step": 40 + }, + { + "epoch": 0.050505050505050504, + "eval_loss": 2.7477543354034424, + "eval_runtime": 37.0669, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 40 + }, + { + "epoch": 0.05176767676767677, + "grad_norm": 2.6264421939849854, + "learning_rate": 5.672268907563026e-07, + "loss": 2.7651, + "step": 41 + }, + { + "epoch": 0.05303030303030303, + "grad_norm": 2.518627405166626, + "learning_rate": 5.882352941176471e-07, + "loss": 2.7241, + "step": 42 + }, + { + "epoch": 0.054292929292929296, + "grad_norm": 2.521034002304077, + "learning_rate": 6.092436974789917e-07, + "loss": 2.6691, + "step": 43 + }, + { + "epoch": 0.05555555555555555, + "grad_norm": 2.4679362773895264, + "learning_rate": 6.302521008403362e-07, + "loss": 2.743, + "step": 44 + }, + { + "epoch": 0.056818181818181816, + "grad_norm": 2.5817418098449707, + "learning_rate": 6.512605042016808e-07, + "loss": 2.7067, + "step": 45 + }, + { + "epoch": 0.05808080808080808, + "grad_norm": 2.3704075813293457, + "learning_rate": 6.722689075630252e-07, + "loss": 2.6495, + "step": 46 + }, + { + "epoch": 0.059343434343434344, + "grad_norm": 2.2321906089782715, + "learning_rate": 6.932773109243699e-07, + "loss": 2.6606, + "step": 47 + }, + { + "epoch": 0.06060606060606061, + "grad_norm": 1.985407829284668, + "learning_rate": 7.142857142857143e-07, + "loss": 2.662, + "step": 48 + }, + { + "epoch": 0.06186868686868687, + "grad_norm": 1.945688247680664, + "learning_rate": 7.352941176470589e-07, + "loss": 2.6319, + "step": 49 + }, + { + "epoch": 0.06313131313131314, + "grad_norm": 1.9392870664596558, + "learning_rate": 7.563025210084034e-07, + "loss": 2.6094, + "step": 50 + }, + { + "epoch": 0.06313131313131314, + "eval_loss": 2.6554172039031982, + "eval_runtime": 37.0569, + "eval_samples_per_second": 26.986, + "eval_steps_per_second": 1.133, + "step": 50 + }, + { + "epoch": 0.06439393939393939, + "grad_norm": 1.9975577592849731, + "learning_rate": 7.77310924369748e-07, + "loss": 2.6045, + "step": 51 + }, + { + "epoch": 0.06565656565656566, + "grad_norm": 1.9733693599700928, + "learning_rate": 7.983193277310924e-07, + "loss": 2.649, + "step": 52 + }, + { + "epoch": 0.06691919191919192, + "grad_norm": 1.7324917316436768, + "learning_rate": 8.193277310924371e-07, + "loss": 2.5689, + "step": 53 + }, + { + "epoch": 0.06818181818181818, + "grad_norm": 1.7553645372390747, + "learning_rate": 8.403361344537816e-07, + "loss": 2.644, + "step": 54 + }, + { + "epoch": 0.06944444444444445, + "grad_norm": 1.8163801431655884, + "learning_rate": 8.613445378151261e-07, + "loss": 2.4943, + "step": 55 + }, + { + "epoch": 0.0707070707070707, + "grad_norm": 1.7717015743255615, + "learning_rate": 8.823529411764707e-07, + "loss": 2.6093, + "step": 56 + }, + { + "epoch": 0.07196969696969698, + "grad_norm": 1.7013663053512573, + "learning_rate": 9.033613445378152e-07, + "loss": 2.6376, + "step": 57 + }, + { + "epoch": 0.07323232323232323, + "grad_norm": 1.752943992614746, + "learning_rate": 9.243697478991598e-07, + "loss": 2.5172, + "step": 58 + }, + { + "epoch": 0.07449494949494949, + "grad_norm": 1.5362415313720703, + "learning_rate": 9.453781512605043e-07, + "loss": 2.6129, + "step": 59 + }, + { + "epoch": 0.07575757575757576, + "grad_norm": 1.514885425567627, + "learning_rate": 9.66386554621849e-07, + "loss": 2.6064, + "step": 60 + }, + { + "epoch": 0.07575757575757576, + "eval_loss": 2.5934340953826904, + "eval_runtime": 37.0782, + "eval_samples_per_second": 26.97, + "eval_steps_per_second": 1.133, + "step": 60 + }, + { + "epoch": 0.07702020202020202, + "grad_norm": 1.5915579795837402, + "learning_rate": 9.873949579831934e-07, + "loss": 2.5491, + "step": 61 + }, + { + "epoch": 0.07828282828282829, + "grad_norm": 1.601773977279663, + "learning_rate": 1.0084033613445378e-06, + "loss": 2.5685, + "step": 62 + }, + { + "epoch": 0.07954545454545454, + "grad_norm": 1.8327804803848267, + "learning_rate": 1.0294117647058825e-06, + "loss": 2.5763, + "step": 63 + }, + { + "epoch": 0.08080808080808081, + "grad_norm": 1.5276741981506348, + "learning_rate": 1.050420168067227e-06, + "loss": 2.5494, + "step": 64 + }, + { + "epoch": 0.08207070707070707, + "grad_norm": 1.7065422534942627, + "learning_rate": 1.0714285714285714e-06, + "loss": 2.449, + "step": 65 + }, + { + "epoch": 0.08333333333333333, + "grad_norm": 2.0551910400390625, + "learning_rate": 1.092436974789916e-06, + "loss": 2.4371, + "step": 66 + }, + { + "epoch": 0.0845959595959596, + "grad_norm": 2.154855728149414, + "learning_rate": 1.1134453781512607e-06, + "loss": 2.5417, + "step": 67 + }, + { + "epoch": 0.08585858585858586, + "grad_norm": 1.7922390699386597, + "learning_rate": 1.1344537815126051e-06, + "loss": 2.4754, + "step": 68 + }, + { + "epoch": 0.08712121212121213, + "grad_norm": 1.745632290840149, + "learning_rate": 1.1554621848739498e-06, + "loss": 2.4669, + "step": 69 + }, + { + "epoch": 0.08838383838383838, + "grad_norm": 1.8907713890075684, + "learning_rate": 1.1764705882352942e-06, + "loss": 2.4943, + "step": 70 + }, + { + "epoch": 0.08838383838383838, + "eval_loss": 2.518237829208374, + "eval_runtime": 37.074, + "eval_samples_per_second": 26.973, + "eval_steps_per_second": 1.133, + "step": 70 + }, + { + "epoch": 0.08964646464646464, + "grad_norm": 1.8178164958953857, + "learning_rate": 1.1974789915966389e-06, + "loss": 2.4891, + "step": 71 + }, + { + "epoch": 0.09090909090909091, + "grad_norm": 1.970969796180725, + "learning_rate": 1.2184873949579833e-06, + "loss": 2.4797, + "step": 72 + }, + { + "epoch": 0.09217171717171717, + "grad_norm": 2.0211431980133057, + "learning_rate": 1.2394957983193278e-06, + "loss": 2.4716, + "step": 73 + }, + { + "epoch": 0.09343434343434344, + "grad_norm": 2.650787591934204, + "learning_rate": 1.2605042016806724e-06, + "loss": 2.4915, + "step": 74 + }, + { + "epoch": 0.0946969696969697, + "grad_norm": 2.235499143600464, + "learning_rate": 1.2815126050420169e-06, + "loss": 2.3917, + "step": 75 + }, + { + "epoch": 0.09595959595959595, + "grad_norm": 2.245999813079834, + "learning_rate": 1.3025210084033615e-06, + "loss": 2.4942, + "step": 76 + }, + { + "epoch": 0.09722222222222222, + "grad_norm": 2.4414420127868652, + "learning_rate": 1.323529411764706e-06, + "loss": 2.4004, + "step": 77 + }, + { + "epoch": 0.09848484848484848, + "grad_norm": 2.279937982559204, + "learning_rate": 1.3445378151260504e-06, + "loss": 2.4043, + "step": 78 + }, + { + "epoch": 0.09974747474747475, + "grad_norm": 2.3830068111419678, + "learning_rate": 1.3655462184873949e-06, + "loss": 2.4096, + "step": 79 + }, + { + "epoch": 0.10101010101010101, + "grad_norm": 2.0119733810424805, + "learning_rate": 1.3865546218487397e-06, + "loss": 2.4227, + "step": 80 + }, + { + "epoch": 0.10101010101010101, + "eval_loss": 2.4316186904907227, + "eval_runtime": 37.0837, + "eval_samples_per_second": 26.966, + "eval_steps_per_second": 1.133, + "step": 80 + }, + { + "epoch": 0.10227272727272728, + "grad_norm": 2.1273374557495117, + "learning_rate": 1.4075630252100842e-06, + "loss": 2.3541, + "step": 81 + }, + { + "epoch": 0.10353535353535354, + "grad_norm": 2.6852898597717285, + "learning_rate": 1.4285714285714286e-06, + "loss": 2.3321, + "step": 82 + }, + { + "epoch": 0.10479797979797979, + "grad_norm": 2.2622721195220947, + "learning_rate": 1.4495798319327733e-06, + "loss": 2.3166, + "step": 83 + }, + { + "epoch": 0.10606060606060606, + "grad_norm": 2.98258900642395, + "learning_rate": 1.4705882352941177e-06, + "loss": 2.3733, + "step": 84 + }, + { + "epoch": 0.10732323232323232, + "grad_norm": 2.5140459537506104, + "learning_rate": 1.4915966386554624e-06, + "loss": 2.314, + "step": 85 + }, + { + "epoch": 0.10858585858585859, + "grad_norm": 2.637608051300049, + "learning_rate": 1.5126050420168068e-06, + "loss": 2.3479, + "step": 86 + }, + { + "epoch": 0.10984848484848485, + "grad_norm": 2.399508476257324, + "learning_rate": 1.5336134453781515e-06, + "loss": 2.2932, + "step": 87 + }, + { + "epoch": 0.1111111111111111, + "grad_norm": 3.0835793018341064, + "learning_rate": 1.554621848739496e-06, + "loss": 2.2886, + "step": 88 + }, + { + "epoch": 0.11237373737373738, + "grad_norm": 3.288769483566284, + "learning_rate": 1.5756302521008404e-06, + "loss": 2.2914, + "step": 89 + }, + { + "epoch": 0.11363636363636363, + "grad_norm": 2.538361072540283, + "learning_rate": 1.5966386554621848e-06, + "loss": 2.3247, + "step": 90 + }, + { + "epoch": 0.11363636363636363, + "eval_loss": 2.3510072231292725, + "eval_runtime": 37.0847, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 90 + }, + { + "epoch": 0.1148989898989899, + "grad_norm": 2.8016724586486816, + "learning_rate": 1.6176470588235297e-06, + "loss": 2.281, + "step": 91 + }, + { + "epoch": 0.11616161616161616, + "grad_norm": 2.9251859188079834, + "learning_rate": 1.6386554621848741e-06, + "loss": 2.2715, + "step": 92 + }, + { + "epoch": 0.11742424242424243, + "grad_norm": 2.595750570297241, + "learning_rate": 1.6596638655462186e-06, + "loss": 2.2127, + "step": 93 + }, + { + "epoch": 0.11868686868686869, + "grad_norm": 2.6989378929138184, + "learning_rate": 1.6806722689075632e-06, + "loss": 2.2623, + "step": 94 + }, + { + "epoch": 0.11994949494949494, + "grad_norm": 2.554863929748535, + "learning_rate": 1.7016806722689077e-06, + "loss": 2.2649, + "step": 95 + }, + { + "epoch": 0.12121212121212122, + "grad_norm": 2.7013020515441895, + "learning_rate": 1.7226890756302521e-06, + "loss": 2.2428, + "step": 96 + }, + { + "epoch": 0.12247474747474747, + "grad_norm": 2.7415151596069336, + "learning_rate": 1.7436974789915968e-06, + "loss": 2.1882, + "step": 97 + }, + { + "epoch": 0.12373737373737374, + "grad_norm": 2.672806739807129, + "learning_rate": 1.7647058823529414e-06, + "loss": 2.2302, + "step": 98 + }, + { + "epoch": 0.125, + "grad_norm": 2.9508326053619385, + "learning_rate": 1.7857142857142859e-06, + "loss": 2.1894, + "step": 99 + }, + { + "epoch": 0.12626262626262627, + "grad_norm": 2.4668166637420654, + "learning_rate": 1.8067226890756303e-06, + "loss": 2.2043, + "step": 100 + }, + { + "epoch": 0.12626262626262627, + "eval_loss": 2.2788307666778564, + "eval_runtime": 37.1156, + "eval_samples_per_second": 26.943, + "eval_steps_per_second": 1.132, + "step": 100 + }, + { + "epoch": 0.1275252525252525, + "grad_norm": 2.8241467475891113, + "learning_rate": 1.8277310924369748e-06, + "loss": 2.2109, + "step": 101 + }, + { + "epoch": 0.12878787878787878, + "grad_norm": 2.4179229736328125, + "learning_rate": 1.8487394957983196e-06, + "loss": 2.2193, + "step": 102 + }, + { + "epoch": 0.13005050505050506, + "grad_norm": 2.714315891265869, + "learning_rate": 1.869747899159664e-06, + "loss": 2.2474, + "step": 103 + }, + { + "epoch": 0.13131313131313133, + "grad_norm": 2.8560359477996826, + "learning_rate": 1.8907563025210085e-06, + "loss": 2.2205, + "step": 104 + }, + { + "epoch": 0.13257575757575757, + "grad_norm": 2.7780051231384277, + "learning_rate": 1.9117647058823528e-06, + "loss": 2.1834, + "step": 105 + }, + { + "epoch": 0.13383838383838384, + "grad_norm": 3.3773467540740967, + "learning_rate": 1.932773109243698e-06, + "loss": 2.2063, + "step": 106 + }, + { + "epoch": 0.1351010101010101, + "grad_norm": 2.927551746368408, + "learning_rate": 1.953781512605042e-06, + "loss": 2.1592, + "step": 107 + }, + { + "epoch": 0.13636363636363635, + "grad_norm": 2.8094165325164795, + "learning_rate": 1.9747899159663867e-06, + "loss": 2.2619, + "step": 108 + }, + { + "epoch": 0.13762626262626262, + "grad_norm": 2.5312564373016357, + "learning_rate": 1.9957983193277314e-06, + "loss": 2.2323, + "step": 109 + }, + { + "epoch": 0.1388888888888889, + "grad_norm": 3.175746202468872, + "learning_rate": 2.0168067226890756e-06, + "loss": 2.1776, + "step": 110 + }, + { + "epoch": 0.1388888888888889, + "eval_loss": 2.2208564281463623, + "eval_runtime": 37.3885, + "eval_samples_per_second": 26.746, + "eval_steps_per_second": 1.123, + "step": 110 + }, + { + "epoch": 0.14015151515151514, + "grad_norm": 2.660886287689209, + "learning_rate": 2.0378151260504203e-06, + "loss": 2.1626, + "step": 111 + }, + { + "epoch": 0.1414141414141414, + "grad_norm": 2.6948511600494385, + "learning_rate": 2.058823529411765e-06, + "loss": 2.1731, + "step": 112 + }, + { + "epoch": 0.14267676767676768, + "grad_norm": 2.826976776123047, + "learning_rate": 2.0798319327731096e-06, + "loss": 2.1865, + "step": 113 + }, + { + "epoch": 0.14393939393939395, + "grad_norm": 3.210033655166626, + "learning_rate": 2.100840336134454e-06, + "loss": 2.1325, + "step": 114 + }, + { + "epoch": 0.1452020202020202, + "grad_norm": 2.509060859680176, + "learning_rate": 2.1218487394957985e-06, + "loss": 2.0926, + "step": 115 + }, + { + "epoch": 0.14646464646464646, + "grad_norm": 3.51330304145813, + "learning_rate": 2.1428571428571427e-06, + "loss": 2.1577, + "step": 116 + }, + { + "epoch": 0.14772727272727273, + "grad_norm": 2.468395471572876, + "learning_rate": 2.163865546218488e-06, + "loss": 2.1, + "step": 117 + }, + { + "epoch": 0.14898989898989898, + "grad_norm": 3.081529378890991, + "learning_rate": 2.184873949579832e-06, + "loss": 2.1082, + "step": 118 + }, + { + "epoch": 0.15025252525252525, + "grad_norm": 2.8856358528137207, + "learning_rate": 2.2058823529411767e-06, + "loss": 2.1288, + "step": 119 + }, + { + "epoch": 0.15151515151515152, + "grad_norm": 3.323089838027954, + "learning_rate": 2.2268907563025214e-06, + "loss": 2.0933, + "step": 120 + }, + { + "epoch": 0.15151515151515152, + "eval_loss": 2.1834487915039062, + "eval_runtime": 37.4344, + "eval_samples_per_second": 26.713, + "eval_steps_per_second": 1.122, + "step": 120 + }, + { + "epoch": 0.1527777777777778, + "grad_norm": 2.8871560096740723, + "learning_rate": 2.2478991596638656e-06, + "loss": 2.1519, + "step": 121 + }, + { + "epoch": 0.15404040404040403, + "grad_norm": 3.0718588829040527, + "learning_rate": 2.2689075630252102e-06, + "loss": 2.0321, + "step": 122 + }, + { + "epoch": 0.1553030303030303, + "grad_norm": 2.820739507675171, + "learning_rate": 2.289915966386555e-06, + "loss": 2.0944, + "step": 123 + }, + { + "epoch": 0.15656565656565657, + "grad_norm": 3.1259310245513916, + "learning_rate": 2.3109243697478996e-06, + "loss": 2.0141, + "step": 124 + }, + { + "epoch": 0.15782828282828282, + "grad_norm": 3.303004741668701, + "learning_rate": 2.331932773109244e-06, + "loss": 2.1095, + "step": 125 + }, + { + "epoch": 0.1590909090909091, + "grad_norm": 3.1838674545288086, + "learning_rate": 2.3529411764705885e-06, + "loss": 2.0727, + "step": 126 + }, + { + "epoch": 0.16035353535353536, + "grad_norm": 3.1013786792755127, + "learning_rate": 2.3739495798319327e-06, + "loss": 2.0963, + "step": 127 + }, + { + "epoch": 0.16161616161616163, + "grad_norm": 2.803283214569092, + "learning_rate": 2.3949579831932778e-06, + "loss": 2.0551, + "step": 128 + }, + { + "epoch": 0.16287878787878787, + "grad_norm": 3.1616814136505127, + "learning_rate": 2.415966386554622e-06, + "loss": 2.0485, + "step": 129 + }, + { + "epoch": 0.16414141414141414, + "grad_norm": 3.333782196044922, + "learning_rate": 2.4369747899159667e-06, + "loss": 2.041, + "step": 130 + }, + { + "epoch": 0.16414141414141414, + "eval_loss": 2.150329351425171, + "eval_runtime": 37.494, + "eval_samples_per_second": 26.671, + "eval_steps_per_second": 1.12, + "step": 130 + }, + { + "epoch": 0.16540404040404041, + "grad_norm": 2.9714694023132324, + "learning_rate": 2.457983193277311e-06, + "loss": 2.0686, + "step": 131 + }, + { + "epoch": 0.16666666666666666, + "grad_norm": 3.216895818710327, + "learning_rate": 2.4789915966386555e-06, + "loss": 2.003, + "step": 132 + }, + { + "epoch": 0.16792929292929293, + "grad_norm": 2.9636809825897217, + "learning_rate": 2.5e-06, + "loss": 2.0534, + "step": 133 + }, + { + "epoch": 0.1691919191919192, + "grad_norm": 3.6477177143096924, + "learning_rate": 2.521008403361345e-06, + "loss": 2.077, + "step": 134 + }, + { + "epoch": 0.17045454545454544, + "grad_norm": 2.7437169551849365, + "learning_rate": 2.542016806722689e-06, + "loss": 2.0501, + "step": 135 + }, + { + "epoch": 0.1717171717171717, + "grad_norm": 3.2992663383483887, + "learning_rate": 2.5630252100840338e-06, + "loss": 2.0284, + "step": 136 + }, + { + "epoch": 0.17297979797979798, + "grad_norm": 3.0735950469970703, + "learning_rate": 2.5840336134453784e-06, + "loss": 2.025, + "step": 137 + }, + { + "epoch": 0.17424242424242425, + "grad_norm": 3.588418960571289, + "learning_rate": 2.605042016806723e-06, + "loss": 2.002, + "step": 138 + }, + { + "epoch": 0.1755050505050505, + "grad_norm": 2.8772335052490234, + "learning_rate": 2.6260504201680677e-06, + "loss": 1.9873, + "step": 139 + }, + { + "epoch": 0.17676767676767677, + "grad_norm": 2.9078969955444336, + "learning_rate": 2.647058823529412e-06, + "loss": 2.011, + "step": 140 + }, + { + "epoch": 0.17676767676767677, + "eval_loss": 2.1225531101226807, + "eval_runtime": 37.4531, + "eval_samples_per_second": 26.7, + "eval_steps_per_second": 1.121, + "step": 140 + }, + { + "epoch": 0.17803030303030304, + "grad_norm": 3.556654214859009, + "learning_rate": 2.6680672268907566e-06, + "loss": 2.0401, + "step": 141 + }, + { + "epoch": 0.17929292929292928, + "grad_norm": 2.9375200271606445, + "learning_rate": 2.689075630252101e-06, + "loss": 2.0277, + "step": 142 + }, + { + "epoch": 0.18055555555555555, + "grad_norm": 3.4019014835357666, + "learning_rate": 2.7100840336134455e-06, + "loss": 1.975, + "step": 143 + }, + { + "epoch": 0.18181818181818182, + "grad_norm": 2.7653748989105225, + "learning_rate": 2.7310924369747897e-06, + "loss": 2.0102, + "step": 144 + }, + { + "epoch": 0.1830808080808081, + "grad_norm": 3.4416797161102295, + "learning_rate": 2.752100840336135e-06, + "loss": 2.0093, + "step": 145 + }, + { + "epoch": 0.18434343434343434, + "grad_norm": 3.2531726360321045, + "learning_rate": 2.7731092436974795e-06, + "loss": 1.9852, + "step": 146 + }, + { + "epoch": 0.1856060606060606, + "grad_norm": 3.171502113342285, + "learning_rate": 2.7941176470588237e-06, + "loss": 1.9165, + "step": 147 + }, + { + "epoch": 0.18686868686868688, + "grad_norm": 3.114550828933716, + "learning_rate": 2.8151260504201684e-06, + "loss": 1.9516, + "step": 148 + }, + { + "epoch": 0.18813131313131312, + "grad_norm": 3.3601152896881104, + "learning_rate": 2.8361344537815126e-06, + "loss": 2.0183, + "step": 149 + }, + { + "epoch": 0.1893939393939394, + "grad_norm": 3.059622049331665, + "learning_rate": 2.8571428571428573e-06, + "loss": 1.9682, + "step": 150 + }, + { + "epoch": 0.1893939393939394, + "eval_loss": 2.0962483882904053, + "eval_runtime": 37.2364, + "eval_samples_per_second": 26.855, + "eval_steps_per_second": 1.128, + "step": 150 + }, + { + "epoch": 0.19065656565656566, + "grad_norm": 2.8749427795410156, + "learning_rate": 2.8781512605042015e-06, + "loss": 2.0089, + "step": 151 + }, + { + "epoch": 0.1919191919191919, + "grad_norm": 3.067267894744873, + "learning_rate": 2.8991596638655466e-06, + "loss": 1.9661, + "step": 152 + }, + { + "epoch": 0.19318181818181818, + "grad_norm": 2.7126266956329346, + "learning_rate": 2.9201680672268912e-06, + "loss": 1.982, + "step": 153 + }, + { + "epoch": 0.19444444444444445, + "grad_norm": 3.1330652236938477, + "learning_rate": 2.9411764705882355e-06, + "loss": 1.9753, + "step": 154 + }, + { + "epoch": 0.19570707070707072, + "grad_norm": 3.158313274383545, + "learning_rate": 2.96218487394958e-06, + "loss": 1.9431, + "step": 155 + }, + { + "epoch": 0.19696969696969696, + "grad_norm": 3.30806303024292, + "learning_rate": 2.9831932773109248e-06, + "loss": 1.9812, + "step": 156 + }, + { + "epoch": 0.19823232323232323, + "grad_norm": 3.0344884395599365, + "learning_rate": 3.004201680672269e-06, + "loss": 1.9677, + "step": 157 + }, + { + "epoch": 0.1994949494949495, + "grad_norm": 2.9032909870147705, + "learning_rate": 3.0252100840336137e-06, + "loss": 1.9132, + "step": 158 + }, + { + "epoch": 0.20075757575757575, + "grad_norm": 3.459934711456299, + "learning_rate": 3.046218487394958e-06, + "loss": 1.9581, + "step": 159 + }, + { + "epoch": 0.20202020202020202, + "grad_norm": 2.9455478191375732, + "learning_rate": 3.067226890756303e-06, + "loss": 1.9509, + "step": 160 + }, + { + "epoch": 0.20202020202020202, + "eval_loss": 2.074032783508301, + "eval_runtime": 37.1824, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 160 + }, + { + "epoch": 0.2032828282828283, + "grad_norm": 3.2347211837768555, + "learning_rate": 3.0882352941176476e-06, + "loss": 1.9743, + "step": 161 + }, + { + "epoch": 0.20454545454545456, + "grad_norm": 3.3152947425842285, + "learning_rate": 3.109243697478992e-06, + "loss": 1.8971, + "step": 162 + }, + { + "epoch": 0.2058080808080808, + "grad_norm": 3.035548210144043, + "learning_rate": 3.1302521008403365e-06, + "loss": 1.8762, + "step": 163 + }, + { + "epoch": 0.20707070707070707, + "grad_norm": 3.1004066467285156, + "learning_rate": 3.1512605042016808e-06, + "loss": 1.9512, + "step": 164 + }, + { + "epoch": 0.20833333333333334, + "grad_norm": 3.2280282974243164, + "learning_rate": 3.1722689075630254e-06, + "loss": 1.8856, + "step": 165 + }, + { + "epoch": 0.20959595959595959, + "grad_norm": 3.021969795227051, + "learning_rate": 3.1932773109243696e-06, + "loss": 1.9127, + "step": 166 + }, + { + "epoch": 0.21085858585858586, + "grad_norm": 3.1647000312805176, + "learning_rate": 3.2142857142857147e-06, + "loss": 2.038, + "step": 167 + }, + { + "epoch": 0.21212121212121213, + "grad_norm": 3.0084872245788574, + "learning_rate": 3.2352941176470594e-06, + "loss": 1.9536, + "step": 168 + }, + { + "epoch": 0.21338383838383837, + "grad_norm": 2.951509952545166, + "learning_rate": 3.2563025210084036e-06, + "loss": 1.933, + "step": 169 + }, + { + "epoch": 0.21464646464646464, + "grad_norm": 3.472856283187866, + "learning_rate": 3.2773109243697483e-06, + "loss": 1.934, + "step": 170 + }, + { + "epoch": 0.21464646464646464, + "eval_loss": 2.0450048446655273, + "eval_runtime": 37.1669, + "eval_samples_per_second": 26.906, + "eval_steps_per_second": 1.13, + "step": 170 + }, + { + "epoch": 0.2159090909090909, + "grad_norm": 3.1104896068573, + "learning_rate": 3.2983193277310925e-06, + "loss": 1.9008, + "step": 171 + }, + { + "epoch": 0.21717171717171718, + "grad_norm": 2.847781181335449, + "learning_rate": 3.319327731092437e-06, + "loss": 1.9592, + "step": 172 + }, + { + "epoch": 0.21843434343434343, + "grad_norm": 3.384554624557495, + "learning_rate": 3.3403361344537814e-06, + "loss": 1.9087, + "step": 173 + }, + { + "epoch": 0.2196969696969697, + "grad_norm": 3.221564292907715, + "learning_rate": 3.3613445378151265e-06, + "loss": 1.9107, + "step": 174 + }, + { + "epoch": 0.22095959595959597, + "grad_norm": 3.30204439163208, + "learning_rate": 3.382352941176471e-06, + "loss": 1.8879, + "step": 175 + }, + { + "epoch": 0.2222222222222222, + "grad_norm": 3.549259662628174, + "learning_rate": 3.4033613445378154e-06, + "loss": 1.9592, + "step": 176 + }, + { + "epoch": 0.22348484848484848, + "grad_norm": 3.498199462890625, + "learning_rate": 3.42436974789916e-06, + "loss": 1.984, + "step": 177 + }, + { + "epoch": 0.22474747474747475, + "grad_norm": 3.576317548751831, + "learning_rate": 3.4453781512605043e-06, + "loss": 1.8897, + "step": 178 + }, + { + "epoch": 0.22601010101010102, + "grad_norm": 3.5337026119232178, + "learning_rate": 3.466386554621849e-06, + "loss": 1.9607, + "step": 179 + }, + { + "epoch": 0.22727272727272727, + "grad_norm": 3.342886209487915, + "learning_rate": 3.4873949579831936e-06, + "loss": 1.8797, + "step": 180 + }, + { + "epoch": 0.22727272727272727, + "eval_loss": 2.02158522605896, + "eval_runtime": 37.1382, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 180 + }, + { + "epoch": 0.22853535353535354, + "grad_norm": 2.941906213760376, + "learning_rate": 3.508403361344538e-06, + "loss": 1.9, + "step": 181 + }, + { + "epoch": 0.2297979797979798, + "grad_norm": 3.1913063526153564, + "learning_rate": 3.529411764705883e-06, + "loss": 1.8704, + "step": 182 + }, + { + "epoch": 0.23106060606060605, + "grad_norm": 3.0619900226593018, + "learning_rate": 3.550420168067227e-06, + "loss": 1.8851, + "step": 183 + }, + { + "epoch": 0.23232323232323232, + "grad_norm": 3.051288604736328, + "learning_rate": 3.5714285714285718e-06, + "loss": 1.9422, + "step": 184 + }, + { + "epoch": 0.2335858585858586, + "grad_norm": 2.999845266342163, + "learning_rate": 3.5924369747899164e-06, + "loss": 1.8757, + "step": 185 + }, + { + "epoch": 0.23484848484848486, + "grad_norm": 3.262355089187622, + "learning_rate": 3.6134453781512607e-06, + "loss": 1.9114, + "step": 186 + }, + { + "epoch": 0.2361111111111111, + "grad_norm": 3.0354340076446533, + "learning_rate": 3.6344537815126053e-06, + "loss": 1.8547, + "step": 187 + }, + { + "epoch": 0.23737373737373738, + "grad_norm": 2.9401369094848633, + "learning_rate": 3.6554621848739496e-06, + "loss": 1.8974, + "step": 188 + }, + { + "epoch": 0.23863636363636365, + "grad_norm": 3.0544443130493164, + "learning_rate": 3.6764705882352946e-06, + "loss": 1.8662, + "step": 189 + }, + { + "epoch": 0.2398989898989899, + "grad_norm": 2.7257750034332275, + "learning_rate": 3.6974789915966393e-06, + "loss": 1.8737, + "step": 190 + }, + { + "epoch": 0.2398989898989899, + "eval_loss": 2.0065128803253174, + "eval_runtime": 37.1296, + "eval_samples_per_second": 26.933, + "eval_steps_per_second": 1.131, + "step": 190 + }, + { + "epoch": 0.24116161616161616, + "grad_norm": 2.8592522144317627, + "learning_rate": 3.7184873949579835e-06, + "loss": 1.8984, + "step": 191 + }, + { + "epoch": 0.24242424242424243, + "grad_norm": 2.833958864212036, + "learning_rate": 3.739495798319328e-06, + "loss": 1.9084, + "step": 192 + }, + { + "epoch": 0.24368686868686867, + "grad_norm": 2.7842230796813965, + "learning_rate": 3.7605042016806724e-06, + "loss": 1.8843, + "step": 193 + }, + { + "epoch": 0.24494949494949494, + "grad_norm": 3.1309409141540527, + "learning_rate": 3.781512605042017e-06, + "loss": 1.8259, + "step": 194 + }, + { + "epoch": 0.24621212121212122, + "grad_norm": 3.120788335800171, + "learning_rate": 3.8025210084033613e-06, + "loss": 1.8951, + "step": 195 + }, + { + "epoch": 0.2474747474747475, + "grad_norm": 3.2228143215179443, + "learning_rate": 3.8235294117647055e-06, + "loss": 1.8608, + "step": 196 + }, + { + "epoch": 0.24873737373737373, + "grad_norm": 3.3362114429473877, + "learning_rate": 3.844537815126051e-06, + "loss": 1.9163, + "step": 197 + }, + { + "epoch": 0.25, + "grad_norm": 3.0973799228668213, + "learning_rate": 3.865546218487396e-06, + "loss": 1.911, + "step": 198 + }, + { + "epoch": 0.25126262626262624, + "grad_norm": 3.5567545890808105, + "learning_rate": 3.8865546218487395e-06, + "loss": 1.8795, + "step": 199 + }, + { + "epoch": 0.25252525252525254, + "grad_norm": 3.4758830070495605, + "learning_rate": 3.907563025210084e-06, + "loss": 1.8813, + "step": 200 + }, + { + "epoch": 0.25252525252525254, + "eval_loss": 1.9892102479934692, + "eval_runtime": 37.1311, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 200 + }, + { + "epoch": 0.2537878787878788, + "grad_norm": 3.5564475059509277, + "learning_rate": 3.928571428571429e-06, + "loss": 1.8469, + "step": 201 + }, + { + "epoch": 0.255050505050505, + "grad_norm": 3.7004899978637695, + "learning_rate": 3.9495798319327735e-06, + "loss": 1.873, + "step": 202 + }, + { + "epoch": 0.2563131313131313, + "grad_norm": 3.9285225868225098, + "learning_rate": 3.970588235294118e-06, + "loss": 1.8596, + "step": 203 + }, + { + "epoch": 0.25757575757575757, + "grad_norm": 3.1388907432556152, + "learning_rate": 3.991596638655463e-06, + "loss": 1.7991, + "step": 204 + }, + { + "epoch": 0.2588383838383838, + "grad_norm": 3.3797011375427246, + "learning_rate": 4.0126050420168075e-06, + "loss": 1.8682, + "step": 205 + }, + { + "epoch": 0.2601010101010101, + "grad_norm": 3.32183575630188, + "learning_rate": 4.033613445378151e-06, + "loss": 1.8396, + "step": 206 + }, + { + "epoch": 0.26136363636363635, + "grad_norm": 3.457291603088379, + "learning_rate": 4.054621848739496e-06, + "loss": 1.8314, + "step": 207 + }, + { + "epoch": 0.26262626262626265, + "grad_norm": 3.2906265258789062, + "learning_rate": 4.075630252100841e-06, + "loss": 1.844, + "step": 208 + }, + { + "epoch": 0.2638888888888889, + "grad_norm": 3.081913948059082, + "learning_rate": 4.096638655462185e-06, + "loss": 1.8198, + "step": 209 + }, + { + "epoch": 0.26515151515151514, + "grad_norm": 2.9697105884552, + "learning_rate": 4.11764705882353e-06, + "loss": 1.7885, + "step": 210 + }, + { + "epoch": 0.26515151515151514, + "eval_loss": 1.9749963283538818, + "eval_runtime": 36.962, + "eval_samples_per_second": 27.055, + "eval_steps_per_second": 1.136, + "step": 210 + }, + { + "epoch": 0.26641414141414144, + "grad_norm": 3.0732104778289795, + "learning_rate": 4.138655462184874e-06, + "loss": 1.8463, + "step": 211 + }, + { + "epoch": 0.2676767676767677, + "grad_norm": 3.4714958667755127, + "learning_rate": 4.159663865546219e-06, + "loss": 1.8537, + "step": 212 + }, + { + "epoch": 0.2689393939393939, + "grad_norm": 2.955864429473877, + "learning_rate": 4.180672268907564e-06, + "loss": 1.8566, + "step": 213 + }, + { + "epoch": 0.2702020202020202, + "grad_norm": 3.5314948558807373, + "learning_rate": 4.201680672268908e-06, + "loss": 1.7625, + "step": 214 + }, + { + "epoch": 0.27146464646464646, + "grad_norm": 2.8855810165405273, + "learning_rate": 4.222689075630252e-06, + "loss": 1.7968, + "step": 215 + }, + { + "epoch": 0.2727272727272727, + "grad_norm": 3.7866363525390625, + "learning_rate": 4.243697478991597e-06, + "loss": 1.8341, + "step": 216 + }, + { + "epoch": 0.273989898989899, + "grad_norm": 3.0352272987365723, + "learning_rate": 4.264705882352942e-06, + "loss": 1.8276, + "step": 217 + }, + { + "epoch": 0.27525252525252525, + "grad_norm": 3.4880247116088867, + "learning_rate": 4.2857142857142855e-06, + "loss": 1.8194, + "step": 218 + }, + { + "epoch": 0.2765151515151515, + "grad_norm": 3.6807212829589844, + "learning_rate": 4.306722689075631e-06, + "loss": 1.8495, + "step": 219 + }, + { + "epoch": 0.2777777777777778, + "grad_norm": 3.029017448425293, + "learning_rate": 4.327731092436976e-06, + "loss": 1.8125, + "step": 220 + }, + { + "epoch": 0.2777777777777778, + "eval_loss": 1.9633830785751343, + "eval_runtime": 37.1259, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 220 + }, + { + "epoch": 0.27904040404040403, + "grad_norm": 3.4820573329925537, + "learning_rate": 4.3487394957983194e-06, + "loss": 1.8159, + "step": 221 + }, + { + "epoch": 0.2803030303030303, + "grad_norm": 3.2057833671569824, + "learning_rate": 4.369747899159664e-06, + "loss": 1.8215, + "step": 222 + }, + { + "epoch": 0.2815656565656566, + "grad_norm": 3.477302312850952, + "learning_rate": 4.390756302521009e-06, + "loss": 1.7837, + "step": 223 + }, + { + "epoch": 0.2828282828282828, + "grad_norm": 2.8349595069885254, + "learning_rate": 4.411764705882353e-06, + "loss": 1.8555, + "step": 224 + }, + { + "epoch": 0.2840909090909091, + "grad_norm": 3.616478443145752, + "learning_rate": 4.432773109243698e-06, + "loss": 1.8067, + "step": 225 + }, + { + "epoch": 0.28535353535353536, + "grad_norm": 3.111969470977783, + "learning_rate": 4.453781512605043e-06, + "loss": 1.8322, + "step": 226 + }, + { + "epoch": 0.2866161616161616, + "grad_norm": 3.248695135116577, + "learning_rate": 4.474789915966387e-06, + "loss": 1.8004, + "step": 227 + }, + { + "epoch": 0.2878787878787879, + "grad_norm": 3.017082691192627, + "learning_rate": 4.495798319327731e-06, + "loss": 1.776, + "step": 228 + }, + { + "epoch": 0.28914141414141414, + "grad_norm": 3.6790099143981934, + "learning_rate": 4.516806722689076e-06, + "loss": 1.8176, + "step": 229 + }, + { + "epoch": 0.2904040404040404, + "grad_norm": 3.1513001918792725, + "learning_rate": 4.5378151260504205e-06, + "loss": 1.8016, + "step": 230 + }, + { + "epoch": 0.2904040404040404, + "eval_loss": 1.949769377708435, + "eval_runtime": 37.2573, + "eval_samples_per_second": 26.84, + "eval_steps_per_second": 1.127, + "step": 230 + }, + { + "epoch": 0.2916666666666667, + "grad_norm": 3.0559096336364746, + "learning_rate": 4.558823529411765e-06, + "loss": 1.7915, + "step": 231 + }, + { + "epoch": 0.29292929292929293, + "grad_norm": 3.0974984169006348, + "learning_rate": 4.57983193277311e-06, + "loss": 1.8277, + "step": 232 + }, + { + "epoch": 0.29419191919191917, + "grad_norm": 2.8514139652252197, + "learning_rate": 4.600840336134454e-06, + "loss": 1.7616, + "step": 233 + }, + { + "epoch": 0.29545454545454547, + "grad_norm": 3.355198383331299, + "learning_rate": 4.621848739495799e-06, + "loss": 1.841, + "step": 234 + }, + { + "epoch": 0.2967171717171717, + "grad_norm": 3.1467435359954834, + "learning_rate": 4.642857142857144e-06, + "loss": 1.7393, + "step": 235 + }, + { + "epoch": 0.29797979797979796, + "grad_norm": 4.065819263458252, + "learning_rate": 4.663865546218488e-06, + "loss": 1.7508, + "step": 236 + }, + { + "epoch": 0.29924242424242425, + "grad_norm": 3.2602460384368896, + "learning_rate": 4.684873949579832e-06, + "loss": 1.7415, + "step": 237 + }, + { + "epoch": 0.3005050505050505, + "grad_norm": 3.6884894371032715, + "learning_rate": 4.705882352941177e-06, + "loss": 1.7608, + "step": 238 + }, + { + "epoch": 0.30176767676767674, + "grad_norm": 3.9843380451202393, + "learning_rate": 4.7268907563025216e-06, + "loss": 1.8519, + "step": 239 + }, + { + "epoch": 0.30303030303030304, + "grad_norm": 3.2823405265808105, + "learning_rate": 4.747899159663865e-06, + "loss": 1.7689, + "step": 240 + }, + { + "epoch": 0.30303030303030304, + "eval_loss": 1.939724326133728, + "eval_runtime": 37.465, + "eval_samples_per_second": 26.692, + "eval_steps_per_second": 1.121, + "step": 240 + }, + { + "epoch": 0.3042929292929293, + "grad_norm": 3.4200189113616943, + "learning_rate": 4.768907563025211e-06, + "loss": 1.7944, + "step": 241 + }, + { + "epoch": 0.3055555555555556, + "grad_norm": 3.7541189193725586, + "learning_rate": 4.7899159663865555e-06, + "loss": 1.8151, + "step": 242 + }, + { + "epoch": 0.3068181818181818, + "grad_norm": 3.9103832244873047, + "learning_rate": 4.810924369747899e-06, + "loss": 1.7921, + "step": 243 + }, + { + "epoch": 0.30808080808080807, + "grad_norm": 3.2769975662231445, + "learning_rate": 4.831932773109244e-06, + "loss": 1.8408, + "step": 244 + }, + { + "epoch": 0.30934343434343436, + "grad_norm": 3.414168357849121, + "learning_rate": 4.852941176470589e-06, + "loss": 1.7937, + "step": 245 + }, + { + "epoch": 0.3106060606060606, + "grad_norm": 3.219625473022461, + "learning_rate": 4.873949579831933e-06, + "loss": 1.796, + "step": 246 + }, + { + "epoch": 0.31186868686868685, + "grad_norm": 3.2573912143707275, + "learning_rate": 4.894957983193277e-06, + "loss": 1.8052, + "step": 247 + }, + { + "epoch": 0.31313131313131315, + "grad_norm": 3.4201040267944336, + "learning_rate": 4.915966386554622e-06, + "loss": 1.8529, + "step": 248 + }, + { + "epoch": 0.3143939393939394, + "grad_norm": 3.7448058128356934, + "learning_rate": 4.936974789915967e-06, + "loss": 1.8183, + "step": 249 + }, + { + "epoch": 0.31565656565656564, + "grad_norm": 3.4108386039733887, + "learning_rate": 4.957983193277311e-06, + "loss": 1.7926, + "step": 250 + }, + { + "epoch": 0.31565656565656564, + "eval_loss": 1.9319047927856445, + "eval_runtime": 37.4745, + "eval_samples_per_second": 26.685, + "eval_steps_per_second": 1.121, + "step": 250 + }, + { + "epoch": 0.31691919191919193, + "grad_norm": 3.50185227394104, + "learning_rate": 4.978991596638656e-06, + "loss": 1.792, + "step": 251 + }, + { + "epoch": 0.3181818181818182, + "grad_norm": 3.225581407546997, + "learning_rate": 5e-06, + "loss": 1.7286, + "step": 252 + }, + { + "epoch": 0.3194444444444444, + "grad_norm": 3.3070669174194336, + "learning_rate": 4.997661365762395e-06, + "loss": 1.7957, + "step": 253 + }, + { + "epoch": 0.3207070707070707, + "grad_norm": 3.498079776763916, + "learning_rate": 4.99532273152479e-06, + "loss": 1.823, + "step": 254 + }, + { + "epoch": 0.32196969696969696, + "grad_norm": 3.316534996032715, + "learning_rate": 4.9929840972871845e-06, + "loss": 1.8249, + "step": 255 + }, + { + "epoch": 0.32323232323232326, + "grad_norm": 3.1641833782196045, + "learning_rate": 4.990645463049579e-06, + "loss": 1.7658, + "step": 256 + }, + { + "epoch": 0.3244949494949495, + "grad_norm": 3.0794270038604736, + "learning_rate": 4.988306828811974e-06, + "loss": 1.8146, + "step": 257 + }, + { + "epoch": 0.32575757575757575, + "grad_norm": 3.2410383224487305, + "learning_rate": 4.985968194574369e-06, + "loss": 1.7894, + "step": 258 + }, + { + "epoch": 0.32702020202020204, + "grad_norm": 3.0899274349212646, + "learning_rate": 4.983629560336764e-06, + "loss": 1.7381, + "step": 259 + }, + { + "epoch": 0.3282828282828283, + "grad_norm": 3.0132803916931152, + "learning_rate": 4.981290926099159e-06, + "loss": 1.7713, + "step": 260 + }, + { + "epoch": 0.3282828282828283, + "eval_loss": 1.9123589992523193, + "eval_runtime": 37.3201, + "eval_samples_per_second": 26.795, + "eval_steps_per_second": 1.125, + "step": 260 + }, + { + "epoch": 0.32954545454545453, + "grad_norm": 2.92564058303833, + "learning_rate": 4.9789522918615536e-06, + "loss": 1.7659, + "step": 261 + }, + { + "epoch": 0.33080808080808083, + "grad_norm": 3.1402878761291504, + "learning_rate": 4.976613657623948e-06, + "loss": 1.8439, + "step": 262 + }, + { + "epoch": 0.33207070707070707, + "grad_norm": 3.170881509780884, + "learning_rate": 4.974275023386343e-06, + "loss": 1.7605, + "step": 263 + }, + { + "epoch": 0.3333333333333333, + "grad_norm": 3.014850378036499, + "learning_rate": 4.971936389148738e-06, + "loss": 1.7899, + "step": 264 + }, + { + "epoch": 0.3345959595959596, + "grad_norm": 3.169818878173828, + "learning_rate": 4.969597754911132e-06, + "loss": 1.8263, + "step": 265 + }, + { + "epoch": 0.33585858585858586, + "grad_norm": 2.8709678649902344, + "learning_rate": 4.967259120673527e-06, + "loss": 1.7834, + "step": 266 + }, + { + "epoch": 0.3371212121212121, + "grad_norm": 2.7527096271514893, + "learning_rate": 4.964920486435922e-06, + "loss": 1.8106, + "step": 267 + }, + { + "epoch": 0.3383838383838384, + "grad_norm": 2.9960227012634277, + "learning_rate": 4.962581852198316e-06, + "loss": 1.7452, + "step": 268 + }, + { + "epoch": 0.33964646464646464, + "grad_norm": 2.979163885116577, + "learning_rate": 4.9602432179607115e-06, + "loss": 1.6943, + "step": 269 + }, + { + "epoch": 0.3409090909090909, + "grad_norm": 3.2396063804626465, + "learning_rate": 4.957904583723106e-06, + "loss": 1.7798, + "step": 270 + }, + { + "epoch": 0.3409090909090909, + "eval_loss": 1.9126806259155273, + "eval_runtime": 37.123, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 270 + }, + { + "epoch": 0.3421717171717172, + "grad_norm": 2.7868010997772217, + "learning_rate": 4.955565949485501e-06, + "loss": 1.8053, + "step": 271 + }, + { + "epoch": 0.3434343434343434, + "grad_norm": 2.8895349502563477, + "learning_rate": 4.9532273152478956e-06, + "loss": 1.7672, + "step": 272 + }, + { + "epoch": 0.3446969696969697, + "grad_norm": 3.0839710235595703, + "learning_rate": 4.95088868101029e-06, + "loss": 1.7414, + "step": 273 + }, + { + "epoch": 0.34595959595959597, + "grad_norm": 2.81974458694458, + "learning_rate": 4.948550046772685e-06, + "loss": 1.7527, + "step": 274 + }, + { + "epoch": 0.3472222222222222, + "grad_norm": 3.7469775676727295, + "learning_rate": 4.9462114125350805e-06, + "loss": 1.7689, + "step": 275 + }, + { + "epoch": 0.3484848484848485, + "grad_norm": 3.434573173522949, + "learning_rate": 4.943872778297475e-06, + "loss": 1.739, + "step": 276 + }, + { + "epoch": 0.34974747474747475, + "grad_norm": 2.985041379928589, + "learning_rate": 4.941534144059869e-06, + "loss": 1.7894, + "step": 277 + }, + { + "epoch": 0.351010101010101, + "grad_norm": 3.441204309463501, + "learning_rate": 4.939195509822265e-06, + "loss": 1.7804, + "step": 278 + }, + { + "epoch": 0.3522727272727273, + "grad_norm": 2.8371102809906006, + "learning_rate": 4.936856875584659e-06, + "loss": 1.7415, + "step": 279 + }, + { + "epoch": 0.35353535353535354, + "grad_norm": 3.662773847579956, + "learning_rate": 4.9345182413470535e-06, + "loss": 1.7894, + "step": 280 + }, + { + "epoch": 0.35353535353535354, + "eval_loss": 1.915844440460205, + "eval_runtime": 37.0941, + "eval_samples_per_second": 26.958, + "eval_steps_per_second": 1.132, + "step": 280 + }, + { + "epoch": 0.3547979797979798, + "grad_norm": 3.3254036903381348, + "learning_rate": 4.932179607109449e-06, + "loss": 1.7238, + "step": 281 + }, + { + "epoch": 0.3560606060606061, + "grad_norm": 2.9872665405273438, + "learning_rate": 4.929840972871843e-06, + "loss": 1.769, + "step": 282 + }, + { + "epoch": 0.3573232323232323, + "grad_norm": 3.0163521766662598, + "learning_rate": 4.9275023386342376e-06, + "loss": 1.7702, + "step": 283 + }, + { + "epoch": 0.35858585858585856, + "grad_norm": 2.9687387943267822, + "learning_rate": 4.925163704396633e-06, + "loss": 1.7927, + "step": 284 + }, + { + "epoch": 0.35984848484848486, + "grad_norm": 2.7064011096954346, + "learning_rate": 4.922825070159027e-06, + "loss": 1.7638, + "step": 285 + }, + { + "epoch": 0.3611111111111111, + "grad_norm": 3.0043134689331055, + "learning_rate": 4.9204864359214225e-06, + "loss": 1.8041, + "step": 286 + }, + { + "epoch": 0.36237373737373735, + "grad_norm": 3.037100315093994, + "learning_rate": 4.918147801683817e-06, + "loss": 1.7579, + "step": 287 + }, + { + "epoch": 0.36363636363636365, + "grad_norm": 2.7528676986694336, + "learning_rate": 4.915809167446211e-06, + "loss": 1.7629, + "step": 288 + }, + { + "epoch": 0.3648989898989899, + "grad_norm": 3.075815439224243, + "learning_rate": 4.913470533208607e-06, + "loss": 1.7966, + "step": 289 + }, + { + "epoch": 0.3661616161616162, + "grad_norm": 2.964785575866699, + "learning_rate": 4.911131898971001e-06, + "loss": 1.801, + "step": 290 + }, + { + "epoch": 0.3661616161616162, + "eval_loss": 1.8839410543441772, + "eval_runtime": 37.0844, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 290 + }, + { + "epoch": 0.36742424242424243, + "grad_norm": 2.668747663497925, + "learning_rate": 4.908793264733396e-06, + "loss": 1.7703, + "step": 291 + }, + { + "epoch": 0.3686868686868687, + "grad_norm": 2.814751148223877, + "learning_rate": 4.906454630495791e-06, + "loss": 1.7075, + "step": 292 + }, + { + "epoch": 0.369949494949495, + "grad_norm": 3.005922794342041, + "learning_rate": 4.904115996258186e-06, + "loss": 1.7418, + "step": 293 + }, + { + "epoch": 0.3712121212121212, + "grad_norm": 3.050936460494995, + "learning_rate": 4.90177736202058e-06, + "loss": 1.793, + "step": 294 + }, + { + "epoch": 0.37247474747474746, + "grad_norm": 2.822004556655884, + "learning_rate": 4.899438727782975e-06, + "loss": 1.8303, + "step": 295 + }, + { + "epoch": 0.37373737373737376, + "grad_norm": 2.950254201889038, + "learning_rate": 4.89710009354537e-06, + "loss": 1.808, + "step": 296 + }, + { + "epoch": 0.375, + "grad_norm": 2.8140666484832764, + "learning_rate": 4.8947614593077645e-06, + "loss": 1.7262, + "step": 297 + }, + { + "epoch": 0.37626262626262624, + "grad_norm": 2.899620532989502, + "learning_rate": 4.89242282507016e-06, + "loss": 1.7157, + "step": 298 + }, + { + "epoch": 0.37752525252525254, + "grad_norm": 2.805753469467163, + "learning_rate": 4.890084190832554e-06, + "loss": 1.7294, + "step": 299 + }, + { + "epoch": 0.3787878787878788, + "grad_norm": 3.110246181488037, + "learning_rate": 4.887745556594949e-06, + "loss": 1.7415, + "step": 300 + }, + { + "epoch": 0.3787878787878788, + "eval_loss": 1.892285943031311, + "eval_runtime": 37.0732, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 300 + }, + { + "epoch": 0.380050505050505, + "grad_norm": 2.9197490215301514, + "learning_rate": 4.885406922357344e-06, + "loss": 1.7132, + "step": 301 + }, + { + "epoch": 0.3813131313131313, + "grad_norm": 3.3165123462677, + "learning_rate": 4.883068288119738e-06, + "loss": 1.777, + "step": 302 + }, + { + "epoch": 0.38257575757575757, + "grad_norm": 2.8409054279327393, + "learning_rate": 4.880729653882133e-06, + "loss": 1.7092, + "step": 303 + }, + { + "epoch": 0.3838383838383838, + "grad_norm": 2.6800901889801025, + "learning_rate": 4.878391019644528e-06, + "loss": 1.7622, + "step": 304 + }, + { + "epoch": 0.3851010101010101, + "grad_norm": 3.2847490310668945, + "learning_rate": 4.876052385406922e-06, + "loss": 1.729, + "step": 305 + }, + { + "epoch": 0.38636363636363635, + "grad_norm": 2.786057949066162, + "learning_rate": 4.873713751169318e-06, + "loss": 1.7168, + "step": 306 + }, + { + "epoch": 0.38762626262626265, + "grad_norm": 3.236050605773926, + "learning_rate": 4.871375116931712e-06, + "loss": 1.6809, + "step": 307 + }, + { + "epoch": 0.3888888888888889, + "grad_norm": 2.851910352706909, + "learning_rate": 4.869036482694107e-06, + "loss": 1.7336, + "step": 308 + }, + { + "epoch": 0.39015151515151514, + "grad_norm": 2.892421007156372, + "learning_rate": 4.866697848456502e-06, + "loss": 1.7451, + "step": 309 + }, + { + "epoch": 0.39141414141414144, + "grad_norm": 2.8175048828125, + "learning_rate": 4.864359214218897e-06, + "loss": 1.7734, + "step": 310 + }, + { + "epoch": 0.39141414141414144, + "eval_loss": 1.8777843713760376, + "eval_runtime": 36.9321, + "eval_samples_per_second": 27.077, + "eval_steps_per_second": 1.137, + "step": 310 + }, + { + "epoch": 0.3926767676767677, + "grad_norm": 3.005329132080078, + "learning_rate": 4.8620205799812915e-06, + "loss": 1.7476, + "step": 311 + }, + { + "epoch": 0.3939393939393939, + "grad_norm": 2.849017858505249, + "learning_rate": 4.859681945743686e-06, + "loss": 1.7854, + "step": 312 + }, + { + "epoch": 0.3952020202020202, + "grad_norm": 2.801048517227173, + "learning_rate": 4.857343311506081e-06, + "loss": 1.7186, + "step": 313 + }, + { + "epoch": 0.39646464646464646, + "grad_norm": 2.5772831439971924, + "learning_rate": 4.855004677268476e-06, + "loss": 1.7279, + "step": 314 + }, + { + "epoch": 0.3977272727272727, + "grad_norm": 2.7969934940338135, + "learning_rate": 4.85266604303087e-06, + "loss": 1.7475, + "step": 315 + }, + { + "epoch": 0.398989898989899, + "grad_norm": 2.7067770957946777, + "learning_rate": 4.850327408793265e-06, + "loss": 1.7055, + "step": 316 + }, + { + "epoch": 0.40025252525252525, + "grad_norm": 2.70928955078125, + "learning_rate": 4.84798877455566e-06, + "loss": 1.7203, + "step": 317 + }, + { + "epoch": 0.4015151515151515, + "grad_norm": 2.7704474925994873, + "learning_rate": 4.845650140318054e-06, + "loss": 1.8058, + "step": 318 + }, + { + "epoch": 0.4027777777777778, + "grad_norm": 2.7680182456970215, + "learning_rate": 4.843311506080449e-06, + "loss": 1.7162, + "step": 319 + }, + { + "epoch": 0.40404040404040403, + "grad_norm": 3.0953330993652344, + "learning_rate": 4.840972871842844e-06, + "loss": 1.7617, + "step": 320 + }, + { + "epoch": 0.40404040404040403, + "eval_loss": 1.8680965900421143, + "eval_runtime": 37.0382, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 320 + }, + { + "epoch": 0.4053030303030303, + "grad_norm": 2.8340392112731934, + "learning_rate": 4.838634237605239e-06, + "loss": 1.7767, + "step": 321 + }, + { + "epoch": 0.4065656565656566, + "grad_norm": 3.057861804962158, + "learning_rate": 4.836295603367634e-06, + "loss": 1.8043, + "step": 322 + }, + { + "epoch": 0.4078282828282828, + "grad_norm": 3.0123419761657715, + "learning_rate": 4.833956969130029e-06, + "loss": 1.7561, + "step": 323 + }, + { + "epoch": 0.4090909090909091, + "grad_norm": 3.0873124599456787, + "learning_rate": 4.831618334892423e-06, + "loss": 1.7581, + "step": 324 + }, + { + "epoch": 0.41035353535353536, + "grad_norm": 2.6600539684295654, + "learning_rate": 4.8292797006548184e-06, + "loss": 1.7005, + "step": 325 + }, + { + "epoch": 0.4116161616161616, + "grad_norm": 2.905280590057373, + "learning_rate": 4.826941066417213e-06, + "loss": 1.7592, + "step": 326 + }, + { + "epoch": 0.4128787878787879, + "grad_norm": 3.0642189979553223, + "learning_rate": 4.824602432179607e-06, + "loss": 1.6456, + "step": 327 + }, + { + "epoch": 0.41414141414141414, + "grad_norm": 2.606531858444214, + "learning_rate": 4.8222637979420025e-06, + "loss": 1.6731, + "step": 328 + }, + { + "epoch": 0.4154040404040404, + "grad_norm": 2.9918158054351807, + "learning_rate": 4.819925163704397e-06, + "loss": 1.6854, + "step": 329 + }, + { + "epoch": 0.4166666666666667, + "grad_norm": 2.721940040588379, + "learning_rate": 4.817586529466791e-06, + "loss": 1.7346, + "step": 330 + }, + { + "epoch": 0.4166666666666667, + "eval_loss": 1.863736629486084, + "eval_runtime": 37.0887, + "eval_samples_per_second": 26.962, + "eval_steps_per_second": 1.132, + "step": 330 + }, + { + "epoch": 0.41792929292929293, + "grad_norm": 3.1076390743255615, + "learning_rate": 4.815247895229187e-06, + "loss": 1.6286, + "step": 331 + }, + { + "epoch": 0.41919191919191917, + "grad_norm": 3.1235783100128174, + "learning_rate": 4.812909260991581e-06, + "loss": 1.7404, + "step": 332 + }, + { + "epoch": 0.42045454545454547, + "grad_norm": 3.1211330890655518, + "learning_rate": 4.8105706267539755e-06, + "loss": 1.7143, + "step": 333 + }, + { + "epoch": 0.4217171717171717, + "grad_norm": 3.072317600250244, + "learning_rate": 4.808231992516371e-06, + "loss": 1.7514, + "step": 334 + }, + { + "epoch": 0.42297979797979796, + "grad_norm": 2.880671739578247, + "learning_rate": 4.805893358278765e-06, + "loss": 1.7018, + "step": 335 + }, + { + "epoch": 0.42424242424242425, + "grad_norm": 2.700063943862915, + "learning_rate": 4.8035547240411604e-06, + "loss": 1.7089, + "step": 336 + }, + { + "epoch": 0.4255050505050505, + "grad_norm": 2.9038445949554443, + "learning_rate": 4.801216089803556e-06, + "loss": 1.7483, + "step": 337 + }, + { + "epoch": 0.42676767676767674, + "grad_norm": 2.7822766304016113, + "learning_rate": 4.79887745556595e-06, + "loss": 1.6927, + "step": 338 + }, + { + "epoch": 0.42803030303030304, + "grad_norm": 2.957554340362549, + "learning_rate": 4.7965388213283445e-06, + "loss": 1.7162, + "step": 339 + }, + { + "epoch": 0.4292929292929293, + "grad_norm": 3.1333892345428467, + "learning_rate": 4.79420018709074e-06, + "loss": 1.731, + "step": 340 + }, + { + "epoch": 0.4292929292929293, + "eval_loss": 1.852413535118103, + "eval_runtime": 37.1809, + "eval_samples_per_second": 26.896, + "eval_steps_per_second": 1.13, + "step": 340 + }, + { + "epoch": 0.4305555555555556, + "grad_norm": 2.9801084995269775, + "learning_rate": 4.791861552853134e-06, + "loss": 1.6892, + "step": 341 + }, + { + "epoch": 0.4318181818181818, + "grad_norm": 2.5077271461486816, + "learning_rate": 4.789522918615529e-06, + "loss": 1.637, + "step": 342 + }, + { + "epoch": 0.43308080808080807, + "grad_norm": 2.8654391765594482, + "learning_rate": 4.787184284377924e-06, + "loss": 1.6915, + "step": 343 + }, + { + "epoch": 0.43434343434343436, + "grad_norm": 2.6204264163970947, + "learning_rate": 4.784845650140318e-06, + "loss": 1.6722, + "step": 344 + }, + { + "epoch": 0.4356060606060606, + "grad_norm": 2.633991003036499, + "learning_rate": 4.782507015902713e-06, + "loss": 1.6797, + "step": 345 + }, + { + "epoch": 0.43686868686868685, + "grad_norm": 2.7203924655914307, + "learning_rate": 4.780168381665108e-06, + "loss": 1.7018, + "step": 346 + }, + { + "epoch": 0.43813131313131315, + "grad_norm": 2.6789209842681885, + "learning_rate": 4.7778297474275024e-06, + "loss": 1.7153, + "step": 347 + }, + { + "epoch": 0.4393939393939394, + "grad_norm": 2.6285433769226074, + "learning_rate": 4.775491113189898e-06, + "loss": 1.7082, + "step": 348 + }, + { + "epoch": 0.44065656565656564, + "grad_norm": 2.623793840408325, + "learning_rate": 4.773152478952292e-06, + "loss": 1.7094, + "step": 349 + }, + { + "epoch": 0.44191919191919193, + "grad_norm": 2.647160291671753, + "learning_rate": 4.7708138447146865e-06, + "loss": 1.6528, + "step": 350 + }, + { + "epoch": 0.44191919191919193, + "eval_loss": 1.8506718873977661, + "eval_runtime": 37.4232, + "eval_samples_per_second": 26.721, + "eval_steps_per_second": 1.122, + "step": 350 + }, + { + "epoch": 0.4431818181818182, + "grad_norm": 2.664553642272949, + "learning_rate": 4.768475210477082e-06, + "loss": 1.7008, + "step": 351 + }, + { + "epoch": 0.4444444444444444, + "grad_norm": 2.8486251831054688, + "learning_rate": 4.766136576239477e-06, + "loss": 1.6811, + "step": 352 + }, + { + "epoch": 0.4457070707070707, + "grad_norm": 2.7514266967773438, + "learning_rate": 4.7637979420018715e-06, + "loss": 1.6969, + "step": 353 + }, + { + "epoch": 0.44696969696969696, + "grad_norm": 2.848083257675171, + "learning_rate": 4.761459307764266e-06, + "loss": 1.7058, + "step": 354 + }, + { + "epoch": 0.44823232323232326, + "grad_norm": 2.759218454360962, + "learning_rate": 4.759120673526661e-06, + "loss": 1.6109, + "step": 355 + }, + { + "epoch": 0.4494949494949495, + "grad_norm": 2.905318021774292, + "learning_rate": 4.756782039289056e-06, + "loss": 1.7689, + "step": 356 + }, + { + "epoch": 0.45075757575757575, + "grad_norm": 2.845646381378174, + "learning_rate": 4.75444340505145e-06, + "loss": 1.6687, + "step": 357 + }, + { + "epoch": 0.45202020202020204, + "grad_norm": 2.8609306812286377, + "learning_rate": 4.752104770813845e-06, + "loss": 1.6332, + "step": 358 + }, + { + "epoch": 0.4532828282828283, + "grad_norm": 3.016479969024658, + "learning_rate": 4.74976613657624e-06, + "loss": 1.6692, + "step": 359 + }, + { + "epoch": 0.45454545454545453, + "grad_norm": 3.2790887355804443, + "learning_rate": 4.747427502338635e-06, + "loss": 1.704, + "step": 360 + }, + { + "epoch": 0.45454545454545453, + "eval_loss": 1.848663330078125, + "eval_runtime": 37.4817, + "eval_samples_per_second": 26.68, + "eval_steps_per_second": 1.121, + "step": 360 + }, + { + "epoch": 0.45580808080808083, + "grad_norm": 3.1137545108795166, + "learning_rate": 4.745088868101029e-06, + "loss": 1.6797, + "step": 361 + }, + { + "epoch": 0.45707070707070707, + "grad_norm": 2.7598798274993896, + "learning_rate": 4.742750233863424e-06, + "loss": 1.6477, + "step": 362 + }, + { + "epoch": 0.4583333333333333, + "grad_norm": 3.4851579666137695, + "learning_rate": 4.740411599625819e-06, + "loss": 1.6481, + "step": 363 + }, + { + "epoch": 0.4595959595959596, + "grad_norm": 2.9285919666290283, + "learning_rate": 4.7380729653882135e-06, + "loss": 1.7094, + "step": 364 + }, + { + "epoch": 0.46085858585858586, + "grad_norm": 2.901458501815796, + "learning_rate": 4.735734331150608e-06, + "loss": 1.6906, + "step": 365 + }, + { + "epoch": 0.4621212121212121, + "grad_norm": 2.8737902641296387, + "learning_rate": 4.733395696913003e-06, + "loss": 1.6819, + "step": 366 + }, + { + "epoch": 0.4633838383838384, + "grad_norm": 3.2480952739715576, + "learning_rate": 4.7310570626753984e-06, + "loss": 1.6827, + "step": 367 + }, + { + "epoch": 0.46464646464646464, + "grad_norm": 2.8714661598205566, + "learning_rate": 4.728718428437793e-06, + "loss": 1.7349, + "step": 368 + }, + { + "epoch": 0.4659090909090909, + "grad_norm": 2.8410098552703857, + "learning_rate": 4.726379794200187e-06, + "loss": 1.7214, + "step": 369 + }, + { + "epoch": 0.4671717171717172, + "grad_norm": 2.744807720184326, + "learning_rate": 4.7240411599625826e-06, + "loss": 1.6742, + "step": 370 + }, + { + "epoch": 0.4671717171717172, + "eval_loss": 1.842191457748413, + "eval_runtime": 37.454, + "eval_samples_per_second": 26.699, + "eval_steps_per_second": 1.121, + "step": 370 + }, + { + "epoch": 0.4684343434343434, + "grad_norm": 2.8135552406311035, + "learning_rate": 4.721702525724977e-06, + "loss": 1.6957, + "step": 371 + }, + { + "epoch": 0.4696969696969697, + "grad_norm": 2.856764078140259, + "learning_rate": 4.719363891487372e-06, + "loss": 1.6802, + "step": 372 + }, + { + "epoch": 0.47095959595959597, + "grad_norm": 3.0707688331604004, + "learning_rate": 4.717025257249767e-06, + "loss": 1.7383, + "step": 373 + }, + { + "epoch": 0.4722222222222222, + "grad_norm": 2.8832876682281494, + "learning_rate": 4.714686623012161e-06, + "loss": 1.6309, + "step": 374 + }, + { + "epoch": 0.4734848484848485, + "grad_norm": 2.6840622425079346, + "learning_rate": 4.712347988774556e-06, + "loss": 1.6624, + "step": 375 + }, + { + "epoch": 0.47474747474747475, + "grad_norm": 3.2920548915863037, + "learning_rate": 4.710009354536951e-06, + "loss": 1.6887, + "step": 376 + }, + { + "epoch": 0.476010101010101, + "grad_norm": 3.0317189693450928, + "learning_rate": 4.707670720299345e-06, + "loss": 1.6514, + "step": 377 + }, + { + "epoch": 0.4772727272727273, + "grad_norm": 2.7284793853759766, + "learning_rate": 4.7053320860617404e-06, + "loss": 1.6595, + "step": 378 + }, + { + "epoch": 0.47853535353535354, + "grad_norm": 2.86665415763855, + "learning_rate": 4.702993451824135e-06, + "loss": 1.7219, + "step": 379 + }, + { + "epoch": 0.4797979797979798, + "grad_norm": 3.041191577911377, + "learning_rate": 4.700654817586529e-06, + "loss": 1.6959, + "step": 380 + }, + { + "epoch": 0.4797979797979798, + "eval_loss": 1.8414148092269897, + "eval_runtime": 37.2537, + "eval_samples_per_second": 26.843, + "eval_steps_per_second": 1.127, + "step": 380 + }, + { + "epoch": 0.4810606060606061, + "grad_norm": 3.1729421615600586, + "learning_rate": 4.6983161833489246e-06, + "loss": 1.6538, + "step": 381 + }, + { + "epoch": 0.4823232323232323, + "grad_norm": 2.8562991619110107, + "learning_rate": 4.69597754911132e-06, + "loss": 1.6667, + "step": 382 + }, + { + "epoch": 0.48358585858585856, + "grad_norm": 3.0348711013793945, + "learning_rate": 4.693638914873714e-06, + "loss": 1.6888, + "step": 383 + }, + { + "epoch": 0.48484848484848486, + "grad_norm": 2.9156477451324463, + "learning_rate": 4.6913002806361095e-06, + "loss": 1.628, + "step": 384 + }, + { + "epoch": 0.4861111111111111, + "grad_norm": 2.832689046859741, + "learning_rate": 4.688961646398504e-06, + "loss": 1.664, + "step": 385 + }, + { + "epoch": 0.48737373737373735, + "grad_norm": 2.9858946800231934, + "learning_rate": 4.686623012160898e-06, + "loss": 1.6696, + "step": 386 + }, + { + "epoch": 0.48863636363636365, + "grad_norm": 2.732468843460083, + "learning_rate": 4.684284377923294e-06, + "loss": 1.6444, + "step": 387 + }, + { + "epoch": 0.4898989898989899, + "grad_norm": 2.7801132202148438, + "learning_rate": 4.681945743685688e-06, + "loss": 1.7084, + "step": 388 + }, + { + "epoch": 0.4911616161616162, + "grad_norm": 2.646303415298462, + "learning_rate": 4.6796071094480824e-06, + "loss": 1.6963, + "step": 389 + }, + { + "epoch": 0.49242424242424243, + "grad_norm": 2.8750250339508057, + "learning_rate": 4.677268475210478e-06, + "loss": 1.6814, + "step": 390 + }, + { + "epoch": 0.49242424242424243, + "eval_loss": 1.8283894062042236, + "eval_runtime": 37.2058, + "eval_samples_per_second": 26.877, + "eval_steps_per_second": 1.129, + "step": 390 + }, + { + "epoch": 0.4936868686868687, + "grad_norm": 2.5373270511627197, + "learning_rate": 4.674929840972872e-06, + "loss": 1.5956, + "step": 391 + }, + { + "epoch": 0.494949494949495, + "grad_norm": 2.837170124053955, + "learning_rate": 4.6725912067352666e-06, + "loss": 1.6237, + "step": 392 + }, + { + "epoch": 0.4962121212121212, + "grad_norm": 3.199463367462158, + "learning_rate": 4.670252572497662e-06, + "loss": 1.7087, + "step": 393 + }, + { + "epoch": 0.49747474747474746, + "grad_norm": 2.9850947856903076, + "learning_rate": 4.667913938260056e-06, + "loss": 1.5913, + "step": 394 + }, + { + "epoch": 0.49873737373737376, + "grad_norm": 2.954148292541504, + "learning_rate": 4.665575304022451e-06, + "loss": 1.6272, + "step": 395 + }, + { + "epoch": 0.5, + "grad_norm": 2.8625757694244385, + "learning_rate": 4.663236669784846e-06, + "loss": 1.6623, + "step": 396 + }, + { + "epoch": 0.5012626262626263, + "grad_norm": 3.196354627609253, + "learning_rate": 4.66089803554724e-06, + "loss": 1.7249, + "step": 397 + }, + { + "epoch": 0.5025252525252525, + "grad_norm": 2.7257323265075684, + "learning_rate": 4.658559401309636e-06, + "loss": 1.665, + "step": 398 + }, + { + "epoch": 0.5037878787878788, + "grad_norm": 3.0976016521453857, + "learning_rate": 4.656220767072031e-06, + "loss": 1.6763, + "step": 399 + }, + { + "epoch": 0.5050505050505051, + "grad_norm": 3.0398128032684326, + "learning_rate": 4.653882132834425e-06, + "loss": 1.6375, + "step": 400 + }, + { + "epoch": 0.5050505050505051, + "eval_loss": 1.8324880599975586, + "eval_runtime": 37.1754, + "eval_samples_per_second": 26.9, + "eval_steps_per_second": 1.13, + "step": 400 + }, + { + "epoch": 0.5063131313131313, + "grad_norm": 3.314958095550537, + "learning_rate": 4.65154349859682e-06, + "loss": 1.6699, + "step": 401 + }, + { + "epoch": 0.5075757575757576, + "grad_norm": 2.8478870391845703, + "learning_rate": 4.649204864359215e-06, + "loss": 1.6495, + "step": 402 + }, + { + "epoch": 0.5088383838383839, + "grad_norm": 2.8476510047912598, + "learning_rate": 4.646866230121609e-06, + "loss": 1.633, + "step": 403 + }, + { + "epoch": 0.51010101010101, + "grad_norm": 2.916667938232422, + "learning_rate": 4.644527595884004e-06, + "loss": 1.6092, + "step": 404 + }, + { + "epoch": 0.5113636363636364, + "grad_norm": 2.977226495742798, + "learning_rate": 4.642188961646399e-06, + "loss": 1.6588, + "step": 405 + }, + { + "epoch": 0.5126262626262627, + "grad_norm": 3.0540225505828857, + "learning_rate": 4.6398503274087935e-06, + "loss": 1.6529, + "step": 406 + }, + { + "epoch": 0.5138888888888888, + "grad_norm": 2.902035713195801, + "learning_rate": 4.637511693171188e-06, + "loss": 1.5983, + "step": 407 + }, + { + "epoch": 0.5151515151515151, + "grad_norm": 2.9009487628936768, + "learning_rate": 4.635173058933583e-06, + "loss": 1.6389, + "step": 408 + }, + { + "epoch": 0.5164141414141414, + "grad_norm": 2.904982805252075, + "learning_rate": 4.632834424695978e-06, + "loss": 1.6971, + "step": 409 + }, + { + "epoch": 0.5176767676767676, + "grad_norm": 2.8340721130371094, + "learning_rate": 4.630495790458373e-06, + "loss": 1.6336, + "step": 410 + }, + { + "epoch": 0.5176767676767676, + "eval_loss": 1.8218064308166504, + "eval_runtime": 36.9963, + "eval_samples_per_second": 27.03, + "eval_steps_per_second": 1.135, + "step": 410 + }, + { + "epoch": 0.5189393939393939, + "grad_norm": 2.9507815837860107, + "learning_rate": 4.628157156220767e-06, + "loss": 1.6795, + "step": 411 + }, + { + "epoch": 0.5202020202020202, + "grad_norm": 2.868793249130249, + "learning_rate": 4.625818521983162e-06, + "loss": 1.6494, + "step": 412 + }, + { + "epoch": 0.5214646464646465, + "grad_norm": 2.786395788192749, + "learning_rate": 4.623479887745557e-06, + "loss": 1.7156, + "step": 413 + }, + { + "epoch": 0.5227272727272727, + "grad_norm": 2.7456934452056885, + "learning_rate": 4.621141253507952e-06, + "loss": 1.6528, + "step": 414 + }, + { + "epoch": 0.523989898989899, + "grad_norm": 3.1211695671081543, + "learning_rate": 4.618802619270347e-06, + "loss": 1.6891, + "step": 415 + }, + { + "epoch": 0.5252525252525253, + "grad_norm": 3.2319254875183105, + "learning_rate": 4.616463985032741e-06, + "loss": 1.6452, + "step": 416 + }, + { + "epoch": 0.5265151515151515, + "grad_norm": 2.6718053817749023, + "learning_rate": 4.614125350795136e-06, + "loss": 1.6442, + "step": 417 + }, + { + "epoch": 0.5277777777777778, + "grad_norm": 2.985020637512207, + "learning_rate": 4.611786716557531e-06, + "loss": 1.6935, + "step": 418 + }, + { + "epoch": 0.5290404040404041, + "grad_norm": 2.798720598220825, + "learning_rate": 4.609448082319925e-06, + "loss": 1.6303, + "step": 419 + }, + { + "epoch": 0.5303030303030303, + "grad_norm": 2.860403299331665, + "learning_rate": 4.6071094480823205e-06, + "loss": 1.6445, + "step": 420 + }, + { + "epoch": 0.5303030303030303, + "eval_loss": 1.8173401355743408, + "eval_runtime": 37.1334, + "eval_samples_per_second": 26.93, + "eval_steps_per_second": 1.131, + "step": 420 + }, + { + "epoch": 0.5315656565656566, + "grad_norm": 2.883209705352783, + "learning_rate": 4.604770813844715e-06, + "loss": 1.6182, + "step": 421 + }, + { + "epoch": 0.5328282828282829, + "grad_norm": 3.0330183506011963, + "learning_rate": 4.60243217960711e-06, + "loss": 1.6208, + "step": 422 + }, + { + "epoch": 0.5340909090909091, + "grad_norm": 3.224184274673462, + "learning_rate": 4.6000935453695046e-06, + "loss": 1.6687, + "step": 423 + }, + { + "epoch": 0.5353535353535354, + "grad_norm": 3.0345349311828613, + "learning_rate": 4.597754911131899e-06, + "loss": 1.6654, + "step": 424 + }, + { + "epoch": 0.5366161616161617, + "grad_norm": 3.1425750255584717, + "learning_rate": 4.595416276894294e-06, + "loss": 1.6569, + "step": 425 + }, + { + "epoch": 0.5378787878787878, + "grad_norm": 2.6387979984283447, + "learning_rate": 4.593077642656689e-06, + "loss": 1.6868, + "step": 426 + }, + { + "epoch": 0.5391414141414141, + "grad_norm": 2.959446430206299, + "learning_rate": 4.590739008419083e-06, + "loss": 1.621, + "step": 427 + }, + { + "epoch": 0.5404040404040404, + "grad_norm": 3.009673833847046, + "learning_rate": 4.588400374181478e-06, + "loss": 1.6414, + "step": 428 + }, + { + "epoch": 0.5416666666666666, + "grad_norm": 2.827531576156616, + "learning_rate": 4.586061739943874e-06, + "loss": 1.6561, + "step": 429 + }, + { + "epoch": 0.5429292929292929, + "grad_norm": 2.8770735263824463, + "learning_rate": 4.583723105706268e-06, + "loss": 1.6685, + "step": 430 + }, + { + "epoch": 0.5429292929292929, + "eval_loss": 1.8234716653823853, + "eval_runtime": 37.1647, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 430 + }, + { + "epoch": 0.5441919191919192, + "grad_norm": 3.1215782165527344, + "learning_rate": 4.5813844714686625e-06, + "loss": 1.6548, + "step": 431 + }, + { + "epoch": 0.5454545454545454, + "grad_norm": 2.6587517261505127, + "learning_rate": 4.579045837231058e-06, + "loss": 1.649, + "step": 432 + }, + { + "epoch": 0.5467171717171717, + "grad_norm": 2.7059030532836914, + "learning_rate": 4.576707202993452e-06, + "loss": 1.6293, + "step": 433 + }, + { + "epoch": 0.547979797979798, + "grad_norm": 2.7390925884246826, + "learning_rate": 4.574368568755847e-06, + "loss": 1.6957, + "step": 434 + }, + { + "epoch": 0.5492424242424242, + "grad_norm": 2.9457757472991943, + "learning_rate": 4.572029934518242e-06, + "loss": 1.6714, + "step": 435 + }, + { + "epoch": 0.5505050505050505, + "grad_norm": 2.6871347427368164, + "learning_rate": 4.569691300280636e-06, + "loss": 1.6648, + "step": 436 + }, + { + "epoch": 0.5517676767676768, + "grad_norm": 3.1708056926727295, + "learning_rate": 4.5673526660430315e-06, + "loss": 1.6415, + "step": 437 + }, + { + "epoch": 0.553030303030303, + "grad_norm": 2.8767378330230713, + "learning_rate": 4.565014031805426e-06, + "loss": 1.6475, + "step": 438 + }, + { + "epoch": 0.5542929292929293, + "grad_norm": 3.083566188812256, + "learning_rate": 4.56267539756782e-06, + "loss": 1.632, + "step": 439 + }, + { + "epoch": 0.5555555555555556, + "grad_norm": 2.584460735321045, + "learning_rate": 4.560336763330216e-06, + "loss": 1.6469, + "step": 440 + }, + { + "epoch": 0.5555555555555556, + "eval_loss": 1.8140095472335815, + "eval_runtime": 37.1986, + "eval_samples_per_second": 26.883, + "eval_steps_per_second": 1.129, + "step": 440 + }, + { + "epoch": 0.5568181818181818, + "grad_norm": 2.5885088443756104, + "learning_rate": 4.55799812909261e-06, + "loss": 1.6058, + "step": 441 + }, + { + "epoch": 0.5580808080808081, + "grad_norm": 2.8467836380004883, + "learning_rate": 4.5556594948550045e-06, + "loss": 1.6959, + "step": 442 + }, + { + "epoch": 0.5593434343434344, + "grad_norm": 2.764326333999634, + "learning_rate": 4.5533208606174e-06, + "loss": 1.6565, + "step": 443 + }, + { + "epoch": 0.5606060606060606, + "grad_norm": 2.88089919090271, + "learning_rate": 4.550982226379795e-06, + "loss": 1.6651, + "step": 444 + }, + { + "epoch": 0.5618686868686869, + "grad_norm": 2.6239469051361084, + "learning_rate": 4.548643592142189e-06, + "loss": 1.6255, + "step": 445 + }, + { + "epoch": 0.5631313131313131, + "grad_norm": 2.7748732566833496, + "learning_rate": 4.546304957904585e-06, + "loss": 1.6587, + "step": 446 + }, + { + "epoch": 0.5643939393939394, + "grad_norm": 2.7075202465057373, + "learning_rate": 4.543966323666979e-06, + "loss": 1.6448, + "step": 447 + }, + { + "epoch": 0.5656565656565656, + "grad_norm": 2.9291634559631348, + "learning_rate": 4.5416276894293735e-06, + "loss": 1.6564, + "step": 448 + }, + { + "epoch": 0.5669191919191919, + "grad_norm": 3.0039470195770264, + "learning_rate": 4.539289055191769e-06, + "loss": 1.623, + "step": 449 + }, + { + "epoch": 0.5681818181818182, + "grad_norm": 3.085277557373047, + "learning_rate": 4.536950420954163e-06, + "loss": 1.6176, + "step": 450 + }, + { + "epoch": 0.5681818181818182, + "eval_loss": 1.8022958040237427, + "eval_runtime": 37.2204, + "eval_samples_per_second": 26.867, + "eval_steps_per_second": 1.128, + "step": 450 + }, + { + "epoch": 0.5694444444444444, + "grad_norm": 2.93575119972229, + "learning_rate": 4.534611786716558e-06, + "loss": 1.615, + "step": 451 + }, + { + "epoch": 0.5707070707070707, + "grad_norm": 2.847141742706299, + "learning_rate": 4.532273152478953e-06, + "loss": 1.6474, + "step": 452 + }, + { + "epoch": 0.571969696969697, + "grad_norm": 3.0299603939056396, + "learning_rate": 4.529934518241347e-06, + "loss": 1.6424, + "step": 453 + }, + { + "epoch": 0.5732323232323232, + "grad_norm": 3.0307669639587402, + "learning_rate": 4.527595884003742e-06, + "loss": 1.6473, + "step": 454 + }, + { + "epoch": 0.5744949494949495, + "grad_norm": 2.864269971847534, + "learning_rate": 4.525257249766137e-06, + "loss": 1.6631, + "step": 455 + }, + { + "epoch": 0.5757575757575758, + "grad_norm": 2.606637716293335, + "learning_rate": 4.522918615528531e-06, + "loss": 1.6118, + "step": 456 + }, + { + "epoch": 0.577020202020202, + "grad_norm": 3.127891778945923, + "learning_rate": 4.520579981290926e-06, + "loss": 1.6286, + "step": 457 + }, + { + "epoch": 0.5782828282828283, + "grad_norm": 2.8809893131256104, + "learning_rate": 4.518241347053321e-06, + "loss": 1.6158, + "step": 458 + }, + { + "epoch": 0.5795454545454546, + "grad_norm": 3.0662436485290527, + "learning_rate": 4.515902712815716e-06, + "loss": 1.7336, + "step": 459 + }, + { + "epoch": 0.5808080808080808, + "grad_norm": 3.0795772075653076, + "learning_rate": 4.513564078578111e-06, + "loss": 1.6974, + "step": 460 + }, + { + "epoch": 0.5808080808080808, + "eval_loss": 1.7981404066085815, + "eval_runtime": 37.299, + "eval_samples_per_second": 26.81, + "eval_steps_per_second": 1.126, + "step": 460 + }, + { + "epoch": 0.5820707070707071, + "grad_norm": 3.091688871383667, + "learning_rate": 4.511225444340506e-06, + "loss": 1.6803, + "step": 461 + }, + { + "epoch": 0.5833333333333334, + "grad_norm": 2.5597071647644043, + "learning_rate": 4.5088868101029005e-06, + "loss": 1.6104, + "step": 462 + }, + { + "epoch": 0.5845959595959596, + "grad_norm": 2.849806070327759, + "learning_rate": 4.506548175865295e-06, + "loss": 1.6147, + "step": 463 + }, + { + "epoch": 0.5858585858585859, + "grad_norm": 2.9791204929351807, + "learning_rate": 4.50420954162769e-06, + "loss": 1.6376, + "step": 464 + }, + { + "epoch": 0.5871212121212122, + "grad_norm": 2.8582305908203125, + "learning_rate": 4.501870907390085e-06, + "loss": 1.6831, + "step": 465 + }, + { + "epoch": 0.5883838383838383, + "grad_norm": 2.8445184230804443, + "learning_rate": 4.499532273152479e-06, + "loss": 1.6463, + "step": 466 + }, + { + "epoch": 0.5896464646464646, + "grad_norm": 2.858168125152588, + "learning_rate": 4.497193638914874e-06, + "loss": 1.6662, + "step": 467 + }, + { + "epoch": 0.5909090909090909, + "grad_norm": 2.829702854156494, + "learning_rate": 4.494855004677269e-06, + "loss": 1.6058, + "step": 468 + }, + { + "epoch": 0.5921717171717171, + "grad_norm": 2.9292726516723633, + "learning_rate": 4.492516370439663e-06, + "loss": 1.6341, + "step": 469 + }, + { + "epoch": 0.5934343434343434, + "grad_norm": 2.9006292819976807, + "learning_rate": 4.490177736202058e-06, + "loss": 1.5726, + "step": 470 + }, + { + "epoch": 0.5934343434343434, + "eval_loss": 1.7953916788101196, + "eval_runtime": 37.5058, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 470 + }, + { + "epoch": 0.5946969696969697, + "grad_norm": 2.8116230964660645, + "learning_rate": 4.487839101964453e-06, + "loss": 1.6083, + "step": 471 + }, + { + "epoch": 0.5959595959595959, + "grad_norm": 2.540607452392578, + "learning_rate": 4.485500467726848e-06, + "loss": 1.6208, + "step": 472 + }, + { + "epoch": 0.5972222222222222, + "grad_norm": 2.7339096069335938, + "learning_rate": 4.4831618334892425e-06, + "loss": 1.6839, + "step": 473 + }, + { + "epoch": 0.5984848484848485, + "grad_norm": 2.9094762802124023, + "learning_rate": 4.480823199251638e-06, + "loss": 1.6962, + "step": 474 + }, + { + "epoch": 0.5997474747474747, + "grad_norm": 2.7733027935028076, + "learning_rate": 4.478484565014032e-06, + "loss": 1.6252, + "step": 475 + }, + { + "epoch": 0.601010101010101, + "grad_norm": 2.5709428787231445, + "learning_rate": 4.4761459307764274e-06, + "loss": 1.5896, + "step": 476 + }, + { + "epoch": 0.6022727272727273, + "grad_norm": 2.8358278274536133, + "learning_rate": 4.473807296538822e-06, + "loss": 1.6247, + "step": 477 + }, + { + "epoch": 0.6035353535353535, + "grad_norm": 2.7504897117614746, + "learning_rate": 4.471468662301216e-06, + "loss": 1.7138, + "step": 478 + }, + { + "epoch": 0.6047979797979798, + "grad_norm": 2.6965620517730713, + "learning_rate": 4.4691300280636115e-06, + "loss": 1.6054, + "step": 479 + }, + { + "epoch": 0.6060606060606061, + "grad_norm": 2.9401891231536865, + "learning_rate": 4.466791393826006e-06, + "loss": 1.6172, + "step": 480 + }, + { + "epoch": 0.6060606060606061, + "eval_loss": 1.7946635484695435, + "eval_runtime": 37.5366, + "eval_samples_per_second": 26.641, + "eval_steps_per_second": 1.119, + "step": 480 + }, + { + "epoch": 0.6073232323232324, + "grad_norm": 2.71535587310791, + "learning_rate": 4.4644527595884e-06, + "loss": 1.61, + "step": 481 + }, + { + "epoch": 0.6085858585858586, + "grad_norm": 2.7910654544830322, + "learning_rate": 4.462114125350796e-06, + "loss": 1.6193, + "step": 482 + }, + { + "epoch": 0.6098484848484849, + "grad_norm": 2.809241533279419, + "learning_rate": 4.45977549111319e-06, + "loss": 1.602, + "step": 483 + }, + { + "epoch": 0.6111111111111112, + "grad_norm": 2.7152137756347656, + "learning_rate": 4.457436856875585e-06, + "loss": 1.6516, + "step": 484 + }, + { + "epoch": 0.6123737373737373, + "grad_norm": 2.7653465270996094, + "learning_rate": 4.45509822263798e-06, + "loss": 1.5904, + "step": 485 + }, + { + "epoch": 0.6136363636363636, + "grad_norm": 2.729876756668091, + "learning_rate": 4.452759588400374e-06, + "loss": 1.6259, + "step": 486 + }, + { + "epoch": 0.61489898989899, + "grad_norm": 2.6290180683135986, + "learning_rate": 4.4504209541627694e-06, + "loss": 1.5796, + "step": 487 + }, + { + "epoch": 0.6161616161616161, + "grad_norm": 2.939884662628174, + "learning_rate": 4.448082319925164e-06, + "loss": 1.6774, + "step": 488 + }, + { + "epoch": 0.6174242424242424, + "grad_norm": 2.7554702758789062, + "learning_rate": 4.445743685687559e-06, + "loss": 1.6226, + "step": 489 + }, + { + "epoch": 0.6186868686868687, + "grad_norm": 2.890458106994629, + "learning_rate": 4.4434050514499535e-06, + "loss": 1.592, + "step": 490 + }, + { + "epoch": 0.6186868686868687, + "eval_loss": 1.7868437767028809, + "eval_runtime": 37.5267, + "eval_samples_per_second": 26.648, + "eval_steps_per_second": 1.119, + "step": 490 + }, + { + "epoch": 0.6199494949494949, + "grad_norm": 2.757178783416748, + "learning_rate": 4.441066417212349e-06, + "loss": 1.6297, + "step": 491 + }, + { + "epoch": 0.6212121212121212, + "grad_norm": 2.7179954051971436, + "learning_rate": 4.438727782974743e-06, + "loss": 1.5685, + "step": 492 + }, + { + "epoch": 0.6224747474747475, + "grad_norm": 2.824514389038086, + "learning_rate": 4.436389148737138e-06, + "loss": 1.6625, + "step": 493 + }, + { + "epoch": 0.6237373737373737, + "grad_norm": 3.0062615871429443, + "learning_rate": 4.434050514499533e-06, + "loss": 1.5889, + "step": 494 + }, + { + "epoch": 0.625, + "grad_norm": 2.962176561355591, + "learning_rate": 4.431711880261927e-06, + "loss": 1.6813, + "step": 495 + }, + { + "epoch": 0.6262626262626263, + "grad_norm": 2.7531204223632812, + "learning_rate": 4.429373246024323e-06, + "loss": 1.6152, + "step": 496 + }, + { + "epoch": 0.6275252525252525, + "grad_norm": 2.764021873474121, + "learning_rate": 4.427034611786717e-06, + "loss": 1.6164, + "step": 497 + }, + { + "epoch": 0.6287878787878788, + "grad_norm": 2.836826801300049, + "learning_rate": 4.4246959775491114e-06, + "loss": 1.6192, + "step": 498 + }, + { + "epoch": 0.6300505050505051, + "grad_norm": 2.871220827102661, + "learning_rate": 4.422357343311507e-06, + "loss": 1.6096, + "step": 499 + }, + { + "epoch": 0.6313131313131313, + "grad_norm": 2.8129167556762695, + "learning_rate": 4.420018709073901e-06, + "loss": 1.6674, + "step": 500 + }, + { + "epoch": 0.6313131313131313, + "eval_loss": 1.785156011581421, + "eval_runtime": 37.2971, + "eval_samples_per_second": 26.812, + "eval_steps_per_second": 1.126, + "step": 500 + }, + { + "epoch": 0.6325757575757576, + "grad_norm": 2.775108814239502, + "learning_rate": 4.4176800748362955e-06, + "loss": 1.5851, + "step": 501 + }, + { + "epoch": 0.6338383838383839, + "grad_norm": 2.7775518894195557, + "learning_rate": 4.415341440598691e-06, + "loss": 1.6839, + "step": 502 + }, + { + "epoch": 0.63510101010101, + "grad_norm": 2.7997918128967285, + "learning_rate": 4.413002806361085e-06, + "loss": 1.5849, + "step": 503 + }, + { + "epoch": 0.6363636363636364, + "grad_norm": 2.91103196144104, + "learning_rate": 4.4106641721234805e-06, + "loss": 1.6369, + "step": 504 + }, + { + "epoch": 0.6376262626262627, + "grad_norm": 2.6800882816314697, + "learning_rate": 4.408325537885875e-06, + "loss": 1.6049, + "step": 505 + }, + { + "epoch": 0.6388888888888888, + "grad_norm": 2.7451021671295166, + "learning_rate": 4.40598690364827e-06, + "loss": 1.658, + "step": 506 + }, + { + "epoch": 0.6401515151515151, + "grad_norm": 2.794813632965088, + "learning_rate": 4.403648269410665e-06, + "loss": 1.6015, + "step": 507 + }, + { + "epoch": 0.6414141414141414, + "grad_norm": 2.769373655319214, + "learning_rate": 4.401309635173059e-06, + "loss": 1.6217, + "step": 508 + }, + { + "epoch": 0.6426767676767676, + "grad_norm": 2.9081523418426514, + "learning_rate": 4.398971000935454e-06, + "loss": 1.6235, + "step": 509 + }, + { + "epoch": 0.6439393939393939, + "grad_norm": 2.707618236541748, + "learning_rate": 4.396632366697849e-06, + "loss": 1.6003, + "step": 510 + }, + { + "epoch": 0.6439393939393939, + "eval_loss": 1.7876425981521606, + "eval_runtime": 36.9832, + "eval_samples_per_second": 27.039, + "eval_steps_per_second": 1.136, + "step": 510 + }, + { + "epoch": 0.6452020202020202, + "grad_norm": 2.7923099994659424, + "learning_rate": 4.394293732460244e-06, + "loss": 1.5647, + "step": 511 + }, + { + "epoch": 0.6464646464646465, + "grad_norm": 2.8449175357818604, + "learning_rate": 4.391955098222638e-06, + "loss": 1.6456, + "step": 512 + }, + { + "epoch": 0.6477272727272727, + "grad_norm": 2.755856990814209, + "learning_rate": 4.389616463985033e-06, + "loss": 1.6962, + "step": 513 + }, + { + "epoch": 0.648989898989899, + "grad_norm": 2.5578861236572266, + "learning_rate": 4.387277829747428e-06, + "loss": 1.6097, + "step": 514 + }, + { + "epoch": 0.6502525252525253, + "grad_norm": 2.8804502487182617, + "learning_rate": 4.3849391955098225e-06, + "loss": 1.6212, + "step": 515 + }, + { + "epoch": 0.6515151515151515, + "grad_norm": 2.684215545654297, + "learning_rate": 4.382600561272217e-06, + "loss": 1.5529, + "step": 516 + }, + { + "epoch": 0.6527777777777778, + "grad_norm": 2.9128026962280273, + "learning_rate": 4.380261927034612e-06, + "loss": 1.6424, + "step": 517 + }, + { + "epoch": 0.6540404040404041, + "grad_norm": 2.733415126800537, + "learning_rate": 4.377923292797007e-06, + "loss": 1.6201, + "step": 518 + }, + { + "epoch": 0.6553030303030303, + "grad_norm": 2.7211062908172607, + "learning_rate": 4.375584658559401e-06, + "loss": 1.5833, + "step": 519 + }, + { + "epoch": 0.6565656565656566, + "grad_norm": 2.7308316230773926, + "learning_rate": 4.373246024321796e-06, + "loss": 1.6288, + "step": 520 + }, + { + "epoch": 0.6565656565656566, + "eval_loss": 1.7787725925445557, + "eval_runtime": 37.1089, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 520 + }, + { + "epoch": 0.6578282828282829, + "grad_norm": 2.760362148284912, + "learning_rate": 4.3709073900841915e-06, + "loss": 1.5814, + "step": 521 + }, + { + "epoch": 0.6590909090909091, + "grad_norm": 2.7625153064727783, + "learning_rate": 4.368568755846586e-06, + "loss": 1.6239, + "step": 522 + }, + { + "epoch": 0.6603535353535354, + "grad_norm": 2.715710401535034, + "learning_rate": 4.366230121608981e-06, + "loss": 1.6069, + "step": 523 + }, + { + "epoch": 0.6616161616161617, + "grad_norm": 2.7753846645355225, + "learning_rate": 4.363891487371376e-06, + "loss": 1.6178, + "step": 524 + }, + { + "epoch": 0.6628787878787878, + "grad_norm": 2.650716543197632, + "learning_rate": 4.36155285313377e-06, + "loss": 1.6534, + "step": 525 + }, + { + "epoch": 0.6641414141414141, + "grad_norm": 2.573969602584839, + "learning_rate": 4.359214218896165e-06, + "loss": 1.6721, + "step": 526 + }, + { + "epoch": 0.6654040404040404, + "grad_norm": 2.9017715454101562, + "learning_rate": 4.35687558465856e-06, + "loss": 1.6273, + "step": 527 + }, + { + "epoch": 0.6666666666666666, + "grad_norm": 2.691643714904785, + "learning_rate": 4.354536950420954e-06, + "loss": 1.6199, + "step": 528 + }, + { + "epoch": 0.6679292929292929, + "grad_norm": 2.818662166595459, + "learning_rate": 4.3521983161833494e-06, + "loss": 1.5758, + "step": 529 + }, + { + "epoch": 0.6691919191919192, + "grad_norm": 3.017738103866577, + "learning_rate": 4.349859681945744e-06, + "loss": 1.6437, + "step": 530 + }, + { + "epoch": 0.6691919191919192, + "eval_loss": 1.7835458517074585, + "eval_runtime": 37.1359, + "eval_samples_per_second": 26.928, + "eval_steps_per_second": 1.131, + "step": 530 + }, + { + "epoch": 0.6704545454545454, + "grad_norm": 2.8027284145355225, + "learning_rate": 4.347521047708138e-06, + "loss": 1.6688, + "step": 531 + }, + { + "epoch": 0.6717171717171717, + "grad_norm": 2.6642649173736572, + "learning_rate": 4.3451824134705335e-06, + "loss": 1.6327, + "step": 532 + }, + { + "epoch": 0.672979797979798, + "grad_norm": 2.7406396865844727, + "learning_rate": 4.342843779232928e-06, + "loss": 1.6297, + "step": 533 + }, + { + "epoch": 0.6742424242424242, + "grad_norm": 2.758538007736206, + "learning_rate": 4.340505144995323e-06, + "loss": 1.5276, + "step": 534 + }, + { + "epoch": 0.6755050505050505, + "grad_norm": 2.815817356109619, + "learning_rate": 4.3381665107577185e-06, + "loss": 1.6192, + "step": 535 + }, + { + "epoch": 0.6767676767676768, + "grad_norm": 2.8416755199432373, + "learning_rate": 4.335827876520113e-06, + "loss": 1.5966, + "step": 536 + }, + { + "epoch": 0.678030303030303, + "grad_norm": 3.014401435852051, + "learning_rate": 4.333489242282507e-06, + "loss": 1.6104, + "step": 537 + }, + { + "epoch": 0.6792929292929293, + "grad_norm": 2.731356143951416, + "learning_rate": 4.331150608044903e-06, + "loss": 1.6177, + "step": 538 + }, + { + "epoch": 0.6805555555555556, + "grad_norm": 2.633310317993164, + "learning_rate": 4.328811973807297e-06, + "loss": 1.5004, + "step": 539 + }, + { + "epoch": 0.6818181818181818, + "grad_norm": 2.977476119995117, + "learning_rate": 4.3264733395696914e-06, + "loss": 1.6171, + "step": 540 + }, + { + "epoch": 0.6818181818181818, + "eval_loss": 1.7790141105651855, + "eval_runtime": 37.12, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 540 + }, + { + "epoch": 0.6830808080808081, + "grad_norm": 2.769422769546509, + "learning_rate": 4.324134705332087e-06, + "loss": 1.5591, + "step": 541 + }, + { + "epoch": 0.6843434343434344, + "grad_norm": 3.088426351547241, + "learning_rate": 4.321796071094481e-06, + "loss": 1.5603, + "step": 542 + }, + { + "epoch": 0.6856060606060606, + "grad_norm": 2.6784212589263916, + "learning_rate": 4.3194574368568755e-06, + "loss": 1.6061, + "step": 543 + }, + { + "epoch": 0.6868686868686869, + "grad_norm": 3.209821939468384, + "learning_rate": 4.317118802619271e-06, + "loss": 1.585, + "step": 544 + }, + { + "epoch": 0.6881313131313131, + "grad_norm": 2.940791368484497, + "learning_rate": 4.314780168381665e-06, + "loss": 1.6092, + "step": 545 + }, + { + "epoch": 0.6893939393939394, + "grad_norm": 2.679884672164917, + "learning_rate": 4.3124415341440605e-06, + "loss": 1.6207, + "step": 546 + }, + { + "epoch": 0.6906565656565656, + "grad_norm": 2.853809118270874, + "learning_rate": 4.310102899906455e-06, + "loss": 1.5674, + "step": 547 + }, + { + "epoch": 0.6919191919191919, + "grad_norm": 2.693838119506836, + "learning_rate": 4.307764265668849e-06, + "loss": 1.5886, + "step": 548 + }, + { + "epoch": 0.6931818181818182, + "grad_norm": 2.871680736541748, + "learning_rate": 4.305425631431245e-06, + "loss": 1.6309, + "step": 549 + }, + { + "epoch": 0.6944444444444444, + "grad_norm": 2.9908292293548584, + "learning_rate": 4.30308699719364e-06, + "loss": 1.5577, + "step": 550 + }, + { + "epoch": 0.6944444444444444, + "eval_loss": 1.7790372371673584, + "eval_runtime": 37.114, + "eval_samples_per_second": 26.944, + "eval_steps_per_second": 1.132, + "step": 550 + }, + { + "epoch": 0.6957070707070707, + "grad_norm": 2.883481740951538, + "learning_rate": 4.300748362956034e-06, + "loss": 1.5749, + "step": 551 + }, + { + "epoch": 0.696969696969697, + "grad_norm": 2.7523458003997803, + "learning_rate": 4.298409728718429e-06, + "loss": 1.624, + "step": 552 + }, + { + "epoch": 0.6982323232323232, + "grad_norm": 2.69509220123291, + "learning_rate": 4.296071094480824e-06, + "loss": 1.5856, + "step": 553 + }, + { + "epoch": 0.6994949494949495, + "grad_norm": 2.6340222358703613, + "learning_rate": 4.293732460243218e-06, + "loss": 1.5857, + "step": 554 + }, + { + "epoch": 0.7007575757575758, + "grad_norm": 2.4998607635498047, + "learning_rate": 4.291393826005613e-06, + "loss": 1.6038, + "step": 555 + }, + { + "epoch": 0.702020202020202, + "grad_norm": 2.7913858890533447, + "learning_rate": 4.289055191768008e-06, + "loss": 1.6411, + "step": 556 + }, + { + "epoch": 0.7032828282828283, + "grad_norm": 2.7149479389190674, + "learning_rate": 4.2867165575304025e-06, + "loss": 1.592, + "step": 557 + }, + { + "epoch": 0.7045454545454546, + "grad_norm": 2.7187132835388184, + "learning_rate": 4.284377923292797e-06, + "loss": 1.5413, + "step": 558 + }, + { + "epoch": 0.7058080808080808, + "grad_norm": 2.8913674354553223, + "learning_rate": 4.282039289055192e-06, + "loss": 1.6105, + "step": 559 + }, + { + "epoch": 0.7070707070707071, + "grad_norm": 2.8470897674560547, + "learning_rate": 4.279700654817587e-06, + "loss": 1.6481, + "step": 560 + }, + { + "epoch": 0.7070707070707071, + "eval_loss": 1.7711386680603027, + "eval_runtime": 37.1243, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 560 + }, + { + "epoch": 0.7083333333333334, + "grad_norm": 2.7525687217712402, + "learning_rate": 4.277362020579982e-06, + "loss": 1.5749, + "step": 561 + }, + { + "epoch": 0.7095959595959596, + "grad_norm": 2.6935207843780518, + "learning_rate": 4.275023386342376e-06, + "loss": 1.637, + "step": 562 + }, + { + "epoch": 0.7108585858585859, + "grad_norm": 2.6935207843780518, + "learning_rate": 4.275023386342376e-06, + "loss": 1.5903, + "step": 563 + }, + { + "epoch": 0.7121212121212122, + "grad_norm": 2.992992401123047, + "learning_rate": 4.272684752104771e-06, + "loss": 1.6654, + "step": 564 + }, + { + "epoch": 0.7133838383838383, + "grad_norm": 2.7314348220825195, + "learning_rate": 4.270346117867166e-06, + "loss": 1.6775, + "step": 565 + }, + { + "epoch": 0.7146464646464646, + "grad_norm": 2.6952013969421387, + "learning_rate": 4.26800748362956e-06, + "loss": 1.5936, + "step": 566 + }, + { + "epoch": 0.7159090909090909, + "grad_norm": 2.8244829177856445, + "learning_rate": 4.265668849391956e-06, + "loss": 1.6228, + "step": 567 + }, + { + "epoch": 0.7171717171717171, + "grad_norm": 2.9110143184661865, + "learning_rate": 4.26333021515435e-06, + "loss": 1.6016, + "step": 568 + }, + { + "epoch": 0.7184343434343434, + "grad_norm": 2.7953436374664307, + "learning_rate": 4.260991580916745e-06, + "loss": 1.611, + "step": 569 + }, + { + "epoch": 0.7196969696969697, + "grad_norm": 2.854447364807129, + "learning_rate": 4.25865294667914e-06, + "loss": 1.6365, + "step": 570 + }, + { + "epoch": 0.7196969696969697, + "eval_loss": 1.7711069583892822, + "eval_runtime": 37.1374, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 570 + }, + { + "epoch": 0.7209595959595959, + "grad_norm": 2.7885239124298096, + "learning_rate": 4.256314312441534e-06, + "loss": 1.6106, + "step": 571 + }, + { + "epoch": 0.7222222222222222, + "grad_norm": 2.8012993335723877, + "learning_rate": 4.2539756782039295e-06, + "loss": 1.6081, + "step": 572 + }, + { + "epoch": 0.7234848484848485, + "grad_norm": 2.9223906993865967, + "learning_rate": 4.251637043966324e-06, + "loss": 1.5836, + "step": 573 + }, + { + "epoch": 0.7247474747474747, + "grad_norm": 2.885830879211426, + "learning_rate": 4.249298409728719e-06, + "loss": 1.5616, + "step": 574 + }, + { + "epoch": 0.726010101010101, + "grad_norm": 2.687372922897339, + "learning_rate": 4.2469597754911136e-06, + "loss": 1.5873, + "step": 575 + }, + { + "epoch": 0.7272727272727273, + "grad_norm": 3.0156168937683105, + "learning_rate": 4.244621141253508e-06, + "loss": 1.5763, + "step": 576 + }, + { + "epoch": 0.7285353535353535, + "grad_norm": 3.061842918395996, + "learning_rate": 4.242282507015903e-06, + "loss": 1.6148, + "step": 577 + }, + { + "epoch": 0.7297979797979798, + "grad_norm": 2.7467715740203857, + "learning_rate": 4.239943872778298e-06, + "loss": 1.577, + "step": 578 + }, + { + "epoch": 0.7310606060606061, + "grad_norm": 2.669501543045044, + "learning_rate": 4.237605238540692e-06, + "loss": 1.5653, + "step": 579 + }, + { + "epoch": 0.7323232323232324, + "grad_norm": 2.874349355697632, + "learning_rate": 4.235266604303087e-06, + "loss": 1.6523, + "step": 580 + }, + { + "epoch": 0.7323232323232324, + "eval_loss": 1.7680890560150146, + "eval_runtime": 37.2931, + "eval_samples_per_second": 26.815, + "eval_steps_per_second": 1.126, + "step": 580 + }, + { + "epoch": 0.7335858585858586, + "grad_norm": 2.7468504905700684, + "learning_rate": 4.232927970065482e-06, + "loss": 1.5935, + "step": 581 + }, + { + "epoch": 0.7348484848484849, + "grad_norm": 2.936697244644165, + "learning_rate": 4.230589335827877e-06, + "loss": 1.6335, + "step": 582 + }, + { + "epoch": 0.7361111111111112, + "grad_norm": 2.6878974437713623, + "learning_rate": 4.2282507015902715e-06, + "loss": 1.5699, + "step": 583 + }, + { + "epoch": 0.7373737373737373, + "grad_norm": 3.3578853607177734, + "learning_rate": 4.225912067352667e-06, + "loss": 1.5979, + "step": 584 + }, + { + "epoch": 0.7386363636363636, + "grad_norm": 2.8441176414489746, + "learning_rate": 4.223573433115061e-06, + "loss": 1.5736, + "step": 585 + }, + { + "epoch": 0.73989898989899, + "grad_norm": 2.7186129093170166, + "learning_rate": 4.221234798877456e-06, + "loss": 1.636, + "step": 586 + }, + { + "epoch": 0.7411616161616161, + "grad_norm": 2.940002918243408, + "learning_rate": 4.218896164639851e-06, + "loss": 1.5721, + "step": 587 + }, + { + "epoch": 0.7424242424242424, + "grad_norm": 2.676284074783325, + "learning_rate": 4.216557530402245e-06, + "loss": 1.5972, + "step": 588 + }, + { + "epoch": 0.7436868686868687, + "grad_norm": 2.9176571369171143, + "learning_rate": 4.2142188961646405e-06, + "loss": 1.6045, + "step": 589 + }, + { + "epoch": 0.7449494949494949, + "grad_norm": 2.7716288566589355, + "learning_rate": 4.211880261927035e-06, + "loss": 1.5988, + "step": 590 + }, + { + "epoch": 0.7449494949494949, + "eval_loss": 1.775631070137024, + "eval_runtime": 37.4113, + "eval_samples_per_second": 26.73, + "eval_steps_per_second": 1.123, + "step": 590 + }, + { + "epoch": 0.7462121212121212, + "grad_norm": 2.7201945781707764, + "learning_rate": 4.209541627689429e-06, + "loss": 1.5767, + "step": 591 + }, + { + "epoch": 0.7474747474747475, + "grad_norm": 2.838362455368042, + "learning_rate": 4.207202993451825e-06, + "loss": 1.5793, + "step": 592 + }, + { + "epoch": 0.7487373737373737, + "grad_norm": 2.6511716842651367, + "learning_rate": 4.204864359214219e-06, + "loss": 1.5776, + "step": 593 + }, + { + "epoch": 0.75, + "grad_norm": 2.7062957286834717, + "learning_rate": 4.2025257249766135e-06, + "loss": 1.5593, + "step": 594 + }, + { + "epoch": 0.7512626262626263, + "grad_norm": 2.808570384979248, + "learning_rate": 4.200187090739009e-06, + "loss": 1.6249, + "step": 595 + }, + { + "epoch": 0.7525252525252525, + "grad_norm": 2.600278377532959, + "learning_rate": 4.197848456501403e-06, + "loss": 1.6171, + "step": 596 + }, + { + "epoch": 0.7537878787878788, + "grad_norm": 2.9669382572174072, + "learning_rate": 4.195509822263798e-06, + "loss": 1.608, + "step": 597 + }, + { + "epoch": 0.7550505050505051, + "grad_norm": 2.6576340198516846, + "learning_rate": 4.193171188026194e-06, + "loss": 1.5986, + "step": 598 + }, + { + "epoch": 0.7563131313131313, + "grad_norm": 2.787060022354126, + "learning_rate": 4.190832553788588e-06, + "loss": 1.548, + "step": 599 + }, + { + "epoch": 0.7575757575757576, + "grad_norm": 2.854971408843994, + "learning_rate": 4.1884939195509825e-06, + "loss": 1.6421, + "step": 600 + }, + { + "epoch": 0.7575757575757576, + "eval_loss": 1.7721046209335327, + "eval_runtime": 37.4471, + "eval_samples_per_second": 26.704, + "eval_steps_per_second": 1.122, + "step": 600 + }, + { + "epoch": 0.7588383838383839, + "grad_norm": 2.7417967319488525, + "learning_rate": 4.186155285313378e-06, + "loss": 1.6399, + "step": 601 + }, + { + "epoch": 0.76010101010101, + "grad_norm": 3.0036182403564453, + "learning_rate": 4.183816651075772e-06, + "loss": 1.5953, + "step": 602 + }, + { + "epoch": 0.7613636363636364, + "grad_norm": 2.502861261367798, + "learning_rate": 4.181478016838167e-06, + "loss": 1.5284, + "step": 603 + }, + { + "epoch": 0.7626262626262627, + "grad_norm": 2.7262094020843506, + "learning_rate": 4.179139382600562e-06, + "loss": 1.619, + "step": 604 + }, + { + "epoch": 0.7638888888888888, + "grad_norm": 2.8748724460601807, + "learning_rate": 4.176800748362956e-06, + "loss": 1.5625, + "step": 605 + }, + { + "epoch": 0.7651515151515151, + "grad_norm": 2.7272109985351562, + "learning_rate": 4.174462114125351e-06, + "loss": 1.5935, + "step": 606 + }, + { + "epoch": 0.7664141414141414, + "grad_norm": 2.7930479049682617, + "learning_rate": 4.172123479887746e-06, + "loss": 1.61, + "step": 607 + }, + { + "epoch": 0.7676767676767676, + "grad_norm": 2.952056407928467, + "learning_rate": 4.16978484565014e-06, + "loss": 1.6369, + "step": 608 + }, + { + "epoch": 0.7689393939393939, + "grad_norm": 2.979564905166626, + "learning_rate": 4.167446211412535e-06, + "loss": 1.5748, + "step": 609 + }, + { + "epoch": 0.7702020202020202, + "grad_norm": 2.7945773601531982, + "learning_rate": 4.16510757717493e-06, + "loss": 1.6096, + "step": 610 + }, + { + "epoch": 0.7702020202020202, + "eval_loss": 1.7568180561065674, + "eval_runtime": 36.969, + "eval_samples_per_second": 27.05, + "eval_steps_per_second": 1.136, + "step": 610 + }, + { + "epoch": 0.7714646464646465, + "grad_norm": 2.9217135906219482, + "learning_rate": 4.1627689429373245e-06, + "loss": 1.5902, + "step": 611 + }, + { + "epoch": 0.7727272727272727, + "grad_norm": 2.8184452056884766, + "learning_rate": 4.16043030869972e-06, + "loss": 1.5775, + "step": 612 + }, + { + "epoch": 0.773989898989899, + "grad_norm": 2.7121212482452393, + "learning_rate": 4.158091674462115e-06, + "loss": 1.5778, + "step": 613 + }, + { + "epoch": 0.7752525252525253, + "grad_norm": 2.7600955963134766, + "learning_rate": 4.1557530402245095e-06, + "loss": 1.6145, + "step": 614 + }, + { + "epoch": 0.7765151515151515, + "grad_norm": 2.6388845443725586, + "learning_rate": 4.153414405986904e-06, + "loss": 1.6475, + "step": 615 + }, + { + "epoch": 0.7777777777777778, + "grad_norm": 2.684379816055298, + "learning_rate": 4.151075771749299e-06, + "loss": 1.6187, + "step": 616 + }, + { + "epoch": 0.7790404040404041, + "grad_norm": 2.6907055377960205, + "learning_rate": 4.1487371375116936e-06, + "loss": 1.6108, + "step": 617 + }, + { + "epoch": 0.7803030303030303, + "grad_norm": 2.8458995819091797, + "learning_rate": 4.146398503274088e-06, + "loss": 1.5596, + "step": 618 + }, + { + "epoch": 0.7815656565656566, + "grad_norm": 2.68843674659729, + "learning_rate": 4.144059869036483e-06, + "loss": 1.5912, + "step": 619 + }, + { + "epoch": 0.7828282828282829, + "grad_norm": 2.7620413303375244, + "learning_rate": 4.141721234798878e-06, + "loss": 1.627, + "step": 620 + }, + { + "epoch": 0.7828282828282829, + "eval_loss": 1.757076621055603, + "eval_runtime": 37.0489, + "eval_samples_per_second": 26.991, + "eval_steps_per_second": 1.134, + "step": 620 + }, + { + "epoch": 0.7840909090909091, + "grad_norm": 2.7084949016571045, + "learning_rate": 4.139382600561272e-06, + "loss": 1.5249, + "step": 621 + }, + { + "epoch": 0.7853535353535354, + "grad_norm": 3.0410406589508057, + "learning_rate": 4.137043966323667e-06, + "loss": 1.6458, + "step": 622 + }, + { + "epoch": 0.7866161616161617, + "grad_norm": 2.8445630073547363, + "learning_rate": 4.134705332086062e-06, + "loss": 1.6404, + "step": 623 + }, + { + "epoch": 0.7878787878787878, + "grad_norm": 3.0764353275299072, + "learning_rate": 4.132366697848457e-06, + "loss": 1.5041, + "step": 624 + }, + { + "epoch": 0.7891414141414141, + "grad_norm": 2.626544237136841, + "learning_rate": 4.1300280636108515e-06, + "loss": 1.6145, + "step": 625 + }, + { + "epoch": 0.7904040404040404, + "grad_norm": 2.9735796451568604, + "learning_rate": 4.127689429373246e-06, + "loss": 1.5672, + "step": 626 + }, + { + "epoch": 0.7916666666666666, + "grad_norm": 2.7798562049865723, + "learning_rate": 4.125350795135641e-06, + "loss": 1.6272, + "step": 627 + }, + { + "epoch": 0.7929292929292929, + "grad_norm": 2.7438931465148926, + "learning_rate": 4.123012160898036e-06, + "loss": 1.5318, + "step": 628 + }, + { + "epoch": 0.7941919191919192, + "grad_norm": 3.103142499923706, + "learning_rate": 4.120673526660431e-06, + "loss": 1.6, + "step": 629 + }, + { + "epoch": 0.7954545454545454, + "grad_norm": 2.9056239128112793, + "learning_rate": 4.118334892422825e-06, + "loss": 1.6581, + "step": 630 + }, + { + "epoch": 0.7954545454545454, + "eval_loss": 1.7552870512008667, + "eval_runtime": 37.1223, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 630 + }, + { + "epoch": 0.7967171717171717, + "grad_norm": 2.7320070266723633, + "learning_rate": 4.1159962581852205e-06, + "loss": 1.5808, + "step": 631 + }, + { + "epoch": 0.797979797979798, + "grad_norm": 2.9518637657165527, + "learning_rate": 4.113657623947615e-06, + "loss": 1.6034, + "step": 632 + }, + { + "epoch": 0.7992424242424242, + "grad_norm": 2.7135164737701416, + "learning_rate": 4.111318989710009e-06, + "loss": 1.5684, + "step": 633 + }, + { + "epoch": 0.8005050505050505, + "grad_norm": 2.948230266571045, + "learning_rate": 4.108980355472405e-06, + "loss": 1.6378, + "step": 634 + }, + { + "epoch": 0.8017676767676768, + "grad_norm": 2.816206216812134, + "learning_rate": 4.106641721234799e-06, + "loss": 1.6118, + "step": 635 + }, + { + "epoch": 0.803030303030303, + "grad_norm": 2.8032639026641846, + "learning_rate": 4.104303086997194e-06, + "loss": 1.5627, + "step": 636 + }, + { + "epoch": 0.8042929292929293, + "grad_norm": 3.032348871231079, + "learning_rate": 4.101964452759589e-06, + "loss": 1.6192, + "step": 637 + }, + { + "epoch": 0.8055555555555556, + "grad_norm": 2.9512710571289062, + "learning_rate": 4.099625818521983e-06, + "loss": 1.5655, + "step": 638 + }, + { + "epoch": 0.8068181818181818, + "grad_norm": 2.738973379135132, + "learning_rate": 4.097287184284378e-06, + "loss": 1.5686, + "step": 639 + }, + { + "epoch": 0.8080808080808081, + "grad_norm": 2.6510915756225586, + "learning_rate": 4.094948550046773e-06, + "loss": 1.5889, + "step": 640 + }, + { + "epoch": 0.8080808080808081, + "eval_loss": 1.762879490852356, + "eval_runtime": 37.1457, + "eval_samples_per_second": 26.921, + "eval_steps_per_second": 1.131, + "step": 640 + }, + { + "epoch": 0.8093434343434344, + "grad_norm": 2.6933884620666504, + "learning_rate": 4.092609915809167e-06, + "loss": 1.5436, + "step": 641 + }, + { + "epoch": 0.8106060606060606, + "grad_norm": 2.703784227371216, + "learning_rate": 4.0902712815715625e-06, + "loss": 1.5869, + "step": 642 + }, + { + "epoch": 0.8118686868686869, + "grad_norm": 2.7742321491241455, + "learning_rate": 4.087932647333958e-06, + "loss": 1.5779, + "step": 643 + }, + { + "epoch": 0.8131313131313131, + "grad_norm": 2.836832284927368, + "learning_rate": 4.085594013096352e-06, + "loss": 1.5511, + "step": 644 + }, + { + "epoch": 0.8143939393939394, + "grad_norm": 2.9212028980255127, + "learning_rate": 4.083255378858747e-06, + "loss": 1.5604, + "step": 645 + }, + { + "epoch": 0.8156565656565656, + "grad_norm": 2.8608527183532715, + "learning_rate": 4.080916744621142e-06, + "loss": 1.5975, + "step": 646 + }, + { + "epoch": 0.8169191919191919, + "grad_norm": 2.8246703147888184, + "learning_rate": 4.078578110383536e-06, + "loss": 1.5962, + "step": 647 + }, + { + "epoch": 0.8181818181818182, + "grad_norm": 3.0648651123046875, + "learning_rate": 4.076239476145932e-06, + "loss": 1.6267, + "step": 648 + }, + { + "epoch": 0.8194444444444444, + "grad_norm": 2.849240303039551, + "learning_rate": 4.073900841908326e-06, + "loss": 1.6271, + "step": 649 + }, + { + "epoch": 0.8207070707070707, + "grad_norm": 2.8817555904388428, + "learning_rate": 4.07156220767072e-06, + "loss": 1.5916, + "step": 650 + }, + { + "epoch": 0.8207070707070707, + "eval_loss": 1.7502237558364868, + "eval_runtime": 37.1497, + "eval_samples_per_second": 26.918, + "eval_steps_per_second": 1.131, + "step": 650 + }, + { + "epoch": 0.821969696969697, + "grad_norm": 2.5747952461242676, + "learning_rate": 4.069223573433116e-06, + "loss": 1.5661, + "step": 651 + }, + { + "epoch": 0.8232323232323232, + "grad_norm": 2.7721974849700928, + "learning_rate": 4.06688493919551e-06, + "loss": 1.5903, + "step": 652 + }, + { + "epoch": 0.8244949494949495, + "grad_norm": 2.9606690406799316, + "learning_rate": 4.0645463049579045e-06, + "loss": 1.5431, + "step": 653 + }, + { + "epoch": 0.8257575757575758, + "grad_norm": 2.941169261932373, + "learning_rate": 4.0622076707203e-06, + "loss": 1.5711, + "step": 654 + }, + { + "epoch": 0.827020202020202, + "grad_norm": 2.8408572673797607, + "learning_rate": 4.059869036482694e-06, + "loss": 1.5975, + "step": 655 + }, + { + "epoch": 0.8282828282828283, + "grad_norm": 2.619875431060791, + "learning_rate": 4.057530402245089e-06, + "loss": 1.5194, + "step": 656 + }, + { + "epoch": 0.8295454545454546, + "grad_norm": 2.910982847213745, + "learning_rate": 4.055191768007484e-06, + "loss": 1.6318, + "step": 657 + }, + { + "epoch": 0.8308080808080808, + "grad_norm": 2.747422456741333, + "learning_rate": 4.052853133769879e-06, + "loss": 1.6205, + "step": 658 + }, + { + "epoch": 0.8320707070707071, + "grad_norm": 2.6889796257019043, + "learning_rate": 4.050514499532274e-06, + "loss": 1.5537, + "step": 659 + }, + { + "epoch": 0.8333333333333334, + "grad_norm": 2.9849202632904053, + "learning_rate": 4.048175865294669e-06, + "loss": 1.5438, + "step": 660 + }, + { + "epoch": 0.8333333333333334, + "eval_loss": 1.7484922409057617, + "eval_runtime": 37.161, + "eval_samples_per_second": 26.91, + "eval_steps_per_second": 1.13, + "step": 660 + }, + { + "epoch": 0.8345959595959596, + "grad_norm": 2.816930055618286, + "learning_rate": 4.045837231057063e-06, + "loss": 1.555, + "step": 661 + }, + { + "epoch": 0.8358585858585859, + "grad_norm": 2.990126371383667, + "learning_rate": 4.043498596819458e-06, + "loss": 1.6263, + "step": 662 + }, + { + "epoch": 0.8371212121212122, + "grad_norm": 2.898263692855835, + "learning_rate": 4.041159962581853e-06, + "loss": 1.5522, + "step": 663 + }, + { + "epoch": 0.8383838383838383, + "grad_norm": 2.7824151515960693, + "learning_rate": 4.038821328344247e-06, + "loss": 1.5366, + "step": 664 + }, + { + "epoch": 0.8396464646464646, + "grad_norm": 2.9943954944610596, + "learning_rate": 4.036482694106642e-06, + "loss": 1.6214, + "step": 665 + }, + { + "epoch": 0.8409090909090909, + "grad_norm": 2.628308057785034, + "learning_rate": 4.034144059869037e-06, + "loss": 1.5797, + "step": 666 + }, + { + "epoch": 0.8421717171717171, + "grad_norm": 2.7565176486968994, + "learning_rate": 4.0318054256314315e-06, + "loss": 1.5567, + "step": 667 + }, + { + "epoch": 0.8434343434343434, + "grad_norm": 2.688354969024658, + "learning_rate": 4.029466791393826e-06, + "loss": 1.5516, + "step": 668 + }, + { + "epoch": 0.8446969696969697, + "grad_norm": 2.899620771408081, + "learning_rate": 4.027128157156221e-06, + "loss": 1.6014, + "step": 669 + }, + { + "epoch": 0.8459595959595959, + "grad_norm": 2.7894718647003174, + "learning_rate": 4.024789522918616e-06, + "loss": 1.5688, + "step": 670 + }, + { + "epoch": 0.8459595959595959, + "eval_loss": 1.7486555576324463, + "eval_runtime": 37.1824, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 670 + }, + { + "epoch": 0.8472222222222222, + "grad_norm": 2.7441444396972656, + "learning_rate": 4.02245088868101e-06, + "loss": 1.5949, + "step": 671 + }, + { + "epoch": 0.8484848484848485, + "grad_norm": 2.7088451385498047, + "learning_rate": 4.020112254443405e-06, + "loss": 1.5878, + "step": 672 + }, + { + "epoch": 0.8497474747474747, + "grad_norm": 2.756920576095581, + "learning_rate": 4.0177736202058005e-06, + "loss": 1.6356, + "step": 673 + }, + { + "epoch": 0.851010101010101, + "grad_norm": 2.5605697631835938, + "learning_rate": 4.015434985968195e-06, + "loss": 1.5586, + "step": 674 + }, + { + "epoch": 0.8522727272727273, + "grad_norm": 2.946573257446289, + "learning_rate": 4.01309635173059e-06, + "loss": 1.5537, + "step": 675 + }, + { + "epoch": 0.8535353535353535, + "grad_norm": 2.65500545501709, + "learning_rate": 4.010757717492985e-06, + "loss": 1.5875, + "step": 676 + }, + { + "epoch": 0.8547979797979798, + "grad_norm": 2.6632118225097656, + "learning_rate": 4.008419083255379e-06, + "loss": 1.5329, + "step": 677 + }, + { + "epoch": 0.8560606060606061, + "grad_norm": 2.8909032344818115, + "learning_rate": 4.006080449017774e-06, + "loss": 1.6396, + "step": 678 + }, + { + "epoch": 0.8573232323232324, + "grad_norm": 2.8780951499938965, + "learning_rate": 4.003741814780169e-06, + "loss": 1.6189, + "step": 679 + }, + { + "epoch": 0.8585858585858586, + "grad_norm": 2.7757010459899902, + "learning_rate": 4.001403180542563e-06, + "loss": 1.5968, + "step": 680 + }, + { + "epoch": 0.8585858585858586, + "eval_loss": 1.7423925399780273, + "eval_runtime": 37.1537, + "eval_samples_per_second": 26.915, + "eval_steps_per_second": 1.13, + "step": 680 + }, + { + "epoch": 0.8598484848484849, + "grad_norm": 2.6610794067382812, + "learning_rate": 3.9990645463049584e-06, + "loss": 1.5258, + "step": 681 + }, + { + "epoch": 0.8611111111111112, + "grad_norm": 2.9292140007019043, + "learning_rate": 3.996725912067353e-06, + "loss": 1.5669, + "step": 682 + }, + { + "epoch": 0.8623737373737373, + "grad_norm": 2.7885360717773438, + "learning_rate": 3.994387277829747e-06, + "loss": 1.5737, + "step": 683 + }, + { + "epoch": 0.8636363636363636, + "grad_norm": 2.7511544227600098, + "learning_rate": 3.9920486435921425e-06, + "loss": 1.5715, + "step": 684 + }, + { + "epoch": 0.86489898989899, + "grad_norm": 2.8561315536499023, + "learning_rate": 3.989710009354537e-06, + "loss": 1.5965, + "step": 685 + }, + { + "epoch": 0.8661616161616161, + "grad_norm": 2.9127237796783447, + "learning_rate": 3.987371375116932e-06, + "loss": 1.5407, + "step": 686 + }, + { + "epoch": 0.8674242424242424, + "grad_norm": 2.735577344894409, + "learning_rate": 3.985032740879327e-06, + "loss": 1.5449, + "step": 687 + }, + { + "epoch": 0.8686868686868687, + "grad_norm": 2.6626689434051514, + "learning_rate": 3.982694106641721e-06, + "loss": 1.5173, + "step": 688 + }, + { + "epoch": 0.8699494949494949, + "grad_norm": 2.901592969894409, + "learning_rate": 3.980355472404116e-06, + "loss": 1.5757, + "step": 689 + }, + { + "epoch": 0.8712121212121212, + "grad_norm": 2.8801231384277344, + "learning_rate": 3.978016838166512e-06, + "loss": 1.5633, + "step": 690 + }, + { + "epoch": 0.8712121212121212, + "eval_loss": 1.7453432083129883, + "eval_runtime": 37.4368, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 690 + }, + { + "epoch": 0.8724747474747475, + "grad_norm": 2.7463626861572266, + "learning_rate": 3.975678203928906e-06, + "loss": 1.5786, + "step": 691 + }, + { + "epoch": 0.8737373737373737, + "grad_norm": 2.659820318222046, + "learning_rate": 3.9733395696913004e-06, + "loss": 1.5923, + "step": 692 + }, + { + "epoch": 0.875, + "grad_norm": 2.759030342102051, + "learning_rate": 3.971000935453696e-06, + "loss": 1.6217, + "step": 693 + }, + { + "epoch": 0.8762626262626263, + "grad_norm": 2.689967632293701, + "learning_rate": 3.96866230121609e-06, + "loss": 1.5533, + "step": 694 + }, + { + "epoch": 0.8775252525252525, + "grad_norm": 3.1230740547180176, + "learning_rate": 3.9663236669784845e-06, + "loss": 1.6538, + "step": 695 + }, + { + "epoch": 0.8787878787878788, + "grad_norm": 2.657022476196289, + "learning_rate": 3.96398503274088e-06, + "loss": 1.5432, + "step": 696 + }, + { + "epoch": 0.8800505050505051, + "grad_norm": 2.919379711151123, + "learning_rate": 3.961646398503274e-06, + "loss": 1.5867, + "step": 697 + }, + { + "epoch": 0.8813131313131313, + "grad_norm": 2.607135057449341, + "learning_rate": 3.9593077642656695e-06, + "loss": 1.5079, + "step": 698 + }, + { + "epoch": 0.8825757575757576, + "grad_norm": 2.9978976249694824, + "learning_rate": 3.956969130028064e-06, + "loss": 1.5905, + "step": 699 + }, + { + "epoch": 0.8838383838383839, + "grad_norm": 2.8117148876190186, + "learning_rate": 3.954630495790458e-06, + "loss": 1.5588, + "step": 700 + }, + { + "epoch": 0.8838383838383839, + "eval_loss": 1.7376645803451538, + "eval_runtime": 37.5045, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 700 + }, + { + "epoch": 0.88510101010101, + "grad_norm": 2.7135095596313477, + "learning_rate": 3.952291861552854e-06, + "loss": 1.6013, + "step": 701 + }, + { + "epoch": 0.8863636363636364, + "grad_norm": 2.812061071395874, + "learning_rate": 3.949953227315248e-06, + "loss": 1.5518, + "step": 702 + }, + { + "epoch": 0.8876262626262627, + "grad_norm": 3.0532422065734863, + "learning_rate": 3.9476145930776424e-06, + "loss": 1.5579, + "step": 703 + }, + { + "epoch": 0.8888888888888888, + "grad_norm": 2.8306567668914795, + "learning_rate": 3.945275958840038e-06, + "loss": 1.5478, + "step": 704 + }, + { + "epoch": 0.8901515151515151, + "grad_norm": 2.737921953201294, + "learning_rate": 3.942937324602433e-06, + "loss": 1.4958, + "step": 705 + }, + { + "epoch": 0.8914141414141414, + "grad_norm": 2.902211904525757, + "learning_rate": 3.940598690364827e-06, + "loss": 1.5715, + "step": 706 + }, + { + "epoch": 0.8926767676767676, + "grad_norm": 2.725818395614624, + "learning_rate": 3.938260056127222e-06, + "loss": 1.5502, + "step": 707 + }, + { + "epoch": 0.8939393939393939, + "grad_norm": 2.9530506134033203, + "learning_rate": 3.935921421889617e-06, + "loss": 1.564, + "step": 708 + }, + { + "epoch": 0.8952020202020202, + "grad_norm": 2.798231363296509, + "learning_rate": 3.9335827876520115e-06, + "loss": 1.5445, + "step": 709 + }, + { + "epoch": 0.8964646464646465, + "grad_norm": 3.0233471393585205, + "learning_rate": 3.931244153414407e-06, + "loss": 1.6193, + "step": 710 + }, + { + "epoch": 0.8964646464646465, + "eval_loss": 1.742077350616455, + "eval_runtime": 37.0255, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 710 + }, + { + "epoch": 0.8977272727272727, + "grad_norm": 2.84309458732605, + "learning_rate": 3.928905519176801e-06, + "loss": 1.5379, + "step": 711 + }, + { + "epoch": 0.898989898989899, + "grad_norm": 2.7412185668945312, + "learning_rate": 3.926566884939196e-06, + "loss": 1.5837, + "step": 712 + }, + { + "epoch": 0.9002525252525253, + "grad_norm": 3.1094822883605957, + "learning_rate": 3.924228250701591e-06, + "loss": 1.5621, + "step": 713 + }, + { + "epoch": 0.9015151515151515, + "grad_norm": 2.776059627532959, + "learning_rate": 3.921889616463985e-06, + "loss": 1.5714, + "step": 714 + }, + { + "epoch": 0.9027777777777778, + "grad_norm": 2.7650489807128906, + "learning_rate": 3.91955098222638e-06, + "loss": 1.5418, + "step": 715 + }, + { + "epoch": 0.9040404040404041, + "grad_norm": 2.8586974143981934, + "learning_rate": 3.917212347988775e-06, + "loss": 1.5891, + "step": 716 + }, + { + "epoch": 0.9053030303030303, + "grad_norm": 2.74332594871521, + "learning_rate": 3.914873713751169e-06, + "loss": 1.5492, + "step": 717 + }, + { + "epoch": 0.9065656565656566, + "grad_norm": 2.931682586669922, + "learning_rate": 3.912535079513564e-06, + "loss": 1.54, + "step": 718 + }, + { + "epoch": 0.9078282828282829, + "grad_norm": 2.678812026977539, + "learning_rate": 3.910196445275959e-06, + "loss": 1.5677, + "step": 719 + }, + { + "epoch": 0.9090909090909091, + "grad_norm": 2.7282893657684326, + "learning_rate": 3.907857811038354e-06, + "loss": 1.5413, + "step": 720 + }, + { + "epoch": 0.9090909090909091, + "eval_loss": 1.7409753799438477, + "eval_runtime": 37.1883, + "eval_samples_per_second": 26.89, + "eval_steps_per_second": 1.129, + "step": 720 + }, + { + "epoch": 0.9103535353535354, + "grad_norm": 3.073997735977173, + "learning_rate": 3.905519176800749e-06, + "loss": 1.5734, + "step": 721 + }, + { + "epoch": 0.9116161616161617, + "grad_norm": 2.6249935626983643, + "learning_rate": 3.903180542563144e-06, + "loss": 1.5063, + "step": 722 + }, + { + "epoch": 0.9128787878787878, + "grad_norm": 2.7758615016937256, + "learning_rate": 3.9008419083255384e-06, + "loss": 1.5112, + "step": 723 + }, + { + "epoch": 0.9141414141414141, + "grad_norm": 2.8954596519470215, + "learning_rate": 3.898503274087933e-06, + "loss": 1.5508, + "step": 724 + }, + { + "epoch": 0.9154040404040404, + "grad_norm": 2.79073166847229, + "learning_rate": 3.896164639850328e-06, + "loss": 1.5773, + "step": 725 + }, + { + "epoch": 0.9166666666666666, + "grad_norm": 2.687027931213379, + "learning_rate": 3.8938260056127226e-06, + "loss": 1.5243, + "step": 726 + }, + { + "epoch": 0.9179292929292929, + "grad_norm": 2.5704851150512695, + "learning_rate": 3.891487371375117e-06, + "loss": 1.5075, + "step": 727 + }, + { + "epoch": 0.9191919191919192, + "grad_norm": 2.8684048652648926, + "learning_rate": 3.889148737137512e-06, + "loss": 1.6471, + "step": 728 + }, + { + "epoch": 0.9204545454545454, + "grad_norm": 2.8995590209960938, + "learning_rate": 3.886810102899907e-06, + "loss": 1.5429, + "step": 729 + }, + { + "epoch": 0.9217171717171717, + "grad_norm": 2.7575254440307617, + "learning_rate": 3.884471468662301e-06, + "loss": 1.5241, + "step": 730 + }, + { + "epoch": 0.9217171717171717, + "eval_loss": 1.7294232845306396, + "eval_runtime": 37.1941, + "eval_samples_per_second": 26.886, + "eval_steps_per_second": 1.129, + "step": 730 + }, + { + "epoch": 0.922979797979798, + "grad_norm": 2.7057321071624756, + "learning_rate": 3.882132834424696e-06, + "loss": 1.5498, + "step": 731 + }, + { + "epoch": 0.9242424242424242, + "grad_norm": 2.7679436206817627, + "learning_rate": 3.879794200187091e-06, + "loss": 1.563, + "step": 732 + }, + { + "epoch": 0.9255050505050505, + "grad_norm": 2.7258687019348145, + "learning_rate": 3.877455565949485e-06, + "loss": 1.5016, + "step": 733 + }, + { + "epoch": 0.9267676767676768, + "grad_norm": 2.8474221229553223, + "learning_rate": 3.8751169317118804e-06, + "loss": 1.5318, + "step": 734 + }, + { + "epoch": 0.928030303030303, + "grad_norm": 2.7718396186828613, + "learning_rate": 3.872778297474276e-06, + "loss": 1.5919, + "step": 735 + }, + { + "epoch": 0.9292929292929293, + "grad_norm": 2.7785022258758545, + "learning_rate": 3.87043966323667e-06, + "loss": 1.5399, + "step": 736 + }, + { + "epoch": 0.9305555555555556, + "grad_norm": 2.742476463317871, + "learning_rate": 3.868101028999065e-06, + "loss": 1.5428, + "step": 737 + }, + { + "epoch": 0.9318181818181818, + "grad_norm": 2.7050726413726807, + "learning_rate": 3.86576239476146e-06, + "loss": 1.5604, + "step": 738 + }, + { + "epoch": 0.9330808080808081, + "grad_norm": 2.6755242347717285, + "learning_rate": 3.863423760523854e-06, + "loss": 1.5409, + "step": 739 + }, + { + "epoch": 0.9343434343434344, + "grad_norm": 2.6947829723358154, + "learning_rate": 3.8610851262862495e-06, + "loss": 1.5537, + "step": 740 + }, + { + "epoch": 0.9343434343434344, + "eval_loss": 1.7396037578582764, + "eval_runtime": 37.1912, + "eval_samples_per_second": 26.888, + "eval_steps_per_second": 1.129, + "step": 740 + }, + { + "epoch": 0.9356060606060606, + "grad_norm": 2.729816436767578, + "learning_rate": 3.858746492048644e-06, + "loss": 1.5581, + "step": 741 + }, + { + "epoch": 0.9368686868686869, + "grad_norm": 2.6951887607574463, + "learning_rate": 3.856407857811038e-06, + "loss": 1.5929, + "step": 742 + }, + { + "epoch": 0.9381313131313131, + "grad_norm": 2.6601064205169678, + "learning_rate": 3.854069223573434e-06, + "loss": 1.4881, + "step": 743 + }, + { + "epoch": 0.9393939393939394, + "grad_norm": 2.623898983001709, + "learning_rate": 3.851730589335828e-06, + "loss": 1.5499, + "step": 744 + }, + { + "epoch": 0.9406565656565656, + "grad_norm": 2.703871250152588, + "learning_rate": 3.8493919550982224e-06, + "loss": 1.5585, + "step": 745 + }, + { + "epoch": 0.9419191919191919, + "grad_norm": 2.674893379211426, + "learning_rate": 3.847053320860618e-06, + "loss": 1.5586, + "step": 746 + }, + { + "epoch": 0.9431818181818182, + "grad_norm": 2.8353004455566406, + "learning_rate": 3.844714686623012e-06, + "loss": 1.5937, + "step": 747 + }, + { + "epoch": 0.9444444444444444, + "grad_norm": 2.68009614944458, + "learning_rate": 3.842376052385407e-06, + "loss": 1.5192, + "step": 748 + }, + { + "epoch": 0.9457070707070707, + "grad_norm": 2.6847407817840576, + "learning_rate": 3.840037418147802e-06, + "loss": 1.512, + "step": 749 + }, + { + "epoch": 0.946969696969697, + "grad_norm": 2.624290943145752, + "learning_rate": 3.837698783910197e-06, + "loss": 1.5348, + "step": 750 + }, + { + "epoch": 0.946969696969697, + "eval_loss": 1.738332748413086, + "eval_runtime": 37.1782, + "eval_samples_per_second": 26.898, + "eval_steps_per_second": 1.13, + "step": 750 + }, + { + "epoch": 0.9482323232323232, + "grad_norm": 2.7467777729034424, + "learning_rate": 3.8353601496725915e-06, + "loss": 1.5603, + "step": 751 + }, + { + "epoch": 0.9494949494949495, + "grad_norm": 2.851987600326538, + "learning_rate": 3.833021515434987e-06, + "loss": 1.5974, + "step": 752 + }, + { + "epoch": 0.9507575757575758, + "grad_norm": 2.6642227172851562, + "learning_rate": 3.830682881197381e-06, + "loss": 1.5138, + "step": 753 + }, + { + "epoch": 0.952020202020202, + "grad_norm": 2.7351491451263428, + "learning_rate": 3.828344246959776e-06, + "loss": 1.5656, + "step": 754 + }, + { + "epoch": 0.9532828282828283, + "grad_norm": 2.593748092651367, + "learning_rate": 3.826005612722171e-06, + "loss": 1.5395, + "step": 755 + }, + { + "epoch": 0.9545454545454546, + "grad_norm": 2.639180898666382, + "learning_rate": 3.823666978484565e-06, + "loss": 1.5349, + "step": 756 + }, + { + "epoch": 0.9558080808080808, + "grad_norm": 2.805148124694824, + "learning_rate": 3.82132834424696e-06, + "loss": 1.5805, + "step": 757 + }, + { + "epoch": 0.9570707070707071, + "grad_norm": 2.5619425773620605, + "learning_rate": 3.818989710009355e-06, + "loss": 1.5389, + "step": 758 + }, + { + "epoch": 0.9583333333333334, + "grad_norm": 2.7224011421203613, + "learning_rate": 3.816651075771749e-06, + "loss": 1.5316, + "step": 759 + }, + { + "epoch": 0.9595959595959596, + "grad_norm": 2.4667656421661377, + "learning_rate": 3.8143124415341442e-06, + "loss": 1.4942, + "step": 760 + }, + { + "epoch": 0.9595959595959596, + "eval_loss": 1.7205606698989868, + "eval_runtime": 37.1654, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 760 + }, + { + "epoch": 0.9608585858585859, + "grad_norm": 2.456643581390381, + "learning_rate": 3.811973807296539e-06, + "loss": 1.538, + "step": 761 + }, + { + "epoch": 0.9621212121212122, + "grad_norm": 2.696720838546753, + "learning_rate": 3.8096351730589335e-06, + "loss": 1.5462, + "step": 762 + }, + { + "epoch": 0.9633838383838383, + "grad_norm": 2.7314131259918213, + "learning_rate": 3.8072965388213284e-06, + "loss": 1.5931, + "step": 763 + }, + { + "epoch": 0.9646464646464646, + "grad_norm": 2.616532325744629, + "learning_rate": 3.804957904583723e-06, + "loss": 1.5469, + "step": 764 + }, + { + "epoch": 0.9659090909090909, + "grad_norm": 2.7471885681152344, + "learning_rate": 3.8026192703461185e-06, + "loss": 1.5431, + "step": 765 + }, + { + "epoch": 0.9671717171717171, + "grad_norm": 2.7478253841400146, + "learning_rate": 3.8002806361085133e-06, + "loss": 1.5452, + "step": 766 + }, + { + "epoch": 0.9684343434343434, + "grad_norm": 2.543790340423584, + "learning_rate": 3.7979420018709077e-06, + "loss": 1.6091, + "step": 767 + }, + { + "epoch": 0.9696969696969697, + "grad_norm": 2.6169509887695312, + "learning_rate": 3.7956033676333026e-06, + "loss": 1.5557, + "step": 768 + }, + { + "epoch": 0.9709595959595959, + "grad_norm": 2.4768588542938232, + "learning_rate": 3.7932647333956974e-06, + "loss": 1.5247, + "step": 769 + }, + { + "epoch": 0.9722222222222222, + "grad_norm": 2.654648780822754, + "learning_rate": 3.7909260991580923e-06, + "loss": 1.5281, + "step": 770 + }, + { + "epoch": 0.9722222222222222, + "eval_loss": 1.7289395332336426, + "eval_runtime": 37.1374, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 770 + }, + { + "epoch": 0.9734848484848485, + "grad_norm": 2.893073081970215, + "learning_rate": 3.7885874649204867e-06, + "loss": 1.5749, + "step": 771 + }, + { + "epoch": 0.9747474747474747, + "grad_norm": 2.696341037750244, + "learning_rate": 3.7862488306828815e-06, + "loss": 1.5025, + "step": 772 + }, + { + "epoch": 0.976010101010101, + "grad_norm": 2.678222417831421, + "learning_rate": 3.7839101964452764e-06, + "loss": 1.538, + "step": 773 + }, + { + "epoch": 0.9772727272727273, + "grad_norm": 2.7772746086120605, + "learning_rate": 3.7815715622076708e-06, + "loss": 1.5648, + "step": 774 + }, + { + "epoch": 0.9785353535353535, + "grad_norm": 3.076620101928711, + "learning_rate": 3.7792329279700656e-06, + "loss": 1.6114, + "step": 775 + }, + { + "epoch": 0.9797979797979798, + "grad_norm": 2.700056791305542, + "learning_rate": 3.7768942937324605e-06, + "loss": 1.5881, + "step": 776 + }, + { + "epoch": 0.9810606060606061, + "grad_norm": 2.601419687271118, + "learning_rate": 3.7745556594948553e-06, + "loss": 1.5319, + "step": 777 + }, + { + "epoch": 0.9823232323232324, + "grad_norm": 2.6664950847625732, + "learning_rate": 3.7722170252572497e-06, + "loss": 1.5497, + "step": 778 + }, + { + "epoch": 0.9835858585858586, + "grad_norm": 2.7374589443206787, + "learning_rate": 3.7698783910196446e-06, + "loss": 1.5488, + "step": 779 + }, + { + "epoch": 0.9848484848484849, + "grad_norm": 2.797938823699951, + "learning_rate": 3.76753975678204e-06, + "loss": 1.5376, + "step": 780 + }, + { + "epoch": 0.9848484848484849, + "eval_loss": 1.7334709167480469, + "eval_runtime": 37.1372, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 780 + }, + { + "epoch": 0.9861111111111112, + "grad_norm": 2.881591796875, + "learning_rate": 3.7652011225444347e-06, + "loss": 1.5405, + "step": 781 + }, + { + "epoch": 0.9873737373737373, + "grad_norm": 2.6965644359588623, + "learning_rate": 3.7628624883068295e-06, + "loss": 1.5066, + "step": 782 + }, + { + "epoch": 0.9886363636363636, + "grad_norm": 2.7286899089813232, + "learning_rate": 3.760523854069224e-06, + "loss": 1.5277, + "step": 783 + }, + { + "epoch": 0.98989898989899, + "grad_norm": 2.795644998550415, + "learning_rate": 3.7581852198316188e-06, + "loss": 1.5335, + "step": 784 + }, + { + "epoch": 0.9911616161616161, + "grad_norm": 2.8160440921783447, + "learning_rate": 3.7558465855940136e-06, + "loss": 1.585, + "step": 785 + }, + { + "epoch": 0.9924242424242424, + "grad_norm": 2.7984936237335205, + "learning_rate": 3.753507951356408e-06, + "loss": 1.5569, + "step": 786 + }, + { + "epoch": 0.9936868686868687, + "grad_norm": 2.7596278190612793, + "learning_rate": 3.751169317118803e-06, + "loss": 1.4838, + "step": 787 + }, + { + "epoch": 0.9949494949494949, + "grad_norm": 2.666383743286133, + "learning_rate": 3.7488306828811977e-06, + "loss": 1.5656, + "step": 788 + }, + { + "epoch": 0.9962121212121212, + "grad_norm": 2.8139331340789795, + "learning_rate": 3.7464920486435926e-06, + "loss": 1.5753, + "step": 789 + }, + { + "epoch": 0.9974747474747475, + "grad_norm": 2.7714269161224365, + "learning_rate": 3.744153414405987e-06, + "loss": 1.5187, + "step": 790 + }, + { + "epoch": 0.9974747474747475, + "eval_loss": 1.7257990837097168, + "eval_runtime": 37.1065, + "eval_samples_per_second": 26.949, + "eval_steps_per_second": 1.132, + "step": 790 + }, + { + "epoch": 0.9987373737373737, + "grad_norm": 2.697606325149536, + "learning_rate": 3.741814780168382e-06, + "loss": 1.56, + "step": 791 + }, + { + "epoch": 1.0, + "grad_norm": 2.7685229778289795, + "learning_rate": 3.7394761459307767e-06, + "loss": 1.5705, + "step": 792 + }, + { + "epoch": 1.0012626262626263, + "grad_norm": 2.778200626373291, + "learning_rate": 3.737137511693171e-06, + "loss": 1.5288, + "step": 793 + }, + { + "epoch": 1.0025252525252526, + "grad_norm": 2.7211337089538574, + "learning_rate": 3.734798877455566e-06, + "loss": 1.5348, + "step": 794 + }, + { + "epoch": 1.003787878787879, + "grad_norm": 2.690739154815674, + "learning_rate": 3.7324602432179608e-06, + "loss": 1.4734, + "step": 795 + }, + { + "epoch": 1.005050505050505, + "grad_norm": 2.6809420585632324, + "learning_rate": 3.730121608980356e-06, + "loss": 1.4852, + "step": 796 + }, + { + "epoch": 1.0063131313131313, + "grad_norm": 2.8550117015838623, + "learning_rate": 3.727782974742751e-06, + "loss": 1.4562, + "step": 797 + }, + { + "epoch": 1.0075757575757576, + "grad_norm": 2.713935613632202, + "learning_rate": 3.7254443405051453e-06, + "loss": 1.5005, + "step": 798 + }, + { + "epoch": 1.0088383838383839, + "grad_norm": 2.668363332748413, + "learning_rate": 3.72310570626754e-06, + "loss": 1.485, + "step": 799 + }, + { + "epoch": 1.0101010101010102, + "grad_norm": 2.840769052505493, + "learning_rate": 3.720767072029935e-06, + "loss": 1.4531, + "step": 800 + }, + { + "epoch": 1.0101010101010102, + "eval_loss": 1.725937008857727, + "eval_runtime": 37.2471, + "eval_samples_per_second": 26.848, + "eval_steps_per_second": 1.128, + "step": 800 + }, + { + "epoch": 1.0113636363636365, + "grad_norm": 2.7345798015594482, + "learning_rate": 3.71842843779233e-06, + "loss": 1.5091, + "step": 801 + }, + { + "epoch": 1.0126262626262625, + "grad_norm": 2.9052624702453613, + "learning_rate": 3.7160898035547243e-06, + "loss": 1.5039, + "step": 802 + }, + { + "epoch": 1.0138888888888888, + "grad_norm": 2.893073558807373, + "learning_rate": 3.713751169317119e-06, + "loss": 1.4613, + "step": 803 + }, + { + "epoch": 1.0151515151515151, + "grad_norm": 2.8505427837371826, + "learning_rate": 3.711412535079514e-06, + "loss": 1.4932, + "step": 804 + }, + { + "epoch": 1.0164141414141414, + "grad_norm": 2.9065170288085938, + "learning_rate": 3.7090739008419084e-06, + "loss": 1.5082, + "step": 805 + }, + { + "epoch": 1.0176767676767677, + "grad_norm": 2.888068675994873, + "learning_rate": 3.706735266604303e-06, + "loss": 1.5675, + "step": 806 + }, + { + "epoch": 1.018939393939394, + "grad_norm": 2.7325260639190674, + "learning_rate": 3.704396632366698e-06, + "loss": 1.4649, + "step": 807 + }, + { + "epoch": 1.02020202020202, + "grad_norm": 2.8861136436462402, + "learning_rate": 3.702057998129093e-06, + "loss": 1.4517, + "step": 808 + }, + { + "epoch": 1.0214646464646464, + "grad_norm": 2.8837692737579346, + "learning_rate": 3.6997193638914873e-06, + "loss": 1.5198, + "step": 809 + }, + { + "epoch": 1.0227272727272727, + "grad_norm": 2.790330410003662, + "learning_rate": 3.697380729653882e-06, + "loss": 1.5066, + "step": 810 + }, + { + "epoch": 1.0227272727272727, + "eval_loss": 1.7269084453582764, + "eval_runtime": 36.89, + "eval_samples_per_second": 27.108, + "eval_steps_per_second": 1.139, + "step": 810 + }, + { + "epoch": 1.023989898989899, + "grad_norm": 3.03643536567688, + "learning_rate": 3.6950420954162774e-06, + "loss": 1.4928, + "step": 811 + }, + { + "epoch": 1.0252525252525253, + "grad_norm": 2.841055154800415, + "learning_rate": 3.6927034611786723e-06, + "loss": 1.5273, + "step": 812 + }, + { + "epoch": 1.0265151515151516, + "grad_norm": 2.84134578704834, + "learning_rate": 3.690364826941067e-06, + "loss": 1.5443, + "step": 813 + }, + { + "epoch": 1.0277777777777777, + "grad_norm": 2.837758779525757, + "learning_rate": 3.6880261927034615e-06, + "loss": 1.5586, + "step": 814 + }, + { + "epoch": 1.029040404040404, + "grad_norm": 2.731412410736084, + "learning_rate": 3.6856875584658564e-06, + "loss": 1.4465, + "step": 815 + }, + { + "epoch": 1.0303030303030303, + "grad_norm": 2.8979599475860596, + "learning_rate": 3.683348924228251e-06, + "loss": 1.5663, + "step": 816 + }, + { + "epoch": 1.0315656565656566, + "grad_norm": 2.7740373611450195, + "learning_rate": 3.6810102899906456e-06, + "loss": 1.4578, + "step": 817 + }, + { + "epoch": 1.0328282828282829, + "grad_norm": 2.8984005451202393, + "learning_rate": 3.6786716557530405e-06, + "loss": 1.493, + "step": 818 + }, + { + "epoch": 1.0340909090909092, + "grad_norm": 2.880680561065674, + "learning_rate": 3.6763330215154353e-06, + "loss": 1.4858, + "step": 819 + }, + { + "epoch": 1.0353535353535352, + "grad_norm": 3.0220961570739746, + "learning_rate": 3.67399438727783e-06, + "loss": 1.5009, + "step": 820 + }, + { + "epoch": 1.0353535353535352, + "eval_loss": 1.7245674133300781, + "eval_runtime": 37.0235, + "eval_samples_per_second": 27.01, + "eval_steps_per_second": 1.134, + "step": 820 + }, + { + "epoch": 1.0366161616161615, + "grad_norm": 2.7905962467193604, + "learning_rate": 3.6716557530402246e-06, + "loss": 1.4728, + "step": 821 + }, + { + "epoch": 1.0378787878787878, + "grad_norm": 2.856935501098633, + "learning_rate": 3.6693171188026194e-06, + "loss": 1.521, + "step": 822 + }, + { + "epoch": 1.0391414141414141, + "grad_norm": 2.956584930419922, + "learning_rate": 3.6669784845650143e-06, + "loss": 1.4725, + "step": 823 + }, + { + "epoch": 1.0404040404040404, + "grad_norm": 3.130960702896118, + "learning_rate": 3.6646398503274087e-06, + "loss": 1.6127, + "step": 824 + }, + { + "epoch": 1.0416666666666667, + "grad_norm": 2.8855528831481934, + "learning_rate": 3.6623012160898035e-06, + "loss": 1.5078, + "step": 825 + }, + { + "epoch": 1.0429292929292928, + "grad_norm": 2.8030972480773926, + "learning_rate": 3.659962581852199e-06, + "loss": 1.4635, + "step": 826 + }, + { + "epoch": 1.0441919191919191, + "grad_norm": 3.061593770980835, + "learning_rate": 3.6576239476145936e-06, + "loss": 1.4944, + "step": 827 + }, + { + "epoch": 1.0454545454545454, + "grad_norm": 2.9983572959899902, + "learning_rate": 3.6552853133769885e-06, + "loss": 1.505, + "step": 828 + }, + { + "epoch": 1.0467171717171717, + "grad_norm": 2.7681095600128174, + "learning_rate": 3.652946679139383e-06, + "loss": 1.4774, + "step": 829 + }, + { + "epoch": 1.047979797979798, + "grad_norm": 3.0530946254730225, + "learning_rate": 3.6506080449017777e-06, + "loss": 1.4676, + "step": 830 + }, + { + "epoch": 1.047979797979798, + "eval_loss": 1.7242811918258667, + "eval_runtime": 37.0752, + "eval_samples_per_second": 26.972, + "eval_steps_per_second": 1.133, + "step": 830 + }, + { + "epoch": 1.0492424242424243, + "grad_norm": 2.8383126258850098, + "learning_rate": 3.6482694106641726e-06, + "loss": 1.486, + "step": 831 + }, + { + "epoch": 1.0505050505050506, + "grad_norm": 3.095928907394409, + "learning_rate": 3.6459307764265674e-06, + "loss": 1.5331, + "step": 832 + }, + { + "epoch": 1.0517676767676767, + "grad_norm": 2.9225456714630127, + "learning_rate": 3.643592142188962e-06, + "loss": 1.556, + "step": 833 + }, + { + "epoch": 1.053030303030303, + "grad_norm": 2.9037210941314697, + "learning_rate": 3.6412535079513567e-06, + "loss": 1.4834, + "step": 834 + }, + { + "epoch": 1.0542929292929293, + "grad_norm": 2.915022373199463, + "learning_rate": 3.6389148737137515e-06, + "loss": 1.5346, + "step": 835 + }, + { + "epoch": 1.0555555555555556, + "grad_norm": 3.250411033630371, + "learning_rate": 3.636576239476146e-06, + "loss": 1.4885, + "step": 836 + }, + { + "epoch": 1.0568181818181819, + "grad_norm": 2.759061336517334, + "learning_rate": 3.634237605238541e-06, + "loss": 1.456, + "step": 837 + }, + { + "epoch": 1.0580808080808082, + "grad_norm": 2.955981731414795, + "learning_rate": 3.6318989710009356e-06, + "loss": 1.4889, + "step": 838 + }, + { + "epoch": 1.0593434343434343, + "grad_norm": 3.243338108062744, + "learning_rate": 3.6295603367633305e-06, + "loss": 1.5207, + "step": 839 + }, + { + "epoch": 1.0606060606060606, + "grad_norm": 3.075840950012207, + "learning_rate": 3.627221702525725e-06, + "loss": 1.4906, + "step": 840 + }, + { + "epoch": 1.0606060606060606, + "eval_loss": 1.7290290594100952, + "eval_runtime": 37.097, + "eval_samples_per_second": 26.956, + "eval_steps_per_second": 1.132, + "step": 840 + }, + { + "epoch": 1.0618686868686869, + "grad_norm": 3.149125099182129, + "learning_rate": 3.62488306828812e-06, + "loss": 1.4755, + "step": 841 + }, + { + "epoch": 1.0631313131313131, + "grad_norm": 3.099318504333496, + "learning_rate": 3.622544434050515e-06, + "loss": 1.534, + "step": 842 + }, + { + "epoch": 1.0643939393939394, + "grad_norm": 2.8942148685455322, + "learning_rate": 3.62020579981291e-06, + "loss": 1.4844, + "step": 843 + }, + { + "epoch": 1.0656565656565657, + "grad_norm": 3.142664670944214, + "learning_rate": 3.6178671655753047e-06, + "loss": 1.5179, + "step": 844 + }, + { + "epoch": 1.0669191919191918, + "grad_norm": 3.106552839279175, + "learning_rate": 3.615528531337699e-06, + "loss": 1.4667, + "step": 845 + }, + { + "epoch": 1.0681818181818181, + "grad_norm": 2.8507936000823975, + "learning_rate": 3.613189897100094e-06, + "loss": 1.4463, + "step": 846 + }, + { + "epoch": 1.0694444444444444, + "grad_norm": 3.2198901176452637, + "learning_rate": 3.610851262862489e-06, + "loss": 1.5146, + "step": 847 + }, + { + "epoch": 1.0707070707070707, + "grad_norm": 3.146388292312622, + "learning_rate": 3.6085126286248832e-06, + "loss": 1.4927, + "step": 848 + }, + { + "epoch": 1.071969696969697, + "grad_norm": 2.8304760456085205, + "learning_rate": 3.606173994387278e-06, + "loss": 1.4971, + "step": 849 + }, + { + "epoch": 1.0732323232323233, + "grad_norm": 3.1714529991149902, + "learning_rate": 3.603835360149673e-06, + "loss": 1.5418, + "step": 850 + }, + { + "epoch": 1.0732323232323233, + "eval_loss": 1.725188970565796, + "eval_runtime": 37.117, + "eval_samples_per_second": 26.942, + "eval_steps_per_second": 1.132, + "step": 850 + }, + { + "epoch": 1.0744949494949494, + "grad_norm": 3.06162428855896, + "learning_rate": 3.6014967259120677e-06, + "loss": 1.4959, + "step": 851 + }, + { + "epoch": 1.0757575757575757, + "grad_norm": 2.812206268310547, + "learning_rate": 3.599158091674462e-06, + "loss": 1.5249, + "step": 852 + }, + { + "epoch": 1.077020202020202, + "grad_norm": 3.236210584640503, + "learning_rate": 3.596819457436857e-06, + "loss": 1.5166, + "step": 853 + }, + { + "epoch": 1.0782828282828283, + "grad_norm": 2.8397765159606934, + "learning_rate": 3.594480823199252e-06, + "loss": 1.5043, + "step": 854 + }, + { + "epoch": 1.0795454545454546, + "grad_norm": 2.9215919971466064, + "learning_rate": 3.5921421889616463e-06, + "loss": 1.5235, + "step": 855 + }, + { + "epoch": 1.0808080808080809, + "grad_norm": 2.8219058513641357, + "learning_rate": 3.589803554724041e-06, + "loss": 1.5091, + "step": 856 + }, + { + "epoch": 1.0820707070707072, + "grad_norm": 2.9344675540924072, + "learning_rate": 3.5874649204864364e-06, + "loss": 1.5355, + "step": 857 + }, + { + "epoch": 1.0833333333333333, + "grad_norm": 2.949094533920288, + "learning_rate": 3.5851262862488312e-06, + "loss": 1.5273, + "step": 858 + }, + { + "epoch": 1.0845959595959596, + "grad_norm": 2.9801411628723145, + "learning_rate": 3.582787652011226e-06, + "loss": 1.5078, + "step": 859 + }, + { + "epoch": 1.0858585858585859, + "grad_norm": 2.802182912826538, + "learning_rate": 3.5804490177736205e-06, + "loss": 1.5015, + "step": 860 + }, + { + "epoch": 1.0858585858585859, + "eval_loss": 1.7209502458572388, + "eval_runtime": 37.1238, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 860 + }, + { + "epoch": 1.0871212121212122, + "grad_norm": 2.857457160949707, + "learning_rate": 3.5781103835360153e-06, + "loss": 1.5057, + "step": 861 + }, + { + "epoch": 1.0883838383838385, + "grad_norm": 2.794478178024292, + "learning_rate": 3.57577174929841e-06, + "loss": 1.4594, + "step": 862 + }, + { + "epoch": 1.0896464646464645, + "grad_norm": 3.0028598308563232, + "learning_rate": 3.573433115060805e-06, + "loss": 1.4666, + "step": 863 + }, + { + "epoch": 1.0909090909090908, + "grad_norm": 2.7701406478881836, + "learning_rate": 3.5710944808231994e-06, + "loss": 1.4596, + "step": 864 + }, + { + "epoch": 1.0921717171717171, + "grad_norm": 2.9705960750579834, + "learning_rate": 3.5687558465855943e-06, + "loss": 1.5092, + "step": 865 + }, + { + "epoch": 1.0934343434343434, + "grad_norm": 2.9820821285247803, + "learning_rate": 3.566417212347989e-06, + "loss": 1.5246, + "step": 866 + }, + { + "epoch": 1.0946969696969697, + "grad_norm": 2.988356113433838, + "learning_rate": 3.5640785781103835e-06, + "loss": 1.5035, + "step": 867 + }, + { + "epoch": 1.095959595959596, + "grad_norm": 2.901599884033203, + "learning_rate": 3.5617399438727784e-06, + "loss": 1.5097, + "step": 868 + }, + { + "epoch": 1.0972222222222223, + "grad_norm": 2.9128799438476562, + "learning_rate": 3.5594013096351732e-06, + "loss": 1.5424, + "step": 869 + }, + { + "epoch": 1.0984848484848484, + "grad_norm": 2.719289779663086, + "learning_rate": 3.557062675397568e-06, + "loss": 1.4313, + "step": 870 + }, + { + "epoch": 1.0984848484848484, + "eval_loss": 1.7159698009490967, + "eval_runtime": 37.122, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 870 + }, + { + "epoch": 1.0997474747474747, + "grad_norm": 3.0080976486206055, + "learning_rate": 3.5547240411599625e-06, + "loss": 1.4483, + "step": 871 + }, + { + "epoch": 1.101010101010101, + "grad_norm": 2.856523036956787, + "learning_rate": 3.5523854069223578e-06, + "loss": 1.5463, + "step": 872 + }, + { + "epoch": 1.1022727272727273, + "grad_norm": 2.8585214614868164, + "learning_rate": 3.5500467726847526e-06, + "loss": 1.5391, + "step": 873 + }, + { + "epoch": 1.1035353535353536, + "grad_norm": 2.8440098762512207, + "learning_rate": 3.5477081384471474e-06, + "loss": 1.499, + "step": 874 + }, + { + "epoch": 1.10479797979798, + "grad_norm": 2.8114542961120605, + "learning_rate": 3.545369504209542e-06, + "loss": 1.4691, + "step": 875 + }, + { + "epoch": 1.106060606060606, + "grad_norm": 2.980273962020874, + "learning_rate": 3.5430308699719367e-06, + "loss": 1.4541, + "step": 876 + }, + { + "epoch": 1.1073232323232323, + "grad_norm": 3.0053417682647705, + "learning_rate": 3.5406922357343315e-06, + "loss": 1.4895, + "step": 877 + }, + { + "epoch": 1.1085858585858586, + "grad_norm": 2.924046754837036, + "learning_rate": 3.5383536014967264e-06, + "loss": 1.5186, + "step": 878 + }, + { + "epoch": 1.1098484848484849, + "grad_norm": 2.969461679458618, + "learning_rate": 3.536014967259121e-06, + "loss": 1.5213, + "step": 879 + }, + { + "epoch": 1.1111111111111112, + "grad_norm": 2.8936445713043213, + "learning_rate": 3.5336763330215156e-06, + "loss": 1.5585, + "step": 880 + }, + { + "epoch": 1.1111111111111112, + "eval_loss": 1.7150688171386719, + "eval_runtime": 37.1105, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 880 + }, + { + "epoch": 1.1123737373737375, + "grad_norm": 2.992034435272217, + "learning_rate": 3.5313376987839105e-06, + "loss": 1.5264, + "step": 881 + }, + { + "epoch": 1.1136363636363635, + "grad_norm": 3.0453898906707764, + "learning_rate": 3.5289990645463053e-06, + "loss": 1.5482, + "step": 882 + }, + { + "epoch": 1.1148989898989898, + "grad_norm": 2.7480897903442383, + "learning_rate": 3.5266604303086998e-06, + "loss": 1.4851, + "step": 883 + }, + { + "epoch": 1.1161616161616161, + "grad_norm": 2.9408013820648193, + "learning_rate": 3.5243217960710946e-06, + "loss": 1.536, + "step": 884 + }, + { + "epoch": 1.1174242424242424, + "grad_norm": 2.854438543319702, + "learning_rate": 3.5219831618334894e-06, + "loss": 1.5814, + "step": 885 + }, + { + "epoch": 1.1186868686868687, + "grad_norm": 2.9060397148132324, + "learning_rate": 3.519644527595884e-06, + "loss": 1.5019, + "step": 886 + }, + { + "epoch": 1.119949494949495, + "grad_norm": 2.8741681575775146, + "learning_rate": 3.517305893358279e-06, + "loss": 1.4713, + "step": 887 + }, + { + "epoch": 1.121212121212121, + "grad_norm": 2.972510576248169, + "learning_rate": 3.514967259120674e-06, + "loss": 1.5104, + "step": 888 + }, + { + "epoch": 1.1224747474747474, + "grad_norm": 2.8113558292388916, + "learning_rate": 3.512628624883069e-06, + "loss": 1.4973, + "step": 889 + }, + { + "epoch": 1.1237373737373737, + "grad_norm": 2.9687416553497314, + "learning_rate": 3.5102899906454637e-06, + "loss": 1.4774, + "step": 890 + }, + { + "epoch": 1.1237373737373737, + "eval_loss": 1.7169743776321411, + "eval_runtime": 37.0911, + "eval_samples_per_second": 26.961, + "eval_steps_per_second": 1.132, + "step": 890 + }, + { + "epoch": 1.125, + "grad_norm": 2.911409854888916, + "learning_rate": 3.507951356407858e-06, + "loss": 1.4536, + "step": 891 + }, + { + "epoch": 1.1262626262626263, + "grad_norm": 2.8209526538848877, + "learning_rate": 3.505612722170253e-06, + "loss": 1.5357, + "step": 892 + }, + { + "epoch": 1.1275252525252526, + "grad_norm": 2.811971426010132, + "learning_rate": 3.5032740879326478e-06, + "loss": 1.4889, + "step": 893 + }, + { + "epoch": 1.128787878787879, + "grad_norm": 2.969878673553467, + "learning_rate": 3.5009354536950426e-06, + "loss": 1.4758, + "step": 894 + }, + { + "epoch": 1.130050505050505, + "grad_norm": 2.858635425567627, + "learning_rate": 3.498596819457437e-06, + "loss": 1.5355, + "step": 895 + }, + { + "epoch": 1.1313131313131313, + "grad_norm": 3.018164873123169, + "learning_rate": 3.496258185219832e-06, + "loss": 1.5421, + "step": 896 + }, + { + "epoch": 1.1325757575757576, + "grad_norm": 2.846277952194214, + "learning_rate": 3.4939195509822267e-06, + "loss": 1.4936, + "step": 897 + }, + { + "epoch": 1.1338383838383839, + "grad_norm": 2.965245246887207, + "learning_rate": 3.491580916744621e-06, + "loss": 1.4903, + "step": 898 + }, + { + "epoch": 1.1351010101010102, + "grad_norm": 2.8500895500183105, + "learning_rate": 3.489242282507016e-06, + "loss": 1.5152, + "step": 899 + }, + { + "epoch": 1.1363636363636362, + "grad_norm": 2.8242156505584717, + "learning_rate": 3.486903648269411e-06, + "loss": 1.5531, + "step": 900 + }, + { + "epoch": 1.1363636363636362, + "eval_loss": 1.7180876731872559, + "eval_runtime": 37.4145, + "eval_samples_per_second": 26.728, + "eval_steps_per_second": 1.123, + "step": 900 + }, + { + "epoch": 1.1376262626262625, + "grad_norm": 2.8846635818481445, + "learning_rate": 3.4845650140318057e-06, + "loss": 1.5314, + "step": 901 + }, + { + "epoch": 1.1388888888888888, + "grad_norm": 2.9443984031677246, + "learning_rate": 3.4822263797942e-06, + "loss": 1.4699, + "step": 902 + }, + { + "epoch": 1.1401515151515151, + "grad_norm": 2.7435739040374756, + "learning_rate": 3.4798877455565953e-06, + "loss": 1.4949, + "step": 903 + }, + { + "epoch": 1.1414141414141414, + "grad_norm": 2.979102373123169, + "learning_rate": 3.47754911131899e-06, + "loss": 1.5689, + "step": 904 + }, + { + "epoch": 1.1426767676767677, + "grad_norm": 2.9814586639404297, + "learning_rate": 3.475210477081385e-06, + "loss": 1.5183, + "step": 905 + }, + { + "epoch": 1.143939393939394, + "grad_norm": 2.8380966186523438, + "learning_rate": 3.4728718428437794e-06, + "loss": 1.5177, + "step": 906 + }, + { + "epoch": 1.14520202020202, + "grad_norm": 2.8478410243988037, + "learning_rate": 3.4705332086061743e-06, + "loss": 1.4787, + "step": 907 + }, + { + "epoch": 1.1464646464646464, + "grad_norm": 2.8247780799865723, + "learning_rate": 3.468194574368569e-06, + "loss": 1.425, + "step": 908 + }, + { + "epoch": 1.1477272727272727, + "grad_norm": 2.869649887084961, + "learning_rate": 3.465855940130964e-06, + "loss": 1.4619, + "step": 909 + }, + { + "epoch": 1.148989898989899, + "grad_norm": 2.9439175128936768, + "learning_rate": 3.4635173058933584e-06, + "loss": 1.5194, + "step": 910 + }, + { + "epoch": 1.148989898989899, + "eval_loss": 1.7227450609207153, + "eval_runtime": 36.9723, + "eval_samples_per_second": 27.047, + "eval_steps_per_second": 1.136, + "step": 910 + }, + { + "epoch": 1.1502525252525253, + "grad_norm": 3.0219292640686035, + "learning_rate": 3.4611786716557532e-06, + "loss": 1.5878, + "step": 911 + }, + { + "epoch": 1.1515151515151516, + "grad_norm": 2.988475799560547, + "learning_rate": 3.458840037418148e-06, + "loss": 1.5349, + "step": 912 + }, + { + "epoch": 1.1527777777777777, + "grad_norm": 2.708740234375, + "learning_rate": 3.456501403180543e-06, + "loss": 1.4637, + "step": 913 + }, + { + "epoch": 1.154040404040404, + "grad_norm": 3.0844545364379883, + "learning_rate": 3.4541627689429373e-06, + "loss": 1.546, + "step": 914 + }, + { + "epoch": 1.1553030303030303, + "grad_norm": 2.800487518310547, + "learning_rate": 3.451824134705332e-06, + "loss": 1.5374, + "step": 915 + }, + { + "epoch": 1.1565656565656566, + "grad_norm": 2.8701722621917725, + "learning_rate": 3.449485500467727e-06, + "loss": 1.4549, + "step": 916 + }, + { + "epoch": 1.1578282828282829, + "grad_norm": 3.036418914794922, + "learning_rate": 3.4471468662301214e-06, + "loss": 1.5253, + "step": 917 + }, + { + "epoch": 1.1590909090909092, + "grad_norm": 3.0300986766815186, + "learning_rate": 3.4448082319925167e-06, + "loss": 1.4727, + "step": 918 + }, + { + "epoch": 1.1603535353535355, + "grad_norm": 2.8652496337890625, + "learning_rate": 3.4424695977549116e-06, + "loss": 1.4842, + "step": 919 + }, + { + "epoch": 1.1616161616161615, + "grad_norm": 2.9140491485595703, + "learning_rate": 3.4401309635173064e-06, + "loss": 1.4968, + "step": 920 + }, + { + "epoch": 1.1616161616161615, + "eval_loss": 1.7184255123138428, + "eval_runtime": 37.1186, + "eval_samples_per_second": 26.941, + "eval_steps_per_second": 1.132, + "step": 920 + }, + { + "epoch": 1.1628787878787878, + "grad_norm": 2.9708127975463867, + "learning_rate": 3.4377923292797012e-06, + "loss": 1.4817, + "step": 921 + }, + { + "epoch": 1.1641414141414141, + "grad_norm": 2.954814910888672, + "learning_rate": 3.4354536950420957e-06, + "loss": 1.4937, + "step": 922 + }, + { + "epoch": 1.1654040404040404, + "grad_norm": 2.803945779800415, + "learning_rate": 3.4331150608044905e-06, + "loss": 1.527, + "step": 923 + }, + { + "epoch": 1.1666666666666667, + "grad_norm": 2.863250255584717, + "learning_rate": 3.4307764265668853e-06, + "loss": 1.4922, + "step": 924 + }, + { + "epoch": 1.1679292929292928, + "grad_norm": 2.8030712604522705, + "learning_rate": 3.4284377923292798e-06, + "loss": 1.4427, + "step": 925 + }, + { + "epoch": 1.1691919191919191, + "grad_norm": 2.944110155105591, + "learning_rate": 3.4260991580916746e-06, + "loss": 1.4435, + "step": 926 + }, + { + "epoch": 1.1704545454545454, + "grad_norm": 2.9672467708587646, + "learning_rate": 3.4237605238540695e-06, + "loss": 1.5005, + "step": 927 + }, + { + "epoch": 1.1717171717171717, + "grad_norm": 2.920205593109131, + "learning_rate": 3.4214218896164643e-06, + "loss": 1.5303, + "step": 928 + }, + { + "epoch": 1.172979797979798, + "grad_norm": 3.1062121391296387, + "learning_rate": 3.4190832553788587e-06, + "loss": 1.5395, + "step": 929 + }, + { + "epoch": 1.1742424242424243, + "grad_norm": 3.1106555461883545, + "learning_rate": 3.4167446211412536e-06, + "loss": 1.5038, + "step": 930 + }, + { + "epoch": 1.1742424242424243, + "eval_loss": 1.7127314805984497, + "eval_runtime": 37.1193, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 930 + }, + { + "epoch": 1.1755050505050506, + "grad_norm": 2.7898054122924805, + "learning_rate": 3.4144059869036484e-06, + "loss": 1.5096, + "step": 931 + }, + { + "epoch": 1.1767676767676767, + "grad_norm": 2.881694793701172, + "learning_rate": 3.4120673526660432e-06, + "loss": 1.4556, + "step": 932 + }, + { + "epoch": 1.178030303030303, + "grad_norm": 3.131551742553711, + "learning_rate": 3.4097287184284385e-06, + "loss": 1.5487, + "step": 933 + }, + { + "epoch": 1.1792929292929293, + "grad_norm": 2.9734153747558594, + "learning_rate": 3.407390084190833e-06, + "loss": 1.439, + "step": 934 + }, + { + "epoch": 1.1805555555555556, + "grad_norm": 3.009657144546509, + "learning_rate": 3.4050514499532278e-06, + "loss": 1.462, + "step": 935 + }, + { + "epoch": 1.1818181818181819, + "grad_norm": 2.8378758430480957, + "learning_rate": 3.4027128157156226e-06, + "loss": 1.5083, + "step": 936 + }, + { + "epoch": 1.1830808080808082, + "grad_norm": 2.923969268798828, + "learning_rate": 3.400374181478017e-06, + "loss": 1.5844, + "step": 937 + }, + { + "epoch": 1.1843434343434343, + "grad_norm": 3.2384307384490967, + "learning_rate": 3.398035547240412e-06, + "loss": 1.5175, + "step": 938 + }, + { + "epoch": 1.1856060606060606, + "grad_norm": 3.024385929107666, + "learning_rate": 3.3956969130028067e-06, + "loss": 1.4756, + "step": 939 + }, + { + "epoch": 1.1868686868686869, + "grad_norm": 3.088745594024658, + "learning_rate": 3.3933582787652016e-06, + "loss": 1.4914, + "step": 940 + }, + { + "epoch": 1.1868686868686869, + "eval_loss": 1.705490231513977, + "eval_runtime": 37.1291, + "eval_samples_per_second": 26.933, + "eval_steps_per_second": 1.131, + "step": 940 + }, + { + "epoch": 1.1881313131313131, + "grad_norm": 2.898893117904663, + "learning_rate": 3.391019644527596e-06, + "loss": 1.5002, + "step": 941 + }, + { + "epoch": 1.1893939393939394, + "grad_norm": 2.800103187561035, + "learning_rate": 3.388681010289991e-06, + "loss": 1.4842, + "step": 942 + }, + { + "epoch": 1.1906565656565657, + "grad_norm": 3.0287230014801025, + "learning_rate": 3.3863423760523857e-06, + "loss": 1.5454, + "step": 943 + }, + { + "epoch": 1.1919191919191918, + "grad_norm": 2.9379546642303467, + "learning_rate": 3.3840037418147805e-06, + "loss": 1.5427, + "step": 944 + }, + { + "epoch": 1.1931818181818181, + "grad_norm": 2.879699468612671, + "learning_rate": 3.381665107577175e-06, + "loss": 1.4804, + "step": 945 + }, + { + "epoch": 1.1944444444444444, + "grad_norm": 2.8708608150482178, + "learning_rate": 3.3793264733395698e-06, + "loss": 1.479, + "step": 946 + }, + { + "epoch": 1.1957070707070707, + "grad_norm": 2.8513312339782715, + "learning_rate": 3.3769878391019646e-06, + "loss": 1.5343, + "step": 947 + }, + { + "epoch": 1.196969696969697, + "grad_norm": 2.8715100288391113, + "learning_rate": 3.37464920486436e-06, + "loss": 1.5114, + "step": 948 + }, + { + "epoch": 1.1982323232323233, + "grad_norm": 2.819213628768921, + "learning_rate": 3.3723105706267543e-06, + "loss": 1.4873, + "step": 949 + }, + { + "epoch": 1.1994949494949494, + "grad_norm": 2.946578025817871, + "learning_rate": 3.369971936389149e-06, + "loss": 1.5378, + "step": 950 + }, + { + "epoch": 1.1994949494949494, + "eval_loss": 1.7094330787658691, + "eval_runtime": 37.1453, + "eval_samples_per_second": 26.921, + "eval_steps_per_second": 1.131, + "step": 950 + }, + { + "epoch": 1.2007575757575757, + "grad_norm": 2.8170437812805176, + "learning_rate": 3.367633302151544e-06, + "loss": 1.4416, + "step": 951 + }, + { + "epoch": 1.202020202020202, + "grad_norm": 2.9654042720794678, + "learning_rate": 3.365294667913939e-06, + "loss": 1.5115, + "step": 952 + }, + { + "epoch": 1.2032828282828283, + "grad_norm": 2.9138922691345215, + "learning_rate": 3.3629560336763333e-06, + "loss": 1.5106, + "step": 953 + }, + { + "epoch": 1.2045454545454546, + "grad_norm": 2.989961862564087, + "learning_rate": 3.360617399438728e-06, + "loss": 1.5278, + "step": 954 + }, + { + "epoch": 1.2058080808080809, + "grad_norm": 2.9620680809020996, + "learning_rate": 3.358278765201123e-06, + "loss": 1.4949, + "step": 955 + }, + { + "epoch": 1.2070707070707072, + "grad_norm": 2.945505380630493, + "learning_rate": 3.3559401309635174e-06, + "loss": 1.5018, + "step": 956 + }, + { + "epoch": 1.2083333333333333, + "grad_norm": 2.806464433670044, + "learning_rate": 3.353601496725912e-06, + "loss": 1.4951, + "step": 957 + }, + { + "epoch": 1.2095959595959596, + "grad_norm": 2.7847559452056885, + "learning_rate": 3.351262862488307e-06, + "loss": 1.4259, + "step": 958 + }, + { + "epoch": 1.2108585858585859, + "grad_norm": 2.760003089904785, + "learning_rate": 3.348924228250702e-06, + "loss": 1.4757, + "step": 959 + }, + { + "epoch": 1.2121212121212122, + "grad_norm": 3.037522792816162, + "learning_rate": 3.3465855940130963e-06, + "loss": 1.5462, + "step": 960 + }, + { + "epoch": 1.2121212121212122, + "eval_loss": 1.7071956396102905, + "eval_runtime": 37.1351, + "eval_samples_per_second": 26.929, + "eval_steps_per_second": 1.131, + "step": 960 + }, + { + "epoch": 1.2133838383838385, + "grad_norm": 2.7755539417266846, + "learning_rate": 3.344246959775491e-06, + "loss": 1.453, + "step": 961 + }, + { + "epoch": 1.2146464646464645, + "grad_norm": 2.910369634628296, + "learning_rate": 3.341908325537886e-06, + "loss": 1.4914, + "step": 962 + }, + { + "epoch": 1.2159090909090908, + "grad_norm": 2.8999392986297607, + "learning_rate": 3.339569691300281e-06, + "loss": 1.5389, + "step": 963 + }, + { + "epoch": 1.2171717171717171, + "grad_norm": 2.891995906829834, + "learning_rate": 3.337231057062676e-06, + "loss": 1.4915, + "step": 964 + }, + { + "epoch": 1.2184343434343434, + "grad_norm": 2.814624309539795, + "learning_rate": 3.3348924228250705e-06, + "loss": 1.4936, + "step": 965 + }, + { + "epoch": 1.2196969696969697, + "grad_norm": 2.971163034439087, + "learning_rate": 3.3325537885874654e-06, + "loss": 1.4813, + "step": 966 + }, + { + "epoch": 1.220959595959596, + "grad_norm": 2.934155225753784, + "learning_rate": 3.33021515434986e-06, + "loss": 1.5162, + "step": 967 + }, + { + "epoch": 1.2222222222222223, + "grad_norm": 3.007847309112549, + "learning_rate": 3.3278765201122546e-06, + "loss": 1.4741, + "step": 968 + }, + { + "epoch": 1.2234848484848484, + "grad_norm": 2.977207660675049, + "learning_rate": 3.3255378858746495e-06, + "loss": 1.5186, + "step": 969 + }, + { + "epoch": 1.2247474747474747, + "grad_norm": 2.926567316055298, + "learning_rate": 3.3231992516370443e-06, + "loss": 1.5111, + "step": 970 + }, + { + "epoch": 1.2247474747474747, + "eval_loss": 1.7068836688995361, + "eval_runtime": 37.1419, + "eval_samples_per_second": 26.924, + "eval_steps_per_second": 1.131, + "step": 970 + }, + { + "epoch": 1.226010101010101, + "grad_norm": 3.063671350479126, + "learning_rate": 3.320860617399439e-06, + "loss": 1.4376, + "step": 971 + }, + { + "epoch": 1.2272727272727273, + "grad_norm": 2.7263271808624268, + "learning_rate": 3.3185219831618336e-06, + "loss": 1.4555, + "step": 972 + }, + { + "epoch": 1.2285353535353536, + "grad_norm": 2.912658929824829, + "learning_rate": 3.3161833489242284e-06, + "loss": 1.5116, + "step": 973 + }, + { + "epoch": 1.22979797979798, + "grad_norm": 3.2216858863830566, + "learning_rate": 3.3138447146866233e-06, + "loss": 1.5124, + "step": 974 + }, + { + "epoch": 1.231060606060606, + "grad_norm": 2.9138946533203125, + "learning_rate": 3.311506080449018e-06, + "loss": 1.4917, + "step": 975 + }, + { + "epoch": 1.2323232323232323, + "grad_norm": 3.0519120693206787, + "learning_rate": 3.3091674462114125e-06, + "loss": 1.4986, + "step": 976 + }, + { + "epoch": 1.2335858585858586, + "grad_norm": 2.8186402320861816, + "learning_rate": 3.3068288119738074e-06, + "loss": 1.4835, + "step": 977 + }, + { + "epoch": 1.2348484848484849, + "grad_norm": 2.799189329147339, + "learning_rate": 3.304490177736202e-06, + "loss": 1.5069, + "step": 978 + }, + { + "epoch": 1.2361111111111112, + "grad_norm": 3.010010004043579, + "learning_rate": 3.3021515434985975e-06, + "loss": 1.4276, + "step": 979 + }, + { + "epoch": 1.2373737373737375, + "grad_norm": 3.153592109680176, + "learning_rate": 3.299812909260992e-06, + "loss": 1.5459, + "step": 980 + }, + { + "epoch": 1.2373737373737375, + "eval_loss": 1.7070198059082031, + "eval_runtime": 37.1397, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 980 + }, + { + "epoch": 1.2386363636363638, + "grad_norm": 2.7818424701690674, + "learning_rate": 3.2974742750233867e-06, + "loss": 1.5045, + "step": 981 + }, + { + "epoch": 1.2398989898989898, + "grad_norm": 2.9211390018463135, + "learning_rate": 3.2951356407857816e-06, + "loss": 1.4966, + "step": 982 + }, + { + "epoch": 1.2411616161616161, + "grad_norm": 2.979743242263794, + "learning_rate": 3.2927970065481764e-06, + "loss": 1.4754, + "step": 983 + }, + { + "epoch": 1.2424242424242424, + "grad_norm": 2.8476738929748535, + "learning_rate": 3.290458372310571e-06, + "loss": 1.4619, + "step": 984 + }, + { + "epoch": 1.2436868686868687, + "grad_norm": 2.8304896354675293, + "learning_rate": 3.2881197380729657e-06, + "loss": 1.4192, + "step": 985 + }, + { + "epoch": 1.244949494949495, + "grad_norm": 3.3150880336761475, + "learning_rate": 3.2857811038353605e-06, + "loss": 1.515, + "step": 986 + }, + { + "epoch": 1.246212121212121, + "grad_norm": 3.049398422241211, + "learning_rate": 3.283442469597755e-06, + "loss": 1.5382, + "step": 987 + }, + { + "epoch": 1.2474747474747474, + "grad_norm": 3.0910539627075195, + "learning_rate": 3.2811038353601498e-06, + "loss": 1.5458, + "step": 988 + }, + { + "epoch": 1.2487373737373737, + "grad_norm": 3.002788782119751, + "learning_rate": 3.2787652011225446e-06, + "loss": 1.4668, + "step": 989 + }, + { + "epoch": 1.25, + "grad_norm": 2.9342103004455566, + "learning_rate": 3.2764265668849395e-06, + "loss": 1.5187, + "step": 990 + }, + { + "epoch": 1.25, + "eval_loss": 1.7082287073135376, + "eval_runtime": 37.1399, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 990 + }, + { + "epoch": 1.2512626262626263, + "grad_norm": 2.7596492767333984, + "learning_rate": 3.274087932647334e-06, + "loss": 1.5191, + "step": 991 + }, + { + "epoch": 1.2525252525252526, + "grad_norm": 2.949368953704834, + "learning_rate": 3.2717492984097287e-06, + "loss": 1.4776, + "step": 992 + }, + { + "epoch": 1.253787878787879, + "grad_norm": 2.7099363803863525, + "learning_rate": 3.2694106641721236e-06, + "loss": 1.455, + "step": 993 + }, + { + "epoch": 1.255050505050505, + "grad_norm": 2.8631885051727295, + "learning_rate": 3.267072029934519e-06, + "loss": 1.4394, + "step": 994 + }, + { + "epoch": 1.2563131313131313, + "grad_norm": 3.09389591217041, + "learning_rate": 3.2647333956969137e-06, + "loss": 1.4958, + "step": 995 + }, + { + "epoch": 1.2575757575757576, + "grad_norm": 3.108271360397339, + "learning_rate": 3.262394761459308e-06, + "loss": 1.5108, + "step": 996 + }, + { + "epoch": 1.2588383838383839, + "grad_norm": 3.079176902770996, + "learning_rate": 3.260056127221703e-06, + "loss": 1.4749, + "step": 997 + }, + { + "epoch": 1.2601010101010102, + "grad_norm": 3.0458056926727295, + "learning_rate": 3.257717492984098e-06, + "loss": 1.4911, + "step": 998 + }, + { + "epoch": 1.2613636363636362, + "grad_norm": 2.952151298522949, + "learning_rate": 3.2553788587464922e-06, + "loss": 1.4842, + "step": 999 + }, + { + "epoch": 1.2626262626262625, + "grad_norm": 2.7918665409088135, + "learning_rate": 3.253040224508887e-06, + "loss": 1.4684, + "step": 1000 + }, + { + "epoch": 1.2626262626262625, + "eval_loss": 1.7100834846496582, + "eval_runtime": 37.3594, + "eval_samples_per_second": 26.767, + "eval_steps_per_second": 1.124, + "step": 1000 + }, + { + "epoch": 1.2638888888888888, + "grad_norm": 2.798064947128296, + "learning_rate": 3.250701590271282e-06, + "loss": 1.4295, + "step": 1001 + }, + { + "epoch": 1.2651515151515151, + "grad_norm": 2.852234363555908, + "learning_rate": 3.2483629560336767e-06, + "loss": 1.5039, + "step": 1002 + }, + { + "epoch": 1.2664141414141414, + "grad_norm": 2.762645721435547, + "learning_rate": 3.246024321796071e-06, + "loss": 1.4365, + "step": 1003 + }, + { + "epoch": 1.2676767676767677, + "grad_norm": 2.944326162338257, + "learning_rate": 3.243685687558466e-06, + "loss": 1.4849, + "step": 1004 + }, + { + "epoch": 1.268939393939394, + "grad_norm": 2.9298877716064453, + "learning_rate": 3.241347053320861e-06, + "loss": 1.4397, + "step": 1005 + }, + { + "epoch": 1.2702020202020203, + "grad_norm": 3.1259615421295166, + "learning_rate": 3.2390084190832553e-06, + "loss": 1.4819, + "step": 1006 + }, + { + "epoch": 1.2714646464646464, + "grad_norm": 3.114792823791504, + "learning_rate": 3.23666978484565e-06, + "loss": 1.4728, + "step": 1007 + }, + { + "epoch": 1.2727272727272727, + "grad_norm": 2.9800360202789307, + "learning_rate": 3.234331150608045e-06, + "loss": 1.4426, + "step": 1008 + }, + { + "epoch": 1.273989898989899, + "grad_norm": 2.8487257957458496, + "learning_rate": 3.2319925163704402e-06, + "loss": 1.461, + "step": 1009 + }, + { + "epoch": 1.2752525252525253, + "grad_norm": 2.800760507583618, + "learning_rate": 3.229653882132835e-06, + "loss": 1.4764, + "step": 1010 + }, + { + "epoch": 1.2752525252525253, + "eval_loss": 1.701302409172058, + "eval_runtime": 36.9359, + "eval_samples_per_second": 27.074, + "eval_steps_per_second": 1.137, + "step": 1010 + }, + { + "epoch": 1.2765151515151514, + "grad_norm": 2.851012706756592, + "learning_rate": 3.2273152478952295e-06, + "loss": 1.5124, + "step": 1011 + }, + { + "epoch": 1.2777777777777777, + "grad_norm": 2.9498190879821777, + "learning_rate": 3.2249766136576243e-06, + "loss": 1.4789, + "step": 1012 + }, + { + "epoch": 1.279040404040404, + "grad_norm": 2.80710768699646, + "learning_rate": 3.222637979420019e-06, + "loss": 1.5592, + "step": 1013 + }, + { + "epoch": 1.2803030303030303, + "grad_norm": 2.9525952339172363, + "learning_rate": 3.220299345182414e-06, + "loss": 1.4941, + "step": 1014 + }, + { + "epoch": 1.2815656565656566, + "grad_norm": 2.9770407676696777, + "learning_rate": 3.2179607109448084e-06, + "loss": 1.4437, + "step": 1015 + }, + { + "epoch": 1.2828282828282829, + "grad_norm": 3.2553279399871826, + "learning_rate": 3.2156220767072033e-06, + "loss": 1.5248, + "step": 1016 + }, + { + "epoch": 1.2840909090909092, + "grad_norm": 3.0591554641723633, + "learning_rate": 3.213283442469598e-06, + "loss": 1.5704, + "step": 1017 + }, + { + "epoch": 1.2853535353535355, + "grad_norm": 2.776968002319336, + "learning_rate": 3.2109448082319925e-06, + "loss": 1.4239, + "step": 1018 + }, + { + "epoch": 1.2866161616161615, + "grad_norm": 2.8728222846984863, + "learning_rate": 3.2086061739943874e-06, + "loss": 1.5115, + "step": 1019 + }, + { + "epoch": 1.2878787878787878, + "grad_norm": 2.907308340072632, + "learning_rate": 3.2062675397567822e-06, + "loss": 1.4764, + "step": 1020 + }, + { + "epoch": 1.2878787878787878, + "eval_loss": 1.7016223669052124, + "eval_runtime": 37.0728, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 1020 + }, + { + "epoch": 1.2891414141414141, + "grad_norm": 3.0686309337615967, + "learning_rate": 3.203928905519177e-06, + "loss": 1.5175, + "step": 1021 + }, + { + "epoch": 1.2904040404040404, + "grad_norm": 2.8636863231658936, + "learning_rate": 3.2015902712815715e-06, + "loss": 1.5025, + "step": 1022 + }, + { + "epoch": 1.2916666666666667, + "grad_norm": 2.8539068698883057, + "learning_rate": 3.1992516370439663e-06, + "loss": 1.4774, + "step": 1023 + }, + { + "epoch": 1.2929292929292928, + "grad_norm": 2.9830520153045654, + "learning_rate": 3.196913002806361e-06, + "loss": 1.533, + "step": 1024 + }, + { + "epoch": 1.2941919191919191, + "grad_norm": 2.8806374073028564, + "learning_rate": 3.1945743685687564e-06, + "loss": 1.5298, + "step": 1025 + }, + { + "epoch": 1.2954545454545454, + "grad_norm": 2.932870626449585, + "learning_rate": 3.1922357343311513e-06, + "loss": 1.5028, + "step": 1026 + }, + { + "epoch": 1.2967171717171717, + "grad_norm": 2.889601945877075, + "learning_rate": 3.1898971000935457e-06, + "loss": 1.4534, + "step": 1027 + }, + { + "epoch": 1.297979797979798, + "grad_norm": 2.930546522140503, + "learning_rate": 3.1875584658559405e-06, + "loss": 1.5047, + "step": 1028 + }, + { + "epoch": 1.2992424242424243, + "grad_norm": 3.0944340229034424, + "learning_rate": 3.1852198316183354e-06, + "loss": 1.5198, + "step": 1029 + }, + { + "epoch": 1.3005050505050506, + "grad_norm": 2.8930089473724365, + "learning_rate": 3.18288119738073e-06, + "loss": 1.4655, + "step": 1030 + }, + { + "epoch": 1.3005050505050506, + "eval_loss": 1.7008370161056519, + "eval_runtime": 37.1263, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 1030 + }, + { + "epoch": 1.3017676767676767, + "grad_norm": 2.888303518295288, + "learning_rate": 3.1805425631431246e-06, + "loss": 1.4249, + "step": 1031 + }, + { + "epoch": 1.303030303030303, + "grad_norm": 3.0144314765930176, + "learning_rate": 3.1782039289055195e-06, + "loss": 1.4963, + "step": 1032 + }, + { + "epoch": 1.3042929292929293, + "grad_norm": 3.0118296146392822, + "learning_rate": 3.1758652946679143e-06, + "loss": 1.5119, + "step": 1033 + }, + { + "epoch": 1.3055555555555556, + "grad_norm": 2.8148205280303955, + "learning_rate": 3.1735266604303087e-06, + "loss": 1.4921, + "step": 1034 + }, + { + "epoch": 1.3068181818181819, + "grad_norm": 2.811166286468506, + "learning_rate": 3.1711880261927036e-06, + "loss": 1.4689, + "step": 1035 + }, + { + "epoch": 1.308080808080808, + "grad_norm": 2.924088478088379, + "learning_rate": 3.1688493919550984e-06, + "loss": 1.5682, + "step": 1036 + }, + { + "epoch": 1.3093434343434343, + "grad_norm": 2.905988931655884, + "learning_rate": 3.166510757717493e-06, + "loss": 1.5104, + "step": 1037 + }, + { + "epoch": 1.3106060606060606, + "grad_norm": 2.972963571548462, + "learning_rate": 3.1641721234798877e-06, + "loss": 1.4543, + "step": 1038 + }, + { + "epoch": 1.3118686868686869, + "grad_norm": 2.8034229278564453, + "learning_rate": 3.1618334892422825e-06, + "loss": 1.45, + "step": 1039 + }, + { + "epoch": 1.3131313131313131, + "grad_norm": 2.8977267742156982, + "learning_rate": 3.159494855004678e-06, + "loss": 1.4531, + "step": 1040 + }, + { + "epoch": 1.3131313131313131, + "eval_loss": 1.7027597427368164, + "eval_runtime": 37.0629, + "eval_samples_per_second": 26.981, + "eval_steps_per_second": 1.133, + "step": 1040 + }, + { + "epoch": 1.3143939393939394, + "grad_norm": 3.106973886489868, + "learning_rate": 3.1571562207670726e-06, + "loss": 1.4723, + "step": 1041 + }, + { + "epoch": 1.3156565656565657, + "grad_norm": 2.7628490924835205, + "learning_rate": 3.154817586529467e-06, + "loss": 1.4902, + "step": 1042 + }, + { + "epoch": 1.316919191919192, + "grad_norm": 3.119751453399658, + "learning_rate": 3.152478952291862e-06, + "loss": 1.5326, + "step": 1043 + }, + { + "epoch": 1.3181818181818181, + "grad_norm": 3.1470835208892822, + "learning_rate": 3.1501403180542568e-06, + "loss": 1.4626, + "step": 1044 + }, + { + "epoch": 1.3194444444444444, + "grad_norm": 2.866586208343506, + "learning_rate": 3.1478016838166516e-06, + "loss": 1.5006, + "step": 1045 + }, + { + "epoch": 1.3207070707070707, + "grad_norm": 2.878464698791504, + "learning_rate": 3.145463049579046e-06, + "loss": 1.4727, + "step": 1046 + }, + { + "epoch": 1.321969696969697, + "grad_norm": 2.8567826747894287, + "learning_rate": 3.143124415341441e-06, + "loss": 1.4202, + "step": 1047 + }, + { + "epoch": 1.3232323232323233, + "grad_norm": 2.8718442916870117, + "learning_rate": 3.1407857811038357e-06, + "loss": 1.473, + "step": 1048 + }, + { + "epoch": 1.3244949494949494, + "grad_norm": 3.003040313720703, + "learning_rate": 3.13844714686623e-06, + "loss": 1.4784, + "step": 1049 + }, + { + "epoch": 1.3257575757575757, + "grad_norm": 3.080288887023926, + "learning_rate": 3.136108512628625e-06, + "loss": 1.4514, + "step": 1050 + }, + { + "epoch": 1.3257575757575757, + "eval_loss": 1.704396367073059, + "eval_runtime": 37.0757, + "eval_samples_per_second": 26.972, + "eval_steps_per_second": 1.133, + "step": 1050 + }, + { + "epoch": 1.327020202020202, + "grad_norm": 2.8699166774749756, + "learning_rate": 3.13376987839102e-06, + "loss": 1.4966, + "step": 1051 + }, + { + "epoch": 1.3282828282828283, + "grad_norm": 2.979238271713257, + "learning_rate": 3.1314312441534146e-06, + "loss": 1.4422, + "step": 1052 + }, + { + "epoch": 1.3295454545454546, + "grad_norm": 2.8612024784088135, + "learning_rate": 3.129092609915809e-06, + "loss": 1.4982, + "step": 1053 + }, + { + "epoch": 1.3308080808080809, + "grad_norm": 2.7759032249450684, + "learning_rate": 3.126753975678204e-06, + "loss": 1.413, + "step": 1054 + }, + { + "epoch": 1.3320707070707072, + "grad_norm": 2.9772675037384033, + "learning_rate": 3.124415341440599e-06, + "loss": 1.4789, + "step": 1055 + }, + { + "epoch": 1.3333333333333333, + "grad_norm": 2.9382526874542236, + "learning_rate": 3.122076707202994e-06, + "loss": 1.4547, + "step": 1056 + }, + { + "epoch": 1.3345959595959596, + "grad_norm": 2.85050630569458, + "learning_rate": 3.119738072965389e-06, + "loss": 1.4193, + "step": 1057 + }, + { + "epoch": 1.3358585858585859, + "grad_norm": 3.1148695945739746, + "learning_rate": 3.1173994387277833e-06, + "loss": 1.4527, + "step": 1058 + }, + { + "epoch": 1.3371212121212122, + "grad_norm": 3.3906028270721436, + "learning_rate": 3.115060804490178e-06, + "loss": 1.5273, + "step": 1059 + }, + { + "epoch": 1.3383838383838385, + "grad_norm": 2.9665908813476562, + "learning_rate": 3.112722170252573e-06, + "loss": 1.5237, + "step": 1060 + }, + { + "epoch": 1.3383838383838385, + "eval_loss": 1.7004541158676147, + "eval_runtime": 37.081, + "eval_samples_per_second": 26.968, + "eval_steps_per_second": 1.133, + "step": 1060 + }, + { + "epoch": 1.3396464646464645, + "grad_norm": 2.9809224605560303, + "learning_rate": 3.1103835360149674e-06, + "loss": 1.498, + "step": 1061 + }, + { + "epoch": 1.3409090909090908, + "grad_norm": 2.95265793800354, + "learning_rate": 3.1080449017773622e-06, + "loss": 1.511, + "step": 1062 + }, + { + "epoch": 1.3421717171717171, + "grad_norm": 3.0814359188079834, + "learning_rate": 3.105706267539757e-06, + "loss": 1.4448, + "step": 1063 + }, + { + "epoch": 1.3434343434343434, + "grad_norm": 2.8267405033111572, + "learning_rate": 3.103367633302152e-06, + "loss": 1.4648, + "step": 1064 + }, + { + "epoch": 1.3446969696969697, + "grad_norm": 3.08182954788208, + "learning_rate": 3.1010289990645463e-06, + "loss": 1.5245, + "step": 1065 + }, + { + "epoch": 1.345959595959596, + "grad_norm": 3.183209180831909, + "learning_rate": 3.098690364826941e-06, + "loss": 1.4388, + "step": 1066 + }, + { + "epoch": 1.3472222222222223, + "grad_norm": 3.2121284008026123, + "learning_rate": 3.096351730589336e-06, + "loss": 1.4328, + "step": 1067 + }, + { + "epoch": 1.3484848484848486, + "grad_norm": 3.093921661376953, + "learning_rate": 3.0940130963517304e-06, + "loss": 1.5176, + "step": 1068 + }, + { + "epoch": 1.3497474747474747, + "grad_norm": 2.977710008621216, + "learning_rate": 3.0916744621141253e-06, + "loss": 1.4153, + "step": 1069 + }, + { + "epoch": 1.351010101010101, + "grad_norm": 3.119001865386963, + "learning_rate": 3.08933582787652e-06, + "loss": 1.4879, + "step": 1070 + }, + { + "epoch": 1.351010101010101, + "eval_loss": 1.7006300687789917, + "eval_runtime": 37.0959, + "eval_samples_per_second": 26.957, + "eval_steps_per_second": 1.132, + "step": 1070 + }, + { + "epoch": 1.3522727272727273, + "grad_norm": 3.026405096054077, + "learning_rate": 3.0869971936389154e-06, + "loss": 1.483, + "step": 1071 + }, + { + "epoch": 1.3535353535353536, + "grad_norm": 2.770728826522827, + "learning_rate": 3.0846585594013102e-06, + "loss": 1.5175, + "step": 1072 + }, + { + "epoch": 1.3547979797979797, + "grad_norm": 2.7930657863616943, + "learning_rate": 3.0823199251637047e-06, + "loss": 1.4833, + "step": 1073 + }, + { + "epoch": 1.356060606060606, + "grad_norm": 2.9128804206848145, + "learning_rate": 3.0799812909260995e-06, + "loss": 1.5016, + "step": 1074 + }, + { + "epoch": 1.3573232323232323, + "grad_norm": 3.0432205200195312, + "learning_rate": 3.0776426566884943e-06, + "loss": 1.5211, + "step": 1075 + }, + { + "epoch": 1.3585858585858586, + "grad_norm": 2.9197583198547363, + "learning_rate": 3.075304022450889e-06, + "loss": 1.5065, + "step": 1076 + }, + { + "epoch": 1.3598484848484849, + "grad_norm": 2.944986581802368, + "learning_rate": 3.0729653882132836e-06, + "loss": 1.5015, + "step": 1077 + }, + { + "epoch": 1.3611111111111112, + "grad_norm": 2.7553489208221436, + "learning_rate": 3.0706267539756784e-06, + "loss": 1.4692, + "step": 1078 + }, + { + "epoch": 1.3623737373737375, + "grad_norm": 2.7607858180999756, + "learning_rate": 3.0682881197380733e-06, + "loss": 1.4494, + "step": 1079 + }, + { + "epoch": 1.3636363636363638, + "grad_norm": 2.7896173000335693, + "learning_rate": 3.0659494855004677e-06, + "loss": 1.4625, + "step": 1080 + }, + { + "epoch": 1.3636363636363638, + "eval_loss": 1.6998732089996338, + "eval_runtime": 37.1079, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 1080 + }, + { + "epoch": 1.3648989898989898, + "grad_norm": 3.020059823989868, + "learning_rate": 3.0636108512628626e-06, + "loss": 1.5331, + "step": 1081 + }, + { + "epoch": 1.3661616161616161, + "grad_norm": 2.805593729019165, + "learning_rate": 3.0612722170252574e-06, + "loss": 1.4867, + "step": 1082 + }, + { + "epoch": 1.3674242424242424, + "grad_norm": 2.965871810913086, + "learning_rate": 3.0589335827876522e-06, + "loss": 1.4494, + "step": 1083 + }, + { + "epoch": 1.3686868686868687, + "grad_norm": 2.8010783195495605, + "learning_rate": 3.0565949485500467e-06, + "loss": 1.4484, + "step": 1084 + }, + { + "epoch": 1.369949494949495, + "grad_norm": 2.985258102416992, + "learning_rate": 3.0542563143124415e-06, + "loss": 1.5029, + "step": 1085 + }, + { + "epoch": 1.371212121212121, + "grad_norm": 2.7528188228607178, + "learning_rate": 3.0519176800748368e-06, + "loss": 1.4695, + "step": 1086 + }, + { + "epoch": 1.3724747474747474, + "grad_norm": 2.9123780727386475, + "learning_rate": 3.0495790458372316e-06, + "loss": 1.5527, + "step": 1087 + }, + { + "epoch": 1.3737373737373737, + "grad_norm": 2.8589160442352295, + "learning_rate": 3.0472404115996265e-06, + "loss": 1.5104, + "step": 1088 + }, + { + "epoch": 1.375, + "grad_norm": 2.853546142578125, + "learning_rate": 3.044901777362021e-06, + "loss": 1.4498, + "step": 1089 + }, + { + "epoch": 1.3762626262626263, + "grad_norm": 2.7989797592163086, + "learning_rate": 3.0425631431244157e-06, + "loss": 1.4494, + "step": 1090 + }, + { + "epoch": 1.3762626262626263, + "eval_loss": 1.6923459768295288, + "eval_runtime": 37.2917, + "eval_samples_per_second": 26.816, + "eval_steps_per_second": 1.126, + "step": 1090 + }, + { + "epoch": 1.3775252525252526, + "grad_norm": 2.8750412464141846, + "learning_rate": 3.0402245088868106e-06, + "loss": 1.5026, + "step": 1091 + }, + { + "epoch": 1.378787878787879, + "grad_norm": 2.8185842037200928, + "learning_rate": 3.037885874649205e-06, + "loss": 1.4859, + "step": 1092 + }, + { + "epoch": 1.380050505050505, + "grad_norm": 2.736332893371582, + "learning_rate": 3.0355472404116e-06, + "loss": 1.4944, + "step": 1093 + }, + { + "epoch": 1.3813131313131313, + "grad_norm": 2.93930983543396, + "learning_rate": 3.0332086061739947e-06, + "loss": 1.4975, + "step": 1094 + }, + { + "epoch": 1.3825757575757576, + "grad_norm": 2.964146852493286, + "learning_rate": 3.0308699719363895e-06, + "loss": 1.4861, + "step": 1095 + }, + { + "epoch": 1.3838383838383839, + "grad_norm": 2.8321545124053955, + "learning_rate": 3.028531337698784e-06, + "loss": 1.4535, + "step": 1096 + }, + { + "epoch": 1.3851010101010102, + "grad_norm": 2.921372413635254, + "learning_rate": 3.0261927034611788e-06, + "loss": 1.5242, + "step": 1097 + }, + { + "epoch": 1.3863636363636362, + "grad_norm": 2.7309911251068115, + "learning_rate": 3.0238540692235736e-06, + "loss": 1.4105, + "step": 1098 + }, + { + "epoch": 1.3876262626262625, + "grad_norm": 2.768899917602539, + "learning_rate": 3.021515434985968e-06, + "loss": 1.4183, + "step": 1099 + }, + { + "epoch": 1.3888888888888888, + "grad_norm": 2.883399248123169, + "learning_rate": 3.019176800748363e-06, + "loss": 1.5307, + "step": 1100 + }, + { + "epoch": 1.3888888888888888, + "eval_loss": 1.6948082447052002, + "eval_runtime": 37.488, + "eval_samples_per_second": 26.675, + "eval_steps_per_second": 1.12, + "step": 1100 + } + ], + "logging_steps": 1, + "max_steps": 2376, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 3.429040350298112e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1100/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/latest b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/latest new file mode 100644 index 0000000000000000000000000000000000000000..873b08355ae55e303606c474ab5fba25f9e8262a --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/latest @@ -0,0 +1 @@ +global_step1200 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..8991b49e9c2a43fc527dab9e09ad8171f0cc5943 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/tokenizer_config.json @@ -0,0 +1,2086 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128257": { + "content": "🅂", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + }, + "128258": { + "content": "🄿", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..cc6ebf5f984e8c3f4419ec7a16712f6c18b01050 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/trainer_state.json @@ -0,0 +1,9393 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.5151515151515151, + "eval_steps": 10, + "global_step": 1200, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0012626262626262627, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8571, + "step": 1 + }, + { + "epoch": 0.0025252525252525255, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7758, + "step": 2 + }, + { + "epoch": 0.003787878787878788, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8383, + "step": 3 + }, + { + "epoch": 0.005050505050505051, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.9088, + "step": 4 + }, + { + "epoch": 0.006313131313131313, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8748, + "step": 5 + }, + { + "epoch": 0.007575757575757576, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8575, + "step": 6 + }, + { + "epoch": 0.008838383838383838, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8271, + "step": 7 + }, + { + "epoch": 0.010101010101010102, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8336, + "step": 8 + }, + { + "epoch": 0.011363636363636364, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8037, + "step": 9 + }, + { + "epoch": 0.012626262626262626, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8323, + "step": 10 + }, + { + "epoch": 0.012626262626262626, + "eval_loss": 2.8736937046051025, + "eval_runtime": 37.7802, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 10 + }, + { + "epoch": 0.013888888888888888, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8653, + "step": 11 + }, + { + "epoch": 0.015151515151515152, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8104, + "step": 12 + }, + { + "epoch": 0.016414141414141416, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8489, + "step": 13 + }, + { + "epoch": 0.017676767676767676, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8442, + "step": 14 + }, + { + "epoch": 0.01893939393939394, + "grad_norm": 5.356770992279053, + "learning_rate": 2.1008403361344538e-08, + "loss": 2.8207, + "step": 15 + }, + { + "epoch": 0.020202020202020204, + "grad_norm": 4.954462051391602, + "learning_rate": 4.2016806722689076e-08, + "loss": 2.8104, + "step": 16 + }, + { + "epoch": 0.021464646464646464, + "grad_norm": 5.248059272766113, + "learning_rate": 6.302521008403361e-08, + "loss": 2.8178, + "step": 17 + }, + { + "epoch": 0.022727272727272728, + "grad_norm": 5.277204513549805, + "learning_rate": 8.403361344537815e-08, + "loss": 2.8255, + "step": 18 + }, + { + "epoch": 0.023989898989898988, + "grad_norm": 5.344228267669678, + "learning_rate": 1.050420168067227e-07, + "loss": 2.8018, + "step": 19 + }, + { + "epoch": 0.025252525252525252, + "grad_norm": 5.004310131072998, + "learning_rate": 1.2605042016806723e-07, + "loss": 2.8337, + "step": 20 + }, + { + "epoch": 0.025252525252525252, + "eval_loss": 2.872413158416748, + "eval_runtime": 37.7426, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 20 + }, + { + "epoch": 0.026515151515151516, + "grad_norm": 5.209287166595459, + "learning_rate": 1.4705882352941178e-07, + "loss": 2.8246, + "step": 21 + }, + { + "epoch": 0.027777777777777776, + "grad_norm": 5.369327545166016, + "learning_rate": 1.680672268907563e-07, + "loss": 2.8791, + "step": 22 + }, + { + "epoch": 0.02904040404040404, + "grad_norm": 4.646446228027344, + "learning_rate": 1.8907563025210085e-07, + "loss": 2.8425, + "step": 23 + }, + { + "epoch": 0.030303030303030304, + "grad_norm": 4.949256420135498, + "learning_rate": 2.100840336134454e-07, + "loss": 2.8629, + "step": 24 + }, + { + "epoch": 0.03156565656565657, + "grad_norm": 4.709571838378906, + "learning_rate": 2.3109243697478996e-07, + "loss": 2.8264, + "step": 25 + }, + { + "epoch": 0.03282828282828283, + "grad_norm": 4.519557952880859, + "learning_rate": 2.5210084033613445e-07, + "loss": 2.8143, + "step": 26 + }, + { + "epoch": 0.03409090909090909, + "grad_norm": 4.445594310760498, + "learning_rate": 2.73109243697479e-07, + "loss": 2.8031, + "step": 27 + }, + { + "epoch": 0.03535353535353535, + "grad_norm": 4.26489782333374, + "learning_rate": 2.9411764705882356e-07, + "loss": 2.7719, + "step": 28 + }, + { + "epoch": 0.036616161616161616, + "grad_norm": 4.153639793395996, + "learning_rate": 3.151260504201681e-07, + "loss": 2.8251, + "step": 29 + }, + { + "epoch": 0.03787878787878788, + "grad_norm": 3.40531325340271, + "learning_rate": 3.361344537815126e-07, + "loss": 2.849, + "step": 30 + }, + { + "epoch": 0.03787878787878788, + "eval_loss": 2.825382709503174, + "eval_runtime": 37.0715, + "eval_samples_per_second": 26.975, + "eval_steps_per_second": 1.133, + "step": 30 + }, + { + "epoch": 0.039141414141414144, + "grad_norm": 3.4759507179260254, + "learning_rate": 3.5714285714285716e-07, + "loss": 2.7653, + "step": 31 + }, + { + "epoch": 0.04040404040404041, + "grad_norm": 3.4147374629974365, + "learning_rate": 3.781512605042017e-07, + "loss": 2.7916, + "step": 32 + }, + { + "epoch": 0.041666666666666664, + "grad_norm": 3.2775769233703613, + "learning_rate": 3.991596638655462e-07, + "loss": 2.8072, + "step": 33 + }, + { + "epoch": 0.04292929292929293, + "grad_norm": 3.3541715145111084, + "learning_rate": 4.201680672268908e-07, + "loss": 2.7944, + "step": 34 + }, + { + "epoch": 0.04419191919191919, + "grad_norm": 2.991971969604492, + "learning_rate": 4.4117647058823536e-07, + "loss": 2.7827, + "step": 35 + }, + { + "epoch": 0.045454545454545456, + "grad_norm": 2.9127862453460693, + "learning_rate": 4.621848739495799e-07, + "loss": 2.7332, + "step": 36 + }, + { + "epoch": 0.04671717171717172, + "grad_norm": 2.839707136154175, + "learning_rate": 4.831932773109245e-07, + "loss": 2.7568, + "step": 37 + }, + { + "epoch": 0.047979797979797977, + "grad_norm": 2.9395432472229004, + "learning_rate": 5.042016806722689e-07, + "loss": 2.7445, + "step": 38 + }, + { + "epoch": 0.04924242424242424, + "grad_norm": 2.869896650314331, + "learning_rate": 5.252100840336135e-07, + "loss": 2.7157, + "step": 39 + }, + { + "epoch": 0.050505050505050504, + "grad_norm": 2.9064571857452393, + "learning_rate": 5.46218487394958e-07, + "loss": 2.7049, + "step": 40 + }, + { + "epoch": 0.050505050505050504, + "eval_loss": 2.7477543354034424, + "eval_runtime": 37.0669, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 40 + }, + { + "epoch": 0.05176767676767677, + "grad_norm": 2.6264421939849854, + "learning_rate": 5.672268907563026e-07, + "loss": 2.7651, + "step": 41 + }, + { + "epoch": 0.05303030303030303, + "grad_norm": 2.518627405166626, + "learning_rate": 5.882352941176471e-07, + "loss": 2.7241, + "step": 42 + }, + { + "epoch": 0.054292929292929296, + "grad_norm": 2.521034002304077, + "learning_rate": 6.092436974789917e-07, + "loss": 2.6691, + "step": 43 + }, + { + "epoch": 0.05555555555555555, + "grad_norm": 2.4679362773895264, + "learning_rate": 6.302521008403362e-07, + "loss": 2.743, + "step": 44 + }, + { + "epoch": 0.056818181818181816, + "grad_norm": 2.5817418098449707, + "learning_rate": 6.512605042016808e-07, + "loss": 2.7067, + "step": 45 + }, + { + "epoch": 0.05808080808080808, + "grad_norm": 2.3704075813293457, + "learning_rate": 6.722689075630252e-07, + "loss": 2.6495, + "step": 46 + }, + { + "epoch": 0.059343434343434344, + "grad_norm": 2.2321906089782715, + "learning_rate": 6.932773109243699e-07, + "loss": 2.6606, + "step": 47 + }, + { + "epoch": 0.06060606060606061, + "grad_norm": 1.985407829284668, + "learning_rate": 7.142857142857143e-07, + "loss": 2.662, + "step": 48 + }, + { + "epoch": 0.06186868686868687, + "grad_norm": 1.945688247680664, + "learning_rate": 7.352941176470589e-07, + "loss": 2.6319, + "step": 49 + }, + { + "epoch": 0.06313131313131314, + "grad_norm": 1.9392870664596558, + "learning_rate": 7.563025210084034e-07, + "loss": 2.6094, + "step": 50 + }, + { + "epoch": 0.06313131313131314, + "eval_loss": 2.6554172039031982, + "eval_runtime": 37.0569, + "eval_samples_per_second": 26.986, + "eval_steps_per_second": 1.133, + "step": 50 + }, + { + "epoch": 0.06439393939393939, + "grad_norm": 1.9975577592849731, + "learning_rate": 7.77310924369748e-07, + "loss": 2.6045, + "step": 51 + }, + { + "epoch": 0.06565656565656566, + "grad_norm": 1.9733693599700928, + "learning_rate": 7.983193277310924e-07, + "loss": 2.649, + "step": 52 + }, + { + "epoch": 0.06691919191919192, + "grad_norm": 1.7324917316436768, + "learning_rate": 8.193277310924371e-07, + "loss": 2.5689, + "step": 53 + }, + { + "epoch": 0.06818181818181818, + "grad_norm": 1.7553645372390747, + "learning_rate": 8.403361344537816e-07, + "loss": 2.644, + "step": 54 + }, + { + "epoch": 0.06944444444444445, + "grad_norm": 1.8163801431655884, + "learning_rate": 8.613445378151261e-07, + "loss": 2.4943, + "step": 55 + }, + { + "epoch": 0.0707070707070707, + "grad_norm": 1.7717015743255615, + "learning_rate": 8.823529411764707e-07, + "loss": 2.6093, + "step": 56 + }, + { + "epoch": 0.07196969696969698, + "grad_norm": 1.7013663053512573, + "learning_rate": 9.033613445378152e-07, + "loss": 2.6376, + "step": 57 + }, + { + "epoch": 0.07323232323232323, + "grad_norm": 1.752943992614746, + "learning_rate": 9.243697478991598e-07, + "loss": 2.5172, + "step": 58 + }, + { + "epoch": 0.07449494949494949, + "grad_norm": 1.5362415313720703, + "learning_rate": 9.453781512605043e-07, + "loss": 2.6129, + "step": 59 + }, + { + "epoch": 0.07575757575757576, + "grad_norm": 1.514885425567627, + "learning_rate": 9.66386554621849e-07, + "loss": 2.6064, + "step": 60 + }, + { + "epoch": 0.07575757575757576, + "eval_loss": 2.5934340953826904, + "eval_runtime": 37.0782, + "eval_samples_per_second": 26.97, + "eval_steps_per_second": 1.133, + "step": 60 + }, + { + "epoch": 0.07702020202020202, + "grad_norm": 1.5915579795837402, + "learning_rate": 9.873949579831934e-07, + "loss": 2.5491, + "step": 61 + }, + { + "epoch": 0.07828282828282829, + "grad_norm": 1.601773977279663, + "learning_rate": 1.0084033613445378e-06, + "loss": 2.5685, + "step": 62 + }, + { + "epoch": 0.07954545454545454, + "grad_norm": 1.8327804803848267, + "learning_rate": 1.0294117647058825e-06, + "loss": 2.5763, + "step": 63 + }, + { + "epoch": 0.08080808080808081, + "grad_norm": 1.5276741981506348, + "learning_rate": 1.050420168067227e-06, + "loss": 2.5494, + "step": 64 + }, + { + "epoch": 0.08207070707070707, + "grad_norm": 1.7065422534942627, + "learning_rate": 1.0714285714285714e-06, + "loss": 2.449, + "step": 65 + }, + { + "epoch": 0.08333333333333333, + "grad_norm": 2.0551910400390625, + "learning_rate": 1.092436974789916e-06, + "loss": 2.4371, + "step": 66 + }, + { + "epoch": 0.0845959595959596, + "grad_norm": 2.154855728149414, + "learning_rate": 1.1134453781512607e-06, + "loss": 2.5417, + "step": 67 + }, + { + "epoch": 0.08585858585858586, + "grad_norm": 1.7922390699386597, + "learning_rate": 1.1344537815126051e-06, + "loss": 2.4754, + "step": 68 + }, + { + "epoch": 0.08712121212121213, + "grad_norm": 1.745632290840149, + "learning_rate": 1.1554621848739498e-06, + "loss": 2.4669, + "step": 69 + }, + { + "epoch": 0.08838383838383838, + "grad_norm": 1.8907713890075684, + "learning_rate": 1.1764705882352942e-06, + "loss": 2.4943, + "step": 70 + }, + { + "epoch": 0.08838383838383838, + "eval_loss": 2.518237829208374, + "eval_runtime": 37.074, + "eval_samples_per_second": 26.973, + "eval_steps_per_second": 1.133, + "step": 70 + }, + { + "epoch": 0.08964646464646464, + "grad_norm": 1.8178164958953857, + "learning_rate": 1.1974789915966389e-06, + "loss": 2.4891, + "step": 71 + }, + { + "epoch": 0.09090909090909091, + "grad_norm": 1.970969796180725, + "learning_rate": 1.2184873949579833e-06, + "loss": 2.4797, + "step": 72 + }, + { + "epoch": 0.09217171717171717, + "grad_norm": 2.0211431980133057, + "learning_rate": 1.2394957983193278e-06, + "loss": 2.4716, + "step": 73 + }, + { + "epoch": 0.09343434343434344, + "grad_norm": 2.650787591934204, + "learning_rate": 1.2605042016806724e-06, + "loss": 2.4915, + "step": 74 + }, + { + "epoch": 0.0946969696969697, + "grad_norm": 2.235499143600464, + "learning_rate": 1.2815126050420169e-06, + "loss": 2.3917, + "step": 75 + }, + { + "epoch": 0.09595959595959595, + "grad_norm": 2.245999813079834, + "learning_rate": 1.3025210084033615e-06, + "loss": 2.4942, + "step": 76 + }, + { + "epoch": 0.09722222222222222, + "grad_norm": 2.4414420127868652, + "learning_rate": 1.323529411764706e-06, + "loss": 2.4004, + "step": 77 + }, + { + "epoch": 0.09848484848484848, + "grad_norm": 2.279937982559204, + "learning_rate": 1.3445378151260504e-06, + "loss": 2.4043, + "step": 78 + }, + { + "epoch": 0.09974747474747475, + "grad_norm": 2.3830068111419678, + "learning_rate": 1.3655462184873949e-06, + "loss": 2.4096, + "step": 79 + }, + { + "epoch": 0.10101010101010101, + "grad_norm": 2.0119733810424805, + "learning_rate": 1.3865546218487397e-06, + "loss": 2.4227, + "step": 80 + }, + { + "epoch": 0.10101010101010101, + "eval_loss": 2.4316186904907227, + "eval_runtime": 37.0837, + "eval_samples_per_second": 26.966, + "eval_steps_per_second": 1.133, + "step": 80 + }, + { + "epoch": 0.10227272727272728, + "grad_norm": 2.1273374557495117, + "learning_rate": 1.4075630252100842e-06, + "loss": 2.3541, + "step": 81 + }, + { + "epoch": 0.10353535353535354, + "grad_norm": 2.6852898597717285, + "learning_rate": 1.4285714285714286e-06, + "loss": 2.3321, + "step": 82 + }, + { + "epoch": 0.10479797979797979, + "grad_norm": 2.2622721195220947, + "learning_rate": 1.4495798319327733e-06, + "loss": 2.3166, + "step": 83 + }, + { + "epoch": 0.10606060606060606, + "grad_norm": 2.98258900642395, + "learning_rate": 1.4705882352941177e-06, + "loss": 2.3733, + "step": 84 + }, + { + "epoch": 0.10732323232323232, + "grad_norm": 2.5140459537506104, + "learning_rate": 1.4915966386554624e-06, + "loss": 2.314, + "step": 85 + }, + { + "epoch": 0.10858585858585859, + "grad_norm": 2.637608051300049, + "learning_rate": 1.5126050420168068e-06, + "loss": 2.3479, + "step": 86 + }, + { + "epoch": 0.10984848484848485, + "grad_norm": 2.399508476257324, + "learning_rate": 1.5336134453781515e-06, + "loss": 2.2932, + "step": 87 + }, + { + "epoch": 0.1111111111111111, + "grad_norm": 3.0835793018341064, + "learning_rate": 1.554621848739496e-06, + "loss": 2.2886, + "step": 88 + }, + { + "epoch": 0.11237373737373738, + "grad_norm": 3.288769483566284, + "learning_rate": 1.5756302521008404e-06, + "loss": 2.2914, + "step": 89 + }, + { + "epoch": 0.11363636363636363, + "grad_norm": 2.538361072540283, + "learning_rate": 1.5966386554621848e-06, + "loss": 2.3247, + "step": 90 + }, + { + "epoch": 0.11363636363636363, + "eval_loss": 2.3510072231292725, + "eval_runtime": 37.0847, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 90 + }, + { + "epoch": 0.1148989898989899, + "grad_norm": 2.8016724586486816, + "learning_rate": 1.6176470588235297e-06, + "loss": 2.281, + "step": 91 + }, + { + "epoch": 0.11616161616161616, + "grad_norm": 2.9251859188079834, + "learning_rate": 1.6386554621848741e-06, + "loss": 2.2715, + "step": 92 + }, + { + "epoch": 0.11742424242424243, + "grad_norm": 2.595750570297241, + "learning_rate": 1.6596638655462186e-06, + "loss": 2.2127, + "step": 93 + }, + { + "epoch": 0.11868686868686869, + "grad_norm": 2.6989378929138184, + "learning_rate": 1.6806722689075632e-06, + "loss": 2.2623, + "step": 94 + }, + { + "epoch": 0.11994949494949494, + "grad_norm": 2.554863929748535, + "learning_rate": 1.7016806722689077e-06, + "loss": 2.2649, + "step": 95 + }, + { + "epoch": 0.12121212121212122, + "grad_norm": 2.7013020515441895, + "learning_rate": 1.7226890756302521e-06, + "loss": 2.2428, + "step": 96 + }, + { + "epoch": 0.12247474747474747, + "grad_norm": 2.7415151596069336, + "learning_rate": 1.7436974789915968e-06, + "loss": 2.1882, + "step": 97 + }, + { + "epoch": 0.12373737373737374, + "grad_norm": 2.672806739807129, + "learning_rate": 1.7647058823529414e-06, + "loss": 2.2302, + "step": 98 + }, + { + "epoch": 0.125, + "grad_norm": 2.9508326053619385, + "learning_rate": 1.7857142857142859e-06, + "loss": 2.1894, + "step": 99 + }, + { + "epoch": 0.12626262626262627, + "grad_norm": 2.4668166637420654, + "learning_rate": 1.8067226890756303e-06, + "loss": 2.2043, + "step": 100 + }, + { + "epoch": 0.12626262626262627, + "eval_loss": 2.2788307666778564, + "eval_runtime": 37.1156, + "eval_samples_per_second": 26.943, + "eval_steps_per_second": 1.132, + "step": 100 + }, + { + "epoch": 0.1275252525252525, + "grad_norm": 2.8241467475891113, + "learning_rate": 1.8277310924369748e-06, + "loss": 2.2109, + "step": 101 + }, + { + "epoch": 0.12878787878787878, + "grad_norm": 2.4179229736328125, + "learning_rate": 1.8487394957983196e-06, + "loss": 2.2193, + "step": 102 + }, + { + "epoch": 0.13005050505050506, + "grad_norm": 2.714315891265869, + "learning_rate": 1.869747899159664e-06, + "loss": 2.2474, + "step": 103 + }, + { + "epoch": 0.13131313131313133, + "grad_norm": 2.8560359477996826, + "learning_rate": 1.8907563025210085e-06, + "loss": 2.2205, + "step": 104 + }, + { + "epoch": 0.13257575757575757, + "grad_norm": 2.7780051231384277, + "learning_rate": 1.9117647058823528e-06, + "loss": 2.1834, + "step": 105 + }, + { + "epoch": 0.13383838383838384, + "grad_norm": 3.3773467540740967, + "learning_rate": 1.932773109243698e-06, + "loss": 2.2063, + "step": 106 + }, + { + "epoch": 0.1351010101010101, + "grad_norm": 2.927551746368408, + "learning_rate": 1.953781512605042e-06, + "loss": 2.1592, + "step": 107 + }, + { + "epoch": 0.13636363636363635, + "grad_norm": 2.8094165325164795, + "learning_rate": 1.9747899159663867e-06, + "loss": 2.2619, + "step": 108 + }, + { + "epoch": 0.13762626262626262, + "grad_norm": 2.5312564373016357, + "learning_rate": 1.9957983193277314e-06, + "loss": 2.2323, + "step": 109 + }, + { + "epoch": 0.1388888888888889, + "grad_norm": 3.175746202468872, + "learning_rate": 2.0168067226890756e-06, + "loss": 2.1776, + "step": 110 + }, + { + "epoch": 0.1388888888888889, + "eval_loss": 2.2208564281463623, + "eval_runtime": 37.3885, + "eval_samples_per_second": 26.746, + "eval_steps_per_second": 1.123, + "step": 110 + }, + { + "epoch": 0.14015151515151514, + "grad_norm": 2.660886287689209, + "learning_rate": 2.0378151260504203e-06, + "loss": 2.1626, + "step": 111 + }, + { + "epoch": 0.1414141414141414, + "grad_norm": 2.6948511600494385, + "learning_rate": 2.058823529411765e-06, + "loss": 2.1731, + "step": 112 + }, + { + "epoch": 0.14267676767676768, + "grad_norm": 2.826976776123047, + "learning_rate": 2.0798319327731096e-06, + "loss": 2.1865, + "step": 113 + }, + { + "epoch": 0.14393939393939395, + "grad_norm": 3.210033655166626, + "learning_rate": 2.100840336134454e-06, + "loss": 2.1325, + "step": 114 + }, + { + "epoch": 0.1452020202020202, + "grad_norm": 2.509060859680176, + "learning_rate": 2.1218487394957985e-06, + "loss": 2.0926, + "step": 115 + }, + { + "epoch": 0.14646464646464646, + "grad_norm": 3.51330304145813, + "learning_rate": 2.1428571428571427e-06, + "loss": 2.1577, + "step": 116 + }, + { + "epoch": 0.14772727272727273, + "grad_norm": 2.468395471572876, + "learning_rate": 2.163865546218488e-06, + "loss": 2.1, + "step": 117 + }, + { + "epoch": 0.14898989898989898, + "grad_norm": 3.081529378890991, + "learning_rate": 2.184873949579832e-06, + "loss": 2.1082, + "step": 118 + }, + { + "epoch": 0.15025252525252525, + "grad_norm": 2.8856358528137207, + "learning_rate": 2.2058823529411767e-06, + "loss": 2.1288, + "step": 119 + }, + { + "epoch": 0.15151515151515152, + "grad_norm": 3.323089838027954, + "learning_rate": 2.2268907563025214e-06, + "loss": 2.0933, + "step": 120 + }, + { + "epoch": 0.15151515151515152, + "eval_loss": 2.1834487915039062, + "eval_runtime": 37.4344, + "eval_samples_per_second": 26.713, + "eval_steps_per_second": 1.122, + "step": 120 + }, + { + "epoch": 0.1527777777777778, + "grad_norm": 2.8871560096740723, + "learning_rate": 2.2478991596638656e-06, + "loss": 2.1519, + "step": 121 + }, + { + "epoch": 0.15404040404040403, + "grad_norm": 3.0718588829040527, + "learning_rate": 2.2689075630252102e-06, + "loss": 2.0321, + "step": 122 + }, + { + "epoch": 0.1553030303030303, + "grad_norm": 2.820739507675171, + "learning_rate": 2.289915966386555e-06, + "loss": 2.0944, + "step": 123 + }, + { + "epoch": 0.15656565656565657, + "grad_norm": 3.1259310245513916, + "learning_rate": 2.3109243697478996e-06, + "loss": 2.0141, + "step": 124 + }, + { + "epoch": 0.15782828282828282, + "grad_norm": 3.303004741668701, + "learning_rate": 2.331932773109244e-06, + "loss": 2.1095, + "step": 125 + }, + { + "epoch": 0.1590909090909091, + "grad_norm": 3.1838674545288086, + "learning_rate": 2.3529411764705885e-06, + "loss": 2.0727, + "step": 126 + }, + { + "epoch": 0.16035353535353536, + "grad_norm": 3.1013786792755127, + "learning_rate": 2.3739495798319327e-06, + "loss": 2.0963, + "step": 127 + }, + { + "epoch": 0.16161616161616163, + "grad_norm": 2.803283214569092, + "learning_rate": 2.3949579831932778e-06, + "loss": 2.0551, + "step": 128 + }, + { + "epoch": 0.16287878787878787, + "grad_norm": 3.1616814136505127, + "learning_rate": 2.415966386554622e-06, + "loss": 2.0485, + "step": 129 + }, + { + "epoch": 0.16414141414141414, + "grad_norm": 3.333782196044922, + "learning_rate": 2.4369747899159667e-06, + "loss": 2.041, + "step": 130 + }, + { + "epoch": 0.16414141414141414, + "eval_loss": 2.150329351425171, + "eval_runtime": 37.494, + "eval_samples_per_second": 26.671, + "eval_steps_per_second": 1.12, + "step": 130 + }, + { + "epoch": 0.16540404040404041, + "grad_norm": 2.9714694023132324, + "learning_rate": 2.457983193277311e-06, + "loss": 2.0686, + "step": 131 + }, + { + "epoch": 0.16666666666666666, + "grad_norm": 3.216895818710327, + "learning_rate": 2.4789915966386555e-06, + "loss": 2.003, + "step": 132 + }, + { + "epoch": 0.16792929292929293, + "grad_norm": 2.9636809825897217, + "learning_rate": 2.5e-06, + "loss": 2.0534, + "step": 133 + }, + { + "epoch": 0.1691919191919192, + "grad_norm": 3.6477177143096924, + "learning_rate": 2.521008403361345e-06, + "loss": 2.077, + "step": 134 + }, + { + "epoch": 0.17045454545454544, + "grad_norm": 2.7437169551849365, + "learning_rate": 2.542016806722689e-06, + "loss": 2.0501, + "step": 135 + }, + { + "epoch": 0.1717171717171717, + "grad_norm": 3.2992663383483887, + "learning_rate": 2.5630252100840338e-06, + "loss": 2.0284, + "step": 136 + }, + { + "epoch": 0.17297979797979798, + "grad_norm": 3.0735950469970703, + "learning_rate": 2.5840336134453784e-06, + "loss": 2.025, + "step": 137 + }, + { + "epoch": 0.17424242424242425, + "grad_norm": 3.588418960571289, + "learning_rate": 2.605042016806723e-06, + "loss": 2.002, + "step": 138 + }, + { + "epoch": 0.1755050505050505, + "grad_norm": 2.8772335052490234, + "learning_rate": 2.6260504201680677e-06, + "loss": 1.9873, + "step": 139 + }, + { + "epoch": 0.17676767676767677, + "grad_norm": 2.9078969955444336, + "learning_rate": 2.647058823529412e-06, + "loss": 2.011, + "step": 140 + }, + { + "epoch": 0.17676767676767677, + "eval_loss": 2.1225531101226807, + "eval_runtime": 37.4531, + "eval_samples_per_second": 26.7, + "eval_steps_per_second": 1.121, + "step": 140 + }, + { + "epoch": 0.17803030303030304, + "grad_norm": 3.556654214859009, + "learning_rate": 2.6680672268907566e-06, + "loss": 2.0401, + "step": 141 + }, + { + "epoch": 0.17929292929292928, + "grad_norm": 2.9375200271606445, + "learning_rate": 2.689075630252101e-06, + "loss": 2.0277, + "step": 142 + }, + { + "epoch": 0.18055555555555555, + "grad_norm": 3.4019014835357666, + "learning_rate": 2.7100840336134455e-06, + "loss": 1.975, + "step": 143 + }, + { + "epoch": 0.18181818181818182, + "grad_norm": 2.7653748989105225, + "learning_rate": 2.7310924369747897e-06, + "loss": 2.0102, + "step": 144 + }, + { + "epoch": 0.1830808080808081, + "grad_norm": 3.4416797161102295, + "learning_rate": 2.752100840336135e-06, + "loss": 2.0093, + "step": 145 + }, + { + "epoch": 0.18434343434343434, + "grad_norm": 3.2531726360321045, + "learning_rate": 2.7731092436974795e-06, + "loss": 1.9852, + "step": 146 + }, + { + "epoch": 0.1856060606060606, + "grad_norm": 3.171502113342285, + "learning_rate": 2.7941176470588237e-06, + "loss": 1.9165, + "step": 147 + }, + { + "epoch": 0.18686868686868688, + "grad_norm": 3.114550828933716, + "learning_rate": 2.8151260504201684e-06, + "loss": 1.9516, + "step": 148 + }, + { + "epoch": 0.18813131313131312, + "grad_norm": 3.3601152896881104, + "learning_rate": 2.8361344537815126e-06, + "loss": 2.0183, + "step": 149 + }, + { + "epoch": 0.1893939393939394, + "grad_norm": 3.059622049331665, + "learning_rate": 2.8571428571428573e-06, + "loss": 1.9682, + "step": 150 + }, + { + "epoch": 0.1893939393939394, + "eval_loss": 2.0962483882904053, + "eval_runtime": 37.2364, + "eval_samples_per_second": 26.855, + "eval_steps_per_second": 1.128, + "step": 150 + }, + { + "epoch": 0.19065656565656566, + "grad_norm": 2.8749427795410156, + "learning_rate": 2.8781512605042015e-06, + "loss": 2.0089, + "step": 151 + }, + { + "epoch": 0.1919191919191919, + "grad_norm": 3.067267894744873, + "learning_rate": 2.8991596638655466e-06, + "loss": 1.9661, + "step": 152 + }, + { + "epoch": 0.19318181818181818, + "grad_norm": 2.7126266956329346, + "learning_rate": 2.9201680672268912e-06, + "loss": 1.982, + "step": 153 + }, + { + "epoch": 0.19444444444444445, + "grad_norm": 3.1330652236938477, + "learning_rate": 2.9411764705882355e-06, + "loss": 1.9753, + "step": 154 + }, + { + "epoch": 0.19570707070707072, + "grad_norm": 3.158313274383545, + "learning_rate": 2.96218487394958e-06, + "loss": 1.9431, + "step": 155 + }, + { + "epoch": 0.19696969696969696, + "grad_norm": 3.30806303024292, + "learning_rate": 2.9831932773109248e-06, + "loss": 1.9812, + "step": 156 + }, + { + "epoch": 0.19823232323232323, + "grad_norm": 3.0344884395599365, + "learning_rate": 3.004201680672269e-06, + "loss": 1.9677, + "step": 157 + }, + { + "epoch": 0.1994949494949495, + "grad_norm": 2.9032909870147705, + "learning_rate": 3.0252100840336137e-06, + "loss": 1.9132, + "step": 158 + }, + { + "epoch": 0.20075757575757575, + "grad_norm": 3.459934711456299, + "learning_rate": 3.046218487394958e-06, + "loss": 1.9581, + "step": 159 + }, + { + "epoch": 0.20202020202020202, + "grad_norm": 2.9455478191375732, + "learning_rate": 3.067226890756303e-06, + "loss": 1.9509, + "step": 160 + }, + { + "epoch": 0.20202020202020202, + "eval_loss": 2.074032783508301, + "eval_runtime": 37.1824, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 160 + }, + { + "epoch": 0.2032828282828283, + "grad_norm": 3.2347211837768555, + "learning_rate": 3.0882352941176476e-06, + "loss": 1.9743, + "step": 161 + }, + { + "epoch": 0.20454545454545456, + "grad_norm": 3.3152947425842285, + "learning_rate": 3.109243697478992e-06, + "loss": 1.8971, + "step": 162 + }, + { + "epoch": 0.2058080808080808, + "grad_norm": 3.035548210144043, + "learning_rate": 3.1302521008403365e-06, + "loss": 1.8762, + "step": 163 + }, + { + "epoch": 0.20707070707070707, + "grad_norm": 3.1004066467285156, + "learning_rate": 3.1512605042016808e-06, + "loss": 1.9512, + "step": 164 + }, + { + "epoch": 0.20833333333333334, + "grad_norm": 3.2280282974243164, + "learning_rate": 3.1722689075630254e-06, + "loss": 1.8856, + "step": 165 + }, + { + "epoch": 0.20959595959595959, + "grad_norm": 3.021969795227051, + "learning_rate": 3.1932773109243696e-06, + "loss": 1.9127, + "step": 166 + }, + { + "epoch": 0.21085858585858586, + "grad_norm": 3.1647000312805176, + "learning_rate": 3.2142857142857147e-06, + "loss": 2.038, + "step": 167 + }, + { + "epoch": 0.21212121212121213, + "grad_norm": 3.0084872245788574, + "learning_rate": 3.2352941176470594e-06, + "loss": 1.9536, + "step": 168 + }, + { + "epoch": 0.21338383838383837, + "grad_norm": 2.951509952545166, + "learning_rate": 3.2563025210084036e-06, + "loss": 1.933, + "step": 169 + }, + { + "epoch": 0.21464646464646464, + "grad_norm": 3.472856283187866, + "learning_rate": 3.2773109243697483e-06, + "loss": 1.934, + "step": 170 + }, + { + "epoch": 0.21464646464646464, + "eval_loss": 2.0450048446655273, + "eval_runtime": 37.1669, + "eval_samples_per_second": 26.906, + "eval_steps_per_second": 1.13, + "step": 170 + }, + { + "epoch": 0.2159090909090909, + "grad_norm": 3.1104896068573, + "learning_rate": 3.2983193277310925e-06, + "loss": 1.9008, + "step": 171 + }, + { + "epoch": 0.21717171717171718, + "grad_norm": 2.847781181335449, + "learning_rate": 3.319327731092437e-06, + "loss": 1.9592, + "step": 172 + }, + { + "epoch": 0.21843434343434343, + "grad_norm": 3.384554624557495, + "learning_rate": 3.3403361344537814e-06, + "loss": 1.9087, + "step": 173 + }, + { + "epoch": 0.2196969696969697, + "grad_norm": 3.221564292907715, + "learning_rate": 3.3613445378151265e-06, + "loss": 1.9107, + "step": 174 + }, + { + "epoch": 0.22095959595959597, + "grad_norm": 3.30204439163208, + "learning_rate": 3.382352941176471e-06, + "loss": 1.8879, + "step": 175 + }, + { + "epoch": 0.2222222222222222, + "grad_norm": 3.549259662628174, + "learning_rate": 3.4033613445378154e-06, + "loss": 1.9592, + "step": 176 + }, + { + "epoch": 0.22348484848484848, + "grad_norm": 3.498199462890625, + "learning_rate": 3.42436974789916e-06, + "loss": 1.984, + "step": 177 + }, + { + "epoch": 0.22474747474747475, + "grad_norm": 3.576317548751831, + "learning_rate": 3.4453781512605043e-06, + "loss": 1.8897, + "step": 178 + }, + { + "epoch": 0.22601010101010102, + "grad_norm": 3.5337026119232178, + "learning_rate": 3.466386554621849e-06, + "loss": 1.9607, + "step": 179 + }, + { + "epoch": 0.22727272727272727, + "grad_norm": 3.342886209487915, + "learning_rate": 3.4873949579831936e-06, + "loss": 1.8797, + "step": 180 + }, + { + "epoch": 0.22727272727272727, + "eval_loss": 2.02158522605896, + "eval_runtime": 37.1382, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 180 + }, + { + "epoch": 0.22853535353535354, + "grad_norm": 2.941906213760376, + "learning_rate": 3.508403361344538e-06, + "loss": 1.9, + "step": 181 + }, + { + "epoch": 0.2297979797979798, + "grad_norm": 3.1913063526153564, + "learning_rate": 3.529411764705883e-06, + "loss": 1.8704, + "step": 182 + }, + { + "epoch": 0.23106060606060605, + "grad_norm": 3.0619900226593018, + "learning_rate": 3.550420168067227e-06, + "loss": 1.8851, + "step": 183 + }, + { + "epoch": 0.23232323232323232, + "grad_norm": 3.051288604736328, + "learning_rate": 3.5714285714285718e-06, + "loss": 1.9422, + "step": 184 + }, + { + "epoch": 0.2335858585858586, + "grad_norm": 2.999845266342163, + "learning_rate": 3.5924369747899164e-06, + "loss": 1.8757, + "step": 185 + }, + { + "epoch": 0.23484848484848486, + "grad_norm": 3.262355089187622, + "learning_rate": 3.6134453781512607e-06, + "loss": 1.9114, + "step": 186 + }, + { + "epoch": 0.2361111111111111, + "grad_norm": 3.0354340076446533, + "learning_rate": 3.6344537815126053e-06, + "loss": 1.8547, + "step": 187 + }, + { + "epoch": 0.23737373737373738, + "grad_norm": 2.9401369094848633, + "learning_rate": 3.6554621848739496e-06, + "loss": 1.8974, + "step": 188 + }, + { + "epoch": 0.23863636363636365, + "grad_norm": 3.0544443130493164, + "learning_rate": 3.6764705882352946e-06, + "loss": 1.8662, + "step": 189 + }, + { + "epoch": 0.2398989898989899, + "grad_norm": 2.7257750034332275, + "learning_rate": 3.6974789915966393e-06, + "loss": 1.8737, + "step": 190 + }, + { + "epoch": 0.2398989898989899, + "eval_loss": 2.0065128803253174, + "eval_runtime": 37.1296, + "eval_samples_per_second": 26.933, + "eval_steps_per_second": 1.131, + "step": 190 + }, + { + "epoch": 0.24116161616161616, + "grad_norm": 2.8592522144317627, + "learning_rate": 3.7184873949579835e-06, + "loss": 1.8984, + "step": 191 + }, + { + "epoch": 0.24242424242424243, + "grad_norm": 2.833958864212036, + "learning_rate": 3.739495798319328e-06, + "loss": 1.9084, + "step": 192 + }, + { + "epoch": 0.24368686868686867, + "grad_norm": 2.7842230796813965, + "learning_rate": 3.7605042016806724e-06, + "loss": 1.8843, + "step": 193 + }, + { + "epoch": 0.24494949494949494, + "grad_norm": 3.1309409141540527, + "learning_rate": 3.781512605042017e-06, + "loss": 1.8259, + "step": 194 + }, + { + "epoch": 0.24621212121212122, + "grad_norm": 3.120788335800171, + "learning_rate": 3.8025210084033613e-06, + "loss": 1.8951, + "step": 195 + }, + { + "epoch": 0.2474747474747475, + "grad_norm": 3.2228143215179443, + "learning_rate": 3.8235294117647055e-06, + "loss": 1.8608, + "step": 196 + }, + { + "epoch": 0.24873737373737373, + "grad_norm": 3.3362114429473877, + "learning_rate": 3.844537815126051e-06, + "loss": 1.9163, + "step": 197 + }, + { + "epoch": 0.25, + "grad_norm": 3.0973799228668213, + "learning_rate": 3.865546218487396e-06, + "loss": 1.911, + "step": 198 + }, + { + "epoch": 0.25126262626262624, + "grad_norm": 3.5567545890808105, + "learning_rate": 3.8865546218487395e-06, + "loss": 1.8795, + "step": 199 + }, + { + "epoch": 0.25252525252525254, + "grad_norm": 3.4758830070495605, + "learning_rate": 3.907563025210084e-06, + "loss": 1.8813, + "step": 200 + }, + { + "epoch": 0.25252525252525254, + "eval_loss": 1.9892102479934692, + "eval_runtime": 37.1311, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 200 + }, + { + "epoch": 0.2537878787878788, + "grad_norm": 3.5564475059509277, + "learning_rate": 3.928571428571429e-06, + "loss": 1.8469, + "step": 201 + }, + { + "epoch": 0.255050505050505, + "grad_norm": 3.7004899978637695, + "learning_rate": 3.9495798319327735e-06, + "loss": 1.873, + "step": 202 + }, + { + "epoch": 0.2563131313131313, + "grad_norm": 3.9285225868225098, + "learning_rate": 3.970588235294118e-06, + "loss": 1.8596, + "step": 203 + }, + { + "epoch": 0.25757575757575757, + "grad_norm": 3.1388907432556152, + "learning_rate": 3.991596638655463e-06, + "loss": 1.7991, + "step": 204 + }, + { + "epoch": 0.2588383838383838, + "grad_norm": 3.3797011375427246, + "learning_rate": 4.0126050420168075e-06, + "loss": 1.8682, + "step": 205 + }, + { + "epoch": 0.2601010101010101, + "grad_norm": 3.32183575630188, + "learning_rate": 4.033613445378151e-06, + "loss": 1.8396, + "step": 206 + }, + { + "epoch": 0.26136363636363635, + "grad_norm": 3.457291603088379, + "learning_rate": 4.054621848739496e-06, + "loss": 1.8314, + "step": 207 + }, + { + "epoch": 0.26262626262626265, + "grad_norm": 3.2906265258789062, + "learning_rate": 4.075630252100841e-06, + "loss": 1.844, + "step": 208 + }, + { + "epoch": 0.2638888888888889, + "grad_norm": 3.081913948059082, + "learning_rate": 4.096638655462185e-06, + "loss": 1.8198, + "step": 209 + }, + { + "epoch": 0.26515151515151514, + "grad_norm": 2.9697105884552, + "learning_rate": 4.11764705882353e-06, + "loss": 1.7885, + "step": 210 + }, + { + "epoch": 0.26515151515151514, + "eval_loss": 1.9749963283538818, + "eval_runtime": 36.962, + "eval_samples_per_second": 27.055, + "eval_steps_per_second": 1.136, + "step": 210 + }, + { + "epoch": 0.26641414141414144, + "grad_norm": 3.0732104778289795, + "learning_rate": 4.138655462184874e-06, + "loss": 1.8463, + "step": 211 + }, + { + "epoch": 0.2676767676767677, + "grad_norm": 3.4714958667755127, + "learning_rate": 4.159663865546219e-06, + "loss": 1.8537, + "step": 212 + }, + { + "epoch": 0.2689393939393939, + "grad_norm": 2.955864429473877, + "learning_rate": 4.180672268907564e-06, + "loss": 1.8566, + "step": 213 + }, + { + "epoch": 0.2702020202020202, + "grad_norm": 3.5314948558807373, + "learning_rate": 4.201680672268908e-06, + "loss": 1.7625, + "step": 214 + }, + { + "epoch": 0.27146464646464646, + "grad_norm": 2.8855810165405273, + "learning_rate": 4.222689075630252e-06, + "loss": 1.7968, + "step": 215 + }, + { + "epoch": 0.2727272727272727, + "grad_norm": 3.7866363525390625, + "learning_rate": 4.243697478991597e-06, + "loss": 1.8341, + "step": 216 + }, + { + "epoch": 0.273989898989899, + "grad_norm": 3.0352272987365723, + "learning_rate": 4.264705882352942e-06, + "loss": 1.8276, + "step": 217 + }, + { + "epoch": 0.27525252525252525, + "grad_norm": 3.4880247116088867, + "learning_rate": 4.2857142857142855e-06, + "loss": 1.8194, + "step": 218 + }, + { + "epoch": 0.2765151515151515, + "grad_norm": 3.6807212829589844, + "learning_rate": 4.306722689075631e-06, + "loss": 1.8495, + "step": 219 + }, + { + "epoch": 0.2777777777777778, + "grad_norm": 3.029017448425293, + "learning_rate": 4.327731092436976e-06, + "loss": 1.8125, + "step": 220 + }, + { + "epoch": 0.2777777777777778, + "eval_loss": 1.9633830785751343, + "eval_runtime": 37.1259, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 220 + }, + { + "epoch": 0.27904040404040403, + "grad_norm": 3.4820573329925537, + "learning_rate": 4.3487394957983194e-06, + "loss": 1.8159, + "step": 221 + }, + { + "epoch": 0.2803030303030303, + "grad_norm": 3.2057833671569824, + "learning_rate": 4.369747899159664e-06, + "loss": 1.8215, + "step": 222 + }, + { + "epoch": 0.2815656565656566, + "grad_norm": 3.477302312850952, + "learning_rate": 4.390756302521009e-06, + "loss": 1.7837, + "step": 223 + }, + { + "epoch": 0.2828282828282828, + "grad_norm": 2.8349595069885254, + "learning_rate": 4.411764705882353e-06, + "loss": 1.8555, + "step": 224 + }, + { + "epoch": 0.2840909090909091, + "grad_norm": 3.616478443145752, + "learning_rate": 4.432773109243698e-06, + "loss": 1.8067, + "step": 225 + }, + { + "epoch": 0.28535353535353536, + "grad_norm": 3.111969470977783, + "learning_rate": 4.453781512605043e-06, + "loss": 1.8322, + "step": 226 + }, + { + "epoch": 0.2866161616161616, + "grad_norm": 3.248695135116577, + "learning_rate": 4.474789915966387e-06, + "loss": 1.8004, + "step": 227 + }, + { + "epoch": 0.2878787878787879, + "grad_norm": 3.017082691192627, + "learning_rate": 4.495798319327731e-06, + "loss": 1.776, + "step": 228 + }, + { + "epoch": 0.28914141414141414, + "grad_norm": 3.6790099143981934, + "learning_rate": 4.516806722689076e-06, + "loss": 1.8176, + "step": 229 + }, + { + "epoch": 0.2904040404040404, + "grad_norm": 3.1513001918792725, + "learning_rate": 4.5378151260504205e-06, + "loss": 1.8016, + "step": 230 + }, + { + "epoch": 0.2904040404040404, + "eval_loss": 1.949769377708435, + "eval_runtime": 37.2573, + "eval_samples_per_second": 26.84, + "eval_steps_per_second": 1.127, + "step": 230 + }, + { + "epoch": 0.2916666666666667, + "grad_norm": 3.0559096336364746, + "learning_rate": 4.558823529411765e-06, + "loss": 1.7915, + "step": 231 + }, + { + "epoch": 0.29292929292929293, + "grad_norm": 3.0974984169006348, + "learning_rate": 4.57983193277311e-06, + "loss": 1.8277, + "step": 232 + }, + { + "epoch": 0.29419191919191917, + "grad_norm": 2.8514139652252197, + "learning_rate": 4.600840336134454e-06, + "loss": 1.7616, + "step": 233 + }, + { + "epoch": 0.29545454545454547, + "grad_norm": 3.355198383331299, + "learning_rate": 4.621848739495799e-06, + "loss": 1.841, + "step": 234 + }, + { + "epoch": 0.2967171717171717, + "grad_norm": 3.1467435359954834, + "learning_rate": 4.642857142857144e-06, + "loss": 1.7393, + "step": 235 + }, + { + "epoch": 0.29797979797979796, + "grad_norm": 4.065819263458252, + "learning_rate": 4.663865546218488e-06, + "loss": 1.7508, + "step": 236 + }, + { + "epoch": 0.29924242424242425, + "grad_norm": 3.2602460384368896, + "learning_rate": 4.684873949579832e-06, + "loss": 1.7415, + "step": 237 + }, + { + "epoch": 0.3005050505050505, + "grad_norm": 3.6884894371032715, + "learning_rate": 4.705882352941177e-06, + "loss": 1.7608, + "step": 238 + }, + { + "epoch": 0.30176767676767674, + "grad_norm": 3.9843380451202393, + "learning_rate": 4.7268907563025216e-06, + "loss": 1.8519, + "step": 239 + }, + { + "epoch": 0.30303030303030304, + "grad_norm": 3.2823405265808105, + "learning_rate": 4.747899159663865e-06, + "loss": 1.7689, + "step": 240 + }, + { + "epoch": 0.30303030303030304, + "eval_loss": 1.939724326133728, + "eval_runtime": 37.465, + "eval_samples_per_second": 26.692, + "eval_steps_per_second": 1.121, + "step": 240 + }, + { + "epoch": 0.3042929292929293, + "grad_norm": 3.4200189113616943, + "learning_rate": 4.768907563025211e-06, + "loss": 1.7944, + "step": 241 + }, + { + "epoch": 0.3055555555555556, + "grad_norm": 3.7541189193725586, + "learning_rate": 4.7899159663865555e-06, + "loss": 1.8151, + "step": 242 + }, + { + "epoch": 0.3068181818181818, + "grad_norm": 3.9103832244873047, + "learning_rate": 4.810924369747899e-06, + "loss": 1.7921, + "step": 243 + }, + { + "epoch": 0.30808080808080807, + "grad_norm": 3.2769975662231445, + "learning_rate": 4.831932773109244e-06, + "loss": 1.8408, + "step": 244 + }, + { + "epoch": 0.30934343434343436, + "grad_norm": 3.414168357849121, + "learning_rate": 4.852941176470589e-06, + "loss": 1.7937, + "step": 245 + }, + { + "epoch": 0.3106060606060606, + "grad_norm": 3.219625473022461, + "learning_rate": 4.873949579831933e-06, + "loss": 1.796, + "step": 246 + }, + { + "epoch": 0.31186868686868685, + "grad_norm": 3.2573912143707275, + "learning_rate": 4.894957983193277e-06, + "loss": 1.8052, + "step": 247 + }, + { + "epoch": 0.31313131313131315, + "grad_norm": 3.4201040267944336, + "learning_rate": 4.915966386554622e-06, + "loss": 1.8529, + "step": 248 + }, + { + "epoch": 0.3143939393939394, + "grad_norm": 3.7448058128356934, + "learning_rate": 4.936974789915967e-06, + "loss": 1.8183, + "step": 249 + }, + { + "epoch": 0.31565656565656564, + "grad_norm": 3.4108386039733887, + "learning_rate": 4.957983193277311e-06, + "loss": 1.7926, + "step": 250 + }, + { + "epoch": 0.31565656565656564, + "eval_loss": 1.9319047927856445, + "eval_runtime": 37.4745, + "eval_samples_per_second": 26.685, + "eval_steps_per_second": 1.121, + "step": 250 + }, + { + "epoch": 0.31691919191919193, + "grad_norm": 3.50185227394104, + "learning_rate": 4.978991596638656e-06, + "loss": 1.792, + "step": 251 + }, + { + "epoch": 0.3181818181818182, + "grad_norm": 3.225581407546997, + "learning_rate": 5e-06, + "loss": 1.7286, + "step": 252 + }, + { + "epoch": 0.3194444444444444, + "grad_norm": 3.3070669174194336, + "learning_rate": 4.997661365762395e-06, + "loss": 1.7957, + "step": 253 + }, + { + "epoch": 0.3207070707070707, + "grad_norm": 3.498079776763916, + "learning_rate": 4.99532273152479e-06, + "loss": 1.823, + "step": 254 + }, + { + "epoch": 0.32196969696969696, + "grad_norm": 3.316534996032715, + "learning_rate": 4.9929840972871845e-06, + "loss": 1.8249, + "step": 255 + }, + { + "epoch": 0.32323232323232326, + "grad_norm": 3.1641833782196045, + "learning_rate": 4.990645463049579e-06, + "loss": 1.7658, + "step": 256 + }, + { + "epoch": 0.3244949494949495, + "grad_norm": 3.0794270038604736, + "learning_rate": 4.988306828811974e-06, + "loss": 1.8146, + "step": 257 + }, + { + "epoch": 0.32575757575757575, + "grad_norm": 3.2410383224487305, + "learning_rate": 4.985968194574369e-06, + "loss": 1.7894, + "step": 258 + }, + { + "epoch": 0.32702020202020204, + "grad_norm": 3.0899274349212646, + "learning_rate": 4.983629560336764e-06, + "loss": 1.7381, + "step": 259 + }, + { + "epoch": 0.3282828282828283, + "grad_norm": 3.0132803916931152, + "learning_rate": 4.981290926099159e-06, + "loss": 1.7713, + "step": 260 + }, + { + "epoch": 0.3282828282828283, + "eval_loss": 1.9123589992523193, + "eval_runtime": 37.3201, + "eval_samples_per_second": 26.795, + "eval_steps_per_second": 1.125, + "step": 260 + }, + { + "epoch": 0.32954545454545453, + "grad_norm": 2.92564058303833, + "learning_rate": 4.9789522918615536e-06, + "loss": 1.7659, + "step": 261 + }, + { + "epoch": 0.33080808080808083, + "grad_norm": 3.1402878761291504, + "learning_rate": 4.976613657623948e-06, + "loss": 1.8439, + "step": 262 + }, + { + "epoch": 0.33207070707070707, + "grad_norm": 3.170881509780884, + "learning_rate": 4.974275023386343e-06, + "loss": 1.7605, + "step": 263 + }, + { + "epoch": 0.3333333333333333, + "grad_norm": 3.014850378036499, + "learning_rate": 4.971936389148738e-06, + "loss": 1.7899, + "step": 264 + }, + { + "epoch": 0.3345959595959596, + "grad_norm": 3.169818878173828, + "learning_rate": 4.969597754911132e-06, + "loss": 1.8263, + "step": 265 + }, + { + "epoch": 0.33585858585858586, + "grad_norm": 2.8709678649902344, + "learning_rate": 4.967259120673527e-06, + "loss": 1.7834, + "step": 266 + }, + { + "epoch": 0.3371212121212121, + "grad_norm": 2.7527096271514893, + "learning_rate": 4.964920486435922e-06, + "loss": 1.8106, + "step": 267 + }, + { + "epoch": 0.3383838383838384, + "grad_norm": 2.9960227012634277, + "learning_rate": 4.962581852198316e-06, + "loss": 1.7452, + "step": 268 + }, + { + "epoch": 0.33964646464646464, + "grad_norm": 2.979163885116577, + "learning_rate": 4.9602432179607115e-06, + "loss": 1.6943, + "step": 269 + }, + { + "epoch": 0.3409090909090909, + "grad_norm": 3.2396063804626465, + "learning_rate": 4.957904583723106e-06, + "loss": 1.7798, + "step": 270 + }, + { + "epoch": 0.3409090909090909, + "eval_loss": 1.9126806259155273, + "eval_runtime": 37.123, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 270 + }, + { + "epoch": 0.3421717171717172, + "grad_norm": 2.7868010997772217, + "learning_rate": 4.955565949485501e-06, + "loss": 1.8053, + "step": 271 + }, + { + "epoch": 0.3434343434343434, + "grad_norm": 2.8895349502563477, + "learning_rate": 4.9532273152478956e-06, + "loss": 1.7672, + "step": 272 + }, + { + "epoch": 0.3446969696969697, + "grad_norm": 3.0839710235595703, + "learning_rate": 4.95088868101029e-06, + "loss": 1.7414, + "step": 273 + }, + { + "epoch": 0.34595959595959597, + "grad_norm": 2.81974458694458, + "learning_rate": 4.948550046772685e-06, + "loss": 1.7527, + "step": 274 + }, + { + "epoch": 0.3472222222222222, + "grad_norm": 3.7469775676727295, + "learning_rate": 4.9462114125350805e-06, + "loss": 1.7689, + "step": 275 + }, + { + "epoch": 0.3484848484848485, + "grad_norm": 3.434573173522949, + "learning_rate": 4.943872778297475e-06, + "loss": 1.739, + "step": 276 + }, + { + "epoch": 0.34974747474747475, + "grad_norm": 2.985041379928589, + "learning_rate": 4.941534144059869e-06, + "loss": 1.7894, + "step": 277 + }, + { + "epoch": 0.351010101010101, + "grad_norm": 3.441204309463501, + "learning_rate": 4.939195509822265e-06, + "loss": 1.7804, + "step": 278 + }, + { + "epoch": 0.3522727272727273, + "grad_norm": 2.8371102809906006, + "learning_rate": 4.936856875584659e-06, + "loss": 1.7415, + "step": 279 + }, + { + "epoch": 0.35353535353535354, + "grad_norm": 3.662773847579956, + "learning_rate": 4.9345182413470535e-06, + "loss": 1.7894, + "step": 280 + }, + { + "epoch": 0.35353535353535354, + "eval_loss": 1.915844440460205, + "eval_runtime": 37.0941, + "eval_samples_per_second": 26.958, + "eval_steps_per_second": 1.132, + "step": 280 + }, + { + "epoch": 0.3547979797979798, + "grad_norm": 3.3254036903381348, + "learning_rate": 4.932179607109449e-06, + "loss": 1.7238, + "step": 281 + }, + { + "epoch": 0.3560606060606061, + "grad_norm": 2.9872665405273438, + "learning_rate": 4.929840972871843e-06, + "loss": 1.769, + "step": 282 + }, + { + "epoch": 0.3573232323232323, + "grad_norm": 3.0163521766662598, + "learning_rate": 4.9275023386342376e-06, + "loss": 1.7702, + "step": 283 + }, + { + "epoch": 0.35858585858585856, + "grad_norm": 2.9687387943267822, + "learning_rate": 4.925163704396633e-06, + "loss": 1.7927, + "step": 284 + }, + { + "epoch": 0.35984848484848486, + "grad_norm": 2.7064011096954346, + "learning_rate": 4.922825070159027e-06, + "loss": 1.7638, + "step": 285 + }, + { + "epoch": 0.3611111111111111, + "grad_norm": 3.0043134689331055, + "learning_rate": 4.9204864359214225e-06, + "loss": 1.8041, + "step": 286 + }, + { + "epoch": 0.36237373737373735, + "grad_norm": 3.037100315093994, + "learning_rate": 4.918147801683817e-06, + "loss": 1.7579, + "step": 287 + }, + { + "epoch": 0.36363636363636365, + "grad_norm": 2.7528676986694336, + "learning_rate": 4.915809167446211e-06, + "loss": 1.7629, + "step": 288 + }, + { + "epoch": 0.3648989898989899, + "grad_norm": 3.075815439224243, + "learning_rate": 4.913470533208607e-06, + "loss": 1.7966, + "step": 289 + }, + { + "epoch": 0.3661616161616162, + "grad_norm": 2.964785575866699, + "learning_rate": 4.911131898971001e-06, + "loss": 1.801, + "step": 290 + }, + { + "epoch": 0.3661616161616162, + "eval_loss": 1.8839410543441772, + "eval_runtime": 37.0844, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 290 + }, + { + "epoch": 0.36742424242424243, + "grad_norm": 2.668747663497925, + "learning_rate": 4.908793264733396e-06, + "loss": 1.7703, + "step": 291 + }, + { + "epoch": 0.3686868686868687, + "grad_norm": 2.814751148223877, + "learning_rate": 4.906454630495791e-06, + "loss": 1.7075, + "step": 292 + }, + { + "epoch": 0.369949494949495, + "grad_norm": 3.005922794342041, + "learning_rate": 4.904115996258186e-06, + "loss": 1.7418, + "step": 293 + }, + { + "epoch": 0.3712121212121212, + "grad_norm": 3.050936460494995, + "learning_rate": 4.90177736202058e-06, + "loss": 1.793, + "step": 294 + }, + { + "epoch": 0.37247474747474746, + "grad_norm": 2.822004556655884, + "learning_rate": 4.899438727782975e-06, + "loss": 1.8303, + "step": 295 + }, + { + "epoch": 0.37373737373737376, + "grad_norm": 2.950254201889038, + "learning_rate": 4.89710009354537e-06, + "loss": 1.808, + "step": 296 + }, + { + "epoch": 0.375, + "grad_norm": 2.8140666484832764, + "learning_rate": 4.8947614593077645e-06, + "loss": 1.7262, + "step": 297 + }, + { + "epoch": 0.37626262626262624, + "grad_norm": 2.899620532989502, + "learning_rate": 4.89242282507016e-06, + "loss": 1.7157, + "step": 298 + }, + { + "epoch": 0.37752525252525254, + "grad_norm": 2.805753469467163, + "learning_rate": 4.890084190832554e-06, + "loss": 1.7294, + "step": 299 + }, + { + "epoch": 0.3787878787878788, + "grad_norm": 3.110246181488037, + "learning_rate": 4.887745556594949e-06, + "loss": 1.7415, + "step": 300 + }, + { + "epoch": 0.3787878787878788, + "eval_loss": 1.892285943031311, + "eval_runtime": 37.0732, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 300 + }, + { + "epoch": 0.380050505050505, + "grad_norm": 2.9197490215301514, + "learning_rate": 4.885406922357344e-06, + "loss": 1.7132, + "step": 301 + }, + { + "epoch": 0.3813131313131313, + "grad_norm": 3.3165123462677, + "learning_rate": 4.883068288119738e-06, + "loss": 1.777, + "step": 302 + }, + { + "epoch": 0.38257575757575757, + "grad_norm": 2.8409054279327393, + "learning_rate": 4.880729653882133e-06, + "loss": 1.7092, + "step": 303 + }, + { + "epoch": 0.3838383838383838, + "grad_norm": 2.6800901889801025, + "learning_rate": 4.878391019644528e-06, + "loss": 1.7622, + "step": 304 + }, + { + "epoch": 0.3851010101010101, + "grad_norm": 3.2847490310668945, + "learning_rate": 4.876052385406922e-06, + "loss": 1.729, + "step": 305 + }, + { + "epoch": 0.38636363636363635, + "grad_norm": 2.786057949066162, + "learning_rate": 4.873713751169318e-06, + "loss": 1.7168, + "step": 306 + }, + { + "epoch": 0.38762626262626265, + "grad_norm": 3.236050605773926, + "learning_rate": 4.871375116931712e-06, + "loss": 1.6809, + "step": 307 + }, + { + "epoch": 0.3888888888888889, + "grad_norm": 2.851910352706909, + "learning_rate": 4.869036482694107e-06, + "loss": 1.7336, + "step": 308 + }, + { + "epoch": 0.39015151515151514, + "grad_norm": 2.892421007156372, + "learning_rate": 4.866697848456502e-06, + "loss": 1.7451, + "step": 309 + }, + { + "epoch": 0.39141414141414144, + "grad_norm": 2.8175048828125, + "learning_rate": 4.864359214218897e-06, + "loss": 1.7734, + "step": 310 + }, + { + "epoch": 0.39141414141414144, + "eval_loss": 1.8777843713760376, + "eval_runtime": 36.9321, + "eval_samples_per_second": 27.077, + "eval_steps_per_second": 1.137, + "step": 310 + }, + { + "epoch": 0.3926767676767677, + "grad_norm": 3.005329132080078, + "learning_rate": 4.8620205799812915e-06, + "loss": 1.7476, + "step": 311 + }, + { + "epoch": 0.3939393939393939, + "grad_norm": 2.849017858505249, + "learning_rate": 4.859681945743686e-06, + "loss": 1.7854, + "step": 312 + }, + { + "epoch": 0.3952020202020202, + "grad_norm": 2.801048517227173, + "learning_rate": 4.857343311506081e-06, + "loss": 1.7186, + "step": 313 + }, + { + "epoch": 0.39646464646464646, + "grad_norm": 2.5772831439971924, + "learning_rate": 4.855004677268476e-06, + "loss": 1.7279, + "step": 314 + }, + { + "epoch": 0.3977272727272727, + "grad_norm": 2.7969934940338135, + "learning_rate": 4.85266604303087e-06, + "loss": 1.7475, + "step": 315 + }, + { + "epoch": 0.398989898989899, + "grad_norm": 2.7067770957946777, + "learning_rate": 4.850327408793265e-06, + "loss": 1.7055, + "step": 316 + }, + { + "epoch": 0.40025252525252525, + "grad_norm": 2.70928955078125, + "learning_rate": 4.84798877455566e-06, + "loss": 1.7203, + "step": 317 + }, + { + "epoch": 0.4015151515151515, + "grad_norm": 2.7704474925994873, + "learning_rate": 4.845650140318054e-06, + "loss": 1.8058, + "step": 318 + }, + { + "epoch": 0.4027777777777778, + "grad_norm": 2.7680182456970215, + "learning_rate": 4.843311506080449e-06, + "loss": 1.7162, + "step": 319 + }, + { + "epoch": 0.40404040404040403, + "grad_norm": 3.0953330993652344, + "learning_rate": 4.840972871842844e-06, + "loss": 1.7617, + "step": 320 + }, + { + "epoch": 0.40404040404040403, + "eval_loss": 1.8680965900421143, + "eval_runtime": 37.0382, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 320 + }, + { + "epoch": 0.4053030303030303, + "grad_norm": 2.8340392112731934, + "learning_rate": 4.838634237605239e-06, + "loss": 1.7767, + "step": 321 + }, + { + "epoch": 0.4065656565656566, + "grad_norm": 3.057861804962158, + "learning_rate": 4.836295603367634e-06, + "loss": 1.8043, + "step": 322 + }, + { + "epoch": 0.4078282828282828, + "grad_norm": 3.0123419761657715, + "learning_rate": 4.833956969130029e-06, + "loss": 1.7561, + "step": 323 + }, + { + "epoch": 0.4090909090909091, + "grad_norm": 3.0873124599456787, + "learning_rate": 4.831618334892423e-06, + "loss": 1.7581, + "step": 324 + }, + { + "epoch": 0.41035353535353536, + "grad_norm": 2.6600539684295654, + "learning_rate": 4.8292797006548184e-06, + "loss": 1.7005, + "step": 325 + }, + { + "epoch": 0.4116161616161616, + "grad_norm": 2.905280590057373, + "learning_rate": 4.826941066417213e-06, + "loss": 1.7592, + "step": 326 + }, + { + "epoch": 0.4128787878787879, + "grad_norm": 3.0642189979553223, + "learning_rate": 4.824602432179607e-06, + "loss": 1.6456, + "step": 327 + }, + { + "epoch": 0.41414141414141414, + "grad_norm": 2.606531858444214, + "learning_rate": 4.8222637979420025e-06, + "loss": 1.6731, + "step": 328 + }, + { + "epoch": 0.4154040404040404, + "grad_norm": 2.9918158054351807, + "learning_rate": 4.819925163704397e-06, + "loss": 1.6854, + "step": 329 + }, + { + "epoch": 0.4166666666666667, + "grad_norm": 2.721940040588379, + "learning_rate": 4.817586529466791e-06, + "loss": 1.7346, + "step": 330 + }, + { + "epoch": 0.4166666666666667, + "eval_loss": 1.863736629486084, + "eval_runtime": 37.0887, + "eval_samples_per_second": 26.962, + "eval_steps_per_second": 1.132, + "step": 330 + }, + { + "epoch": 0.41792929292929293, + "grad_norm": 3.1076390743255615, + "learning_rate": 4.815247895229187e-06, + "loss": 1.6286, + "step": 331 + }, + { + "epoch": 0.41919191919191917, + "grad_norm": 3.1235783100128174, + "learning_rate": 4.812909260991581e-06, + "loss": 1.7404, + "step": 332 + }, + { + "epoch": 0.42045454545454547, + "grad_norm": 3.1211330890655518, + "learning_rate": 4.8105706267539755e-06, + "loss": 1.7143, + "step": 333 + }, + { + "epoch": 0.4217171717171717, + "grad_norm": 3.072317600250244, + "learning_rate": 4.808231992516371e-06, + "loss": 1.7514, + "step": 334 + }, + { + "epoch": 0.42297979797979796, + "grad_norm": 2.880671739578247, + "learning_rate": 4.805893358278765e-06, + "loss": 1.7018, + "step": 335 + }, + { + "epoch": 0.42424242424242425, + "grad_norm": 2.700063943862915, + "learning_rate": 4.8035547240411604e-06, + "loss": 1.7089, + "step": 336 + }, + { + "epoch": 0.4255050505050505, + "grad_norm": 2.9038445949554443, + "learning_rate": 4.801216089803556e-06, + "loss": 1.7483, + "step": 337 + }, + { + "epoch": 0.42676767676767674, + "grad_norm": 2.7822766304016113, + "learning_rate": 4.79887745556595e-06, + "loss": 1.6927, + "step": 338 + }, + { + "epoch": 0.42803030303030304, + "grad_norm": 2.957554340362549, + "learning_rate": 4.7965388213283445e-06, + "loss": 1.7162, + "step": 339 + }, + { + "epoch": 0.4292929292929293, + "grad_norm": 3.1333892345428467, + "learning_rate": 4.79420018709074e-06, + "loss": 1.731, + "step": 340 + }, + { + "epoch": 0.4292929292929293, + "eval_loss": 1.852413535118103, + "eval_runtime": 37.1809, + "eval_samples_per_second": 26.896, + "eval_steps_per_second": 1.13, + "step": 340 + }, + { + "epoch": 0.4305555555555556, + "grad_norm": 2.9801084995269775, + "learning_rate": 4.791861552853134e-06, + "loss": 1.6892, + "step": 341 + }, + { + "epoch": 0.4318181818181818, + "grad_norm": 2.5077271461486816, + "learning_rate": 4.789522918615529e-06, + "loss": 1.637, + "step": 342 + }, + { + "epoch": 0.43308080808080807, + "grad_norm": 2.8654391765594482, + "learning_rate": 4.787184284377924e-06, + "loss": 1.6915, + "step": 343 + }, + { + "epoch": 0.43434343434343436, + "grad_norm": 2.6204264163970947, + "learning_rate": 4.784845650140318e-06, + "loss": 1.6722, + "step": 344 + }, + { + "epoch": 0.4356060606060606, + "grad_norm": 2.633991003036499, + "learning_rate": 4.782507015902713e-06, + "loss": 1.6797, + "step": 345 + }, + { + "epoch": 0.43686868686868685, + "grad_norm": 2.7203924655914307, + "learning_rate": 4.780168381665108e-06, + "loss": 1.7018, + "step": 346 + }, + { + "epoch": 0.43813131313131315, + "grad_norm": 2.6789209842681885, + "learning_rate": 4.7778297474275024e-06, + "loss": 1.7153, + "step": 347 + }, + { + "epoch": 0.4393939393939394, + "grad_norm": 2.6285433769226074, + "learning_rate": 4.775491113189898e-06, + "loss": 1.7082, + "step": 348 + }, + { + "epoch": 0.44065656565656564, + "grad_norm": 2.623793840408325, + "learning_rate": 4.773152478952292e-06, + "loss": 1.7094, + "step": 349 + }, + { + "epoch": 0.44191919191919193, + "grad_norm": 2.647160291671753, + "learning_rate": 4.7708138447146865e-06, + "loss": 1.6528, + "step": 350 + }, + { + "epoch": 0.44191919191919193, + "eval_loss": 1.8506718873977661, + "eval_runtime": 37.4232, + "eval_samples_per_second": 26.721, + "eval_steps_per_second": 1.122, + "step": 350 + }, + { + "epoch": 0.4431818181818182, + "grad_norm": 2.664553642272949, + "learning_rate": 4.768475210477082e-06, + "loss": 1.7008, + "step": 351 + }, + { + "epoch": 0.4444444444444444, + "grad_norm": 2.8486251831054688, + "learning_rate": 4.766136576239477e-06, + "loss": 1.6811, + "step": 352 + }, + { + "epoch": 0.4457070707070707, + "grad_norm": 2.7514266967773438, + "learning_rate": 4.7637979420018715e-06, + "loss": 1.6969, + "step": 353 + }, + { + "epoch": 0.44696969696969696, + "grad_norm": 2.848083257675171, + "learning_rate": 4.761459307764266e-06, + "loss": 1.7058, + "step": 354 + }, + { + "epoch": 0.44823232323232326, + "grad_norm": 2.759218454360962, + "learning_rate": 4.759120673526661e-06, + "loss": 1.6109, + "step": 355 + }, + { + "epoch": 0.4494949494949495, + "grad_norm": 2.905318021774292, + "learning_rate": 4.756782039289056e-06, + "loss": 1.7689, + "step": 356 + }, + { + "epoch": 0.45075757575757575, + "grad_norm": 2.845646381378174, + "learning_rate": 4.75444340505145e-06, + "loss": 1.6687, + "step": 357 + }, + { + "epoch": 0.45202020202020204, + "grad_norm": 2.8609306812286377, + "learning_rate": 4.752104770813845e-06, + "loss": 1.6332, + "step": 358 + }, + { + "epoch": 0.4532828282828283, + "grad_norm": 3.016479969024658, + "learning_rate": 4.74976613657624e-06, + "loss": 1.6692, + "step": 359 + }, + { + "epoch": 0.45454545454545453, + "grad_norm": 3.2790887355804443, + "learning_rate": 4.747427502338635e-06, + "loss": 1.704, + "step": 360 + }, + { + "epoch": 0.45454545454545453, + "eval_loss": 1.848663330078125, + "eval_runtime": 37.4817, + "eval_samples_per_second": 26.68, + "eval_steps_per_second": 1.121, + "step": 360 + }, + { + "epoch": 0.45580808080808083, + "grad_norm": 3.1137545108795166, + "learning_rate": 4.745088868101029e-06, + "loss": 1.6797, + "step": 361 + }, + { + "epoch": 0.45707070707070707, + "grad_norm": 2.7598798274993896, + "learning_rate": 4.742750233863424e-06, + "loss": 1.6477, + "step": 362 + }, + { + "epoch": 0.4583333333333333, + "grad_norm": 3.4851579666137695, + "learning_rate": 4.740411599625819e-06, + "loss": 1.6481, + "step": 363 + }, + { + "epoch": 0.4595959595959596, + "grad_norm": 2.9285919666290283, + "learning_rate": 4.7380729653882135e-06, + "loss": 1.7094, + "step": 364 + }, + { + "epoch": 0.46085858585858586, + "grad_norm": 2.901458501815796, + "learning_rate": 4.735734331150608e-06, + "loss": 1.6906, + "step": 365 + }, + { + "epoch": 0.4621212121212121, + "grad_norm": 2.8737902641296387, + "learning_rate": 4.733395696913003e-06, + "loss": 1.6819, + "step": 366 + }, + { + "epoch": 0.4633838383838384, + "grad_norm": 3.2480952739715576, + "learning_rate": 4.7310570626753984e-06, + "loss": 1.6827, + "step": 367 + }, + { + "epoch": 0.46464646464646464, + "grad_norm": 2.8714661598205566, + "learning_rate": 4.728718428437793e-06, + "loss": 1.7349, + "step": 368 + }, + { + "epoch": 0.4659090909090909, + "grad_norm": 2.8410098552703857, + "learning_rate": 4.726379794200187e-06, + "loss": 1.7214, + "step": 369 + }, + { + "epoch": 0.4671717171717172, + "grad_norm": 2.744807720184326, + "learning_rate": 4.7240411599625826e-06, + "loss": 1.6742, + "step": 370 + }, + { + "epoch": 0.4671717171717172, + "eval_loss": 1.842191457748413, + "eval_runtime": 37.454, + "eval_samples_per_second": 26.699, + "eval_steps_per_second": 1.121, + "step": 370 + }, + { + "epoch": 0.4684343434343434, + "grad_norm": 2.8135552406311035, + "learning_rate": 4.721702525724977e-06, + "loss": 1.6957, + "step": 371 + }, + { + "epoch": 0.4696969696969697, + "grad_norm": 2.856764078140259, + "learning_rate": 4.719363891487372e-06, + "loss": 1.6802, + "step": 372 + }, + { + "epoch": 0.47095959595959597, + "grad_norm": 3.0707688331604004, + "learning_rate": 4.717025257249767e-06, + "loss": 1.7383, + "step": 373 + }, + { + "epoch": 0.4722222222222222, + "grad_norm": 2.8832876682281494, + "learning_rate": 4.714686623012161e-06, + "loss": 1.6309, + "step": 374 + }, + { + "epoch": 0.4734848484848485, + "grad_norm": 2.6840622425079346, + "learning_rate": 4.712347988774556e-06, + "loss": 1.6624, + "step": 375 + }, + { + "epoch": 0.47474747474747475, + "grad_norm": 3.2920548915863037, + "learning_rate": 4.710009354536951e-06, + "loss": 1.6887, + "step": 376 + }, + { + "epoch": 0.476010101010101, + "grad_norm": 3.0317189693450928, + "learning_rate": 4.707670720299345e-06, + "loss": 1.6514, + "step": 377 + }, + { + "epoch": 0.4772727272727273, + "grad_norm": 2.7284793853759766, + "learning_rate": 4.7053320860617404e-06, + "loss": 1.6595, + "step": 378 + }, + { + "epoch": 0.47853535353535354, + "grad_norm": 2.86665415763855, + "learning_rate": 4.702993451824135e-06, + "loss": 1.7219, + "step": 379 + }, + { + "epoch": 0.4797979797979798, + "grad_norm": 3.041191577911377, + "learning_rate": 4.700654817586529e-06, + "loss": 1.6959, + "step": 380 + }, + { + "epoch": 0.4797979797979798, + "eval_loss": 1.8414148092269897, + "eval_runtime": 37.2537, + "eval_samples_per_second": 26.843, + "eval_steps_per_second": 1.127, + "step": 380 + }, + { + "epoch": 0.4810606060606061, + "grad_norm": 3.1729421615600586, + "learning_rate": 4.6983161833489246e-06, + "loss": 1.6538, + "step": 381 + }, + { + "epoch": 0.4823232323232323, + "grad_norm": 2.8562991619110107, + "learning_rate": 4.69597754911132e-06, + "loss": 1.6667, + "step": 382 + }, + { + "epoch": 0.48358585858585856, + "grad_norm": 3.0348711013793945, + "learning_rate": 4.693638914873714e-06, + "loss": 1.6888, + "step": 383 + }, + { + "epoch": 0.48484848484848486, + "grad_norm": 2.9156477451324463, + "learning_rate": 4.6913002806361095e-06, + "loss": 1.628, + "step": 384 + }, + { + "epoch": 0.4861111111111111, + "grad_norm": 2.832689046859741, + "learning_rate": 4.688961646398504e-06, + "loss": 1.664, + "step": 385 + }, + { + "epoch": 0.48737373737373735, + "grad_norm": 2.9858946800231934, + "learning_rate": 4.686623012160898e-06, + "loss": 1.6696, + "step": 386 + }, + { + "epoch": 0.48863636363636365, + "grad_norm": 2.732468843460083, + "learning_rate": 4.684284377923294e-06, + "loss": 1.6444, + "step": 387 + }, + { + "epoch": 0.4898989898989899, + "grad_norm": 2.7801132202148438, + "learning_rate": 4.681945743685688e-06, + "loss": 1.7084, + "step": 388 + }, + { + "epoch": 0.4911616161616162, + "grad_norm": 2.646303415298462, + "learning_rate": 4.6796071094480824e-06, + "loss": 1.6963, + "step": 389 + }, + { + "epoch": 0.49242424242424243, + "grad_norm": 2.8750250339508057, + "learning_rate": 4.677268475210478e-06, + "loss": 1.6814, + "step": 390 + }, + { + "epoch": 0.49242424242424243, + "eval_loss": 1.8283894062042236, + "eval_runtime": 37.2058, + "eval_samples_per_second": 26.877, + "eval_steps_per_second": 1.129, + "step": 390 + }, + { + "epoch": 0.4936868686868687, + "grad_norm": 2.5373270511627197, + "learning_rate": 4.674929840972872e-06, + "loss": 1.5956, + "step": 391 + }, + { + "epoch": 0.494949494949495, + "grad_norm": 2.837170124053955, + "learning_rate": 4.6725912067352666e-06, + "loss": 1.6237, + "step": 392 + }, + { + "epoch": 0.4962121212121212, + "grad_norm": 3.199463367462158, + "learning_rate": 4.670252572497662e-06, + "loss": 1.7087, + "step": 393 + }, + { + "epoch": 0.49747474747474746, + "grad_norm": 2.9850947856903076, + "learning_rate": 4.667913938260056e-06, + "loss": 1.5913, + "step": 394 + }, + { + "epoch": 0.49873737373737376, + "grad_norm": 2.954148292541504, + "learning_rate": 4.665575304022451e-06, + "loss": 1.6272, + "step": 395 + }, + { + "epoch": 0.5, + "grad_norm": 2.8625757694244385, + "learning_rate": 4.663236669784846e-06, + "loss": 1.6623, + "step": 396 + }, + { + "epoch": 0.5012626262626263, + "grad_norm": 3.196354627609253, + "learning_rate": 4.66089803554724e-06, + "loss": 1.7249, + "step": 397 + }, + { + "epoch": 0.5025252525252525, + "grad_norm": 2.7257323265075684, + "learning_rate": 4.658559401309636e-06, + "loss": 1.665, + "step": 398 + }, + { + "epoch": 0.5037878787878788, + "grad_norm": 3.0976016521453857, + "learning_rate": 4.656220767072031e-06, + "loss": 1.6763, + "step": 399 + }, + { + "epoch": 0.5050505050505051, + "grad_norm": 3.0398128032684326, + "learning_rate": 4.653882132834425e-06, + "loss": 1.6375, + "step": 400 + }, + { + "epoch": 0.5050505050505051, + "eval_loss": 1.8324880599975586, + "eval_runtime": 37.1754, + "eval_samples_per_second": 26.9, + "eval_steps_per_second": 1.13, + "step": 400 + }, + { + "epoch": 0.5063131313131313, + "grad_norm": 3.314958095550537, + "learning_rate": 4.65154349859682e-06, + "loss": 1.6699, + "step": 401 + }, + { + "epoch": 0.5075757575757576, + "grad_norm": 2.8478870391845703, + "learning_rate": 4.649204864359215e-06, + "loss": 1.6495, + "step": 402 + }, + { + "epoch": 0.5088383838383839, + "grad_norm": 2.8476510047912598, + "learning_rate": 4.646866230121609e-06, + "loss": 1.633, + "step": 403 + }, + { + "epoch": 0.51010101010101, + "grad_norm": 2.916667938232422, + "learning_rate": 4.644527595884004e-06, + "loss": 1.6092, + "step": 404 + }, + { + "epoch": 0.5113636363636364, + "grad_norm": 2.977226495742798, + "learning_rate": 4.642188961646399e-06, + "loss": 1.6588, + "step": 405 + }, + { + "epoch": 0.5126262626262627, + "grad_norm": 3.0540225505828857, + "learning_rate": 4.6398503274087935e-06, + "loss": 1.6529, + "step": 406 + }, + { + "epoch": 0.5138888888888888, + "grad_norm": 2.902035713195801, + "learning_rate": 4.637511693171188e-06, + "loss": 1.5983, + "step": 407 + }, + { + "epoch": 0.5151515151515151, + "grad_norm": 2.9009487628936768, + "learning_rate": 4.635173058933583e-06, + "loss": 1.6389, + "step": 408 + }, + { + "epoch": 0.5164141414141414, + "grad_norm": 2.904982805252075, + "learning_rate": 4.632834424695978e-06, + "loss": 1.6971, + "step": 409 + }, + { + "epoch": 0.5176767676767676, + "grad_norm": 2.8340721130371094, + "learning_rate": 4.630495790458373e-06, + "loss": 1.6336, + "step": 410 + }, + { + "epoch": 0.5176767676767676, + "eval_loss": 1.8218064308166504, + "eval_runtime": 36.9963, + "eval_samples_per_second": 27.03, + "eval_steps_per_second": 1.135, + "step": 410 + }, + { + "epoch": 0.5189393939393939, + "grad_norm": 2.9507815837860107, + "learning_rate": 4.628157156220767e-06, + "loss": 1.6795, + "step": 411 + }, + { + "epoch": 0.5202020202020202, + "grad_norm": 2.868793249130249, + "learning_rate": 4.625818521983162e-06, + "loss": 1.6494, + "step": 412 + }, + { + "epoch": 0.5214646464646465, + "grad_norm": 2.786395788192749, + "learning_rate": 4.623479887745557e-06, + "loss": 1.7156, + "step": 413 + }, + { + "epoch": 0.5227272727272727, + "grad_norm": 2.7456934452056885, + "learning_rate": 4.621141253507952e-06, + "loss": 1.6528, + "step": 414 + }, + { + "epoch": 0.523989898989899, + "grad_norm": 3.1211695671081543, + "learning_rate": 4.618802619270347e-06, + "loss": 1.6891, + "step": 415 + }, + { + "epoch": 0.5252525252525253, + "grad_norm": 3.2319254875183105, + "learning_rate": 4.616463985032741e-06, + "loss": 1.6452, + "step": 416 + }, + { + "epoch": 0.5265151515151515, + "grad_norm": 2.6718053817749023, + "learning_rate": 4.614125350795136e-06, + "loss": 1.6442, + "step": 417 + }, + { + "epoch": 0.5277777777777778, + "grad_norm": 2.985020637512207, + "learning_rate": 4.611786716557531e-06, + "loss": 1.6935, + "step": 418 + }, + { + "epoch": 0.5290404040404041, + "grad_norm": 2.798720598220825, + "learning_rate": 4.609448082319925e-06, + "loss": 1.6303, + "step": 419 + }, + { + "epoch": 0.5303030303030303, + "grad_norm": 2.860403299331665, + "learning_rate": 4.6071094480823205e-06, + "loss": 1.6445, + "step": 420 + }, + { + "epoch": 0.5303030303030303, + "eval_loss": 1.8173401355743408, + "eval_runtime": 37.1334, + "eval_samples_per_second": 26.93, + "eval_steps_per_second": 1.131, + "step": 420 + }, + { + "epoch": 0.5315656565656566, + "grad_norm": 2.883209705352783, + "learning_rate": 4.604770813844715e-06, + "loss": 1.6182, + "step": 421 + }, + { + "epoch": 0.5328282828282829, + "grad_norm": 3.0330183506011963, + "learning_rate": 4.60243217960711e-06, + "loss": 1.6208, + "step": 422 + }, + { + "epoch": 0.5340909090909091, + "grad_norm": 3.224184274673462, + "learning_rate": 4.6000935453695046e-06, + "loss": 1.6687, + "step": 423 + }, + { + "epoch": 0.5353535353535354, + "grad_norm": 3.0345349311828613, + "learning_rate": 4.597754911131899e-06, + "loss": 1.6654, + "step": 424 + }, + { + "epoch": 0.5366161616161617, + "grad_norm": 3.1425750255584717, + "learning_rate": 4.595416276894294e-06, + "loss": 1.6569, + "step": 425 + }, + { + "epoch": 0.5378787878787878, + "grad_norm": 2.6387979984283447, + "learning_rate": 4.593077642656689e-06, + "loss": 1.6868, + "step": 426 + }, + { + "epoch": 0.5391414141414141, + "grad_norm": 2.959446430206299, + "learning_rate": 4.590739008419083e-06, + "loss": 1.621, + "step": 427 + }, + { + "epoch": 0.5404040404040404, + "grad_norm": 3.009673833847046, + "learning_rate": 4.588400374181478e-06, + "loss": 1.6414, + "step": 428 + }, + { + "epoch": 0.5416666666666666, + "grad_norm": 2.827531576156616, + "learning_rate": 4.586061739943874e-06, + "loss": 1.6561, + "step": 429 + }, + { + "epoch": 0.5429292929292929, + "grad_norm": 2.8770735263824463, + "learning_rate": 4.583723105706268e-06, + "loss": 1.6685, + "step": 430 + }, + { + "epoch": 0.5429292929292929, + "eval_loss": 1.8234716653823853, + "eval_runtime": 37.1647, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 430 + }, + { + "epoch": 0.5441919191919192, + "grad_norm": 3.1215782165527344, + "learning_rate": 4.5813844714686625e-06, + "loss": 1.6548, + "step": 431 + }, + { + "epoch": 0.5454545454545454, + "grad_norm": 2.6587517261505127, + "learning_rate": 4.579045837231058e-06, + "loss": 1.649, + "step": 432 + }, + { + "epoch": 0.5467171717171717, + "grad_norm": 2.7059030532836914, + "learning_rate": 4.576707202993452e-06, + "loss": 1.6293, + "step": 433 + }, + { + "epoch": 0.547979797979798, + "grad_norm": 2.7390925884246826, + "learning_rate": 4.574368568755847e-06, + "loss": 1.6957, + "step": 434 + }, + { + "epoch": 0.5492424242424242, + "grad_norm": 2.9457757472991943, + "learning_rate": 4.572029934518242e-06, + "loss": 1.6714, + "step": 435 + }, + { + "epoch": 0.5505050505050505, + "grad_norm": 2.6871347427368164, + "learning_rate": 4.569691300280636e-06, + "loss": 1.6648, + "step": 436 + }, + { + "epoch": 0.5517676767676768, + "grad_norm": 3.1708056926727295, + "learning_rate": 4.5673526660430315e-06, + "loss": 1.6415, + "step": 437 + }, + { + "epoch": 0.553030303030303, + "grad_norm": 2.8767378330230713, + "learning_rate": 4.565014031805426e-06, + "loss": 1.6475, + "step": 438 + }, + { + "epoch": 0.5542929292929293, + "grad_norm": 3.083566188812256, + "learning_rate": 4.56267539756782e-06, + "loss": 1.632, + "step": 439 + }, + { + "epoch": 0.5555555555555556, + "grad_norm": 2.584460735321045, + "learning_rate": 4.560336763330216e-06, + "loss": 1.6469, + "step": 440 + }, + { + "epoch": 0.5555555555555556, + "eval_loss": 1.8140095472335815, + "eval_runtime": 37.1986, + "eval_samples_per_second": 26.883, + "eval_steps_per_second": 1.129, + "step": 440 + }, + { + "epoch": 0.5568181818181818, + "grad_norm": 2.5885088443756104, + "learning_rate": 4.55799812909261e-06, + "loss": 1.6058, + "step": 441 + }, + { + "epoch": 0.5580808080808081, + "grad_norm": 2.8467836380004883, + "learning_rate": 4.5556594948550045e-06, + "loss": 1.6959, + "step": 442 + }, + { + "epoch": 0.5593434343434344, + "grad_norm": 2.764326333999634, + "learning_rate": 4.5533208606174e-06, + "loss": 1.6565, + "step": 443 + }, + { + "epoch": 0.5606060606060606, + "grad_norm": 2.88089919090271, + "learning_rate": 4.550982226379795e-06, + "loss": 1.6651, + "step": 444 + }, + { + "epoch": 0.5618686868686869, + "grad_norm": 2.6239469051361084, + "learning_rate": 4.548643592142189e-06, + "loss": 1.6255, + "step": 445 + }, + { + "epoch": 0.5631313131313131, + "grad_norm": 2.7748732566833496, + "learning_rate": 4.546304957904585e-06, + "loss": 1.6587, + "step": 446 + }, + { + "epoch": 0.5643939393939394, + "grad_norm": 2.7075202465057373, + "learning_rate": 4.543966323666979e-06, + "loss": 1.6448, + "step": 447 + }, + { + "epoch": 0.5656565656565656, + "grad_norm": 2.9291634559631348, + "learning_rate": 4.5416276894293735e-06, + "loss": 1.6564, + "step": 448 + }, + { + "epoch": 0.5669191919191919, + "grad_norm": 3.0039470195770264, + "learning_rate": 4.539289055191769e-06, + "loss": 1.623, + "step": 449 + }, + { + "epoch": 0.5681818181818182, + "grad_norm": 3.085277557373047, + "learning_rate": 4.536950420954163e-06, + "loss": 1.6176, + "step": 450 + }, + { + "epoch": 0.5681818181818182, + "eval_loss": 1.8022958040237427, + "eval_runtime": 37.2204, + "eval_samples_per_second": 26.867, + "eval_steps_per_second": 1.128, + "step": 450 + }, + { + "epoch": 0.5694444444444444, + "grad_norm": 2.93575119972229, + "learning_rate": 4.534611786716558e-06, + "loss": 1.615, + "step": 451 + }, + { + "epoch": 0.5707070707070707, + "grad_norm": 2.847141742706299, + "learning_rate": 4.532273152478953e-06, + "loss": 1.6474, + "step": 452 + }, + { + "epoch": 0.571969696969697, + "grad_norm": 3.0299603939056396, + "learning_rate": 4.529934518241347e-06, + "loss": 1.6424, + "step": 453 + }, + { + "epoch": 0.5732323232323232, + "grad_norm": 3.0307669639587402, + "learning_rate": 4.527595884003742e-06, + "loss": 1.6473, + "step": 454 + }, + { + "epoch": 0.5744949494949495, + "grad_norm": 2.864269971847534, + "learning_rate": 4.525257249766137e-06, + "loss": 1.6631, + "step": 455 + }, + { + "epoch": 0.5757575757575758, + "grad_norm": 2.606637716293335, + "learning_rate": 4.522918615528531e-06, + "loss": 1.6118, + "step": 456 + }, + { + "epoch": 0.577020202020202, + "grad_norm": 3.127891778945923, + "learning_rate": 4.520579981290926e-06, + "loss": 1.6286, + "step": 457 + }, + { + "epoch": 0.5782828282828283, + "grad_norm": 2.8809893131256104, + "learning_rate": 4.518241347053321e-06, + "loss": 1.6158, + "step": 458 + }, + { + "epoch": 0.5795454545454546, + "grad_norm": 3.0662436485290527, + "learning_rate": 4.515902712815716e-06, + "loss": 1.7336, + "step": 459 + }, + { + "epoch": 0.5808080808080808, + "grad_norm": 3.0795772075653076, + "learning_rate": 4.513564078578111e-06, + "loss": 1.6974, + "step": 460 + }, + { + "epoch": 0.5808080808080808, + "eval_loss": 1.7981404066085815, + "eval_runtime": 37.299, + "eval_samples_per_second": 26.81, + "eval_steps_per_second": 1.126, + "step": 460 + }, + { + "epoch": 0.5820707070707071, + "grad_norm": 3.091688871383667, + "learning_rate": 4.511225444340506e-06, + "loss": 1.6803, + "step": 461 + }, + { + "epoch": 0.5833333333333334, + "grad_norm": 2.5597071647644043, + "learning_rate": 4.5088868101029005e-06, + "loss": 1.6104, + "step": 462 + }, + { + "epoch": 0.5845959595959596, + "grad_norm": 2.849806070327759, + "learning_rate": 4.506548175865295e-06, + "loss": 1.6147, + "step": 463 + }, + { + "epoch": 0.5858585858585859, + "grad_norm": 2.9791204929351807, + "learning_rate": 4.50420954162769e-06, + "loss": 1.6376, + "step": 464 + }, + { + "epoch": 0.5871212121212122, + "grad_norm": 2.8582305908203125, + "learning_rate": 4.501870907390085e-06, + "loss": 1.6831, + "step": 465 + }, + { + "epoch": 0.5883838383838383, + "grad_norm": 2.8445184230804443, + "learning_rate": 4.499532273152479e-06, + "loss": 1.6463, + "step": 466 + }, + { + "epoch": 0.5896464646464646, + "grad_norm": 2.858168125152588, + "learning_rate": 4.497193638914874e-06, + "loss": 1.6662, + "step": 467 + }, + { + "epoch": 0.5909090909090909, + "grad_norm": 2.829702854156494, + "learning_rate": 4.494855004677269e-06, + "loss": 1.6058, + "step": 468 + }, + { + "epoch": 0.5921717171717171, + "grad_norm": 2.9292726516723633, + "learning_rate": 4.492516370439663e-06, + "loss": 1.6341, + "step": 469 + }, + { + "epoch": 0.5934343434343434, + "grad_norm": 2.9006292819976807, + "learning_rate": 4.490177736202058e-06, + "loss": 1.5726, + "step": 470 + }, + { + "epoch": 0.5934343434343434, + "eval_loss": 1.7953916788101196, + "eval_runtime": 37.5058, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 470 + }, + { + "epoch": 0.5946969696969697, + "grad_norm": 2.8116230964660645, + "learning_rate": 4.487839101964453e-06, + "loss": 1.6083, + "step": 471 + }, + { + "epoch": 0.5959595959595959, + "grad_norm": 2.540607452392578, + "learning_rate": 4.485500467726848e-06, + "loss": 1.6208, + "step": 472 + }, + { + "epoch": 0.5972222222222222, + "grad_norm": 2.7339096069335938, + "learning_rate": 4.4831618334892425e-06, + "loss": 1.6839, + "step": 473 + }, + { + "epoch": 0.5984848484848485, + "grad_norm": 2.9094762802124023, + "learning_rate": 4.480823199251638e-06, + "loss": 1.6962, + "step": 474 + }, + { + "epoch": 0.5997474747474747, + "grad_norm": 2.7733027935028076, + "learning_rate": 4.478484565014032e-06, + "loss": 1.6252, + "step": 475 + }, + { + "epoch": 0.601010101010101, + "grad_norm": 2.5709428787231445, + "learning_rate": 4.4761459307764274e-06, + "loss": 1.5896, + "step": 476 + }, + { + "epoch": 0.6022727272727273, + "grad_norm": 2.8358278274536133, + "learning_rate": 4.473807296538822e-06, + "loss": 1.6247, + "step": 477 + }, + { + "epoch": 0.6035353535353535, + "grad_norm": 2.7504897117614746, + "learning_rate": 4.471468662301216e-06, + "loss": 1.7138, + "step": 478 + }, + { + "epoch": 0.6047979797979798, + "grad_norm": 2.6965620517730713, + "learning_rate": 4.4691300280636115e-06, + "loss": 1.6054, + "step": 479 + }, + { + "epoch": 0.6060606060606061, + "grad_norm": 2.9401891231536865, + "learning_rate": 4.466791393826006e-06, + "loss": 1.6172, + "step": 480 + }, + { + "epoch": 0.6060606060606061, + "eval_loss": 1.7946635484695435, + "eval_runtime": 37.5366, + "eval_samples_per_second": 26.641, + "eval_steps_per_second": 1.119, + "step": 480 + }, + { + "epoch": 0.6073232323232324, + "grad_norm": 2.71535587310791, + "learning_rate": 4.4644527595884e-06, + "loss": 1.61, + "step": 481 + }, + { + "epoch": 0.6085858585858586, + "grad_norm": 2.7910654544830322, + "learning_rate": 4.462114125350796e-06, + "loss": 1.6193, + "step": 482 + }, + { + "epoch": 0.6098484848484849, + "grad_norm": 2.809241533279419, + "learning_rate": 4.45977549111319e-06, + "loss": 1.602, + "step": 483 + }, + { + "epoch": 0.6111111111111112, + "grad_norm": 2.7152137756347656, + "learning_rate": 4.457436856875585e-06, + "loss": 1.6516, + "step": 484 + }, + { + "epoch": 0.6123737373737373, + "grad_norm": 2.7653465270996094, + "learning_rate": 4.45509822263798e-06, + "loss": 1.5904, + "step": 485 + }, + { + "epoch": 0.6136363636363636, + "grad_norm": 2.729876756668091, + "learning_rate": 4.452759588400374e-06, + "loss": 1.6259, + "step": 486 + }, + { + "epoch": 0.61489898989899, + "grad_norm": 2.6290180683135986, + "learning_rate": 4.4504209541627694e-06, + "loss": 1.5796, + "step": 487 + }, + { + "epoch": 0.6161616161616161, + "grad_norm": 2.939884662628174, + "learning_rate": 4.448082319925164e-06, + "loss": 1.6774, + "step": 488 + }, + { + "epoch": 0.6174242424242424, + "grad_norm": 2.7554702758789062, + "learning_rate": 4.445743685687559e-06, + "loss": 1.6226, + "step": 489 + }, + { + "epoch": 0.6186868686868687, + "grad_norm": 2.890458106994629, + "learning_rate": 4.4434050514499535e-06, + "loss": 1.592, + "step": 490 + }, + { + "epoch": 0.6186868686868687, + "eval_loss": 1.7868437767028809, + "eval_runtime": 37.5267, + "eval_samples_per_second": 26.648, + "eval_steps_per_second": 1.119, + "step": 490 + }, + { + "epoch": 0.6199494949494949, + "grad_norm": 2.757178783416748, + "learning_rate": 4.441066417212349e-06, + "loss": 1.6297, + "step": 491 + }, + { + "epoch": 0.6212121212121212, + "grad_norm": 2.7179954051971436, + "learning_rate": 4.438727782974743e-06, + "loss": 1.5685, + "step": 492 + }, + { + "epoch": 0.6224747474747475, + "grad_norm": 2.824514389038086, + "learning_rate": 4.436389148737138e-06, + "loss": 1.6625, + "step": 493 + }, + { + "epoch": 0.6237373737373737, + "grad_norm": 3.0062615871429443, + "learning_rate": 4.434050514499533e-06, + "loss": 1.5889, + "step": 494 + }, + { + "epoch": 0.625, + "grad_norm": 2.962176561355591, + "learning_rate": 4.431711880261927e-06, + "loss": 1.6813, + "step": 495 + }, + { + "epoch": 0.6262626262626263, + "grad_norm": 2.7531204223632812, + "learning_rate": 4.429373246024323e-06, + "loss": 1.6152, + "step": 496 + }, + { + "epoch": 0.6275252525252525, + "grad_norm": 2.764021873474121, + "learning_rate": 4.427034611786717e-06, + "loss": 1.6164, + "step": 497 + }, + { + "epoch": 0.6287878787878788, + "grad_norm": 2.836826801300049, + "learning_rate": 4.4246959775491114e-06, + "loss": 1.6192, + "step": 498 + }, + { + "epoch": 0.6300505050505051, + "grad_norm": 2.871220827102661, + "learning_rate": 4.422357343311507e-06, + "loss": 1.6096, + "step": 499 + }, + { + "epoch": 0.6313131313131313, + "grad_norm": 2.8129167556762695, + "learning_rate": 4.420018709073901e-06, + "loss": 1.6674, + "step": 500 + }, + { + "epoch": 0.6313131313131313, + "eval_loss": 1.785156011581421, + "eval_runtime": 37.2971, + "eval_samples_per_second": 26.812, + "eval_steps_per_second": 1.126, + "step": 500 + }, + { + "epoch": 0.6325757575757576, + "grad_norm": 2.775108814239502, + "learning_rate": 4.4176800748362955e-06, + "loss": 1.5851, + "step": 501 + }, + { + "epoch": 0.6338383838383839, + "grad_norm": 2.7775518894195557, + "learning_rate": 4.415341440598691e-06, + "loss": 1.6839, + "step": 502 + }, + { + "epoch": 0.63510101010101, + "grad_norm": 2.7997918128967285, + "learning_rate": 4.413002806361085e-06, + "loss": 1.5849, + "step": 503 + }, + { + "epoch": 0.6363636363636364, + "grad_norm": 2.91103196144104, + "learning_rate": 4.4106641721234805e-06, + "loss": 1.6369, + "step": 504 + }, + { + "epoch": 0.6376262626262627, + "grad_norm": 2.6800882816314697, + "learning_rate": 4.408325537885875e-06, + "loss": 1.6049, + "step": 505 + }, + { + "epoch": 0.6388888888888888, + "grad_norm": 2.7451021671295166, + "learning_rate": 4.40598690364827e-06, + "loss": 1.658, + "step": 506 + }, + { + "epoch": 0.6401515151515151, + "grad_norm": 2.794813632965088, + "learning_rate": 4.403648269410665e-06, + "loss": 1.6015, + "step": 507 + }, + { + "epoch": 0.6414141414141414, + "grad_norm": 2.769373655319214, + "learning_rate": 4.401309635173059e-06, + "loss": 1.6217, + "step": 508 + }, + { + "epoch": 0.6426767676767676, + "grad_norm": 2.9081523418426514, + "learning_rate": 4.398971000935454e-06, + "loss": 1.6235, + "step": 509 + }, + { + "epoch": 0.6439393939393939, + "grad_norm": 2.707618236541748, + "learning_rate": 4.396632366697849e-06, + "loss": 1.6003, + "step": 510 + }, + { + "epoch": 0.6439393939393939, + "eval_loss": 1.7876425981521606, + "eval_runtime": 36.9832, + "eval_samples_per_second": 27.039, + "eval_steps_per_second": 1.136, + "step": 510 + }, + { + "epoch": 0.6452020202020202, + "grad_norm": 2.7923099994659424, + "learning_rate": 4.394293732460244e-06, + "loss": 1.5647, + "step": 511 + }, + { + "epoch": 0.6464646464646465, + "grad_norm": 2.8449175357818604, + "learning_rate": 4.391955098222638e-06, + "loss": 1.6456, + "step": 512 + }, + { + "epoch": 0.6477272727272727, + "grad_norm": 2.755856990814209, + "learning_rate": 4.389616463985033e-06, + "loss": 1.6962, + "step": 513 + }, + { + "epoch": 0.648989898989899, + "grad_norm": 2.5578861236572266, + "learning_rate": 4.387277829747428e-06, + "loss": 1.6097, + "step": 514 + }, + { + "epoch": 0.6502525252525253, + "grad_norm": 2.8804502487182617, + "learning_rate": 4.3849391955098225e-06, + "loss": 1.6212, + "step": 515 + }, + { + "epoch": 0.6515151515151515, + "grad_norm": 2.684215545654297, + "learning_rate": 4.382600561272217e-06, + "loss": 1.5529, + "step": 516 + }, + { + "epoch": 0.6527777777777778, + "grad_norm": 2.9128026962280273, + "learning_rate": 4.380261927034612e-06, + "loss": 1.6424, + "step": 517 + }, + { + "epoch": 0.6540404040404041, + "grad_norm": 2.733415126800537, + "learning_rate": 4.377923292797007e-06, + "loss": 1.6201, + "step": 518 + }, + { + "epoch": 0.6553030303030303, + "grad_norm": 2.7211062908172607, + "learning_rate": 4.375584658559401e-06, + "loss": 1.5833, + "step": 519 + }, + { + "epoch": 0.6565656565656566, + "grad_norm": 2.7308316230773926, + "learning_rate": 4.373246024321796e-06, + "loss": 1.6288, + "step": 520 + }, + { + "epoch": 0.6565656565656566, + "eval_loss": 1.7787725925445557, + "eval_runtime": 37.1089, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 520 + }, + { + "epoch": 0.6578282828282829, + "grad_norm": 2.760362148284912, + "learning_rate": 4.3709073900841915e-06, + "loss": 1.5814, + "step": 521 + }, + { + "epoch": 0.6590909090909091, + "grad_norm": 2.7625153064727783, + "learning_rate": 4.368568755846586e-06, + "loss": 1.6239, + "step": 522 + }, + { + "epoch": 0.6603535353535354, + "grad_norm": 2.715710401535034, + "learning_rate": 4.366230121608981e-06, + "loss": 1.6069, + "step": 523 + }, + { + "epoch": 0.6616161616161617, + "grad_norm": 2.7753846645355225, + "learning_rate": 4.363891487371376e-06, + "loss": 1.6178, + "step": 524 + }, + { + "epoch": 0.6628787878787878, + "grad_norm": 2.650716543197632, + "learning_rate": 4.36155285313377e-06, + "loss": 1.6534, + "step": 525 + }, + { + "epoch": 0.6641414141414141, + "grad_norm": 2.573969602584839, + "learning_rate": 4.359214218896165e-06, + "loss": 1.6721, + "step": 526 + }, + { + "epoch": 0.6654040404040404, + "grad_norm": 2.9017715454101562, + "learning_rate": 4.35687558465856e-06, + "loss": 1.6273, + "step": 527 + }, + { + "epoch": 0.6666666666666666, + "grad_norm": 2.691643714904785, + "learning_rate": 4.354536950420954e-06, + "loss": 1.6199, + "step": 528 + }, + { + "epoch": 0.6679292929292929, + "grad_norm": 2.818662166595459, + "learning_rate": 4.3521983161833494e-06, + "loss": 1.5758, + "step": 529 + }, + { + "epoch": 0.6691919191919192, + "grad_norm": 3.017738103866577, + "learning_rate": 4.349859681945744e-06, + "loss": 1.6437, + "step": 530 + }, + { + "epoch": 0.6691919191919192, + "eval_loss": 1.7835458517074585, + "eval_runtime": 37.1359, + "eval_samples_per_second": 26.928, + "eval_steps_per_second": 1.131, + "step": 530 + }, + { + "epoch": 0.6704545454545454, + "grad_norm": 2.8027284145355225, + "learning_rate": 4.347521047708138e-06, + "loss": 1.6688, + "step": 531 + }, + { + "epoch": 0.6717171717171717, + "grad_norm": 2.6642649173736572, + "learning_rate": 4.3451824134705335e-06, + "loss": 1.6327, + "step": 532 + }, + { + "epoch": 0.672979797979798, + "grad_norm": 2.7406396865844727, + "learning_rate": 4.342843779232928e-06, + "loss": 1.6297, + "step": 533 + }, + { + "epoch": 0.6742424242424242, + "grad_norm": 2.758538007736206, + "learning_rate": 4.340505144995323e-06, + "loss": 1.5276, + "step": 534 + }, + { + "epoch": 0.6755050505050505, + "grad_norm": 2.815817356109619, + "learning_rate": 4.3381665107577185e-06, + "loss": 1.6192, + "step": 535 + }, + { + "epoch": 0.6767676767676768, + "grad_norm": 2.8416755199432373, + "learning_rate": 4.335827876520113e-06, + "loss": 1.5966, + "step": 536 + }, + { + "epoch": 0.678030303030303, + "grad_norm": 3.014401435852051, + "learning_rate": 4.333489242282507e-06, + "loss": 1.6104, + "step": 537 + }, + { + "epoch": 0.6792929292929293, + "grad_norm": 2.731356143951416, + "learning_rate": 4.331150608044903e-06, + "loss": 1.6177, + "step": 538 + }, + { + "epoch": 0.6805555555555556, + "grad_norm": 2.633310317993164, + "learning_rate": 4.328811973807297e-06, + "loss": 1.5004, + "step": 539 + }, + { + "epoch": 0.6818181818181818, + "grad_norm": 2.977476119995117, + "learning_rate": 4.3264733395696914e-06, + "loss": 1.6171, + "step": 540 + }, + { + "epoch": 0.6818181818181818, + "eval_loss": 1.7790141105651855, + "eval_runtime": 37.12, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 540 + }, + { + "epoch": 0.6830808080808081, + "grad_norm": 2.769422769546509, + "learning_rate": 4.324134705332087e-06, + "loss": 1.5591, + "step": 541 + }, + { + "epoch": 0.6843434343434344, + "grad_norm": 3.088426351547241, + "learning_rate": 4.321796071094481e-06, + "loss": 1.5603, + "step": 542 + }, + { + "epoch": 0.6856060606060606, + "grad_norm": 2.6784212589263916, + "learning_rate": 4.3194574368568755e-06, + "loss": 1.6061, + "step": 543 + }, + { + "epoch": 0.6868686868686869, + "grad_norm": 3.209821939468384, + "learning_rate": 4.317118802619271e-06, + "loss": 1.585, + "step": 544 + }, + { + "epoch": 0.6881313131313131, + "grad_norm": 2.940791368484497, + "learning_rate": 4.314780168381665e-06, + "loss": 1.6092, + "step": 545 + }, + { + "epoch": 0.6893939393939394, + "grad_norm": 2.679884672164917, + "learning_rate": 4.3124415341440605e-06, + "loss": 1.6207, + "step": 546 + }, + { + "epoch": 0.6906565656565656, + "grad_norm": 2.853809118270874, + "learning_rate": 4.310102899906455e-06, + "loss": 1.5674, + "step": 547 + }, + { + "epoch": 0.6919191919191919, + "grad_norm": 2.693838119506836, + "learning_rate": 4.307764265668849e-06, + "loss": 1.5886, + "step": 548 + }, + { + "epoch": 0.6931818181818182, + "grad_norm": 2.871680736541748, + "learning_rate": 4.305425631431245e-06, + "loss": 1.6309, + "step": 549 + }, + { + "epoch": 0.6944444444444444, + "grad_norm": 2.9908292293548584, + "learning_rate": 4.30308699719364e-06, + "loss": 1.5577, + "step": 550 + }, + { + "epoch": 0.6944444444444444, + "eval_loss": 1.7790372371673584, + "eval_runtime": 37.114, + "eval_samples_per_second": 26.944, + "eval_steps_per_second": 1.132, + "step": 550 + }, + { + "epoch": 0.6957070707070707, + "grad_norm": 2.883481740951538, + "learning_rate": 4.300748362956034e-06, + "loss": 1.5749, + "step": 551 + }, + { + "epoch": 0.696969696969697, + "grad_norm": 2.7523458003997803, + "learning_rate": 4.298409728718429e-06, + "loss": 1.624, + "step": 552 + }, + { + "epoch": 0.6982323232323232, + "grad_norm": 2.69509220123291, + "learning_rate": 4.296071094480824e-06, + "loss": 1.5856, + "step": 553 + }, + { + "epoch": 0.6994949494949495, + "grad_norm": 2.6340222358703613, + "learning_rate": 4.293732460243218e-06, + "loss": 1.5857, + "step": 554 + }, + { + "epoch": 0.7007575757575758, + "grad_norm": 2.4998607635498047, + "learning_rate": 4.291393826005613e-06, + "loss": 1.6038, + "step": 555 + }, + { + "epoch": 0.702020202020202, + "grad_norm": 2.7913858890533447, + "learning_rate": 4.289055191768008e-06, + "loss": 1.6411, + "step": 556 + }, + { + "epoch": 0.7032828282828283, + "grad_norm": 2.7149479389190674, + "learning_rate": 4.2867165575304025e-06, + "loss": 1.592, + "step": 557 + }, + { + "epoch": 0.7045454545454546, + "grad_norm": 2.7187132835388184, + "learning_rate": 4.284377923292797e-06, + "loss": 1.5413, + "step": 558 + }, + { + "epoch": 0.7058080808080808, + "grad_norm": 2.8913674354553223, + "learning_rate": 4.282039289055192e-06, + "loss": 1.6105, + "step": 559 + }, + { + "epoch": 0.7070707070707071, + "grad_norm": 2.8470897674560547, + "learning_rate": 4.279700654817587e-06, + "loss": 1.6481, + "step": 560 + }, + { + "epoch": 0.7070707070707071, + "eval_loss": 1.7711386680603027, + "eval_runtime": 37.1243, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 560 + }, + { + "epoch": 0.7083333333333334, + "grad_norm": 2.7525687217712402, + "learning_rate": 4.277362020579982e-06, + "loss": 1.5749, + "step": 561 + }, + { + "epoch": 0.7095959595959596, + "grad_norm": 2.6935207843780518, + "learning_rate": 4.275023386342376e-06, + "loss": 1.637, + "step": 562 + }, + { + "epoch": 0.7108585858585859, + "grad_norm": 2.6935207843780518, + "learning_rate": 4.275023386342376e-06, + "loss": 1.5903, + "step": 563 + }, + { + "epoch": 0.7121212121212122, + "grad_norm": 2.992992401123047, + "learning_rate": 4.272684752104771e-06, + "loss": 1.6654, + "step": 564 + }, + { + "epoch": 0.7133838383838383, + "grad_norm": 2.7314348220825195, + "learning_rate": 4.270346117867166e-06, + "loss": 1.6775, + "step": 565 + }, + { + "epoch": 0.7146464646464646, + "grad_norm": 2.6952013969421387, + "learning_rate": 4.26800748362956e-06, + "loss": 1.5936, + "step": 566 + }, + { + "epoch": 0.7159090909090909, + "grad_norm": 2.8244829177856445, + "learning_rate": 4.265668849391956e-06, + "loss": 1.6228, + "step": 567 + }, + { + "epoch": 0.7171717171717171, + "grad_norm": 2.9110143184661865, + "learning_rate": 4.26333021515435e-06, + "loss": 1.6016, + "step": 568 + }, + { + "epoch": 0.7184343434343434, + "grad_norm": 2.7953436374664307, + "learning_rate": 4.260991580916745e-06, + "loss": 1.611, + "step": 569 + }, + { + "epoch": 0.7196969696969697, + "grad_norm": 2.854447364807129, + "learning_rate": 4.25865294667914e-06, + "loss": 1.6365, + "step": 570 + }, + { + "epoch": 0.7196969696969697, + "eval_loss": 1.7711069583892822, + "eval_runtime": 37.1374, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 570 + }, + { + "epoch": 0.7209595959595959, + "grad_norm": 2.7885239124298096, + "learning_rate": 4.256314312441534e-06, + "loss": 1.6106, + "step": 571 + }, + { + "epoch": 0.7222222222222222, + "grad_norm": 2.8012993335723877, + "learning_rate": 4.2539756782039295e-06, + "loss": 1.6081, + "step": 572 + }, + { + "epoch": 0.7234848484848485, + "grad_norm": 2.9223906993865967, + "learning_rate": 4.251637043966324e-06, + "loss": 1.5836, + "step": 573 + }, + { + "epoch": 0.7247474747474747, + "grad_norm": 2.885830879211426, + "learning_rate": 4.249298409728719e-06, + "loss": 1.5616, + "step": 574 + }, + { + "epoch": 0.726010101010101, + "grad_norm": 2.687372922897339, + "learning_rate": 4.2469597754911136e-06, + "loss": 1.5873, + "step": 575 + }, + { + "epoch": 0.7272727272727273, + "grad_norm": 3.0156168937683105, + "learning_rate": 4.244621141253508e-06, + "loss": 1.5763, + "step": 576 + }, + { + "epoch": 0.7285353535353535, + "grad_norm": 3.061842918395996, + "learning_rate": 4.242282507015903e-06, + "loss": 1.6148, + "step": 577 + }, + { + "epoch": 0.7297979797979798, + "grad_norm": 2.7467715740203857, + "learning_rate": 4.239943872778298e-06, + "loss": 1.577, + "step": 578 + }, + { + "epoch": 0.7310606060606061, + "grad_norm": 2.669501543045044, + "learning_rate": 4.237605238540692e-06, + "loss": 1.5653, + "step": 579 + }, + { + "epoch": 0.7323232323232324, + "grad_norm": 2.874349355697632, + "learning_rate": 4.235266604303087e-06, + "loss": 1.6523, + "step": 580 + }, + { + "epoch": 0.7323232323232324, + "eval_loss": 1.7680890560150146, + "eval_runtime": 37.2931, + "eval_samples_per_second": 26.815, + "eval_steps_per_second": 1.126, + "step": 580 + }, + { + "epoch": 0.7335858585858586, + "grad_norm": 2.7468504905700684, + "learning_rate": 4.232927970065482e-06, + "loss": 1.5935, + "step": 581 + }, + { + "epoch": 0.7348484848484849, + "grad_norm": 2.936697244644165, + "learning_rate": 4.230589335827877e-06, + "loss": 1.6335, + "step": 582 + }, + { + "epoch": 0.7361111111111112, + "grad_norm": 2.6878974437713623, + "learning_rate": 4.2282507015902715e-06, + "loss": 1.5699, + "step": 583 + }, + { + "epoch": 0.7373737373737373, + "grad_norm": 3.3578853607177734, + "learning_rate": 4.225912067352667e-06, + "loss": 1.5979, + "step": 584 + }, + { + "epoch": 0.7386363636363636, + "grad_norm": 2.8441176414489746, + "learning_rate": 4.223573433115061e-06, + "loss": 1.5736, + "step": 585 + }, + { + "epoch": 0.73989898989899, + "grad_norm": 2.7186129093170166, + "learning_rate": 4.221234798877456e-06, + "loss": 1.636, + "step": 586 + }, + { + "epoch": 0.7411616161616161, + "grad_norm": 2.940002918243408, + "learning_rate": 4.218896164639851e-06, + "loss": 1.5721, + "step": 587 + }, + { + "epoch": 0.7424242424242424, + "grad_norm": 2.676284074783325, + "learning_rate": 4.216557530402245e-06, + "loss": 1.5972, + "step": 588 + }, + { + "epoch": 0.7436868686868687, + "grad_norm": 2.9176571369171143, + "learning_rate": 4.2142188961646405e-06, + "loss": 1.6045, + "step": 589 + }, + { + "epoch": 0.7449494949494949, + "grad_norm": 2.7716288566589355, + "learning_rate": 4.211880261927035e-06, + "loss": 1.5988, + "step": 590 + }, + { + "epoch": 0.7449494949494949, + "eval_loss": 1.775631070137024, + "eval_runtime": 37.4113, + "eval_samples_per_second": 26.73, + "eval_steps_per_second": 1.123, + "step": 590 + }, + { + "epoch": 0.7462121212121212, + "grad_norm": 2.7201945781707764, + "learning_rate": 4.209541627689429e-06, + "loss": 1.5767, + "step": 591 + }, + { + "epoch": 0.7474747474747475, + "grad_norm": 2.838362455368042, + "learning_rate": 4.207202993451825e-06, + "loss": 1.5793, + "step": 592 + }, + { + "epoch": 0.7487373737373737, + "grad_norm": 2.6511716842651367, + "learning_rate": 4.204864359214219e-06, + "loss": 1.5776, + "step": 593 + }, + { + "epoch": 0.75, + "grad_norm": 2.7062957286834717, + "learning_rate": 4.2025257249766135e-06, + "loss": 1.5593, + "step": 594 + }, + { + "epoch": 0.7512626262626263, + "grad_norm": 2.808570384979248, + "learning_rate": 4.200187090739009e-06, + "loss": 1.6249, + "step": 595 + }, + { + "epoch": 0.7525252525252525, + "grad_norm": 2.600278377532959, + "learning_rate": 4.197848456501403e-06, + "loss": 1.6171, + "step": 596 + }, + { + "epoch": 0.7537878787878788, + "grad_norm": 2.9669382572174072, + "learning_rate": 4.195509822263798e-06, + "loss": 1.608, + "step": 597 + }, + { + "epoch": 0.7550505050505051, + "grad_norm": 2.6576340198516846, + "learning_rate": 4.193171188026194e-06, + "loss": 1.5986, + "step": 598 + }, + { + "epoch": 0.7563131313131313, + "grad_norm": 2.787060022354126, + "learning_rate": 4.190832553788588e-06, + "loss": 1.548, + "step": 599 + }, + { + "epoch": 0.7575757575757576, + "grad_norm": 2.854971408843994, + "learning_rate": 4.1884939195509825e-06, + "loss": 1.6421, + "step": 600 + }, + { + "epoch": 0.7575757575757576, + "eval_loss": 1.7721046209335327, + "eval_runtime": 37.4471, + "eval_samples_per_second": 26.704, + "eval_steps_per_second": 1.122, + "step": 600 + }, + { + "epoch": 0.7588383838383839, + "grad_norm": 2.7417967319488525, + "learning_rate": 4.186155285313378e-06, + "loss": 1.6399, + "step": 601 + }, + { + "epoch": 0.76010101010101, + "grad_norm": 3.0036182403564453, + "learning_rate": 4.183816651075772e-06, + "loss": 1.5953, + "step": 602 + }, + { + "epoch": 0.7613636363636364, + "grad_norm": 2.502861261367798, + "learning_rate": 4.181478016838167e-06, + "loss": 1.5284, + "step": 603 + }, + { + "epoch": 0.7626262626262627, + "grad_norm": 2.7262094020843506, + "learning_rate": 4.179139382600562e-06, + "loss": 1.619, + "step": 604 + }, + { + "epoch": 0.7638888888888888, + "grad_norm": 2.8748724460601807, + "learning_rate": 4.176800748362956e-06, + "loss": 1.5625, + "step": 605 + }, + { + "epoch": 0.7651515151515151, + "grad_norm": 2.7272109985351562, + "learning_rate": 4.174462114125351e-06, + "loss": 1.5935, + "step": 606 + }, + { + "epoch": 0.7664141414141414, + "grad_norm": 2.7930479049682617, + "learning_rate": 4.172123479887746e-06, + "loss": 1.61, + "step": 607 + }, + { + "epoch": 0.7676767676767676, + "grad_norm": 2.952056407928467, + "learning_rate": 4.16978484565014e-06, + "loss": 1.6369, + "step": 608 + }, + { + "epoch": 0.7689393939393939, + "grad_norm": 2.979564905166626, + "learning_rate": 4.167446211412535e-06, + "loss": 1.5748, + "step": 609 + }, + { + "epoch": 0.7702020202020202, + "grad_norm": 2.7945773601531982, + "learning_rate": 4.16510757717493e-06, + "loss": 1.6096, + "step": 610 + }, + { + "epoch": 0.7702020202020202, + "eval_loss": 1.7568180561065674, + "eval_runtime": 36.969, + "eval_samples_per_second": 27.05, + "eval_steps_per_second": 1.136, + "step": 610 + }, + { + "epoch": 0.7714646464646465, + "grad_norm": 2.9217135906219482, + "learning_rate": 4.1627689429373245e-06, + "loss": 1.5902, + "step": 611 + }, + { + "epoch": 0.7727272727272727, + "grad_norm": 2.8184452056884766, + "learning_rate": 4.16043030869972e-06, + "loss": 1.5775, + "step": 612 + }, + { + "epoch": 0.773989898989899, + "grad_norm": 2.7121212482452393, + "learning_rate": 4.158091674462115e-06, + "loss": 1.5778, + "step": 613 + }, + { + "epoch": 0.7752525252525253, + "grad_norm": 2.7600955963134766, + "learning_rate": 4.1557530402245095e-06, + "loss": 1.6145, + "step": 614 + }, + { + "epoch": 0.7765151515151515, + "grad_norm": 2.6388845443725586, + "learning_rate": 4.153414405986904e-06, + "loss": 1.6475, + "step": 615 + }, + { + "epoch": 0.7777777777777778, + "grad_norm": 2.684379816055298, + "learning_rate": 4.151075771749299e-06, + "loss": 1.6187, + "step": 616 + }, + { + "epoch": 0.7790404040404041, + "grad_norm": 2.6907055377960205, + "learning_rate": 4.1487371375116936e-06, + "loss": 1.6108, + "step": 617 + }, + { + "epoch": 0.7803030303030303, + "grad_norm": 2.8458995819091797, + "learning_rate": 4.146398503274088e-06, + "loss": 1.5596, + "step": 618 + }, + { + "epoch": 0.7815656565656566, + "grad_norm": 2.68843674659729, + "learning_rate": 4.144059869036483e-06, + "loss": 1.5912, + "step": 619 + }, + { + "epoch": 0.7828282828282829, + "grad_norm": 2.7620413303375244, + "learning_rate": 4.141721234798878e-06, + "loss": 1.627, + "step": 620 + }, + { + "epoch": 0.7828282828282829, + "eval_loss": 1.757076621055603, + "eval_runtime": 37.0489, + "eval_samples_per_second": 26.991, + "eval_steps_per_second": 1.134, + "step": 620 + }, + { + "epoch": 0.7840909090909091, + "grad_norm": 2.7084949016571045, + "learning_rate": 4.139382600561272e-06, + "loss": 1.5249, + "step": 621 + }, + { + "epoch": 0.7853535353535354, + "grad_norm": 3.0410406589508057, + "learning_rate": 4.137043966323667e-06, + "loss": 1.6458, + "step": 622 + }, + { + "epoch": 0.7866161616161617, + "grad_norm": 2.8445630073547363, + "learning_rate": 4.134705332086062e-06, + "loss": 1.6404, + "step": 623 + }, + { + "epoch": 0.7878787878787878, + "grad_norm": 3.0764353275299072, + "learning_rate": 4.132366697848457e-06, + "loss": 1.5041, + "step": 624 + }, + { + "epoch": 0.7891414141414141, + "grad_norm": 2.626544237136841, + "learning_rate": 4.1300280636108515e-06, + "loss": 1.6145, + "step": 625 + }, + { + "epoch": 0.7904040404040404, + "grad_norm": 2.9735796451568604, + "learning_rate": 4.127689429373246e-06, + "loss": 1.5672, + "step": 626 + }, + { + "epoch": 0.7916666666666666, + "grad_norm": 2.7798562049865723, + "learning_rate": 4.125350795135641e-06, + "loss": 1.6272, + "step": 627 + }, + { + "epoch": 0.7929292929292929, + "grad_norm": 2.7438931465148926, + "learning_rate": 4.123012160898036e-06, + "loss": 1.5318, + "step": 628 + }, + { + "epoch": 0.7941919191919192, + "grad_norm": 3.103142499923706, + "learning_rate": 4.120673526660431e-06, + "loss": 1.6, + "step": 629 + }, + { + "epoch": 0.7954545454545454, + "grad_norm": 2.9056239128112793, + "learning_rate": 4.118334892422825e-06, + "loss": 1.6581, + "step": 630 + }, + { + "epoch": 0.7954545454545454, + "eval_loss": 1.7552870512008667, + "eval_runtime": 37.1223, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 630 + }, + { + "epoch": 0.7967171717171717, + "grad_norm": 2.7320070266723633, + "learning_rate": 4.1159962581852205e-06, + "loss": 1.5808, + "step": 631 + }, + { + "epoch": 0.797979797979798, + "grad_norm": 2.9518637657165527, + "learning_rate": 4.113657623947615e-06, + "loss": 1.6034, + "step": 632 + }, + { + "epoch": 0.7992424242424242, + "grad_norm": 2.7135164737701416, + "learning_rate": 4.111318989710009e-06, + "loss": 1.5684, + "step": 633 + }, + { + "epoch": 0.8005050505050505, + "grad_norm": 2.948230266571045, + "learning_rate": 4.108980355472405e-06, + "loss": 1.6378, + "step": 634 + }, + { + "epoch": 0.8017676767676768, + "grad_norm": 2.816206216812134, + "learning_rate": 4.106641721234799e-06, + "loss": 1.6118, + "step": 635 + }, + { + "epoch": 0.803030303030303, + "grad_norm": 2.8032639026641846, + "learning_rate": 4.104303086997194e-06, + "loss": 1.5627, + "step": 636 + }, + { + "epoch": 0.8042929292929293, + "grad_norm": 3.032348871231079, + "learning_rate": 4.101964452759589e-06, + "loss": 1.6192, + "step": 637 + }, + { + "epoch": 0.8055555555555556, + "grad_norm": 2.9512710571289062, + "learning_rate": 4.099625818521983e-06, + "loss": 1.5655, + "step": 638 + }, + { + "epoch": 0.8068181818181818, + "grad_norm": 2.738973379135132, + "learning_rate": 4.097287184284378e-06, + "loss": 1.5686, + "step": 639 + }, + { + "epoch": 0.8080808080808081, + "grad_norm": 2.6510915756225586, + "learning_rate": 4.094948550046773e-06, + "loss": 1.5889, + "step": 640 + }, + { + "epoch": 0.8080808080808081, + "eval_loss": 1.762879490852356, + "eval_runtime": 37.1457, + "eval_samples_per_second": 26.921, + "eval_steps_per_second": 1.131, + "step": 640 + }, + { + "epoch": 0.8093434343434344, + "grad_norm": 2.6933884620666504, + "learning_rate": 4.092609915809167e-06, + "loss": 1.5436, + "step": 641 + }, + { + "epoch": 0.8106060606060606, + "grad_norm": 2.703784227371216, + "learning_rate": 4.0902712815715625e-06, + "loss": 1.5869, + "step": 642 + }, + { + "epoch": 0.8118686868686869, + "grad_norm": 2.7742321491241455, + "learning_rate": 4.087932647333958e-06, + "loss": 1.5779, + "step": 643 + }, + { + "epoch": 0.8131313131313131, + "grad_norm": 2.836832284927368, + "learning_rate": 4.085594013096352e-06, + "loss": 1.5511, + "step": 644 + }, + { + "epoch": 0.8143939393939394, + "grad_norm": 2.9212028980255127, + "learning_rate": 4.083255378858747e-06, + "loss": 1.5604, + "step": 645 + }, + { + "epoch": 0.8156565656565656, + "grad_norm": 2.8608527183532715, + "learning_rate": 4.080916744621142e-06, + "loss": 1.5975, + "step": 646 + }, + { + "epoch": 0.8169191919191919, + "grad_norm": 2.8246703147888184, + "learning_rate": 4.078578110383536e-06, + "loss": 1.5962, + "step": 647 + }, + { + "epoch": 0.8181818181818182, + "grad_norm": 3.0648651123046875, + "learning_rate": 4.076239476145932e-06, + "loss": 1.6267, + "step": 648 + }, + { + "epoch": 0.8194444444444444, + "grad_norm": 2.849240303039551, + "learning_rate": 4.073900841908326e-06, + "loss": 1.6271, + "step": 649 + }, + { + "epoch": 0.8207070707070707, + "grad_norm": 2.8817555904388428, + "learning_rate": 4.07156220767072e-06, + "loss": 1.5916, + "step": 650 + }, + { + "epoch": 0.8207070707070707, + "eval_loss": 1.7502237558364868, + "eval_runtime": 37.1497, + "eval_samples_per_second": 26.918, + "eval_steps_per_second": 1.131, + "step": 650 + }, + { + "epoch": 0.821969696969697, + "grad_norm": 2.5747952461242676, + "learning_rate": 4.069223573433116e-06, + "loss": 1.5661, + "step": 651 + }, + { + "epoch": 0.8232323232323232, + "grad_norm": 2.7721974849700928, + "learning_rate": 4.06688493919551e-06, + "loss": 1.5903, + "step": 652 + }, + { + "epoch": 0.8244949494949495, + "grad_norm": 2.9606690406799316, + "learning_rate": 4.0645463049579045e-06, + "loss": 1.5431, + "step": 653 + }, + { + "epoch": 0.8257575757575758, + "grad_norm": 2.941169261932373, + "learning_rate": 4.0622076707203e-06, + "loss": 1.5711, + "step": 654 + }, + { + "epoch": 0.827020202020202, + "grad_norm": 2.8408572673797607, + "learning_rate": 4.059869036482694e-06, + "loss": 1.5975, + "step": 655 + }, + { + "epoch": 0.8282828282828283, + "grad_norm": 2.619875431060791, + "learning_rate": 4.057530402245089e-06, + "loss": 1.5194, + "step": 656 + }, + { + "epoch": 0.8295454545454546, + "grad_norm": 2.910982847213745, + "learning_rate": 4.055191768007484e-06, + "loss": 1.6318, + "step": 657 + }, + { + "epoch": 0.8308080808080808, + "grad_norm": 2.747422456741333, + "learning_rate": 4.052853133769879e-06, + "loss": 1.6205, + "step": 658 + }, + { + "epoch": 0.8320707070707071, + "grad_norm": 2.6889796257019043, + "learning_rate": 4.050514499532274e-06, + "loss": 1.5537, + "step": 659 + }, + { + "epoch": 0.8333333333333334, + "grad_norm": 2.9849202632904053, + "learning_rate": 4.048175865294669e-06, + "loss": 1.5438, + "step": 660 + }, + { + "epoch": 0.8333333333333334, + "eval_loss": 1.7484922409057617, + "eval_runtime": 37.161, + "eval_samples_per_second": 26.91, + "eval_steps_per_second": 1.13, + "step": 660 + }, + { + "epoch": 0.8345959595959596, + "grad_norm": 2.816930055618286, + "learning_rate": 4.045837231057063e-06, + "loss": 1.555, + "step": 661 + }, + { + "epoch": 0.8358585858585859, + "grad_norm": 2.990126371383667, + "learning_rate": 4.043498596819458e-06, + "loss": 1.6263, + "step": 662 + }, + { + "epoch": 0.8371212121212122, + "grad_norm": 2.898263692855835, + "learning_rate": 4.041159962581853e-06, + "loss": 1.5522, + "step": 663 + }, + { + "epoch": 0.8383838383838383, + "grad_norm": 2.7824151515960693, + "learning_rate": 4.038821328344247e-06, + "loss": 1.5366, + "step": 664 + }, + { + "epoch": 0.8396464646464646, + "grad_norm": 2.9943954944610596, + "learning_rate": 4.036482694106642e-06, + "loss": 1.6214, + "step": 665 + }, + { + "epoch": 0.8409090909090909, + "grad_norm": 2.628308057785034, + "learning_rate": 4.034144059869037e-06, + "loss": 1.5797, + "step": 666 + }, + { + "epoch": 0.8421717171717171, + "grad_norm": 2.7565176486968994, + "learning_rate": 4.0318054256314315e-06, + "loss": 1.5567, + "step": 667 + }, + { + "epoch": 0.8434343434343434, + "grad_norm": 2.688354969024658, + "learning_rate": 4.029466791393826e-06, + "loss": 1.5516, + "step": 668 + }, + { + "epoch": 0.8446969696969697, + "grad_norm": 2.899620771408081, + "learning_rate": 4.027128157156221e-06, + "loss": 1.6014, + "step": 669 + }, + { + "epoch": 0.8459595959595959, + "grad_norm": 2.7894718647003174, + "learning_rate": 4.024789522918616e-06, + "loss": 1.5688, + "step": 670 + }, + { + "epoch": 0.8459595959595959, + "eval_loss": 1.7486555576324463, + "eval_runtime": 37.1824, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 670 + }, + { + "epoch": 0.8472222222222222, + "grad_norm": 2.7441444396972656, + "learning_rate": 4.02245088868101e-06, + "loss": 1.5949, + "step": 671 + }, + { + "epoch": 0.8484848484848485, + "grad_norm": 2.7088451385498047, + "learning_rate": 4.020112254443405e-06, + "loss": 1.5878, + "step": 672 + }, + { + "epoch": 0.8497474747474747, + "grad_norm": 2.756920576095581, + "learning_rate": 4.0177736202058005e-06, + "loss": 1.6356, + "step": 673 + }, + { + "epoch": 0.851010101010101, + "grad_norm": 2.5605697631835938, + "learning_rate": 4.015434985968195e-06, + "loss": 1.5586, + "step": 674 + }, + { + "epoch": 0.8522727272727273, + "grad_norm": 2.946573257446289, + "learning_rate": 4.01309635173059e-06, + "loss": 1.5537, + "step": 675 + }, + { + "epoch": 0.8535353535353535, + "grad_norm": 2.65500545501709, + "learning_rate": 4.010757717492985e-06, + "loss": 1.5875, + "step": 676 + }, + { + "epoch": 0.8547979797979798, + "grad_norm": 2.6632118225097656, + "learning_rate": 4.008419083255379e-06, + "loss": 1.5329, + "step": 677 + }, + { + "epoch": 0.8560606060606061, + "grad_norm": 2.8909032344818115, + "learning_rate": 4.006080449017774e-06, + "loss": 1.6396, + "step": 678 + }, + { + "epoch": 0.8573232323232324, + "grad_norm": 2.8780951499938965, + "learning_rate": 4.003741814780169e-06, + "loss": 1.6189, + "step": 679 + }, + { + "epoch": 0.8585858585858586, + "grad_norm": 2.7757010459899902, + "learning_rate": 4.001403180542563e-06, + "loss": 1.5968, + "step": 680 + }, + { + "epoch": 0.8585858585858586, + "eval_loss": 1.7423925399780273, + "eval_runtime": 37.1537, + "eval_samples_per_second": 26.915, + "eval_steps_per_second": 1.13, + "step": 680 + }, + { + "epoch": 0.8598484848484849, + "grad_norm": 2.6610794067382812, + "learning_rate": 3.9990645463049584e-06, + "loss": 1.5258, + "step": 681 + }, + { + "epoch": 0.8611111111111112, + "grad_norm": 2.9292140007019043, + "learning_rate": 3.996725912067353e-06, + "loss": 1.5669, + "step": 682 + }, + { + "epoch": 0.8623737373737373, + "grad_norm": 2.7885360717773438, + "learning_rate": 3.994387277829747e-06, + "loss": 1.5737, + "step": 683 + }, + { + "epoch": 0.8636363636363636, + "grad_norm": 2.7511544227600098, + "learning_rate": 3.9920486435921425e-06, + "loss": 1.5715, + "step": 684 + }, + { + "epoch": 0.86489898989899, + "grad_norm": 2.8561315536499023, + "learning_rate": 3.989710009354537e-06, + "loss": 1.5965, + "step": 685 + }, + { + "epoch": 0.8661616161616161, + "grad_norm": 2.9127237796783447, + "learning_rate": 3.987371375116932e-06, + "loss": 1.5407, + "step": 686 + }, + { + "epoch": 0.8674242424242424, + "grad_norm": 2.735577344894409, + "learning_rate": 3.985032740879327e-06, + "loss": 1.5449, + "step": 687 + }, + { + "epoch": 0.8686868686868687, + "grad_norm": 2.6626689434051514, + "learning_rate": 3.982694106641721e-06, + "loss": 1.5173, + "step": 688 + }, + { + "epoch": 0.8699494949494949, + "grad_norm": 2.901592969894409, + "learning_rate": 3.980355472404116e-06, + "loss": 1.5757, + "step": 689 + }, + { + "epoch": 0.8712121212121212, + "grad_norm": 2.8801231384277344, + "learning_rate": 3.978016838166512e-06, + "loss": 1.5633, + "step": 690 + }, + { + "epoch": 0.8712121212121212, + "eval_loss": 1.7453432083129883, + "eval_runtime": 37.4368, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 690 + }, + { + "epoch": 0.8724747474747475, + "grad_norm": 2.7463626861572266, + "learning_rate": 3.975678203928906e-06, + "loss": 1.5786, + "step": 691 + }, + { + "epoch": 0.8737373737373737, + "grad_norm": 2.659820318222046, + "learning_rate": 3.9733395696913004e-06, + "loss": 1.5923, + "step": 692 + }, + { + "epoch": 0.875, + "grad_norm": 2.759030342102051, + "learning_rate": 3.971000935453696e-06, + "loss": 1.6217, + "step": 693 + }, + { + "epoch": 0.8762626262626263, + "grad_norm": 2.689967632293701, + "learning_rate": 3.96866230121609e-06, + "loss": 1.5533, + "step": 694 + }, + { + "epoch": 0.8775252525252525, + "grad_norm": 3.1230740547180176, + "learning_rate": 3.9663236669784845e-06, + "loss": 1.6538, + "step": 695 + }, + { + "epoch": 0.8787878787878788, + "grad_norm": 2.657022476196289, + "learning_rate": 3.96398503274088e-06, + "loss": 1.5432, + "step": 696 + }, + { + "epoch": 0.8800505050505051, + "grad_norm": 2.919379711151123, + "learning_rate": 3.961646398503274e-06, + "loss": 1.5867, + "step": 697 + }, + { + "epoch": 0.8813131313131313, + "grad_norm": 2.607135057449341, + "learning_rate": 3.9593077642656695e-06, + "loss": 1.5079, + "step": 698 + }, + { + "epoch": 0.8825757575757576, + "grad_norm": 2.9978976249694824, + "learning_rate": 3.956969130028064e-06, + "loss": 1.5905, + "step": 699 + }, + { + "epoch": 0.8838383838383839, + "grad_norm": 2.8117148876190186, + "learning_rate": 3.954630495790458e-06, + "loss": 1.5588, + "step": 700 + }, + { + "epoch": 0.8838383838383839, + "eval_loss": 1.7376645803451538, + "eval_runtime": 37.5045, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 700 + }, + { + "epoch": 0.88510101010101, + "grad_norm": 2.7135095596313477, + "learning_rate": 3.952291861552854e-06, + "loss": 1.6013, + "step": 701 + }, + { + "epoch": 0.8863636363636364, + "grad_norm": 2.812061071395874, + "learning_rate": 3.949953227315248e-06, + "loss": 1.5518, + "step": 702 + }, + { + "epoch": 0.8876262626262627, + "grad_norm": 3.0532422065734863, + "learning_rate": 3.9476145930776424e-06, + "loss": 1.5579, + "step": 703 + }, + { + "epoch": 0.8888888888888888, + "grad_norm": 2.8306567668914795, + "learning_rate": 3.945275958840038e-06, + "loss": 1.5478, + "step": 704 + }, + { + "epoch": 0.8901515151515151, + "grad_norm": 2.737921953201294, + "learning_rate": 3.942937324602433e-06, + "loss": 1.4958, + "step": 705 + }, + { + "epoch": 0.8914141414141414, + "grad_norm": 2.902211904525757, + "learning_rate": 3.940598690364827e-06, + "loss": 1.5715, + "step": 706 + }, + { + "epoch": 0.8926767676767676, + "grad_norm": 2.725818395614624, + "learning_rate": 3.938260056127222e-06, + "loss": 1.5502, + "step": 707 + }, + { + "epoch": 0.8939393939393939, + "grad_norm": 2.9530506134033203, + "learning_rate": 3.935921421889617e-06, + "loss": 1.564, + "step": 708 + }, + { + "epoch": 0.8952020202020202, + "grad_norm": 2.798231363296509, + "learning_rate": 3.9335827876520115e-06, + "loss": 1.5445, + "step": 709 + }, + { + "epoch": 0.8964646464646465, + "grad_norm": 3.0233471393585205, + "learning_rate": 3.931244153414407e-06, + "loss": 1.6193, + "step": 710 + }, + { + "epoch": 0.8964646464646465, + "eval_loss": 1.742077350616455, + "eval_runtime": 37.0255, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 710 + }, + { + "epoch": 0.8977272727272727, + "grad_norm": 2.84309458732605, + "learning_rate": 3.928905519176801e-06, + "loss": 1.5379, + "step": 711 + }, + { + "epoch": 0.898989898989899, + "grad_norm": 2.7412185668945312, + "learning_rate": 3.926566884939196e-06, + "loss": 1.5837, + "step": 712 + }, + { + "epoch": 0.9002525252525253, + "grad_norm": 3.1094822883605957, + "learning_rate": 3.924228250701591e-06, + "loss": 1.5621, + "step": 713 + }, + { + "epoch": 0.9015151515151515, + "grad_norm": 2.776059627532959, + "learning_rate": 3.921889616463985e-06, + "loss": 1.5714, + "step": 714 + }, + { + "epoch": 0.9027777777777778, + "grad_norm": 2.7650489807128906, + "learning_rate": 3.91955098222638e-06, + "loss": 1.5418, + "step": 715 + }, + { + "epoch": 0.9040404040404041, + "grad_norm": 2.8586974143981934, + "learning_rate": 3.917212347988775e-06, + "loss": 1.5891, + "step": 716 + }, + { + "epoch": 0.9053030303030303, + "grad_norm": 2.74332594871521, + "learning_rate": 3.914873713751169e-06, + "loss": 1.5492, + "step": 717 + }, + { + "epoch": 0.9065656565656566, + "grad_norm": 2.931682586669922, + "learning_rate": 3.912535079513564e-06, + "loss": 1.54, + "step": 718 + }, + { + "epoch": 0.9078282828282829, + "grad_norm": 2.678812026977539, + "learning_rate": 3.910196445275959e-06, + "loss": 1.5677, + "step": 719 + }, + { + "epoch": 0.9090909090909091, + "grad_norm": 2.7282893657684326, + "learning_rate": 3.907857811038354e-06, + "loss": 1.5413, + "step": 720 + }, + { + "epoch": 0.9090909090909091, + "eval_loss": 1.7409753799438477, + "eval_runtime": 37.1883, + "eval_samples_per_second": 26.89, + "eval_steps_per_second": 1.129, + "step": 720 + }, + { + "epoch": 0.9103535353535354, + "grad_norm": 3.073997735977173, + "learning_rate": 3.905519176800749e-06, + "loss": 1.5734, + "step": 721 + }, + { + "epoch": 0.9116161616161617, + "grad_norm": 2.6249935626983643, + "learning_rate": 3.903180542563144e-06, + "loss": 1.5063, + "step": 722 + }, + { + "epoch": 0.9128787878787878, + "grad_norm": 2.7758615016937256, + "learning_rate": 3.9008419083255384e-06, + "loss": 1.5112, + "step": 723 + }, + { + "epoch": 0.9141414141414141, + "grad_norm": 2.8954596519470215, + "learning_rate": 3.898503274087933e-06, + "loss": 1.5508, + "step": 724 + }, + { + "epoch": 0.9154040404040404, + "grad_norm": 2.79073166847229, + "learning_rate": 3.896164639850328e-06, + "loss": 1.5773, + "step": 725 + }, + { + "epoch": 0.9166666666666666, + "grad_norm": 2.687027931213379, + "learning_rate": 3.8938260056127226e-06, + "loss": 1.5243, + "step": 726 + }, + { + "epoch": 0.9179292929292929, + "grad_norm": 2.5704851150512695, + "learning_rate": 3.891487371375117e-06, + "loss": 1.5075, + "step": 727 + }, + { + "epoch": 0.9191919191919192, + "grad_norm": 2.8684048652648926, + "learning_rate": 3.889148737137512e-06, + "loss": 1.6471, + "step": 728 + }, + { + "epoch": 0.9204545454545454, + "grad_norm": 2.8995590209960938, + "learning_rate": 3.886810102899907e-06, + "loss": 1.5429, + "step": 729 + }, + { + "epoch": 0.9217171717171717, + "grad_norm": 2.7575254440307617, + "learning_rate": 3.884471468662301e-06, + "loss": 1.5241, + "step": 730 + }, + { + "epoch": 0.9217171717171717, + "eval_loss": 1.7294232845306396, + "eval_runtime": 37.1941, + "eval_samples_per_second": 26.886, + "eval_steps_per_second": 1.129, + "step": 730 + }, + { + "epoch": 0.922979797979798, + "grad_norm": 2.7057321071624756, + "learning_rate": 3.882132834424696e-06, + "loss": 1.5498, + "step": 731 + }, + { + "epoch": 0.9242424242424242, + "grad_norm": 2.7679436206817627, + "learning_rate": 3.879794200187091e-06, + "loss": 1.563, + "step": 732 + }, + { + "epoch": 0.9255050505050505, + "grad_norm": 2.7258687019348145, + "learning_rate": 3.877455565949485e-06, + "loss": 1.5016, + "step": 733 + }, + { + "epoch": 0.9267676767676768, + "grad_norm": 2.8474221229553223, + "learning_rate": 3.8751169317118804e-06, + "loss": 1.5318, + "step": 734 + }, + { + "epoch": 0.928030303030303, + "grad_norm": 2.7718396186828613, + "learning_rate": 3.872778297474276e-06, + "loss": 1.5919, + "step": 735 + }, + { + "epoch": 0.9292929292929293, + "grad_norm": 2.7785022258758545, + "learning_rate": 3.87043966323667e-06, + "loss": 1.5399, + "step": 736 + }, + { + "epoch": 0.9305555555555556, + "grad_norm": 2.742476463317871, + "learning_rate": 3.868101028999065e-06, + "loss": 1.5428, + "step": 737 + }, + { + "epoch": 0.9318181818181818, + "grad_norm": 2.7050726413726807, + "learning_rate": 3.86576239476146e-06, + "loss": 1.5604, + "step": 738 + }, + { + "epoch": 0.9330808080808081, + "grad_norm": 2.6755242347717285, + "learning_rate": 3.863423760523854e-06, + "loss": 1.5409, + "step": 739 + }, + { + "epoch": 0.9343434343434344, + "grad_norm": 2.6947829723358154, + "learning_rate": 3.8610851262862495e-06, + "loss": 1.5537, + "step": 740 + }, + { + "epoch": 0.9343434343434344, + "eval_loss": 1.7396037578582764, + "eval_runtime": 37.1912, + "eval_samples_per_second": 26.888, + "eval_steps_per_second": 1.129, + "step": 740 + }, + { + "epoch": 0.9356060606060606, + "grad_norm": 2.729816436767578, + "learning_rate": 3.858746492048644e-06, + "loss": 1.5581, + "step": 741 + }, + { + "epoch": 0.9368686868686869, + "grad_norm": 2.6951887607574463, + "learning_rate": 3.856407857811038e-06, + "loss": 1.5929, + "step": 742 + }, + { + "epoch": 0.9381313131313131, + "grad_norm": 2.6601064205169678, + "learning_rate": 3.854069223573434e-06, + "loss": 1.4881, + "step": 743 + }, + { + "epoch": 0.9393939393939394, + "grad_norm": 2.623898983001709, + "learning_rate": 3.851730589335828e-06, + "loss": 1.5499, + "step": 744 + }, + { + "epoch": 0.9406565656565656, + "grad_norm": 2.703871250152588, + "learning_rate": 3.8493919550982224e-06, + "loss": 1.5585, + "step": 745 + }, + { + "epoch": 0.9419191919191919, + "grad_norm": 2.674893379211426, + "learning_rate": 3.847053320860618e-06, + "loss": 1.5586, + "step": 746 + }, + { + "epoch": 0.9431818181818182, + "grad_norm": 2.8353004455566406, + "learning_rate": 3.844714686623012e-06, + "loss": 1.5937, + "step": 747 + }, + { + "epoch": 0.9444444444444444, + "grad_norm": 2.68009614944458, + "learning_rate": 3.842376052385407e-06, + "loss": 1.5192, + "step": 748 + }, + { + "epoch": 0.9457070707070707, + "grad_norm": 2.6847407817840576, + "learning_rate": 3.840037418147802e-06, + "loss": 1.512, + "step": 749 + }, + { + "epoch": 0.946969696969697, + "grad_norm": 2.624290943145752, + "learning_rate": 3.837698783910197e-06, + "loss": 1.5348, + "step": 750 + }, + { + "epoch": 0.946969696969697, + "eval_loss": 1.738332748413086, + "eval_runtime": 37.1782, + "eval_samples_per_second": 26.898, + "eval_steps_per_second": 1.13, + "step": 750 + }, + { + "epoch": 0.9482323232323232, + "grad_norm": 2.7467777729034424, + "learning_rate": 3.8353601496725915e-06, + "loss": 1.5603, + "step": 751 + }, + { + "epoch": 0.9494949494949495, + "grad_norm": 2.851987600326538, + "learning_rate": 3.833021515434987e-06, + "loss": 1.5974, + "step": 752 + }, + { + "epoch": 0.9507575757575758, + "grad_norm": 2.6642227172851562, + "learning_rate": 3.830682881197381e-06, + "loss": 1.5138, + "step": 753 + }, + { + "epoch": 0.952020202020202, + "grad_norm": 2.7351491451263428, + "learning_rate": 3.828344246959776e-06, + "loss": 1.5656, + "step": 754 + }, + { + "epoch": 0.9532828282828283, + "grad_norm": 2.593748092651367, + "learning_rate": 3.826005612722171e-06, + "loss": 1.5395, + "step": 755 + }, + { + "epoch": 0.9545454545454546, + "grad_norm": 2.639180898666382, + "learning_rate": 3.823666978484565e-06, + "loss": 1.5349, + "step": 756 + }, + { + "epoch": 0.9558080808080808, + "grad_norm": 2.805148124694824, + "learning_rate": 3.82132834424696e-06, + "loss": 1.5805, + "step": 757 + }, + { + "epoch": 0.9570707070707071, + "grad_norm": 2.5619425773620605, + "learning_rate": 3.818989710009355e-06, + "loss": 1.5389, + "step": 758 + }, + { + "epoch": 0.9583333333333334, + "grad_norm": 2.7224011421203613, + "learning_rate": 3.816651075771749e-06, + "loss": 1.5316, + "step": 759 + }, + { + "epoch": 0.9595959595959596, + "grad_norm": 2.4667656421661377, + "learning_rate": 3.8143124415341442e-06, + "loss": 1.4942, + "step": 760 + }, + { + "epoch": 0.9595959595959596, + "eval_loss": 1.7205606698989868, + "eval_runtime": 37.1654, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 760 + }, + { + "epoch": 0.9608585858585859, + "grad_norm": 2.456643581390381, + "learning_rate": 3.811973807296539e-06, + "loss": 1.538, + "step": 761 + }, + { + "epoch": 0.9621212121212122, + "grad_norm": 2.696720838546753, + "learning_rate": 3.8096351730589335e-06, + "loss": 1.5462, + "step": 762 + }, + { + "epoch": 0.9633838383838383, + "grad_norm": 2.7314131259918213, + "learning_rate": 3.8072965388213284e-06, + "loss": 1.5931, + "step": 763 + }, + { + "epoch": 0.9646464646464646, + "grad_norm": 2.616532325744629, + "learning_rate": 3.804957904583723e-06, + "loss": 1.5469, + "step": 764 + }, + { + "epoch": 0.9659090909090909, + "grad_norm": 2.7471885681152344, + "learning_rate": 3.8026192703461185e-06, + "loss": 1.5431, + "step": 765 + }, + { + "epoch": 0.9671717171717171, + "grad_norm": 2.7478253841400146, + "learning_rate": 3.8002806361085133e-06, + "loss": 1.5452, + "step": 766 + }, + { + "epoch": 0.9684343434343434, + "grad_norm": 2.543790340423584, + "learning_rate": 3.7979420018709077e-06, + "loss": 1.6091, + "step": 767 + }, + { + "epoch": 0.9696969696969697, + "grad_norm": 2.6169509887695312, + "learning_rate": 3.7956033676333026e-06, + "loss": 1.5557, + "step": 768 + }, + { + "epoch": 0.9709595959595959, + "grad_norm": 2.4768588542938232, + "learning_rate": 3.7932647333956974e-06, + "loss": 1.5247, + "step": 769 + }, + { + "epoch": 0.9722222222222222, + "grad_norm": 2.654648780822754, + "learning_rate": 3.7909260991580923e-06, + "loss": 1.5281, + "step": 770 + }, + { + "epoch": 0.9722222222222222, + "eval_loss": 1.7289395332336426, + "eval_runtime": 37.1374, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 770 + }, + { + "epoch": 0.9734848484848485, + "grad_norm": 2.893073081970215, + "learning_rate": 3.7885874649204867e-06, + "loss": 1.5749, + "step": 771 + }, + { + "epoch": 0.9747474747474747, + "grad_norm": 2.696341037750244, + "learning_rate": 3.7862488306828815e-06, + "loss": 1.5025, + "step": 772 + }, + { + "epoch": 0.976010101010101, + "grad_norm": 2.678222417831421, + "learning_rate": 3.7839101964452764e-06, + "loss": 1.538, + "step": 773 + }, + { + "epoch": 0.9772727272727273, + "grad_norm": 2.7772746086120605, + "learning_rate": 3.7815715622076708e-06, + "loss": 1.5648, + "step": 774 + }, + { + "epoch": 0.9785353535353535, + "grad_norm": 3.076620101928711, + "learning_rate": 3.7792329279700656e-06, + "loss": 1.6114, + "step": 775 + }, + { + "epoch": 0.9797979797979798, + "grad_norm": 2.700056791305542, + "learning_rate": 3.7768942937324605e-06, + "loss": 1.5881, + "step": 776 + }, + { + "epoch": 0.9810606060606061, + "grad_norm": 2.601419687271118, + "learning_rate": 3.7745556594948553e-06, + "loss": 1.5319, + "step": 777 + }, + { + "epoch": 0.9823232323232324, + "grad_norm": 2.6664950847625732, + "learning_rate": 3.7722170252572497e-06, + "loss": 1.5497, + "step": 778 + }, + { + "epoch": 0.9835858585858586, + "grad_norm": 2.7374589443206787, + "learning_rate": 3.7698783910196446e-06, + "loss": 1.5488, + "step": 779 + }, + { + "epoch": 0.9848484848484849, + "grad_norm": 2.797938823699951, + "learning_rate": 3.76753975678204e-06, + "loss": 1.5376, + "step": 780 + }, + { + "epoch": 0.9848484848484849, + "eval_loss": 1.7334709167480469, + "eval_runtime": 37.1372, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 780 + }, + { + "epoch": 0.9861111111111112, + "grad_norm": 2.881591796875, + "learning_rate": 3.7652011225444347e-06, + "loss": 1.5405, + "step": 781 + }, + { + "epoch": 0.9873737373737373, + "grad_norm": 2.6965644359588623, + "learning_rate": 3.7628624883068295e-06, + "loss": 1.5066, + "step": 782 + }, + { + "epoch": 0.9886363636363636, + "grad_norm": 2.7286899089813232, + "learning_rate": 3.760523854069224e-06, + "loss": 1.5277, + "step": 783 + }, + { + "epoch": 0.98989898989899, + "grad_norm": 2.795644998550415, + "learning_rate": 3.7581852198316188e-06, + "loss": 1.5335, + "step": 784 + }, + { + "epoch": 0.9911616161616161, + "grad_norm": 2.8160440921783447, + "learning_rate": 3.7558465855940136e-06, + "loss": 1.585, + "step": 785 + }, + { + "epoch": 0.9924242424242424, + "grad_norm": 2.7984936237335205, + "learning_rate": 3.753507951356408e-06, + "loss": 1.5569, + "step": 786 + }, + { + "epoch": 0.9936868686868687, + "grad_norm": 2.7596278190612793, + "learning_rate": 3.751169317118803e-06, + "loss": 1.4838, + "step": 787 + }, + { + "epoch": 0.9949494949494949, + "grad_norm": 2.666383743286133, + "learning_rate": 3.7488306828811977e-06, + "loss": 1.5656, + "step": 788 + }, + { + "epoch": 0.9962121212121212, + "grad_norm": 2.8139331340789795, + "learning_rate": 3.7464920486435926e-06, + "loss": 1.5753, + "step": 789 + }, + { + "epoch": 0.9974747474747475, + "grad_norm": 2.7714269161224365, + "learning_rate": 3.744153414405987e-06, + "loss": 1.5187, + "step": 790 + }, + { + "epoch": 0.9974747474747475, + "eval_loss": 1.7257990837097168, + "eval_runtime": 37.1065, + "eval_samples_per_second": 26.949, + "eval_steps_per_second": 1.132, + "step": 790 + }, + { + "epoch": 0.9987373737373737, + "grad_norm": 2.697606325149536, + "learning_rate": 3.741814780168382e-06, + "loss": 1.56, + "step": 791 + }, + { + "epoch": 1.0, + "grad_norm": 2.7685229778289795, + "learning_rate": 3.7394761459307767e-06, + "loss": 1.5705, + "step": 792 + }, + { + "epoch": 1.0012626262626263, + "grad_norm": 2.778200626373291, + "learning_rate": 3.737137511693171e-06, + "loss": 1.5288, + "step": 793 + }, + { + "epoch": 1.0025252525252526, + "grad_norm": 2.7211337089538574, + "learning_rate": 3.734798877455566e-06, + "loss": 1.5348, + "step": 794 + }, + { + "epoch": 1.003787878787879, + "grad_norm": 2.690739154815674, + "learning_rate": 3.7324602432179608e-06, + "loss": 1.4734, + "step": 795 + }, + { + "epoch": 1.005050505050505, + "grad_norm": 2.6809420585632324, + "learning_rate": 3.730121608980356e-06, + "loss": 1.4852, + "step": 796 + }, + { + "epoch": 1.0063131313131313, + "grad_norm": 2.8550117015838623, + "learning_rate": 3.727782974742751e-06, + "loss": 1.4562, + "step": 797 + }, + { + "epoch": 1.0075757575757576, + "grad_norm": 2.713935613632202, + "learning_rate": 3.7254443405051453e-06, + "loss": 1.5005, + "step": 798 + }, + { + "epoch": 1.0088383838383839, + "grad_norm": 2.668363332748413, + "learning_rate": 3.72310570626754e-06, + "loss": 1.485, + "step": 799 + }, + { + "epoch": 1.0101010101010102, + "grad_norm": 2.840769052505493, + "learning_rate": 3.720767072029935e-06, + "loss": 1.4531, + "step": 800 + }, + { + "epoch": 1.0101010101010102, + "eval_loss": 1.725937008857727, + "eval_runtime": 37.2471, + "eval_samples_per_second": 26.848, + "eval_steps_per_second": 1.128, + "step": 800 + }, + { + "epoch": 1.0113636363636365, + "grad_norm": 2.7345798015594482, + "learning_rate": 3.71842843779233e-06, + "loss": 1.5091, + "step": 801 + }, + { + "epoch": 1.0126262626262625, + "grad_norm": 2.9052624702453613, + "learning_rate": 3.7160898035547243e-06, + "loss": 1.5039, + "step": 802 + }, + { + "epoch": 1.0138888888888888, + "grad_norm": 2.893073558807373, + "learning_rate": 3.713751169317119e-06, + "loss": 1.4613, + "step": 803 + }, + { + "epoch": 1.0151515151515151, + "grad_norm": 2.8505427837371826, + "learning_rate": 3.711412535079514e-06, + "loss": 1.4932, + "step": 804 + }, + { + "epoch": 1.0164141414141414, + "grad_norm": 2.9065170288085938, + "learning_rate": 3.7090739008419084e-06, + "loss": 1.5082, + "step": 805 + }, + { + "epoch": 1.0176767676767677, + "grad_norm": 2.888068675994873, + "learning_rate": 3.706735266604303e-06, + "loss": 1.5675, + "step": 806 + }, + { + "epoch": 1.018939393939394, + "grad_norm": 2.7325260639190674, + "learning_rate": 3.704396632366698e-06, + "loss": 1.4649, + "step": 807 + }, + { + "epoch": 1.02020202020202, + "grad_norm": 2.8861136436462402, + "learning_rate": 3.702057998129093e-06, + "loss": 1.4517, + "step": 808 + }, + { + "epoch": 1.0214646464646464, + "grad_norm": 2.8837692737579346, + "learning_rate": 3.6997193638914873e-06, + "loss": 1.5198, + "step": 809 + }, + { + "epoch": 1.0227272727272727, + "grad_norm": 2.790330410003662, + "learning_rate": 3.697380729653882e-06, + "loss": 1.5066, + "step": 810 + }, + { + "epoch": 1.0227272727272727, + "eval_loss": 1.7269084453582764, + "eval_runtime": 36.89, + "eval_samples_per_second": 27.108, + "eval_steps_per_second": 1.139, + "step": 810 + }, + { + "epoch": 1.023989898989899, + "grad_norm": 3.03643536567688, + "learning_rate": 3.6950420954162774e-06, + "loss": 1.4928, + "step": 811 + }, + { + "epoch": 1.0252525252525253, + "grad_norm": 2.841055154800415, + "learning_rate": 3.6927034611786723e-06, + "loss": 1.5273, + "step": 812 + }, + { + "epoch": 1.0265151515151516, + "grad_norm": 2.84134578704834, + "learning_rate": 3.690364826941067e-06, + "loss": 1.5443, + "step": 813 + }, + { + "epoch": 1.0277777777777777, + "grad_norm": 2.837758779525757, + "learning_rate": 3.6880261927034615e-06, + "loss": 1.5586, + "step": 814 + }, + { + "epoch": 1.029040404040404, + "grad_norm": 2.731412410736084, + "learning_rate": 3.6856875584658564e-06, + "loss": 1.4465, + "step": 815 + }, + { + "epoch": 1.0303030303030303, + "grad_norm": 2.8979599475860596, + "learning_rate": 3.683348924228251e-06, + "loss": 1.5663, + "step": 816 + }, + { + "epoch": 1.0315656565656566, + "grad_norm": 2.7740373611450195, + "learning_rate": 3.6810102899906456e-06, + "loss": 1.4578, + "step": 817 + }, + { + "epoch": 1.0328282828282829, + "grad_norm": 2.8984005451202393, + "learning_rate": 3.6786716557530405e-06, + "loss": 1.493, + "step": 818 + }, + { + "epoch": 1.0340909090909092, + "grad_norm": 2.880680561065674, + "learning_rate": 3.6763330215154353e-06, + "loss": 1.4858, + "step": 819 + }, + { + "epoch": 1.0353535353535352, + "grad_norm": 3.0220961570739746, + "learning_rate": 3.67399438727783e-06, + "loss": 1.5009, + "step": 820 + }, + { + "epoch": 1.0353535353535352, + "eval_loss": 1.7245674133300781, + "eval_runtime": 37.0235, + "eval_samples_per_second": 27.01, + "eval_steps_per_second": 1.134, + "step": 820 + }, + { + "epoch": 1.0366161616161615, + "grad_norm": 2.7905962467193604, + "learning_rate": 3.6716557530402246e-06, + "loss": 1.4728, + "step": 821 + }, + { + "epoch": 1.0378787878787878, + "grad_norm": 2.856935501098633, + "learning_rate": 3.6693171188026194e-06, + "loss": 1.521, + "step": 822 + }, + { + "epoch": 1.0391414141414141, + "grad_norm": 2.956584930419922, + "learning_rate": 3.6669784845650143e-06, + "loss": 1.4725, + "step": 823 + }, + { + "epoch": 1.0404040404040404, + "grad_norm": 3.130960702896118, + "learning_rate": 3.6646398503274087e-06, + "loss": 1.6127, + "step": 824 + }, + { + "epoch": 1.0416666666666667, + "grad_norm": 2.8855528831481934, + "learning_rate": 3.6623012160898035e-06, + "loss": 1.5078, + "step": 825 + }, + { + "epoch": 1.0429292929292928, + "grad_norm": 2.8030972480773926, + "learning_rate": 3.659962581852199e-06, + "loss": 1.4635, + "step": 826 + }, + { + "epoch": 1.0441919191919191, + "grad_norm": 3.061593770980835, + "learning_rate": 3.6576239476145936e-06, + "loss": 1.4944, + "step": 827 + }, + { + "epoch": 1.0454545454545454, + "grad_norm": 2.9983572959899902, + "learning_rate": 3.6552853133769885e-06, + "loss": 1.505, + "step": 828 + }, + { + "epoch": 1.0467171717171717, + "grad_norm": 2.7681095600128174, + "learning_rate": 3.652946679139383e-06, + "loss": 1.4774, + "step": 829 + }, + { + "epoch": 1.047979797979798, + "grad_norm": 3.0530946254730225, + "learning_rate": 3.6506080449017777e-06, + "loss": 1.4676, + "step": 830 + }, + { + "epoch": 1.047979797979798, + "eval_loss": 1.7242811918258667, + "eval_runtime": 37.0752, + "eval_samples_per_second": 26.972, + "eval_steps_per_second": 1.133, + "step": 830 + }, + { + "epoch": 1.0492424242424243, + "grad_norm": 2.8383126258850098, + "learning_rate": 3.6482694106641726e-06, + "loss": 1.486, + "step": 831 + }, + { + "epoch": 1.0505050505050506, + "grad_norm": 3.095928907394409, + "learning_rate": 3.6459307764265674e-06, + "loss": 1.5331, + "step": 832 + }, + { + "epoch": 1.0517676767676767, + "grad_norm": 2.9225456714630127, + "learning_rate": 3.643592142188962e-06, + "loss": 1.556, + "step": 833 + }, + { + "epoch": 1.053030303030303, + "grad_norm": 2.9037210941314697, + "learning_rate": 3.6412535079513567e-06, + "loss": 1.4834, + "step": 834 + }, + { + "epoch": 1.0542929292929293, + "grad_norm": 2.915022373199463, + "learning_rate": 3.6389148737137515e-06, + "loss": 1.5346, + "step": 835 + }, + { + "epoch": 1.0555555555555556, + "grad_norm": 3.250411033630371, + "learning_rate": 3.636576239476146e-06, + "loss": 1.4885, + "step": 836 + }, + { + "epoch": 1.0568181818181819, + "grad_norm": 2.759061336517334, + "learning_rate": 3.634237605238541e-06, + "loss": 1.456, + "step": 837 + }, + { + "epoch": 1.0580808080808082, + "grad_norm": 2.955981731414795, + "learning_rate": 3.6318989710009356e-06, + "loss": 1.4889, + "step": 838 + }, + { + "epoch": 1.0593434343434343, + "grad_norm": 3.243338108062744, + "learning_rate": 3.6295603367633305e-06, + "loss": 1.5207, + "step": 839 + }, + { + "epoch": 1.0606060606060606, + "grad_norm": 3.075840950012207, + "learning_rate": 3.627221702525725e-06, + "loss": 1.4906, + "step": 840 + }, + { + "epoch": 1.0606060606060606, + "eval_loss": 1.7290290594100952, + "eval_runtime": 37.097, + "eval_samples_per_second": 26.956, + "eval_steps_per_second": 1.132, + "step": 840 + }, + { + "epoch": 1.0618686868686869, + "grad_norm": 3.149125099182129, + "learning_rate": 3.62488306828812e-06, + "loss": 1.4755, + "step": 841 + }, + { + "epoch": 1.0631313131313131, + "grad_norm": 3.099318504333496, + "learning_rate": 3.622544434050515e-06, + "loss": 1.534, + "step": 842 + }, + { + "epoch": 1.0643939393939394, + "grad_norm": 2.8942148685455322, + "learning_rate": 3.62020579981291e-06, + "loss": 1.4844, + "step": 843 + }, + { + "epoch": 1.0656565656565657, + "grad_norm": 3.142664670944214, + "learning_rate": 3.6178671655753047e-06, + "loss": 1.5179, + "step": 844 + }, + { + "epoch": 1.0669191919191918, + "grad_norm": 3.106552839279175, + "learning_rate": 3.615528531337699e-06, + "loss": 1.4667, + "step": 845 + }, + { + "epoch": 1.0681818181818181, + "grad_norm": 2.8507936000823975, + "learning_rate": 3.613189897100094e-06, + "loss": 1.4463, + "step": 846 + }, + { + "epoch": 1.0694444444444444, + "grad_norm": 3.2198901176452637, + "learning_rate": 3.610851262862489e-06, + "loss": 1.5146, + "step": 847 + }, + { + "epoch": 1.0707070707070707, + "grad_norm": 3.146388292312622, + "learning_rate": 3.6085126286248832e-06, + "loss": 1.4927, + "step": 848 + }, + { + "epoch": 1.071969696969697, + "grad_norm": 2.8304760456085205, + "learning_rate": 3.606173994387278e-06, + "loss": 1.4971, + "step": 849 + }, + { + "epoch": 1.0732323232323233, + "grad_norm": 3.1714529991149902, + "learning_rate": 3.603835360149673e-06, + "loss": 1.5418, + "step": 850 + }, + { + "epoch": 1.0732323232323233, + "eval_loss": 1.725188970565796, + "eval_runtime": 37.117, + "eval_samples_per_second": 26.942, + "eval_steps_per_second": 1.132, + "step": 850 + }, + { + "epoch": 1.0744949494949494, + "grad_norm": 3.06162428855896, + "learning_rate": 3.6014967259120677e-06, + "loss": 1.4959, + "step": 851 + }, + { + "epoch": 1.0757575757575757, + "grad_norm": 2.812206268310547, + "learning_rate": 3.599158091674462e-06, + "loss": 1.5249, + "step": 852 + }, + { + "epoch": 1.077020202020202, + "grad_norm": 3.236210584640503, + "learning_rate": 3.596819457436857e-06, + "loss": 1.5166, + "step": 853 + }, + { + "epoch": 1.0782828282828283, + "grad_norm": 2.8397765159606934, + "learning_rate": 3.594480823199252e-06, + "loss": 1.5043, + "step": 854 + }, + { + "epoch": 1.0795454545454546, + "grad_norm": 2.9215919971466064, + "learning_rate": 3.5921421889616463e-06, + "loss": 1.5235, + "step": 855 + }, + { + "epoch": 1.0808080808080809, + "grad_norm": 2.8219058513641357, + "learning_rate": 3.589803554724041e-06, + "loss": 1.5091, + "step": 856 + }, + { + "epoch": 1.0820707070707072, + "grad_norm": 2.9344675540924072, + "learning_rate": 3.5874649204864364e-06, + "loss": 1.5355, + "step": 857 + }, + { + "epoch": 1.0833333333333333, + "grad_norm": 2.949094533920288, + "learning_rate": 3.5851262862488312e-06, + "loss": 1.5273, + "step": 858 + }, + { + "epoch": 1.0845959595959596, + "grad_norm": 2.9801411628723145, + "learning_rate": 3.582787652011226e-06, + "loss": 1.5078, + "step": 859 + }, + { + "epoch": 1.0858585858585859, + "grad_norm": 2.802182912826538, + "learning_rate": 3.5804490177736205e-06, + "loss": 1.5015, + "step": 860 + }, + { + "epoch": 1.0858585858585859, + "eval_loss": 1.7209502458572388, + "eval_runtime": 37.1238, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 860 + }, + { + "epoch": 1.0871212121212122, + "grad_norm": 2.857457160949707, + "learning_rate": 3.5781103835360153e-06, + "loss": 1.5057, + "step": 861 + }, + { + "epoch": 1.0883838383838385, + "grad_norm": 2.794478178024292, + "learning_rate": 3.57577174929841e-06, + "loss": 1.4594, + "step": 862 + }, + { + "epoch": 1.0896464646464645, + "grad_norm": 3.0028598308563232, + "learning_rate": 3.573433115060805e-06, + "loss": 1.4666, + "step": 863 + }, + { + "epoch": 1.0909090909090908, + "grad_norm": 2.7701406478881836, + "learning_rate": 3.5710944808231994e-06, + "loss": 1.4596, + "step": 864 + }, + { + "epoch": 1.0921717171717171, + "grad_norm": 2.9705960750579834, + "learning_rate": 3.5687558465855943e-06, + "loss": 1.5092, + "step": 865 + }, + { + "epoch": 1.0934343434343434, + "grad_norm": 2.9820821285247803, + "learning_rate": 3.566417212347989e-06, + "loss": 1.5246, + "step": 866 + }, + { + "epoch": 1.0946969696969697, + "grad_norm": 2.988356113433838, + "learning_rate": 3.5640785781103835e-06, + "loss": 1.5035, + "step": 867 + }, + { + "epoch": 1.095959595959596, + "grad_norm": 2.901599884033203, + "learning_rate": 3.5617399438727784e-06, + "loss": 1.5097, + "step": 868 + }, + { + "epoch": 1.0972222222222223, + "grad_norm": 2.9128799438476562, + "learning_rate": 3.5594013096351732e-06, + "loss": 1.5424, + "step": 869 + }, + { + "epoch": 1.0984848484848484, + "grad_norm": 2.719289779663086, + "learning_rate": 3.557062675397568e-06, + "loss": 1.4313, + "step": 870 + }, + { + "epoch": 1.0984848484848484, + "eval_loss": 1.7159698009490967, + "eval_runtime": 37.122, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 870 + }, + { + "epoch": 1.0997474747474747, + "grad_norm": 3.0080976486206055, + "learning_rate": 3.5547240411599625e-06, + "loss": 1.4483, + "step": 871 + }, + { + "epoch": 1.101010101010101, + "grad_norm": 2.856523036956787, + "learning_rate": 3.5523854069223578e-06, + "loss": 1.5463, + "step": 872 + }, + { + "epoch": 1.1022727272727273, + "grad_norm": 2.8585214614868164, + "learning_rate": 3.5500467726847526e-06, + "loss": 1.5391, + "step": 873 + }, + { + "epoch": 1.1035353535353536, + "grad_norm": 2.8440098762512207, + "learning_rate": 3.5477081384471474e-06, + "loss": 1.499, + "step": 874 + }, + { + "epoch": 1.10479797979798, + "grad_norm": 2.8114542961120605, + "learning_rate": 3.545369504209542e-06, + "loss": 1.4691, + "step": 875 + }, + { + "epoch": 1.106060606060606, + "grad_norm": 2.980273962020874, + "learning_rate": 3.5430308699719367e-06, + "loss": 1.4541, + "step": 876 + }, + { + "epoch": 1.1073232323232323, + "grad_norm": 3.0053417682647705, + "learning_rate": 3.5406922357343315e-06, + "loss": 1.4895, + "step": 877 + }, + { + "epoch": 1.1085858585858586, + "grad_norm": 2.924046754837036, + "learning_rate": 3.5383536014967264e-06, + "loss": 1.5186, + "step": 878 + }, + { + "epoch": 1.1098484848484849, + "grad_norm": 2.969461679458618, + "learning_rate": 3.536014967259121e-06, + "loss": 1.5213, + "step": 879 + }, + { + "epoch": 1.1111111111111112, + "grad_norm": 2.8936445713043213, + "learning_rate": 3.5336763330215156e-06, + "loss": 1.5585, + "step": 880 + }, + { + "epoch": 1.1111111111111112, + "eval_loss": 1.7150688171386719, + "eval_runtime": 37.1105, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 880 + }, + { + "epoch": 1.1123737373737375, + "grad_norm": 2.992034435272217, + "learning_rate": 3.5313376987839105e-06, + "loss": 1.5264, + "step": 881 + }, + { + "epoch": 1.1136363636363635, + "grad_norm": 3.0453898906707764, + "learning_rate": 3.5289990645463053e-06, + "loss": 1.5482, + "step": 882 + }, + { + "epoch": 1.1148989898989898, + "grad_norm": 2.7480897903442383, + "learning_rate": 3.5266604303086998e-06, + "loss": 1.4851, + "step": 883 + }, + { + "epoch": 1.1161616161616161, + "grad_norm": 2.9408013820648193, + "learning_rate": 3.5243217960710946e-06, + "loss": 1.536, + "step": 884 + }, + { + "epoch": 1.1174242424242424, + "grad_norm": 2.854438543319702, + "learning_rate": 3.5219831618334894e-06, + "loss": 1.5814, + "step": 885 + }, + { + "epoch": 1.1186868686868687, + "grad_norm": 2.9060397148132324, + "learning_rate": 3.519644527595884e-06, + "loss": 1.5019, + "step": 886 + }, + { + "epoch": 1.119949494949495, + "grad_norm": 2.8741681575775146, + "learning_rate": 3.517305893358279e-06, + "loss": 1.4713, + "step": 887 + }, + { + "epoch": 1.121212121212121, + "grad_norm": 2.972510576248169, + "learning_rate": 3.514967259120674e-06, + "loss": 1.5104, + "step": 888 + }, + { + "epoch": 1.1224747474747474, + "grad_norm": 2.8113558292388916, + "learning_rate": 3.512628624883069e-06, + "loss": 1.4973, + "step": 889 + }, + { + "epoch": 1.1237373737373737, + "grad_norm": 2.9687416553497314, + "learning_rate": 3.5102899906454637e-06, + "loss": 1.4774, + "step": 890 + }, + { + "epoch": 1.1237373737373737, + "eval_loss": 1.7169743776321411, + "eval_runtime": 37.0911, + "eval_samples_per_second": 26.961, + "eval_steps_per_second": 1.132, + "step": 890 + }, + { + "epoch": 1.125, + "grad_norm": 2.911409854888916, + "learning_rate": 3.507951356407858e-06, + "loss": 1.4536, + "step": 891 + }, + { + "epoch": 1.1262626262626263, + "grad_norm": 2.8209526538848877, + "learning_rate": 3.505612722170253e-06, + "loss": 1.5357, + "step": 892 + }, + { + "epoch": 1.1275252525252526, + "grad_norm": 2.811971426010132, + "learning_rate": 3.5032740879326478e-06, + "loss": 1.4889, + "step": 893 + }, + { + "epoch": 1.128787878787879, + "grad_norm": 2.969878673553467, + "learning_rate": 3.5009354536950426e-06, + "loss": 1.4758, + "step": 894 + }, + { + "epoch": 1.130050505050505, + "grad_norm": 2.858635425567627, + "learning_rate": 3.498596819457437e-06, + "loss": 1.5355, + "step": 895 + }, + { + "epoch": 1.1313131313131313, + "grad_norm": 3.018164873123169, + "learning_rate": 3.496258185219832e-06, + "loss": 1.5421, + "step": 896 + }, + { + "epoch": 1.1325757575757576, + "grad_norm": 2.846277952194214, + "learning_rate": 3.4939195509822267e-06, + "loss": 1.4936, + "step": 897 + }, + { + "epoch": 1.1338383838383839, + "grad_norm": 2.965245246887207, + "learning_rate": 3.491580916744621e-06, + "loss": 1.4903, + "step": 898 + }, + { + "epoch": 1.1351010101010102, + "grad_norm": 2.8500895500183105, + "learning_rate": 3.489242282507016e-06, + "loss": 1.5152, + "step": 899 + }, + { + "epoch": 1.1363636363636362, + "grad_norm": 2.8242156505584717, + "learning_rate": 3.486903648269411e-06, + "loss": 1.5531, + "step": 900 + }, + { + "epoch": 1.1363636363636362, + "eval_loss": 1.7180876731872559, + "eval_runtime": 37.4145, + "eval_samples_per_second": 26.728, + "eval_steps_per_second": 1.123, + "step": 900 + }, + { + "epoch": 1.1376262626262625, + "grad_norm": 2.8846635818481445, + "learning_rate": 3.4845650140318057e-06, + "loss": 1.5314, + "step": 901 + }, + { + "epoch": 1.1388888888888888, + "grad_norm": 2.9443984031677246, + "learning_rate": 3.4822263797942e-06, + "loss": 1.4699, + "step": 902 + }, + { + "epoch": 1.1401515151515151, + "grad_norm": 2.7435739040374756, + "learning_rate": 3.4798877455565953e-06, + "loss": 1.4949, + "step": 903 + }, + { + "epoch": 1.1414141414141414, + "grad_norm": 2.979102373123169, + "learning_rate": 3.47754911131899e-06, + "loss": 1.5689, + "step": 904 + }, + { + "epoch": 1.1426767676767677, + "grad_norm": 2.9814586639404297, + "learning_rate": 3.475210477081385e-06, + "loss": 1.5183, + "step": 905 + }, + { + "epoch": 1.143939393939394, + "grad_norm": 2.8380966186523438, + "learning_rate": 3.4728718428437794e-06, + "loss": 1.5177, + "step": 906 + }, + { + "epoch": 1.14520202020202, + "grad_norm": 2.8478410243988037, + "learning_rate": 3.4705332086061743e-06, + "loss": 1.4787, + "step": 907 + }, + { + "epoch": 1.1464646464646464, + "grad_norm": 2.8247780799865723, + "learning_rate": 3.468194574368569e-06, + "loss": 1.425, + "step": 908 + }, + { + "epoch": 1.1477272727272727, + "grad_norm": 2.869649887084961, + "learning_rate": 3.465855940130964e-06, + "loss": 1.4619, + "step": 909 + }, + { + "epoch": 1.148989898989899, + "grad_norm": 2.9439175128936768, + "learning_rate": 3.4635173058933584e-06, + "loss": 1.5194, + "step": 910 + }, + { + "epoch": 1.148989898989899, + "eval_loss": 1.7227450609207153, + "eval_runtime": 36.9723, + "eval_samples_per_second": 27.047, + "eval_steps_per_second": 1.136, + "step": 910 + }, + { + "epoch": 1.1502525252525253, + "grad_norm": 3.0219292640686035, + "learning_rate": 3.4611786716557532e-06, + "loss": 1.5878, + "step": 911 + }, + { + "epoch": 1.1515151515151516, + "grad_norm": 2.988475799560547, + "learning_rate": 3.458840037418148e-06, + "loss": 1.5349, + "step": 912 + }, + { + "epoch": 1.1527777777777777, + "grad_norm": 2.708740234375, + "learning_rate": 3.456501403180543e-06, + "loss": 1.4637, + "step": 913 + }, + { + "epoch": 1.154040404040404, + "grad_norm": 3.0844545364379883, + "learning_rate": 3.4541627689429373e-06, + "loss": 1.546, + "step": 914 + }, + { + "epoch": 1.1553030303030303, + "grad_norm": 2.800487518310547, + "learning_rate": 3.451824134705332e-06, + "loss": 1.5374, + "step": 915 + }, + { + "epoch": 1.1565656565656566, + "grad_norm": 2.8701722621917725, + "learning_rate": 3.449485500467727e-06, + "loss": 1.4549, + "step": 916 + }, + { + "epoch": 1.1578282828282829, + "grad_norm": 3.036418914794922, + "learning_rate": 3.4471468662301214e-06, + "loss": 1.5253, + "step": 917 + }, + { + "epoch": 1.1590909090909092, + "grad_norm": 3.0300986766815186, + "learning_rate": 3.4448082319925167e-06, + "loss": 1.4727, + "step": 918 + }, + { + "epoch": 1.1603535353535355, + "grad_norm": 2.8652496337890625, + "learning_rate": 3.4424695977549116e-06, + "loss": 1.4842, + "step": 919 + }, + { + "epoch": 1.1616161616161615, + "grad_norm": 2.9140491485595703, + "learning_rate": 3.4401309635173064e-06, + "loss": 1.4968, + "step": 920 + }, + { + "epoch": 1.1616161616161615, + "eval_loss": 1.7184255123138428, + "eval_runtime": 37.1186, + "eval_samples_per_second": 26.941, + "eval_steps_per_second": 1.132, + "step": 920 + }, + { + "epoch": 1.1628787878787878, + "grad_norm": 2.9708127975463867, + "learning_rate": 3.4377923292797012e-06, + "loss": 1.4817, + "step": 921 + }, + { + "epoch": 1.1641414141414141, + "grad_norm": 2.954814910888672, + "learning_rate": 3.4354536950420957e-06, + "loss": 1.4937, + "step": 922 + }, + { + "epoch": 1.1654040404040404, + "grad_norm": 2.803945779800415, + "learning_rate": 3.4331150608044905e-06, + "loss": 1.527, + "step": 923 + }, + { + "epoch": 1.1666666666666667, + "grad_norm": 2.863250255584717, + "learning_rate": 3.4307764265668853e-06, + "loss": 1.4922, + "step": 924 + }, + { + "epoch": 1.1679292929292928, + "grad_norm": 2.8030712604522705, + "learning_rate": 3.4284377923292798e-06, + "loss": 1.4427, + "step": 925 + }, + { + "epoch": 1.1691919191919191, + "grad_norm": 2.944110155105591, + "learning_rate": 3.4260991580916746e-06, + "loss": 1.4435, + "step": 926 + }, + { + "epoch": 1.1704545454545454, + "grad_norm": 2.9672467708587646, + "learning_rate": 3.4237605238540695e-06, + "loss": 1.5005, + "step": 927 + }, + { + "epoch": 1.1717171717171717, + "grad_norm": 2.920205593109131, + "learning_rate": 3.4214218896164643e-06, + "loss": 1.5303, + "step": 928 + }, + { + "epoch": 1.172979797979798, + "grad_norm": 3.1062121391296387, + "learning_rate": 3.4190832553788587e-06, + "loss": 1.5395, + "step": 929 + }, + { + "epoch": 1.1742424242424243, + "grad_norm": 3.1106555461883545, + "learning_rate": 3.4167446211412536e-06, + "loss": 1.5038, + "step": 930 + }, + { + "epoch": 1.1742424242424243, + "eval_loss": 1.7127314805984497, + "eval_runtime": 37.1193, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 930 + }, + { + "epoch": 1.1755050505050506, + "grad_norm": 2.7898054122924805, + "learning_rate": 3.4144059869036484e-06, + "loss": 1.5096, + "step": 931 + }, + { + "epoch": 1.1767676767676767, + "grad_norm": 2.881694793701172, + "learning_rate": 3.4120673526660432e-06, + "loss": 1.4556, + "step": 932 + }, + { + "epoch": 1.178030303030303, + "grad_norm": 3.131551742553711, + "learning_rate": 3.4097287184284385e-06, + "loss": 1.5487, + "step": 933 + }, + { + "epoch": 1.1792929292929293, + "grad_norm": 2.9734153747558594, + "learning_rate": 3.407390084190833e-06, + "loss": 1.439, + "step": 934 + }, + { + "epoch": 1.1805555555555556, + "grad_norm": 3.009657144546509, + "learning_rate": 3.4050514499532278e-06, + "loss": 1.462, + "step": 935 + }, + { + "epoch": 1.1818181818181819, + "grad_norm": 2.8378758430480957, + "learning_rate": 3.4027128157156226e-06, + "loss": 1.5083, + "step": 936 + }, + { + "epoch": 1.1830808080808082, + "grad_norm": 2.923969268798828, + "learning_rate": 3.400374181478017e-06, + "loss": 1.5844, + "step": 937 + }, + { + "epoch": 1.1843434343434343, + "grad_norm": 3.2384307384490967, + "learning_rate": 3.398035547240412e-06, + "loss": 1.5175, + "step": 938 + }, + { + "epoch": 1.1856060606060606, + "grad_norm": 3.024385929107666, + "learning_rate": 3.3956969130028067e-06, + "loss": 1.4756, + "step": 939 + }, + { + "epoch": 1.1868686868686869, + "grad_norm": 3.088745594024658, + "learning_rate": 3.3933582787652016e-06, + "loss": 1.4914, + "step": 940 + }, + { + "epoch": 1.1868686868686869, + "eval_loss": 1.705490231513977, + "eval_runtime": 37.1291, + "eval_samples_per_second": 26.933, + "eval_steps_per_second": 1.131, + "step": 940 + }, + { + "epoch": 1.1881313131313131, + "grad_norm": 2.898893117904663, + "learning_rate": 3.391019644527596e-06, + "loss": 1.5002, + "step": 941 + }, + { + "epoch": 1.1893939393939394, + "grad_norm": 2.800103187561035, + "learning_rate": 3.388681010289991e-06, + "loss": 1.4842, + "step": 942 + }, + { + "epoch": 1.1906565656565657, + "grad_norm": 3.0287230014801025, + "learning_rate": 3.3863423760523857e-06, + "loss": 1.5454, + "step": 943 + }, + { + "epoch": 1.1919191919191918, + "grad_norm": 2.9379546642303467, + "learning_rate": 3.3840037418147805e-06, + "loss": 1.5427, + "step": 944 + }, + { + "epoch": 1.1931818181818181, + "grad_norm": 2.879699468612671, + "learning_rate": 3.381665107577175e-06, + "loss": 1.4804, + "step": 945 + }, + { + "epoch": 1.1944444444444444, + "grad_norm": 2.8708608150482178, + "learning_rate": 3.3793264733395698e-06, + "loss": 1.479, + "step": 946 + }, + { + "epoch": 1.1957070707070707, + "grad_norm": 2.8513312339782715, + "learning_rate": 3.3769878391019646e-06, + "loss": 1.5343, + "step": 947 + }, + { + "epoch": 1.196969696969697, + "grad_norm": 2.8715100288391113, + "learning_rate": 3.37464920486436e-06, + "loss": 1.5114, + "step": 948 + }, + { + "epoch": 1.1982323232323233, + "grad_norm": 2.819213628768921, + "learning_rate": 3.3723105706267543e-06, + "loss": 1.4873, + "step": 949 + }, + { + "epoch": 1.1994949494949494, + "grad_norm": 2.946578025817871, + "learning_rate": 3.369971936389149e-06, + "loss": 1.5378, + "step": 950 + }, + { + "epoch": 1.1994949494949494, + "eval_loss": 1.7094330787658691, + "eval_runtime": 37.1453, + "eval_samples_per_second": 26.921, + "eval_steps_per_second": 1.131, + "step": 950 + }, + { + "epoch": 1.2007575757575757, + "grad_norm": 2.8170437812805176, + "learning_rate": 3.367633302151544e-06, + "loss": 1.4416, + "step": 951 + }, + { + "epoch": 1.202020202020202, + "grad_norm": 2.9654042720794678, + "learning_rate": 3.365294667913939e-06, + "loss": 1.5115, + "step": 952 + }, + { + "epoch": 1.2032828282828283, + "grad_norm": 2.9138922691345215, + "learning_rate": 3.3629560336763333e-06, + "loss": 1.5106, + "step": 953 + }, + { + "epoch": 1.2045454545454546, + "grad_norm": 2.989961862564087, + "learning_rate": 3.360617399438728e-06, + "loss": 1.5278, + "step": 954 + }, + { + "epoch": 1.2058080808080809, + "grad_norm": 2.9620680809020996, + "learning_rate": 3.358278765201123e-06, + "loss": 1.4949, + "step": 955 + }, + { + "epoch": 1.2070707070707072, + "grad_norm": 2.945505380630493, + "learning_rate": 3.3559401309635174e-06, + "loss": 1.5018, + "step": 956 + }, + { + "epoch": 1.2083333333333333, + "grad_norm": 2.806464433670044, + "learning_rate": 3.353601496725912e-06, + "loss": 1.4951, + "step": 957 + }, + { + "epoch": 1.2095959595959596, + "grad_norm": 2.7847559452056885, + "learning_rate": 3.351262862488307e-06, + "loss": 1.4259, + "step": 958 + }, + { + "epoch": 1.2108585858585859, + "grad_norm": 2.760003089904785, + "learning_rate": 3.348924228250702e-06, + "loss": 1.4757, + "step": 959 + }, + { + "epoch": 1.2121212121212122, + "grad_norm": 3.037522792816162, + "learning_rate": 3.3465855940130963e-06, + "loss": 1.5462, + "step": 960 + }, + { + "epoch": 1.2121212121212122, + "eval_loss": 1.7071956396102905, + "eval_runtime": 37.1351, + "eval_samples_per_second": 26.929, + "eval_steps_per_second": 1.131, + "step": 960 + }, + { + "epoch": 1.2133838383838385, + "grad_norm": 2.7755539417266846, + "learning_rate": 3.344246959775491e-06, + "loss": 1.453, + "step": 961 + }, + { + "epoch": 1.2146464646464645, + "grad_norm": 2.910369634628296, + "learning_rate": 3.341908325537886e-06, + "loss": 1.4914, + "step": 962 + }, + { + "epoch": 1.2159090909090908, + "grad_norm": 2.8999392986297607, + "learning_rate": 3.339569691300281e-06, + "loss": 1.5389, + "step": 963 + }, + { + "epoch": 1.2171717171717171, + "grad_norm": 2.891995906829834, + "learning_rate": 3.337231057062676e-06, + "loss": 1.4915, + "step": 964 + }, + { + "epoch": 1.2184343434343434, + "grad_norm": 2.814624309539795, + "learning_rate": 3.3348924228250705e-06, + "loss": 1.4936, + "step": 965 + }, + { + "epoch": 1.2196969696969697, + "grad_norm": 2.971163034439087, + "learning_rate": 3.3325537885874654e-06, + "loss": 1.4813, + "step": 966 + }, + { + "epoch": 1.220959595959596, + "grad_norm": 2.934155225753784, + "learning_rate": 3.33021515434986e-06, + "loss": 1.5162, + "step": 967 + }, + { + "epoch": 1.2222222222222223, + "grad_norm": 3.007847309112549, + "learning_rate": 3.3278765201122546e-06, + "loss": 1.4741, + "step": 968 + }, + { + "epoch": 1.2234848484848484, + "grad_norm": 2.977207660675049, + "learning_rate": 3.3255378858746495e-06, + "loss": 1.5186, + "step": 969 + }, + { + "epoch": 1.2247474747474747, + "grad_norm": 2.926567316055298, + "learning_rate": 3.3231992516370443e-06, + "loss": 1.5111, + "step": 970 + }, + { + "epoch": 1.2247474747474747, + "eval_loss": 1.7068836688995361, + "eval_runtime": 37.1419, + "eval_samples_per_second": 26.924, + "eval_steps_per_second": 1.131, + "step": 970 + }, + { + "epoch": 1.226010101010101, + "grad_norm": 3.063671350479126, + "learning_rate": 3.320860617399439e-06, + "loss": 1.4376, + "step": 971 + }, + { + "epoch": 1.2272727272727273, + "grad_norm": 2.7263271808624268, + "learning_rate": 3.3185219831618336e-06, + "loss": 1.4555, + "step": 972 + }, + { + "epoch": 1.2285353535353536, + "grad_norm": 2.912658929824829, + "learning_rate": 3.3161833489242284e-06, + "loss": 1.5116, + "step": 973 + }, + { + "epoch": 1.22979797979798, + "grad_norm": 3.2216858863830566, + "learning_rate": 3.3138447146866233e-06, + "loss": 1.5124, + "step": 974 + }, + { + "epoch": 1.231060606060606, + "grad_norm": 2.9138946533203125, + "learning_rate": 3.311506080449018e-06, + "loss": 1.4917, + "step": 975 + }, + { + "epoch": 1.2323232323232323, + "grad_norm": 3.0519120693206787, + "learning_rate": 3.3091674462114125e-06, + "loss": 1.4986, + "step": 976 + }, + { + "epoch": 1.2335858585858586, + "grad_norm": 2.8186402320861816, + "learning_rate": 3.3068288119738074e-06, + "loss": 1.4835, + "step": 977 + }, + { + "epoch": 1.2348484848484849, + "grad_norm": 2.799189329147339, + "learning_rate": 3.304490177736202e-06, + "loss": 1.5069, + "step": 978 + }, + { + "epoch": 1.2361111111111112, + "grad_norm": 3.010010004043579, + "learning_rate": 3.3021515434985975e-06, + "loss": 1.4276, + "step": 979 + }, + { + "epoch": 1.2373737373737375, + "grad_norm": 3.153592109680176, + "learning_rate": 3.299812909260992e-06, + "loss": 1.5459, + "step": 980 + }, + { + "epoch": 1.2373737373737375, + "eval_loss": 1.7070198059082031, + "eval_runtime": 37.1397, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 980 + }, + { + "epoch": 1.2386363636363638, + "grad_norm": 2.7818424701690674, + "learning_rate": 3.2974742750233867e-06, + "loss": 1.5045, + "step": 981 + }, + { + "epoch": 1.2398989898989898, + "grad_norm": 2.9211390018463135, + "learning_rate": 3.2951356407857816e-06, + "loss": 1.4966, + "step": 982 + }, + { + "epoch": 1.2411616161616161, + "grad_norm": 2.979743242263794, + "learning_rate": 3.2927970065481764e-06, + "loss": 1.4754, + "step": 983 + }, + { + "epoch": 1.2424242424242424, + "grad_norm": 2.8476738929748535, + "learning_rate": 3.290458372310571e-06, + "loss": 1.4619, + "step": 984 + }, + { + "epoch": 1.2436868686868687, + "grad_norm": 2.8304896354675293, + "learning_rate": 3.2881197380729657e-06, + "loss": 1.4192, + "step": 985 + }, + { + "epoch": 1.244949494949495, + "grad_norm": 3.3150880336761475, + "learning_rate": 3.2857811038353605e-06, + "loss": 1.515, + "step": 986 + }, + { + "epoch": 1.246212121212121, + "grad_norm": 3.049398422241211, + "learning_rate": 3.283442469597755e-06, + "loss": 1.5382, + "step": 987 + }, + { + "epoch": 1.2474747474747474, + "grad_norm": 3.0910539627075195, + "learning_rate": 3.2811038353601498e-06, + "loss": 1.5458, + "step": 988 + }, + { + "epoch": 1.2487373737373737, + "grad_norm": 3.002788782119751, + "learning_rate": 3.2787652011225446e-06, + "loss": 1.4668, + "step": 989 + }, + { + "epoch": 1.25, + "grad_norm": 2.9342103004455566, + "learning_rate": 3.2764265668849395e-06, + "loss": 1.5187, + "step": 990 + }, + { + "epoch": 1.25, + "eval_loss": 1.7082287073135376, + "eval_runtime": 37.1399, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 990 + }, + { + "epoch": 1.2512626262626263, + "grad_norm": 2.7596492767333984, + "learning_rate": 3.274087932647334e-06, + "loss": 1.5191, + "step": 991 + }, + { + "epoch": 1.2525252525252526, + "grad_norm": 2.949368953704834, + "learning_rate": 3.2717492984097287e-06, + "loss": 1.4776, + "step": 992 + }, + { + "epoch": 1.253787878787879, + "grad_norm": 2.7099363803863525, + "learning_rate": 3.2694106641721236e-06, + "loss": 1.455, + "step": 993 + }, + { + "epoch": 1.255050505050505, + "grad_norm": 2.8631885051727295, + "learning_rate": 3.267072029934519e-06, + "loss": 1.4394, + "step": 994 + }, + { + "epoch": 1.2563131313131313, + "grad_norm": 3.09389591217041, + "learning_rate": 3.2647333956969137e-06, + "loss": 1.4958, + "step": 995 + }, + { + "epoch": 1.2575757575757576, + "grad_norm": 3.108271360397339, + "learning_rate": 3.262394761459308e-06, + "loss": 1.5108, + "step": 996 + }, + { + "epoch": 1.2588383838383839, + "grad_norm": 3.079176902770996, + "learning_rate": 3.260056127221703e-06, + "loss": 1.4749, + "step": 997 + }, + { + "epoch": 1.2601010101010102, + "grad_norm": 3.0458056926727295, + "learning_rate": 3.257717492984098e-06, + "loss": 1.4911, + "step": 998 + }, + { + "epoch": 1.2613636363636362, + "grad_norm": 2.952151298522949, + "learning_rate": 3.2553788587464922e-06, + "loss": 1.4842, + "step": 999 + }, + { + "epoch": 1.2626262626262625, + "grad_norm": 2.7918665409088135, + "learning_rate": 3.253040224508887e-06, + "loss": 1.4684, + "step": 1000 + }, + { + "epoch": 1.2626262626262625, + "eval_loss": 1.7100834846496582, + "eval_runtime": 37.3594, + "eval_samples_per_second": 26.767, + "eval_steps_per_second": 1.124, + "step": 1000 + }, + { + "epoch": 1.2638888888888888, + "grad_norm": 2.798064947128296, + "learning_rate": 3.250701590271282e-06, + "loss": 1.4295, + "step": 1001 + }, + { + "epoch": 1.2651515151515151, + "grad_norm": 2.852234363555908, + "learning_rate": 3.2483629560336767e-06, + "loss": 1.5039, + "step": 1002 + }, + { + "epoch": 1.2664141414141414, + "grad_norm": 2.762645721435547, + "learning_rate": 3.246024321796071e-06, + "loss": 1.4365, + "step": 1003 + }, + { + "epoch": 1.2676767676767677, + "grad_norm": 2.944326162338257, + "learning_rate": 3.243685687558466e-06, + "loss": 1.4849, + "step": 1004 + }, + { + "epoch": 1.268939393939394, + "grad_norm": 2.9298877716064453, + "learning_rate": 3.241347053320861e-06, + "loss": 1.4397, + "step": 1005 + }, + { + "epoch": 1.2702020202020203, + "grad_norm": 3.1259615421295166, + "learning_rate": 3.2390084190832553e-06, + "loss": 1.4819, + "step": 1006 + }, + { + "epoch": 1.2714646464646464, + "grad_norm": 3.114792823791504, + "learning_rate": 3.23666978484565e-06, + "loss": 1.4728, + "step": 1007 + }, + { + "epoch": 1.2727272727272727, + "grad_norm": 2.9800360202789307, + "learning_rate": 3.234331150608045e-06, + "loss": 1.4426, + "step": 1008 + }, + { + "epoch": 1.273989898989899, + "grad_norm": 2.8487257957458496, + "learning_rate": 3.2319925163704402e-06, + "loss": 1.461, + "step": 1009 + }, + { + "epoch": 1.2752525252525253, + "grad_norm": 2.800760507583618, + "learning_rate": 3.229653882132835e-06, + "loss": 1.4764, + "step": 1010 + }, + { + "epoch": 1.2752525252525253, + "eval_loss": 1.701302409172058, + "eval_runtime": 36.9359, + "eval_samples_per_second": 27.074, + "eval_steps_per_second": 1.137, + "step": 1010 + }, + { + "epoch": 1.2765151515151514, + "grad_norm": 2.851012706756592, + "learning_rate": 3.2273152478952295e-06, + "loss": 1.5124, + "step": 1011 + }, + { + "epoch": 1.2777777777777777, + "grad_norm": 2.9498190879821777, + "learning_rate": 3.2249766136576243e-06, + "loss": 1.4789, + "step": 1012 + }, + { + "epoch": 1.279040404040404, + "grad_norm": 2.80710768699646, + "learning_rate": 3.222637979420019e-06, + "loss": 1.5592, + "step": 1013 + }, + { + "epoch": 1.2803030303030303, + "grad_norm": 2.9525952339172363, + "learning_rate": 3.220299345182414e-06, + "loss": 1.4941, + "step": 1014 + }, + { + "epoch": 1.2815656565656566, + "grad_norm": 2.9770407676696777, + "learning_rate": 3.2179607109448084e-06, + "loss": 1.4437, + "step": 1015 + }, + { + "epoch": 1.2828282828282829, + "grad_norm": 3.2553279399871826, + "learning_rate": 3.2156220767072033e-06, + "loss": 1.5248, + "step": 1016 + }, + { + "epoch": 1.2840909090909092, + "grad_norm": 3.0591554641723633, + "learning_rate": 3.213283442469598e-06, + "loss": 1.5704, + "step": 1017 + }, + { + "epoch": 1.2853535353535355, + "grad_norm": 2.776968002319336, + "learning_rate": 3.2109448082319925e-06, + "loss": 1.4239, + "step": 1018 + }, + { + "epoch": 1.2866161616161615, + "grad_norm": 2.8728222846984863, + "learning_rate": 3.2086061739943874e-06, + "loss": 1.5115, + "step": 1019 + }, + { + "epoch": 1.2878787878787878, + "grad_norm": 2.907308340072632, + "learning_rate": 3.2062675397567822e-06, + "loss": 1.4764, + "step": 1020 + }, + { + "epoch": 1.2878787878787878, + "eval_loss": 1.7016223669052124, + "eval_runtime": 37.0728, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 1020 + }, + { + "epoch": 1.2891414141414141, + "grad_norm": 3.0686309337615967, + "learning_rate": 3.203928905519177e-06, + "loss": 1.5175, + "step": 1021 + }, + { + "epoch": 1.2904040404040404, + "grad_norm": 2.8636863231658936, + "learning_rate": 3.2015902712815715e-06, + "loss": 1.5025, + "step": 1022 + }, + { + "epoch": 1.2916666666666667, + "grad_norm": 2.8539068698883057, + "learning_rate": 3.1992516370439663e-06, + "loss": 1.4774, + "step": 1023 + }, + { + "epoch": 1.2929292929292928, + "grad_norm": 2.9830520153045654, + "learning_rate": 3.196913002806361e-06, + "loss": 1.533, + "step": 1024 + }, + { + "epoch": 1.2941919191919191, + "grad_norm": 2.8806374073028564, + "learning_rate": 3.1945743685687564e-06, + "loss": 1.5298, + "step": 1025 + }, + { + "epoch": 1.2954545454545454, + "grad_norm": 2.932870626449585, + "learning_rate": 3.1922357343311513e-06, + "loss": 1.5028, + "step": 1026 + }, + { + "epoch": 1.2967171717171717, + "grad_norm": 2.889601945877075, + "learning_rate": 3.1898971000935457e-06, + "loss": 1.4534, + "step": 1027 + }, + { + "epoch": 1.297979797979798, + "grad_norm": 2.930546522140503, + "learning_rate": 3.1875584658559405e-06, + "loss": 1.5047, + "step": 1028 + }, + { + "epoch": 1.2992424242424243, + "grad_norm": 3.0944340229034424, + "learning_rate": 3.1852198316183354e-06, + "loss": 1.5198, + "step": 1029 + }, + { + "epoch": 1.3005050505050506, + "grad_norm": 2.8930089473724365, + "learning_rate": 3.18288119738073e-06, + "loss": 1.4655, + "step": 1030 + }, + { + "epoch": 1.3005050505050506, + "eval_loss": 1.7008370161056519, + "eval_runtime": 37.1263, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 1030 + }, + { + "epoch": 1.3017676767676767, + "grad_norm": 2.888303518295288, + "learning_rate": 3.1805425631431246e-06, + "loss": 1.4249, + "step": 1031 + }, + { + "epoch": 1.303030303030303, + "grad_norm": 3.0144314765930176, + "learning_rate": 3.1782039289055195e-06, + "loss": 1.4963, + "step": 1032 + }, + { + "epoch": 1.3042929292929293, + "grad_norm": 3.0118296146392822, + "learning_rate": 3.1758652946679143e-06, + "loss": 1.5119, + "step": 1033 + }, + { + "epoch": 1.3055555555555556, + "grad_norm": 2.8148205280303955, + "learning_rate": 3.1735266604303087e-06, + "loss": 1.4921, + "step": 1034 + }, + { + "epoch": 1.3068181818181819, + "grad_norm": 2.811166286468506, + "learning_rate": 3.1711880261927036e-06, + "loss": 1.4689, + "step": 1035 + }, + { + "epoch": 1.308080808080808, + "grad_norm": 2.924088478088379, + "learning_rate": 3.1688493919550984e-06, + "loss": 1.5682, + "step": 1036 + }, + { + "epoch": 1.3093434343434343, + "grad_norm": 2.905988931655884, + "learning_rate": 3.166510757717493e-06, + "loss": 1.5104, + "step": 1037 + }, + { + "epoch": 1.3106060606060606, + "grad_norm": 2.972963571548462, + "learning_rate": 3.1641721234798877e-06, + "loss": 1.4543, + "step": 1038 + }, + { + "epoch": 1.3118686868686869, + "grad_norm": 2.8034229278564453, + "learning_rate": 3.1618334892422825e-06, + "loss": 1.45, + "step": 1039 + }, + { + "epoch": 1.3131313131313131, + "grad_norm": 2.8977267742156982, + "learning_rate": 3.159494855004678e-06, + "loss": 1.4531, + "step": 1040 + }, + { + "epoch": 1.3131313131313131, + "eval_loss": 1.7027597427368164, + "eval_runtime": 37.0629, + "eval_samples_per_second": 26.981, + "eval_steps_per_second": 1.133, + "step": 1040 + }, + { + "epoch": 1.3143939393939394, + "grad_norm": 3.106973886489868, + "learning_rate": 3.1571562207670726e-06, + "loss": 1.4723, + "step": 1041 + }, + { + "epoch": 1.3156565656565657, + "grad_norm": 2.7628490924835205, + "learning_rate": 3.154817586529467e-06, + "loss": 1.4902, + "step": 1042 + }, + { + "epoch": 1.316919191919192, + "grad_norm": 3.119751453399658, + "learning_rate": 3.152478952291862e-06, + "loss": 1.5326, + "step": 1043 + }, + { + "epoch": 1.3181818181818181, + "grad_norm": 3.1470835208892822, + "learning_rate": 3.1501403180542568e-06, + "loss": 1.4626, + "step": 1044 + }, + { + "epoch": 1.3194444444444444, + "grad_norm": 2.866586208343506, + "learning_rate": 3.1478016838166516e-06, + "loss": 1.5006, + "step": 1045 + }, + { + "epoch": 1.3207070707070707, + "grad_norm": 2.878464698791504, + "learning_rate": 3.145463049579046e-06, + "loss": 1.4727, + "step": 1046 + }, + { + "epoch": 1.321969696969697, + "grad_norm": 2.8567826747894287, + "learning_rate": 3.143124415341441e-06, + "loss": 1.4202, + "step": 1047 + }, + { + "epoch": 1.3232323232323233, + "grad_norm": 2.8718442916870117, + "learning_rate": 3.1407857811038357e-06, + "loss": 1.473, + "step": 1048 + }, + { + "epoch": 1.3244949494949494, + "grad_norm": 3.003040313720703, + "learning_rate": 3.13844714686623e-06, + "loss": 1.4784, + "step": 1049 + }, + { + "epoch": 1.3257575757575757, + "grad_norm": 3.080288887023926, + "learning_rate": 3.136108512628625e-06, + "loss": 1.4514, + "step": 1050 + }, + { + "epoch": 1.3257575757575757, + "eval_loss": 1.704396367073059, + "eval_runtime": 37.0757, + "eval_samples_per_second": 26.972, + "eval_steps_per_second": 1.133, + "step": 1050 + }, + { + "epoch": 1.327020202020202, + "grad_norm": 2.8699166774749756, + "learning_rate": 3.13376987839102e-06, + "loss": 1.4966, + "step": 1051 + }, + { + "epoch": 1.3282828282828283, + "grad_norm": 2.979238271713257, + "learning_rate": 3.1314312441534146e-06, + "loss": 1.4422, + "step": 1052 + }, + { + "epoch": 1.3295454545454546, + "grad_norm": 2.8612024784088135, + "learning_rate": 3.129092609915809e-06, + "loss": 1.4982, + "step": 1053 + }, + { + "epoch": 1.3308080808080809, + "grad_norm": 2.7759032249450684, + "learning_rate": 3.126753975678204e-06, + "loss": 1.413, + "step": 1054 + }, + { + "epoch": 1.3320707070707072, + "grad_norm": 2.9772675037384033, + "learning_rate": 3.124415341440599e-06, + "loss": 1.4789, + "step": 1055 + }, + { + "epoch": 1.3333333333333333, + "grad_norm": 2.9382526874542236, + "learning_rate": 3.122076707202994e-06, + "loss": 1.4547, + "step": 1056 + }, + { + "epoch": 1.3345959595959596, + "grad_norm": 2.85050630569458, + "learning_rate": 3.119738072965389e-06, + "loss": 1.4193, + "step": 1057 + }, + { + "epoch": 1.3358585858585859, + "grad_norm": 3.1148695945739746, + "learning_rate": 3.1173994387277833e-06, + "loss": 1.4527, + "step": 1058 + }, + { + "epoch": 1.3371212121212122, + "grad_norm": 3.3906028270721436, + "learning_rate": 3.115060804490178e-06, + "loss": 1.5273, + "step": 1059 + }, + { + "epoch": 1.3383838383838385, + "grad_norm": 2.9665908813476562, + "learning_rate": 3.112722170252573e-06, + "loss": 1.5237, + "step": 1060 + }, + { + "epoch": 1.3383838383838385, + "eval_loss": 1.7004541158676147, + "eval_runtime": 37.081, + "eval_samples_per_second": 26.968, + "eval_steps_per_second": 1.133, + "step": 1060 + }, + { + "epoch": 1.3396464646464645, + "grad_norm": 2.9809224605560303, + "learning_rate": 3.1103835360149674e-06, + "loss": 1.498, + "step": 1061 + }, + { + "epoch": 1.3409090909090908, + "grad_norm": 2.95265793800354, + "learning_rate": 3.1080449017773622e-06, + "loss": 1.511, + "step": 1062 + }, + { + "epoch": 1.3421717171717171, + "grad_norm": 3.0814359188079834, + "learning_rate": 3.105706267539757e-06, + "loss": 1.4448, + "step": 1063 + }, + { + "epoch": 1.3434343434343434, + "grad_norm": 2.8267405033111572, + "learning_rate": 3.103367633302152e-06, + "loss": 1.4648, + "step": 1064 + }, + { + "epoch": 1.3446969696969697, + "grad_norm": 3.08182954788208, + "learning_rate": 3.1010289990645463e-06, + "loss": 1.5245, + "step": 1065 + }, + { + "epoch": 1.345959595959596, + "grad_norm": 3.183209180831909, + "learning_rate": 3.098690364826941e-06, + "loss": 1.4388, + "step": 1066 + }, + { + "epoch": 1.3472222222222223, + "grad_norm": 3.2121284008026123, + "learning_rate": 3.096351730589336e-06, + "loss": 1.4328, + "step": 1067 + }, + { + "epoch": 1.3484848484848486, + "grad_norm": 3.093921661376953, + "learning_rate": 3.0940130963517304e-06, + "loss": 1.5176, + "step": 1068 + }, + { + "epoch": 1.3497474747474747, + "grad_norm": 2.977710008621216, + "learning_rate": 3.0916744621141253e-06, + "loss": 1.4153, + "step": 1069 + }, + { + "epoch": 1.351010101010101, + "grad_norm": 3.119001865386963, + "learning_rate": 3.08933582787652e-06, + "loss": 1.4879, + "step": 1070 + }, + { + "epoch": 1.351010101010101, + "eval_loss": 1.7006300687789917, + "eval_runtime": 37.0959, + "eval_samples_per_second": 26.957, + "eval_steps_per_second": 1.132, + "step": 1070 + }, + { + "epoch": 1.3522727272727273, + "grad_norm": 3.026405096054077, + "learning_rate": 3.0869971936389154e-06, + "loss": 1.483, + "step": 1071 + }, + { + "epoch": 1.3535353535353536, + "grad_norm": 2.770728826522827, + "learning_rate": 3.0846585594013102e-06, + "loss": 1.5175, + "step": 1072 + }, + { + "epoch": 1.3547979797979797, + "grad_norm": 2.7930657863616943, + "learning_rate": 3.0823199251637047e-06, + "loss": 1.4833, + "step": 1073 + }, + { + "epoch": 1.356060606060606, + "grad_norm": 2.9128804206848145, + "learning_rate": 3.0799812909260995e-06, + "loss": 1.5016, + "step": 1074 + }, + { + "epoch": 1.3573232323232323, + "grad_norm": 3.0432205200195312, + "learning_rate": 3.0776426566884943e-06, + "loss": 1.5211, + "step": 1075 + }, + { + "epoch": 1.3585858585858586, + "grad_norm": 2.9197583198547363, + "learning_rate": 3.075304022450889e-06, + "loss": 1.5065, + "step": 1076 + }, + { + "epoch": 1.3598484848484849, + "grad_norm": 2.944986581802368, + "learning_rate": 3.0729653882132836e-06, + "loss": 1.5015, + "step": 1077 + }, + { + "epoch": 1.3611111111111112, + "grad_norm": 2.7553489208221436, + "learning_rate": 3.0706267539756784e-06, + "loss": 1.4692, + "step": 1078 + }, + { + "epoch": 1.3623737373737375, + "grad_norm": 2.7607858180999756, + "learning_rate": 3.0682881197380733e-06, + "loss": 1.4494, + "step": 1079 + }, + { + "epoch": 1.3636363636363638, + "grad_norm": 2.7896173000335693, + "learning_rate": 3.0659494855004677e-06, + "loss": 1.4625, + "step": 1080 + }, + { + "epoch": 1.3636363636363638, + "eval_loss": 1.6998732089996338, + "eval_runtime": 37.1079, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 1080 + }, + { + "epoch": 1.3648989898989898, + "grad_norm": 3.020059823989868, + "learning_rate": 3.0636108512628626e-06, + "loss": 1.5331, + "step": 1081 + }, + { + "epoch": 1.3661616161616161, + "grad_norm": 2.805593729019165, + "learning_rate": 3.0612722170252574e-06, + "loss": 1.4867, + "step": 1082 + }, + { + "epoch": 1.3674242424242424, + "grad_norm": 2.965871810913086, + "learning_rate": 3.0589335827876522e-06, + "loss": 1.4494, + "step": 1083 + }, + { + "epoch": 1.3686868686868687, + "grad_norm": 2.8010783195495605, + "learning_rate": 3.0565949485500467e-06, + "loss": 1.4484, + "step": 1084 + }, + { + "epoch": 1.369949494949495, + "grad_norm": 2.985258102416992, + "learning_rate": 3.0542563143124415e-06, + "loss": 1.5029, + "step": 1085 + }, + { + "epoch": 1.371212121212121, + "grad_norm": 2.7528188228607178, + "learning_rate": 3.0519176800748368e-06, + "loss": 1.4695, + "step": 1086 + }, + { + "epoch": 1.3724747474747474, + "grad_norm": 2.9123780727386475, + "learning_rate": 3.0495790458372316e-06, + "loss": 1.5527, + "step": 1087 + }, + { + "epoch": 1.3737373737373737, + "grad_norm": 2.8589160442352295, + "learning_rate": 3.0472404115996265e-06, + "loss": 1.5104, + "step": 1088 + }, + { + "epoch": 1.375, + "grad_norm": 2.853546142578125, + "learning_rate": 3.044901777362021e-06, + "loss": 1.4498, + "step": 1089 + }, + { + "epoch": 1.3762626262626263, + "grad_norm": 2.7989797592163086, + "learning_rate": 3.0425631431244157e-06, + "loss": 1.4494, + "step": 1090 + }, + { + "epoch": 1.3762626262626263, + "eval_loss": 1.6923459768295288, + "eval_runtime": 37.2917, + "eval_samples_per_second": 26.816, + "eval_steps_per_second": 1.126, + "step": 1090 + }, + { + "epoch": 1.3775252525252526, + "grad_norm": 2.8750412464141846, + "learning_rate": 3.0402245088868106e-06, + "loss": 1.5026, + "step": 1091 + }, + { + "epoch": 1.378787878787879, + "grad_norm": 2.8185842037200928, + "learning_rate": 3.037885874649205e-06, + "loss": 1.4859, + "step": 1092 + }, + { + "epoch": 1.380050505050505, + "grad_norm": 2.736332893371582, + "learning_rate": 3.0355472404116e-06, + "loss": 1.4944, + "step": 1093 + }, + { + "epoch": 1.3813131313131313, + "grad_norm": 2.93930983543396, + "learning_rate": 3.0332086061739947e-06, + "loss": 1.4975, + "step": 1094 + }, + { + "epoch": 1.3825757575757576, + "grad_norm": 2.964146852493286, + "learning_rate": 3.0308699719363895e-06, + "loss": 1.4861, + "step": 1095 + }, + { + "epoch": 1.3838383838383839, + "grad_norm": 2.8321545124053955, + "learning_rate": 3.028531337698784e-06, + "loss": 1.4535, + "step": 1096 + }, + { + "epoch": 1.3851010101010102, + "grad_norm": 2.921372413635254, + "learning_rate": 3.0261927034611788e-06, + "loss": 1.5242, + "step": 1097 + }, + { + "epoch": 1.3863636363636362, + "grad_norm": 2.7309911251068115, + "learning_rate": 3.0238540692235736e-06, + "loss": 1.4105, + "step": 1098 + }, + { + "epoch": 1.3876262626262625, + "grad_norm": 2.768899917602539, + "learning_rate": 3.021515434985968e-06, + "loss": 1.4183, + "step": 1099 + }, + { + "epoch": 1.3888888888888888, + "grad_norm": 2.883399248123169, + "learning_rate": 3.019176800748363e-06, + "loss": 1.5307, + "step": 1100 + }, + { + "epoch": 1.3888888888888888, + "eval_loss": 1.6948082447052002, + "eval_runtime": 37.488, + "eval_samples_per_second": 26.675, + "eval_steps_per_second": 1.12, + "step": 1100 + }, + { + "epoch": 1.3901515151515151, + "grad_norm": 2.8610544204711914, + "learning_rate": 3.016838166510758e-06, + "loss": 1.5473, + "step": 1101 + }, + { + "epoch": 1.3914141414141414, + "grad_norm": 2.762694835662842, + "learning_rate": 3.014499532273153e-06, + "loss": 1.4179, + "step": 1102 + }, + { + "epoch": 1.3926767676767677, + "grad_norm": 2.975828170776367, + "learning_rate": 3.012160898035548e-06, + "loss": 1.4959, + "step": 1103 + }, + { + "epoch": 1.393939393939394, + "grad_norm": 3.0633926391601562, + "learning_rate": 3.0098222637979422e-06, + "loss": 1.4979, + "step": 1104 + }, + { + "epoch": 1.3952020202020203, + "grad_norm": 2.972900867462158, + "learning_rate": 3.007483629560337e-06, + "loss": 1.5311, + "step": 1105 + }, + { + "epoch": 1.3964646464646464, + "grad_norm": 2.936033248901367, + "learning_rate": 3.005144995322732e-06, + "loss": 1.4856, + "step": 1106 + }, + { + "epoch": 1.3977272727272727, + "grad_norm": 3.0375452041625977, + "learning_rate": 3.0028063610851268e-06, + "loss": 1.5001, + "step": 1107 + }, + { + "epoch": 1.398989898989899, + "grad_norm": 2.998283624649048, + "learning_rate": 3.000467726847521e-06, + "loss": 1.5219, + "step": 1108 + }, + { + "epoch": 1.4002525252525253, + "grad_norm": 2.831354856491089, + "learning_rate": 2.998129092609916e-06, + "loss": 1.4498, + "step": 1109 + }, + { + "epoch": 1.4015151515151514, + "grad_norm": 2.975390911102295, + "learning_rate": 2.995790458372311e-06, + "loss": 1.4655, + "step": 1110 + }, + { + "epoch": 1.4015151515151514, + "eval_loss": 1.6940916776657104, + "eval_runtime": 36.9469, + "eval_samples_per_second": 27.066, + "eval_steps_per_second": 1.137, + "step": 1110 + }, + { + "epoch": 1.4027777777777777, + "grad_norm": 2.9645442962646484, + "learning_rate": 2.9934518241347053e-06, + "loss": 1.4568, + "step": 1111 + }, + { + "epoch": 1.404040404040404, + "grad_norm": 2.873746633529663, + "learning_rate": 2.9911131898971e-06, + "loss": 1.4483, + "step": 1112 + }, + { + "epoch": 1.4053030303030303, + "grad_norm": 3.1041767597198486, + "learning_rate": 2.988774555659495e-06, + "loss": 1.4914, + "step": 1113 + }, + { + "epoch": 1.4065656565656566, + "grad_norm": 3.0122971534729004, + "learning_rate": 2.98643592142189e-06, + "loss": 1.4494, + "step": 1114 + }, + { + "epoch": 1.4078282828282829, + "grad_norm": 2.842233180999756, + "learning_rate": 2.9840972871842842e-06, + "loss": 1.4274, + "step": 1115 + }, + { + "epoch": 1.4090909090909092, + "grad_norm": 3.058265447616577, + "learning_rate": 2.9817586529466795e-06, + "loss": 1.4637, + "step": 1116 + }, + { + "epoch": 1.4103535353535355, + "grad_norm": 3.06136155128479, + "learning_rate": 2.9794200187090744e-06, + "loss": 1.4791, + "step": 1117 + }, + { + "epoch": 1.4116161616161615, + "grad_norm": 2.8931167125701904, + "learning_rate": 2.977081384471469e-06, + "loss": 1.4476, + "step": 1118 + }, + { + "epoch": 1.4128787878787878, + "grad_norm": 2.8025107383728027, + "learning_rate": 2.974742750233864e-06, + "loss": 1.4891, + "step": 1119 + }, + { + "epoch": 1.4141414141414141, + "grad_norm": 3.1213603019714355, + "learning_rate": 2.9724041159962585e-06, + "loss": 1.5513, + "step": 1120 + }, + { + "epoch": 1.4141414141414141, + "eval_loss": 1.6972154378890991, + "eval_runtime": 37.0863, + "eval_samples_per_second": 26.964, + "eval_steps_per_second": 1.132, + "step": 1120 + }, + { + "epoch": 1.4154040404040404, + "grad_norm": 3.126791477203369, + "learning_rate": 2.9700654817586533e-06, + "loss": 1.4952, + "step": 1121 + }, + { + "epoch": 1.4166666666666667, + "grad_norm": 2.959385395050049, + "learning_rate": 2.967726847521048e-06, + "loss": 1.5098, + "step": 1122 + }, + { + "epoch": 1.4179292929292928, + "grad_norm": 2.907566547393799, + "learning_rate": 2.9653882132834426e-06, + "loss": 1.4959, + "step": 1123 + }, + { + "epoch": 1.4191919191919191, + "grad_norm": 2.9181833267211914, + "learning_rate": 2.9630495790458374e-06, + "loss": 1.4607, + "step": 1124 + }, + { + "epoch": 1.4204545454545454, + "grad_norm": 2.8703668117523193, + "learning_rate": 2.9607109448082323e-06, + "loss": 1.5215, + "step": 1125 + }, + { + "epoch": 1.4217171717171717, + "grad_norm": 3.166255235671997, + "learning_rate": 2.958372310570627e-06, + "loss": 1.4451, + "step": 1126 + }, + { + "epoch": 1.422979797979798, + "grad_norm": 2.863943576812744, + "learning_rate": 2.9560336763330215e-06, + "loss": 1.4481, + "step": 1127 + }, + { + "epoch": 1.4242424242424243, + "grad_norm": 2.975482702255249, + "learning_rate": 2.9536950420954164e-06, + "loss": 1.5118, + "step": 1128 + }, + { + "epoch": 1.4255050505050506, + "grad_norm": 3.077357769012451, + "learning_rate": 2.951356407857811e-06, + "loss": 1.4678, + "step": 1129 + }, + { + "epoch": 1.4267676767676767, + "grad_norm": 3.0375332832336426, + "learning_rate": 2.9490177736202056e-06, + "loss": 1.5036, + "step": 1130 + }, + { + "epoch": 1.4267676767676767, + "eval_loss": 1.6973206996917725, + "eval_runtime": 37.1448, + "eval_samples_per_second": 26.922, + "eval_steps_per_second": 1.131, + "step": 1130 + }, + { + "epoch": 1.428030303030303, + "grad_norm": 3.022068500518799, + "learning_rate": 2.9466791393826005e-06, + "loss": 1.442, + "step": 1131 + }, + { + "epoch": 1.4292929292929293, + "grad_norm": 2.9089183807373047, + "learning_rate": 2.9443405051449957e-06, + "loss": 1.4627, + "step": 1132 + }, + { + "epoch": 1.4305555555555556, + "grad_norm": 2.840149402618408, + "learning_rate": 2.9420018709073906e-06, + "loss": 1.4745, + "step": 1133 + }, + { + "epoch": 1.4318181818181819, + "grad_norm": 2.938650608062744, + "learning_rate": 2.9396632366697854e-06, + "loss": 1.4795, + "step": 1134 + }, + { + "epoch": 1.433080808080808, + "grad_norm": 2.9602437019348145, + "learning_rate": 2.93732460243218e-06, + "loss": 1.4228, + "step": 1135 + }, + { + "epoch": 1.4343434343434343, + "grad_norm": 2.929257869720459, + "learning_rate": 2.9349859681945747e-06, + "loss": 1.5016, + "step": 1136 + }, + { + "epoch": 1.4356060606060606, + "grad_norm": 2.8120505809783936, + "learning_rate": 2.9326473339569695e-06, + "loss": 1.474, + "step": 1137 + }, + { + "epoch": 1.4368686868686869, + "grad_norm": 3.0380513668060303, + "learning_rate": 2.9303086997193644e-06, + "loss": 1.4714, + "step": 1138 + }, + { + "epoch": 1.4381313131313131, + "grad_norm": 2.91106915473938, + "learning_rate": 2.9279700654817588e-06, + "loss": 1.4558, + "step": 1139 + }, + { + "epoch": 1.4393939393939394, + "grad_norm": 2.9970154762268066, + "learning_rate": 2.9256314312441536e-06, + "loss": 1.4895, + "step": 1140 + }, + { + "epoch": 1.4393939393939394, + "eval_loss": 1.6916719675064087, + "eval_runtime": 37.1538, + "eval_samples_per_second": 26.915, + "eval_steps_per_second": 1.13, + "step": 1140 + }, + { + "epoch": 1.4406565656565657, + "grad_norm": 3.041616201400757, + "learning_rate": 2.9232927970065485e-06, + "loss": 1.5177, + "step": 1141 + }, + { + "epoch": 1.441919191919192, + "grad_norm": 2.875385046005249, + "learning_rate": 2.920954162768943e-06, + "loss": 1.4697, + "step": 1142 + }, + { + "epoch": 1.4431818181818181, + "grad_norm": 2.7117881774902344, + "learning_rate": 2.9186155285313377e-06, + "loss": 1.4809, + "step": 1143 + }, + { + "epoch": 1.4444444444444444, + "grad_norm": 2.912123203277588, + "learning_rate": 2.9162768942937326e-06, + "loss": 1.4032, + "step": 1144 + }, + { + "epoch": 1.4457070707070707, + "grad_norm": 2.798440933227539, + "learning_rate": 2.9139382600561274e-06, + "loss": 1.4442, + "step": 1145 + }, + { + "epoch": 1.446969696969697, + "grad_norm": 2.9299135208129883, + "learning_rate": 2.911599625818522e-06, + "loss": 1.4412, + "step": 1146 + }, + { + "epoch": 1.4482323232323233, + "grad_norm": 2.988434076309204, + "learning_rate": 2.909260991580917e-06, + "loss": 1.4473, + "step": 1147 + }, + { + "epoch": 1.4494949494949494, + "grad_norm": 3.0697057247161865, + "learning_rate": 2.906922357343312e-06, + "loss": 1.5151, + "step": 1148 + }, + { + "epoch": 1.4507575757575757, + "grad_norm": 2.9009311199188232, + "learning_rate": 2.9045837231057068e-06, + "loss": 1.4434, + "step": 1149 + }, + { + "epoch": 1.452020202020202, + "grad_norm": 2.9949986934661865, + "learning_rate": 2.902245088868101e-06, + "loss": 1.5436, + "step": 1150 + }, + { + "epoch": 1.452020202020202, + "eval_loss": 1.6880143880844116, + "eval_runtime": 37.1212, + "eval_samples_per_second": 26.939, + "eval_steps_per_second": 1.131, + "step": 1150 + }, + { + "epoch": 1.4532828282828283, + "grad_norm": 2.7827775478363037, + "learning_rate": 2.899906454630496e-06, + "loss": 1.4284, + "step": 1151 + }, + { + "epoch": 1.4545454545454546, + "grad_norm": 2.8879382610321045, + "learning_rate": 2.897567820392891e-06, + "loss": 1.4772, + "step": 1152 + }, + { + "epoch": 1.4558080808080809, + "grad_norm": 3.0119130611419678, + "learning_rate": 2.8952291861552857e-06, + "loss": 1.4546, + "step": 1153 + }, + { + "epoch": 1.4570707070707072, + "grad_norm": 2.9002952575683594, + "learning_rate": 2.89289055191768e-06, + "loss": 1.4898, + "step": 1154 + }, + { + "epoch": 1.4583333333333333, + "grad_norm": 2.8384580612182617, + "learning_rate": 2.890551917680075e-06, + "loss": 1.4685, + "step": 1155 + }, + { + "epoch": 1.4595959595959596, + "grad_norm": 2.9149580001831055, + "learning_rate": 2.88821328344247e-06, + "loss": 1.5133, + "step": 1156 + }, + { + "epoch": 1.4608585858585859, + "grad_norm": 3.041875123977661, + "learning_rate": 2.8858746492048647e-06, + "loss": 1.5038, + "step": 1157 + }, + { + "epoch": 1.4621212121212122, + "grad_norm": 2.9623730182647705, + "learning_rate": 2.883536014967259e-06, + "loss": 1.4776, + "step": 1158 + }, + { + "epoch": 1.4633838383838385, + "grad_norm": 2.935295343399048, + "learning_rate": 2.881197380729654e-06, + "loss": 1.4166, + "step": 1159 + }, + { + "epoch": 1.4646464646464645, + "grad_norm": 2.9556751251220703, + "learning_rate": 2.8788587464920488e-06, + "loss": 1.5132, + "step": 1160 + }, + { + "epoch": 1.4646464646464645, + "eval_loss": 1.69142746925354, + "eval_runtime": 37.1425, + "eval_samples_per_second": 26.923, + "eval_steps_per_second": 1.131, + "step": 1160 + }, + { + "epoch": 1.4659090909090908, + "grad_norm": 2.992587089538574, + "learning_rate": 2.876520112254443e-06, + "loss": 1.4695, + "step": 1161 + }, + { + "epoch": 1.4671717171717171, + "grad_norm": 3.005990982055664, + "learning_rate": 2.8741814780168385e-06, + "loss": 1.5122, + "step": 1162 + }, + { + "epoch": 1.4684343434343434, + "grad_norm": 2.9341115951538086, + "learning_rate": 2.8718428437792333e-06, + "loss": 1.4916, + "step": 1163 + }, + { + "epoch": 1.4696969696969697, + "grad_norm": 2.9665164947509766, + "learning_rate": 2.869504209541628e-06, + "loss": 1.4827, + "step": 1164 + }, + { + "epoch": 1.470959595959596, + "grad_norm": 2.9972221851348877, + "learning_rate": 2.867165575304023e-06, + "loss": 1.4845, + "step": 1165 + }, + { + "epoch": 1.4722222222222223, + "grad_norm": 2.981387138366699, + "learning_rate": 2.8648269410664174e-06, + "loss": 1.4374, + "step": 1166 + }, + { + "epoch": 1.4734848484848486, + "grad_norm": 3.04587459564209, + "learning_rate": 2.8624883068288123e-06, + "loss": 1.431, + "step": 1167 + }, + { + "epoch": 1.4747474747474747, + "grad_norm": 3.194389820098877, + "learning_rate": 2.860149672591207e-06, + "loss": 1.4847, + "step": 1168 + }, + { + "epoch": 1.476010101010101, + "grad_norm": 2.949721574783325, + "learning_rate": 2.857811038353602e-06, + "loss": 1.4596, + "step": 1169 + }, + { + "epoch": 1.4772727272727273, + "grad_norm": 3.0032384395599365, + "learning_rate": 2.8554724041159964e-06, + "loss": 1.4821, + "step": 1170 + }, + { + "epoch": 1.4772727272727273, + "eval_loss": 1.6863130331039429, + "eval_runtime": 37.1386, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 1170 + }, + { + "epoch": 1.4785353535353536, + "grad_norm": 2.943141460418701, + "learning_rate": 2.853133769878391e-06, + "loss": 1.4998, + "step": 1171 + }, + { + "epoch": 1.4797979797979797, + "grad_norm": 3.0163350105285645, + "learning_rate": 2.850795135640786e-06, + "loss": 1.4565, + "step": 1172 + }, + { + "epoch": 1.481060606060606, + "grad_norm": 2.9713706970214844, + "learning_rate": 2.8484565014031805e-06, + "loss": 1.4663, + "step": 1173 + }, + { + "epoch": 1.4823232323232323, + "grad_norm": 2.9195785522460938, + "learning_rate": 2.8461178671655753e-06, + "loss": 1.4594, + "step": 1174 + }, + { + "epoch": 1.4835858585858586, + "grad_norm": 3.127619504928589, + "learning_rate": 2.84377923292797e-06, + "loss": 1.4742, + "step": 1175 + }, + { + "epoch": 1.4848484848484849, + "grad_norm": 3.0264782905578613, + "learning_rate": 2.841440598690365e-06, + "loss": 1.4366, + "step": 1176 + }, + { + "epoch": 1.4861111111111112, + "grad_norm": 3.0541677474975586, + "learning_rate": 2.8391019644527603e-06, + "loss": 1.4629, + "step": 1177 + }, + { + "epoch": 1.4873737373737375, + "grad_norm": 2.8967537879943848, + "learning_rate": 2.8367633302151547e-06, + "loss": 1.4281, + "step": 1178 + }, + { + "epoch": 1.4886363636363638, + "grad_norm": 2.975773811340332, + "learning_rate": 2.8344246959775495e-06, + "loss": 1.4272, + "step": 1179 + }, + { + "epoch": 1.4898989898989898, + "grad_norm": 2.9768712520599365, + "learning_rate": 2.8320860617399444e-06, + "loss": 1.453, + "step": 1180 + }, + { + "epoch": 1.4898989898989898, + "eval_loss": 1.6810754537582397, + "eval_runtime": 37.1659, + "eval_samples_per_second": 26.906, + "eval_steps_per_second": 1.13, + "step": 1180 + }, + { + "epoch": 1.4911616161616161, + "grad_norm": 2.8897762298583984, + "learning_rate": 2.829747427502339e-06, + "loss": 1.4577, + "step": 1181 + }, + { + "epoch": 1.4924242424242424, + "grad_norm": 2.848259925842285, + "learning_rate": 2.8274087932647336e-06, + "loss": 1.4323, + "step": 1182 + }, + { + "epoch": 1.4936868686868687, + "grad_norm": 2.9275119304656982, + "learning_rate": 2.8250701590271285e-06, + "loss": 1.4712, + "step": 1183 + }, + { + "epoch": 1.494949494949495, + "grad_norm": 2.8643906116485596, + "learning_rate": 2.8227315247895233e-06, + "loss": 1.4375, + "step": 1184 + }, + { + "epoch": 1.496212121212121, + "grad_norm": 3.0250704288482666, + "learning_rate": 2.8203928905519177e-06, + "loss": 1.5482, + "step": 1185 + }, + { + "epoch": 1.4974747474747474, + "grad_norm": 3.1019556522369385, + "learning_rate": 2.8180542563143126e-06, + "loss": 1.4618, + "step": 1186 + }, + { + "epoch": 1.4987373737373737, + "grad_norm": 3.1480300426483154, + "learning_rate": 2.8157156220767074e-06, + "loss": 1.4152, + "step": 1187 + }, + { + "epoch": 1.5, + "grad_norm": 2.82671856880188, + "learning_rate": 2.8133769878391023e-06, + "loss": 1.4692, + "step": 1188 + }, + { + "epoch": 1.5012626262626263, + "grad_norm": 3.2128899097442627, + "learning_rate": 2.8110383536014967e-06, + "loss": 1.4087, + "step": 1189 + }, + { + "epoch": 1.5025252525252526, + "grad_norm": 3.1939210891723633, + "learning_rate": 2.8086997193638915e-06, + "loss": 1.4919, + "step": 1190 + }, + { + "epoch": 1.5025252525252526, + "eval_loss": 1.6861544847488403, + "eval_runtime": 37.3328, + "eval_samples_per_second": 26.786, + "eval_steps_per_second": 1.125, + "step": 1190 + }, + { + "epoch": 1.503787878787879, + "grad_norm": 3.022059679031372, + "learning_rate": 2.8063610851262864e-06, + "loss": 1.4865, + "step": 1191 + }, + { + "epoch": 1.5050505050505052, + "grad_norm": 3.0068860054016113, + "learning_rate": 2.804022450888681e-06, + "loss": 1.478, + "step": 1192 + }, + { + "epoch": 1.5063131313131313, + "grad_norm": 3.2898383140563965, + "learning_rate": 2.801683816651076e-06, + "loss": 1.4773, + "step": 1193 + }, + { + "epoch": 1.5075757575757576, + "grad_norm": 3.090339422225952, + "learning_rate": 2.799345182413471e-06, + "loss": 1.4012, + "step": 1194 + }, + { + "epoch": 1.5088383838383839, + "grad_norm": 2.977215528488159, + "learning_rate": 2.7970065481758657e-06, + "loss": 1.4651, + "step": 1195 + }, + { + "epoch": 1.51010101010101, + "grad_norm": 3.1349570751190186, + "learning_rate": 2.7946679139382606e-06, + "loss": 1.5572, + "step": 1196 + }, + { + "epoch": 1.5113636363636362, + "grad_norm": 2.8715298175811768, + "learning_rate": 2.792329279700655e-06, + "loss": 1.4558, + "step": 1197 + }, + { + "epoch": 1.5126262626262625, + "grad_norm": 2.9848744869232178, + "learning_rate": 2.78999064546305e-06, + "loss": 1.4383, + "step": 1198 + }, + { + "epoch": 1.5138888888888888, + "grad_norm": 3.0774083137512207, + "learning_rate": 2.7876520112254447e-06, + "loss": 1.4607, + "step": 1199 + }, + { + "epoch": 1.5151515151515151, + "grad_norm": 2.9721343517303467, + "learning_rate": 2.785313376987839e-06, + "loss": 1.4341, + "step": 1200 + }, + { + "epoch": 1.5151515151515151, + "eval_loss": 1.6859924793243408, + "eval_runtime": 37.4589, + "eval_samples_per_second": 26.696, + "eval_steps_per_second": 1.121, + "step": 1200 + } + ], + "logging_steps": 1, + "max_steps": 2376, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 3.740771291234304e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1200/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1600/latest b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1600/latest new file mode 100644 index 0000000000000000000000000000000000000000..10b09fac99bc80ff931649e8b3378aab683b28be --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1600/latest @@ -0,0 +1 @@ +global_step1600 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1600/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1600/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..8991b49e9c2a43fc527dab9e09ad8171f0cc5943 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1600/tokenizer_config.json @@ -0,0 +1,2086 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128257": { + "content": "🅂", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + }, + "128258": { + "content": "🄿", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1600/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1600/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1600/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/latest b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/latest new file mode 100644 index 0000000000000000000000000000000000000000..cb24338af451be6d3ccddc18a950b58a70589761 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/latest @@ -0,0 +1 @@ +global_step1700 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..8991b49e9c2a43fc527dab9e09ad8171f0cc5943 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/tokenizer_config.json @@ -0,0 +1,2086 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128257": { + "content": "🅂", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + }, + "128258": { + "content": "🄿", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..b9efe4d3710651a7fc1e0d36a832d96359df8487 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/trainer_state.json @@ -0,0 +1,13293 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.1464646464646466, + "eval_steps": 10, + "global_step": 1700, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0012626262626262627, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8571, + "step": 1 + }, + { + "epoch": 0.0025252525252525255, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7758, + "step": 2 + }, + { + "epoch": 0.003787878787878788, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8383, + "step": 3 + }, + { + "epoch": 0.005050505050505051, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.9088, + "step": 4 + }, + { + "epoch": 0.006313131313131313, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8748, + "step": 5 + }, + { + "epoch": 0.007575757575757576, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8575, + "step": 6 + }, + { + "epoch": 0.008838383838383838, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8271, + "step": 7 + }, + { + "epoch": 0.010101010101010102, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8336, + "step": 8 + }, + { + "epoch": 0.011363636363636364, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8037, + "step": 9 + }, + { + "epoch": 0.012626262626262626, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8323, + "step": 10 + }, + { + "epoch": 0.012626262626262626, + "eval_loss": 2.8736937046051025, + "eval_runtime": 37.7802, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 10 + }, + { + "epoch": 0.013888888888888888, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8653, + "step": 11 + }, + { + "epoch": 0.015151515151515152, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8104, + "step": 12 + }, + { + "epoch": 0.016414141414141416, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8489, + "step": 13 + }, + { + "epoch": 0.017676767676767676, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8442, + "step": 14 + }, + { + "epoch": 0.01893939393939394, + "grad_norm": 5.356770992279053, + "learning_rate": 2.1008403361344538e-08, + "loss": 2.8207, + "step": 15 + }, + { + "epoch": 0.020202020202020204, + "grad_norm": 4.954462051391602, + "learning_rate": 4.2016806722689076e-08, + "loss": 2.8104, + "step": 16 + }, + { + "epoch": 0.021464646464646464, + "grad_norm": 5.248059272766113, + "learning_rate": 6.302521008403361e-08, + "loss": 2.8178, + "step": 17 + }, + { + "epoch": 0.022727272727272728, + "grad_norm": 5.277204513549805, + "learning_rate": 8.403361344537815e-08, + "loss": 2.8255, + "step": 18 + }, + { + "epoch": 0.023989898989898988, + "grad_norm": 5.344228267669678, + "learning_rate": 1.050420168067227e-07, + "loss": 2.8018, + "step": 19 + }, + { + "epoch": 0.025252525252525252, + "grad_norm": 5.004310131072998, + "learning_rate": 1.2605042016806723e-07, + "loss": 2.8337, + "step": 20 + }, + { + "epoch": 0.025252525252525252, + "eval_loss": 2.872413158416748, + "eval_runtime": 37.7426, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 20 + }, + { + "epoch": 0.026515151515151516, + "grad_norm": 5.209287166595459, + "learning_rate": 1.4705882352941178e-07, + "loss": 2.8246, + "step": 21 + }, + { + "epoch": 0.027777777777777776, + "grad_norm": 5.369327545166016, + "learning_rate": 1.680672268907563e-07, + "loss": 2.8791, + "step": 22 + }, + { + "epoch": 0.02904040404040404, + "grad_norm": 4.646446228027344, + "learning_rate": 1.8907563025210085e-07, + "loss": 2.8425, + "step": 23 + }, + { + "epoch": 0.030303030303030304, + "grad_norm": 4.949256420135498, + "learning_rate": 2.100840336134454e-07, + "loss": 2.8629, + "step": 24 + }, + { + "epoch": 0.03156565656565657, + "grad_norm": 4.709571838378906, + "learning_rate": 2.3109243697478996e-07, + "loss": 2.8264, + "step": 25 + }, + { + "epoch": 0.03282828282828283, + "grad_norm": 4.519557952880859, + "learning_rate": 2.5210084033613445e-07, + "loss": 2.8143, + "step": 26 + }, + { + "epoch": 0.03409090909090909, + "grad_norm": 4.445594310760498, + "learning_rate": 2.73109243697479e-07, + "loss": 2.8031, + "step": 27 + }, + { + "epoch": 0.03535353535353535, + "grad_norm": 4.26489782333374, + "learning_rate": 2.9411764705882356e-07, + "loss": 2.7719, + "step": 28 + }, + { + "epoch": 0.036616161616161616, + "grad_norm": 4.153639793395996, + "learning_rate": 3.151260504201681e-07, + "loss": 2.8251, + "step": 29 + }, + { + "epoch": 0.03787878787878788, + "grad_norm": 3.40531325340271, + "learning_rate": 3.361344537815126e-07, + "loss": 2.849, + "step": 30 + }, + { + "epoch": 0.03787878787878788, + "eval_loss": 2.825382709503174, + "eval_runtime": 37.0715, + "eval_samples_per_second": 26.975, + "eval_steps_per_second": 1.133, + "step": 30 + }, + { + "epoch": 0.039141414141414144, + "grad_norm": 3.4759507179260254, + "learning_rate": 3.5714285714285716e-07, + "loss": 2.7653, + "step": 31 + }, + { + "epoch": 0.04040404040404041, + "grad_norm": 3.4147374629974365, + "learning_rate": 3.781512605042017e-07, + "loss": 2.7916, + "step": 32 + }, + { + "epoch": 0.041666666666666664, + "grad_norm": 3.2775769233703613, + "learning_rate": 3.991596638655462e-07, + "loss": 2.8072, + "step": 33 + }, + { + "epoch": 0.04292929292929293, + "grad_norm": 3.3541715145111084, + "learning_rate": 4.201680672268908e-07, + "loss": 2.7944, + "step": 34 + }, + { + "epoch": 0.04419191919191919, + "grad_norm": 2.991971969604492, + "learning_rate": 4.4117647058823536e-07, + "loss": 2.7827, + "step": 35 + }, + { + "epoch": 0.045454545454545456, + "grad_norm": 2.9127862453460693, + "learning_rate": 4.621848739495799e-07, + "loss": 2.7332, + "step": 36 + }, + { + "epoch": 0.04671717171717172, + "grad_norm": 2.839707136154175, + "learning_rate": 4.831932773109245e-07, + "loss": 2.7568, + "step": 37 + }, + { + "epoch": 0.047979797979797977, + "grad_norm": 2.9395432472229004, + "learning_rate": 5.042016806722689e-07, + "loss": 2.7445, + "step": 38 + }, + { + "epoch": 0.04924242424242424, + "grad_norm": 2.869896650314331, + "learning_rate": 5.252100840336135e-07, + "loss": 2.7157, + "step": 39 + }, + { + "epoch": 0.050505050505050504, + "grad_norm": 2.9064571857452393, + "learning_rate": 5.46218487394958e-07, + "loss": 2.7049, + "step": 40 + }, + { + "epoch": 0.050505050505050504, + "eval_loss": 2.7477543354034424, + "eval_runtime": 37.0669, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 40 + }, + { + "epoch": 0.05176767676767677, + "grad_norm": 2.6264421939849854, + "learning_rate": 5.672268907563026e-07, + "loss": 2.7651, + "step": 41 + }, + { + "epoch": 0.05303030303030303, + "grad_norm": 2.518627405166626, + "learning_rate": 5.882352941176471e-07, + "loss": 2.7241, + "step": 42 + }, + { + "epoch": 0.054292929292929296, + "grad_norm": 2.521034002304077, + "learning_rate": 6.092436974789917e-07, + "loss": 2.6691, + "step": 43 + }, + { + "epoch": 0.05555555555555555, + "grad_norm": 2.4679362773895264, + "learning_rate": 6.302521008403362e-07, + "loss": 2.743, + "step": 44 + }, + { + "epoch": 0.056818181818181816, + "grad_norm": 2.5817418098449707, + "learning_rate": 6.512605042016808e-07, + "loss": 2.7067, + "step": 45 + }, + { + "epoch": 0.05808080808080808, + "grad_norm": 2.3704075813293457, + "learning_rate": 6.722689075630252e-07, + "loss": 2.6495, + "step": 46 + }, + { + "epoch": 0.059343434343434344, + "grad_norm": 2.2321906089782715, + "learning_rate": 6.932773109243699e-07, + "loss": 2.6606, + "step": 47 + }, + { + "epoch": 0.06060606060606061, + "grad_norm": 1.985407829284668, + "learning_rate": 7.142857142857143e-07, + "loss": 2.662, + "step": 48 + }, + { + "epoch": 0.06186868686868687, + "grad_norm": 1.945688247680664, + "learning_rate": 7.352941176470589e-07, + "loss": 2.6319, + "step": 49 + }, + { + "epoch": 0.06313131313131314, + "grad_norm": 1.9392870664596558, + "learning_rate": 7.563025210084034e-07, + "loss": 2.6094, + "step": 50 + }, + { + "epoch": 0.06313131313131314, + "eval_loss": 2.6554172039031982, + "eval_runtime": 37.0569, + "eval_samples_per_second": 26.986, + "eval_steps_per_second": 1.133, + "step": 50 + }, + { + "epoch": 0.06439393939393939, + "grad_norm": 1.9975577592849731, + "learning_rate": 7.77310924369748e-07, + "loss": 2.6045, + "step": 51 + }, + { + "epoch": 0.06565656565656566, + "grad_norm": 1.9733693599700928, + "learning_rate": 7.983193277310924e-07, + "loss": 2.649, + "step": 52 + }, + { + "epoch": 0.06691919191919192, + "grad_norm": 1.7324917316436768, + "learning_rate": 8.193277310924371e-07, + "loss": 2.5689, + "step": 53 + }, + { + "epoch": 0.06818181818181818, + "grad_norm": 1.7553645372390747, + "learning_rate": 8.403361344537816e-07, + "loss": 2.644, + "step": 54 + }, + { + "epoch": 0.06944444444444445, + "grad_norm": 1.8163801431655884, + "learning_rate": 8.613445378151261e-07, + "loss": 2.4943, + "step": 55 + }, + { + "epoch": 0.0707070707070707, + "grad_norm": 1.7717015743255615, + "learning_rate": 8.823529411764707e-07, + "loss": 2.6093, + "step": 56 + }, + { + "epoch": 0.07196969696969698, + "grad_norm": 1.7013663053512573, + "learning_rate": 9.033613445378152e-07, + "loss": 2.6376, + "step": 57 + }, + { + "epoch": 0.07323232323232323, + "grad_norm": 1.752943992614746, + "learning_rate": 9.243697478991598e-07, + "loss": 2.5172, + "step": 58 + }, + { + "epoch": 0.07449494949494949, + "grad_norm": 1.5362415313720703, + "learning_rate": 9.453781512605043e-07, + "loss": 2.6129, + "step": 59 + }, + { + "epoch": 0.07575757575757576, + "grad_norm": 1.514885425567627, + "learning_rate": 9.66386554621849e-07, + "loss": 2.6064, + "step": 60 + }, + { + "epoch": 0.07575757575757576, + "eval_loss": 2.5934340953826904, + "eval_runtime": 37.0782, + "eval_samples_per_second": 26.97, + "eval_steps_per_second": 1.133, + "step": 60 + }, + { + "epoch": 0.07702020202020202, + "grad_norm": 1.5915579795837402, + "learning_rate": 9.873949579831934e-07, + "loss": 2.5491, + "step": 61 + }, + { + "epoch": 0.07828282828282829, + "grad_norm": 1.601773977279663, + "learning_rate": 1.0084033613445378e-06, + "loss": 2.5685, + "step": 62 + }, + { + "epoch": 0.07954545454545454, + "grad_norm": 1.8327804803848267, + "learning_rate": 1.0294117647058825e-06, + "loss": 2.5763, + "step": 63 + }, + { + "epoch": 0.08080808080808081, + "grad_norm": 1.5276741981506348, + "learning_rate": 1.050420168067227e-06, + "loss": 2.5494, + "step": 64 + }, + { + "epoch": 0.08207070707070707, + "grad_norm": 1.7065422534942627, + "learning_rate": 1.0714285714285714e-06, + "loss": 2.449, + "step": 65 + }, + { + "epoch": 0.08333333333333333, + "grad_norm": 2.0551910400390625, + "learning_rate": 1.092436974789916e-06, + "loss": 2.4371, + "step": 66 + }, + { + "epoch": 0.0845959595959596, + "grad_norm": 2.154855728149414, + "learning_rate": 1.1134453781512607e-06, + "loss": 2.5417, + "step": 67 + }, + { + "epoch": 0.08585858585858586, + "grad_norm": 1.7922390699386597, + "learning_rate": 1.1344537815126051e-06, + "loss": 2.4754, + "step": 68 + }, + { + "epoch": 0.08712121212121213, + "grad_norm": 1.745632290840149, + "learning_rate": 1.1554621848739498e-06, + "loss": 2.4669, + "step": 69 + }, + { + "epoch": 0.08838383838383838, + "grad_norm": 1.8907713890075684, + "learning_rate": 1.1764705882352942e-06, + "loss": 2.4943, + "step": 70 + }, + { + "epoch": 0.08838383838383838, + "eval_loss": 2.518237829208374, + "eval_runtime": 37.074, + "eval_samples_per_second": 26.973, + "eval_steps_per_second": 1.133, + "step": 70 + }, + { + "epoch": 0.08964646464646464, + "grad_norm": 1.8178164958953857, + "learning_rate": 1.1974789915966389e-06, + "loss": 2.4891, + "step": 71 + }, + { + "epoch": 0.09090909090909091, + "grad_norm": 1.970969796180725, + "learning_rate": 1.2184873949579833e-06, + "loss": 2.4797, + "step": 72 + }, + { + "epoch": 0.09217171717171717, + "grad_norm": 2.0211431980133057, + "learning_rate": 1.2394957983193278e-06, + "loss": 2.4716, + "step": 73 + }, + { + "epoch": 0.09343434343434344, + "grad_norm": 2.650787591934204, + "learning_rate": 1.2605042016806724e-06, + "loss": 2.4915, + "step": 74 + }, + { + "epoch": 0.0946969696969697, + "grad_norm": 2.235499143600464, + "learning_rate": 1.2815126050420169e-06, + "loss": 2.3917, + "step": 75 + }, + { + "epoch": 0.09595959595959595, + "grad_norm": 2.245999813079834, + "learning_rate": 1.3025210084033615e-06, + "loss": 2.4942, + "step": 76 + }, + { + "epoch": 0.09722222222222222, + "grad_norm": 2.4414420127868652, + "learning_rate": 1.323529411764706e-06, + "loss": 2.4004, + "step": 77 + }, + { + "epoch": 0.09848484848484848, + "grad_norm": 2.279937982559204, + "learning_rate": 1.3445378151260504e-06, + "loss": 2.4043, + "step": 78 + }, + { + "epoch": 0.09974747474747475, + "grad_norm": 2.3830068111419678, + "learning_rate": 1.3655462184873949e-06, + "loss": 2.4096, + "step": 79 + }, + { + "epoch": 0.10101010101010101, + "grad_norm": 2.0119733810424805, + "learning_rate": 1.3865546218487397e-06, + "loss": 2.4227, + "step": 80 + }, + { + "epoch": 0.10101010101010101, + "eval_loss": 2.4316186904907227, + "eval_runtime": 37.0837, + "eval_samples_per_second": 26.966, + "eval_steps_per_second": 1.133, + "step": 80 + }, + { + "epoch": 0.10227272727272728, + "grad_norm": 2.1273374557495117, + "learning_rate": 1.4075630252100842e-06, + "loss": 2.3541, + "step": 81 + }, + { + "epoch": 0.10353535353535354, + "grad_norm": 2.6852898597717285, + "learning_rate": 1.4285714285714286e-06, + "loss": 2.3321, + "step": 82 + }, + { + "epoch": 0.10479797979797979, + "grad_norm": 2.2622721195220947, + "learning_rate": 1.4495798319327733e-06, + "loss": 2.3166, + "step": 83 + }, + { + "epoch": 0.10606060606060606, + "grad_norm": 2.98258900642395, + "learning_rate": 1.4705882352941177e-06, + "loss": 2.3733, + "step": 84 + }, + { + "epoch": 0.10732323232323232, + "grad_norm": 2.5140459537506104, + "learning_rate": 1.4915966386554624e-06, + "loss": 2.314, + "step": 85 + }, + { + "epoch": 0.10858585858585859, + "grad_norm": 2.637608051300049, + "learning_rate": 1.5126050420168068e-06, + "loss": 2.3479, + "step": 86 + }, + { + "epoch": 0.10984848484848485, + "grad_norm": 2.399508476257324, + "learning_rate": 1.5336134453781515e-06, + "loss": 2.2932, + "step": 87 + }, + { + "epoch": 0.1111111111111111, + "grad_norm": 3.0835793018341064, + "learning_rate": 1.554621848739496e-06, + "loss": 2.2886, + "step": 88 + }, + { + "epoch": 0.11237373737373738, + "grad_norm": 3.288769483566284, + "learning_rate": 1.5756302521008404e-06, + "loss": 2.2914, + "step": 89 + }, + { + "epoch": 0.11363636363636363, + "grad_norm": 2.538361072540283, + "learning_rate": 1.5966386554621848e-06, + "loss": 2.3247, + "step": 90 + }, + { + "epoch": 0.11363636363636363, + "eval_loss": 2.3510072231292725, + "eval_runtime": 37.0847, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 90 + }, + { + "epoch": 0.1148989898989899, + "grad_norm": 2.8016724586486816, + "learning_rate": 1.6176470588235297e-06, + "loss": 2.281, + "step": 91 + }, + { + "epoch": 0.11616161616161616, + "grad_norm": 2.9251859188079834, + "learning_rate": 1.6386554621848741e-06, + "loss": 2.2715, + "step": 92 + }, + { + "epoch": 0.11742424242424243, + "grad_norm": 2.595750570297241, + "learning_rate": 1.6596638655462186e-06, + "loss": 2.2127, + "step": 93 + }, + { + "epoch": 0.11868686868686869, + "grad_norm": 2.6989378929138184, + "learning_rate": 1.6806722689075632e-06, + "loss": 2.2623, + "step": 94 + }, + { + "epoch": 0.11994949494949494, + "grad_norm": 2.554863929748535, + "learning_rate": 1.7016806722689077e-06, + "loss": 2.2649, + "step": 95 + }, + { + "epoch": 0.12121212121212122, + "grad_norm": 2.7013020515441895, + "learning_rate": 1.7226890756302521e-06, + "loss": 2.2428, + "step": 96 + }, + { + "epoch": 0.12247474747474747, + "grad_norm": 2.7415151596069336, + "learning_rate": 1.7436974789915968e-06, + "loss": 2.1882, + "step": 97 + }, + { + "epoch": 0.12373737373737374, + "grad_norm": 2.672806739807129, + "learning_rate": 1.7647058823529414e-06, + "loss": 2.2302, + "step": 98 + }, + { + "epoch": 0.125, + "grad_norm": 2.9508326053619385, + "learning_rate": 1.7857142857142859e-06, + "loss": 2.1894, + "step": 99 + }, + { + "epoch": 0.12626262626262627, + "grad_norm": 2.4668166637420654, + "learning_rate": 1.8067226890756303e-06, + "loss": 2.2043, + "step": 100 + }, + { + "epoch": 0.12626262626262627, + "eval_loss": 2.2788307666778564, + "eval_runtime": 37.1156, + "eval_samples_per_second": 26.943, + "eval_steps_per_second": 1.132, + "step": 100 + }, + { + "epoch": 0.1275252525252525, + "grad_norm": 2.8241467475891113, + "learning_rate": 1.8277310924369748e-06, + "loss": 2.2109, + "step": 101 + }, + { + "epoch": 0.12878787878787878, + "grad_norm": 2.4179229736328125, + "learning_rate": 1.8487394957983196e-06, + "loss": 2.2193, + "step": 102 + }, + { + "epoch": 0.13005050505050506, + "grad_norm": 2.714315891265869, + "learning_rate": 1.869747899159664e-06, + "loss": 2.2474, + "step": 103 + }, + { + "epoch": 0.13131313131313133, + "grad_norm": 2.8560359477996826, + "learning_rate": 1.8907563025210085e-06, + "loss": 2.2205, + "step": 104 + }, + { + "epoch": 0.13257575757575757, + "grad_norm": 2.7780051231384277, + "learning_rate": 1.9117647058823528e-06, + "loss": 2.1834, + "step": 105 + }, + { + "epoch": 0.13383838383838384, + "grad_norm": 3.3773467540740967, + "learning_rate": 1.932773109243698e-06, + "loss": 2.2063, + "step": 106 + }, + { + "epoch": 0.1351010101010101, + "grad_norm": 2.927551746368408, + "learning_rate": 1.953781512605042e-06, + "loss": 2.1592, + "step": 107 + }, + { + "epoch": 0.13636363636363635, + "grad_norm": 2.8094165325164795, + "learning_rate": 1.9747899159663867e-06, + "loss": 2.2619, + "step": 108 + }, + { + "epoch": 0.13762626262626262, + "grad_norm": 2.5312564373016357, + "learning_rate": 1.9957983193277314e-06, + "loss": 2.2323, + "step": 109 + }, + { + "epoch": 0.1388888888888889, + "grad_norm": 3.175746202468872, + "learning_rate": 2.0168067226890756e-06, + "loss": 2.1776, + "step": 110 + }, + { + "epoch": 0.1388888888888889, + "eval_loss": 2.2208564281463623, + "eval_runtime": 37.3885, + "eval_samples_per_second": 26.746, + "eval_steps_per_second": 1.123, + "step": 110 + }, + { + "epoch": 0.14015151515151514, + "grad_norm": 2.660886287689209, + "learning_rate": 2.0378151260504203e-06, + "loss": 2.1626, + "step": 111 + }, + { + "epoch": 0.1414141414141414, + "grad_norm": 2.6948511600494385, + "learning_rate": 2.058823529411765e-06, + "loss": 2.1731, + "step": 112 + }, + { + "epoch": 0.14267676767676768, + "grad_norm": 2.826976776123047, + "learning_rate": 2.0798319327731096e-06, + "loss": 2.1865, + "step": 113 + }, + { + "epoch": 0.14393939393939395, + "grad_norm": 3.210033655166626, + "learning_rate": 2.100840336134454e-06, + "loss": 2.1325, + "step": 114 + }, + { + "epoch": 0.1452020202020202, + "grad_norm": 2.509060859680176, + "learning_rate": 2.1218487394957985e-06, + "loss": 2.0926, + "step": 115 + }, + { + "epoch": 0.14646464646464646, + "grad_norm": 3.51330304145813, + "learning_rate": 2.1428571428571427e-06, + "loss": 2.1577, + "step": 116 + }, + { + "epoch": 0.14772727272727273, + "grad_norm": 2.468395471572876, + "learning_rate": 2.163865546218488e-06, + "loss": 2.1, + "step": 117 + }, + { + "epoch": 0.14898989898989898, + "grad_norm": 3.081529378890991, + "learning_rate": 2.184873949579832e-06, + "loss": 2.1082, + "step": 118 + }, + { + "epoch": 0.15025252525252525, + "grad_norm": 2.8856358528137207, + "learning_rate": 2.2058823529411767e-06, + "loss": 2.1288, + "step": 119 + }, + { + "epoch": 0.15151515151515152, + "grad_norm": 3.323089838027954, + "learning_rate": 2.2268907563025214e-06, + "loss": 2.0933, + "step": 120 + }, + { + "epoch": 0.15151515151515152, + "eval_loss": 2.1834487915039062, + "eval_runtime": 37.4344, + "eval_samples_per_second": 26.713, + "eval_steps_per_second": 1.122, + "step": 120 + }, + { + "epoch": 0.1527777777777778, + "grad_norm": 2.8871560096740723, + "learning_rate": 2.2478991596638656e-06, + "loss": 2.1519, + "step": 121 + }, + { + "epoch": 0.15404040404040403, + "grad_norm": 3.0718588829040527, + "learning_rate": 2.2689075630252102e-06, + "loss": 2.0321, + "step": 122 + }, + { + "epoch": 0.1553030303030303, + "grad_norm": 2.820739507675171, + "learning_rate": 2.289915966386555e-06, + "loss": 2.0944, + "step": 123 + }, + { + "epoch": 0.15656565656565657, + "grad_norm": 3.1259310245513916, + "learning_rate": 2.3109243697478996e-06, + "loss": 2.0141, + "step": 124 + }, + { + "epoch": 0.15782828282828282, + "grad_norm": 3.303004741668701, + "learning_rate": 2.331932773109244e-06, + "loss": 2.1095, + "step": 125 + }, + { + "epoch": 0.1590909090909091, + "grad_norm": 3.1838674545288086, + "learning_rate": 2.3529411764705885e-06, + "loss": 2.0727, + "step": 126 + }, + { + "epoch": 0.16035353535353536, + "grad_norm": 3.1013786792755127, + "learning_rate": 2.3739495798319327e-06, + "loss": 2.0963, + "step": 127 + }, + { + "epoch": 0.16161616161616163, + "grad_norm": 2.803283214569092, + "learning_rate": 2.3949579831932778e-06, + "loss": 2.0551, + "step": 128 + }, + { + "epoch": 0.16287878787878787, + "grad_norm": 3.1616814136505127, + "learning_rate": 2.415966386554622e-06, + "loss": 2.0485, + "step": 129 + }, + { + "epoch": 0.16414141414141414, + "grad_norm": 3.333782196044922, + "learning_rate": 2.4369747899159667e-06, + "loss": 2.041, + "step": 130 + }, + { + "epoch": 0.16414141414141414, + "eval_loss": 2.150329351425171, + "eval_runtime": 37.494, + "eval_samples_per_second": 26.671, + "eval_steps_per_second": 1.12, + "step": 130 + }, + { + "epoch": 0.16540404040404041, + "grad_norm": 2.9714694023132324, + "learning_rate": 2.457983193277311e-06, + "loss": 2.0686, + "step": 131 + }, + { + "epoch": 0.16666666666666666, + "grad_norm": 3.216895818710327, + "learning_rate": 2.4789915966386555e-06, + "loss": 2.003, + "step": 132 + }, + { + "epoch": 0.16792929292929293, + "grad_norm": 2.9636809825897217, + "learning_rate": 2.5e-06, + "loss": 2.0534, + "step": 133 + }, + { + "epoch": 0.1691919191919192, + "grad_norm": 3.6477177143096924, + "learning_rate": 2.521008403361345e-06, + "loss": 2.077, + "step": 134 + }, + { + "epoch": 0.17045454545454544, + "grad_norm": 2.7437169551849365, + "learning_rate": 2.542016806722689e-06, + "loss": 2.0501, + "step": 135 + }, + { + "epoch": 0.1717171717171717, + "grad_norm": 3.2992663383483887, + "learning_rate": 2.5630252100840338e-06, + "loss": 2.0284, + "step": 136 + }, + { + "epoch": 0.17297979797979798, + "grad_norm": 3.0735950469970703, + "learning_rate": 2.5840336134453784e-06, + "loss": 2.025, + "step": 137 + }, + { + "epoch": 0.17424242424242425, + "grad_norm": 3.588418960571289, + "learning_rate": 2.605042016806723e-06, + "loss": 2.002, + "step": 138 + }, + { + "epoch": 0.1755050505050505, + "grad_norm": 2.8772335052490234, + "learning_rate": 2.6260504201680677e-06, + "loss": 1.9873, + "step": 139 + }, + { + "epoch": 0.17676767676767677, + "grad_norm": 2.9078969955444336, + "learning_rate": 2.647058823529412e-06, + "loss": 2.011, + "step": 140 + }, + { + "epoch": 0.17676767676767677, + "eval_loss": 2.1225531101226807, + "eval_runtime": 37.4531, + "eval_samples_per_second": 26.7, + "eval_steps_per_second": 1.121, + "step": 140 + }, + { + "epoch": 0.17803030303030304, + "grad_norm": 3.556654214859009, + "learning_rate": 2.6680672268907566e-06, + "loss": 2.0401, + "step": 141 + }, + { + "epoch": 0.17929292929292928, + "grad_norm": 2.9375200271606445, + "learning_rate": 2.689075630252101e-06, + "loss": 2.0277, + "step": 142 + }, + { + "epoch": 0.18055555555555555, + "grad_norm": 3.4019014835357666, + "learning_rate": 2.7100840336134455e-06, + "loss": 1.975, + "step": 143 + }, + { + "epoch": 0.18181818181818182, + "grad_norm": 2.7653748989105225, + "learning_rate": 2.7310924369747897e-06, + "loss": 2.0102, + "step": 144 + }, + { + "epoch": 0.1830808080808081, + "grad_norm": 3.4416797161102295, + "learning_rate": 2.752100840336135e-06, + "loss": 2.0093, + "step": 145 + }, + { + "epoch": 0.18434343434343434, + "grad_norm": 3.2531726360321045, + "learning_rate": 2.7731092436974795e-06, + "loss": 1.9852, + "step": 146 + }, + { + "epoch": 0.1856060606060606, + "grad_norm": 3.171502113342285, + "learning_rate": 2.7941176470588237e-06, + "loss": 1.9165, + "step": 147 + }, + { + "epoch": 0.18686868686868688, + "grad_norm": 3.114550828933716, + "learning_rate": 2.8151260504201684e-06, + "loss": 1.9516, + "step": 148 + }, + { + "epoch": 0.18813131313131312, + "grad_norm": 3.3601152896881104, + "learning_rate": 2.8361344537815126e-06, + "loss": 2.0183, + "step": 149 + }, + { + "epoch": 0.1893939393939394, + "grad_norm": 3.059622049331665, + "learning_rate": 2.8571428571428573e-06, + "loss": 1.9682, + "step": 150 + }, + { + "epoch": 0.1893939393939394, + "eval_loss": 2.0962483882904053, + "eval_runtime": 37.2364, + "eval_samples_per_second": 26.855, + "eval_steps_per_second": 1.128, + "step": 150 + }, + { + "epoch": 0.19065656565656566, + "grad_norm": 2.8749427795410156, + "learning_rate": 2.8781512605042015e-06, + "loss": 2.0089, + "step": 151 + }, + { + "epoch": 0.1919191919191919, + "grad_norm": 3.067267894744873, + "learning_rate": 2.8991596638655466e-06, + "loss": 1.9661, + "step": 152 + }, + { + "epoch": 0.19318181818181818, + "grad_norm": 2.7126266956329346, + "learning_rate": 2.9201680672268912e-06, + "loss": 1.982, + "step": 153 + }, + { + "epoch": 0.19444444444444445, + "grad_norm": 3.1330652236938477, + "learning_rate": 2.9411764705882355e-06, + "loss": 1.9753, + "step": 154 + }, + { + "epoch": 0.19570707070707072, + "grad_norm": 3.158313274383545, + "learning_rate": 2.96218487394958e-06, + "loss": 1.9431, + "step": 155 + }, + { + "epoch": 0.19696969696969696, + "grad_norm": 3.30806303024292, + "learning_rate": 2.9831932773109248e-06, + "loss": 1.9812, + "step": 156 + }, + { + "epoch": 0.19823232323232323, + "grad_norm": 3.0344884395599365, + "learning_rate": 3.004201680672269e-06, + "loss": 1.9677, + "step": 157 + }, + { + "epoch": 0.1994949494949495, + "grad_norm": 2.9032909870147705, + "learning_rate": 3.0252100840336137e-06, + "loss": 1.9132, + "step": 158 + }, + { + "epoch": 0.20075757575757575, + "grad_norm": 3.459934711456299, + "learning_rate": 3.046218487394958e-06, + "loss": 1.9581, + "step": 159 + }, + { + "epoch": 0.20202020202020202, + "grad_norm": 2.9455478191375732, + "learning_rate": 3.067226890756303e-06, + "loss": 1.9509, + "step": 160 + }, + { + "epoch": 0.20202020202020202, + "eval_loss": 2.074032783508301, + "eval_runtime": 37.1824, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 160 + }, + { + "epoch": 0.2032828282828283, + "grad_norm": 3.2347211837768555, + "learning_rate": 3.0882352941176476e-06, + "loss": 1.9743, + "step": 161 + }, + { + "epoch": 0.20454545454545456, + "grad_norm": 3.3152947425842285, + "learning_rate": 3.109243697478992e-06, + "loss": 1.8971, + "step": 162 + }, + { + "epoch": 0.2058080808080808, + "grad_norm": 3.035548210144043, + "learning_rate": 3.1302521008403365e-06, + "loss": 1.8762, + "step": 163 + }, + { + "epoch": 0.20707070707070707, + "grad_norm": 3.1004066467285156, + "learning_rate": 3.1512605042016808e-06, + "loss": 1.9512, + "step": 164 + }, + { + "epoch": 0.20833333333333334, + "grad_norm": 3.2280282974243164, + "learning_rate": 3.1722689075630254e-06, + "loss": 1.8856, + "step": 165 + }, + { + "epoch": 0.20959595959595959, + "grad_norm": 3.021969795227051, + "learning_rate": 3.1932773109243696e-06, + "loss": 1.9127, + "step": 166 + }, + { + "epoch": 0.21085858585858586, + "grad_norm": 3.1647000312805176, + "learning_rate": 3.2142857142857147e-06, + "loss": 2.038, + "step": 167 + }, + { + "epoch": 0.21212121212121213, + "grad_norm": 3.0084872245788574, + "learning_rate": 3.2352941176470594e-06, + "loss": 1.9536, + "step": 168 + }, + { + "epoch": 0.21338383838383837, + "grad_norm": 2.951509952545166, + "learning_rate": 3.2563025210084036e-06, + "loss": 1.933, + "step": 169 + }, + { + "epoch": 0.21464646464646464, + "grad_norm": 3.472856283187866, + "learning_rate": 3.2773109243697483e-06, + "loss": 1.934, + "step": 170 + }, + { + "epoch": 0.21464646464646464, + "eval_loss": 2.0450048446655273, + "eval_runtime": 37.1669, + "eval_samples_per_second": 26.906, + "eval_steps_per_second": 1.13, + "step": 170 + }, + { + "epoch": 0.2159090909090909, + "grad_norm": 3.1104896068573, + "learning_rate": 3.2983193277310925e-06, + "loss": 1.9008, + "step": 171 + }, + { + "epoch": 0.21717171717171718, + "grad_norm": 2.847781181335449, + "learning_rate": 3.319327731092437e-06, + "loss": 1.9592, + "step": 172 + }, + { + "epoch": 0.21843434343434343, + "grad_norm": 3.384554624557495, + "learning_rate": 3.3403361344537814e-06, + "loss": 1.9087, + "step": 173 + }, + { + "epoch": 0.2196969696969697, + "grad_norm": 3.221564292907715, + "learning_rate": 3.3613445378151265e-06, + "loss": 1.9107, + "step": 174 + }, + { + "epoch": 0.22095959595959597, + "grad_norm": 3.30204439163208, + "learning_rate": 3.382352941176471e-06, + "loss": 1.8879, + "step": 175 + }, + { + "epoch": 0.2222222222222222, + "grad_norm": 3.549259662628174, + "learning_rate": 3.4033613445378154e-06, + "loss": 1.9592, + "step": 176 + }, + { + "epoch": 0.22348484848484848, + "grad_norm": 3.498199462890625, + "learning_rate": 3.42436974789916e-06, + "loss": 1.984, + "step": 177 + }, + { + "epoch": 0.22474747474747475, + "grad_norm": 3.576317548751831, + "learning_rate": 3.4453781512605043e-06, + "loss": 1.8897, + "step": 178 + }, + { + "epoch": 0.22601010101010102, + "grad_norm": 3.5337026119232178, + "learning_rate": 3.466386554621849e-06, + "loss": 1.9607, + "step": 179 + }, + { + "epoch": 0.22727272727272727, + "grad_norm": 3.342886209487915, + "learning_rate": 3.4873949579831936e-06, + "loss": 1.8797, + "step": 180 + }, + { + "epoch": 0.22727272727272727, + "eval_loss": 2.02158522605896, + "eval_runtime": 37.1382, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 180 + }, + { + "epoch": 0.22853535353535354, + "grad_norm": 2.941906213760376, + "learning_rate": 3.508403361344538e-06, + "loss": 1.9, + "step": 181 + }, + { + "epoch": 0.2297979797979798, + "grad_norm": 3.1913063526153564, + "learning_rate": 3.529411764705883e-06, + "loss": 1.8704, + "step": 182 + }, + { + "epoch": 0.23106060606060605, + "grad_norm": 3.0619900226593018, + "learning_rate": 3.550420168067227e-06, + "loss": 1.8851, + "step": 183 + }, + { + "epoch": 0.23232323232323232, + "grad_norm": 3.051288604736328, + "learning_rate": 3.5714285714285718e-06, + "loss": 1.9422, + "step": 184 + }, + { + "epoch": 0.2335858585858586, + "grad_norm": 2.999845266342163, + "learning_rate": 3.5924369747899164e-06, + "loss": 1.8757, + "step": 185 + }, + { + "epoch": 0.23484848484848486, + "grad_norm": 3.262355089187622, + "learning_rate": 3.6134453781512607e-06, + "loss": 1.9114, + "step": 186 + }, + { + "epoch": 0.2361111111111111, + "grad_norm": 3.0354340076446533, + "learning_rate": 3.6344537815126053e-06, + "loss": 1.8547, + "step": 187 + }, + { + "epoch": 0.23737373737373738, + "grad_norm": 2.9401369094848633, + "learning_rate": 3.6554621848739496e-06, + "loss": 1.8974, + "step": 188 + }, + { + "epoch": 0.23863636363636365, + "grad_norm": 3.0544443130493164, + "learning_rate": 3.6764705882352946e-06, + "loss": 1.8662, + "step": 189 + }, + { + "epoch": 0.2398989898989899, + "grad_norm": 2.7257750034332275, + "learning_rate": 3.6974789915966393e-06, + "loss": 1.8737, + "step": 190 + }, + { + "epoch": 0.2398989898989899, + "eval_loss": 2.0065128803253174, + "eval_runtime": 37.1296, + "eval_samples_per_second": 26.933, + "eval_steps_per_second": 1.131, + "step": 190 + }, + { + "epoch": 0.24116161616161616, + "grad_norm": 2.8592522144317627, + "learning_rate": 3.7184873949579835e-06, + "loss": 1.8984, + "step": 191 + }, + { + "epoch": 0.24242424242424243, + "grad_norm": 2.833958864212036, + "learning_rate": 3.739495798319328e-06, + "loss": 1.9084, + "step": 192 + }, + { + "epoch": 0.24368686868686867, + "grad_norm": 2.7842230796813965, + "learning_rate": 3.7605042016806724e-06, + "loss": 1.8843, + "step": 193 + }, + { + "epoch": 0.24494949494949494, + "grad_norm": 3.1309409141540527, + "learning_rate": 3.781512605042017e-06, + "loss": 1.8259, + "step": 194 + }, + { + "epoch": 0.24621212121212122, + "grad_norm": 3.120788335800171, + "learning_rate": 3.8025210084033613e-06, + "loss": 1.8951, + "step": 195 + }, + { + "epoch": 0.2474747474747475, + "grad_norm": 3.2228143215179443, + "learning_rate": 3.8235294117647055e-06, + "loss": 1.8608, + "step": 196 + }, + { + "epoch": 0.24873737373737373, + "grad_norm": 3.3362114429473877, + "learning_rate": 3.844537815126051e-06, + "loss": 1.9163, + "step": 197 + }, + { + "epoch": 0.25, + "grad_norm": 3.0973799228668213, + "learning_rate": 3.865546218487396e-06, + "loss": 1.911, + "step": 198 + }, + { + "epoch": 0.25126262626262624, + "grad_norm": 3.5567545890808105, + "learning_rate": 3.8865546218487395e-06, + "loss": 1.8795, + "step": 199 + }, + { + "epoch": 0.25252525252525254, + "grad_norm": 3.4758830070495605, + "learning_rate": 3.907563025210084e-06, + "loss": 1.8813, + "step": 200 + }, + { + "epoch": 0.25252525252525254, + "eval_loss": 1.9892102479934692, + "eval_runtime": 37.1311, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 200 + }, + { + "epoch": 0.2537878787878788, + "grad_norm": 3.5564475059509277, + "learning_rate": 3.928571428571429e-06, + "loss": 1.8469, + "step": 201 + }, + { + "epoch": 0.255050505050505, + "grad_norm": 3.7004899978637695, + "learning_rate": 3.9495798319327735e-06, + "loss": 1.873, + "step": 202 + }, + { + "epoch": 0.2563131313131313, + "grad_norm": 3.9285225868225098, + "learning_rate": 3.970588235294118e-06, + "loss": 1.8596, + "step": 203 + }, + { + "epoch": 0.25757575757575757, + "grad_norm": 3.1388907432556152, + "learning_rate": 3.991596638655463e-06, + "loss": 1.7991, + "step": 204 + }, + { + "epoch": 0.2588383838383838, + "grad_norm": 3.3797011375427246, + "learning_rate": 4.0126050420168075e-06, + "loss": 1.8682, + "step": 205 + }, + { + "epoch": 0.2601010101010101, + "grad_norm": 3.32183575630188, + "learning_rate": 4.033613445378151e-06, + "loss": 1.8396, + "step": 206 + }, + { + "epoch": 0.26136363636363635, + "grad_norm": 3.457291603088379, + "learning_rate": 4.054621848739496e-06, + "loss": 1.8314, + "step": 207 + }, + { + "epoch": 0.26262626262626265, + "grad_norm": 3.2906265258789062, + "learning_rate": 4.075630252100841e-06, + "loss": 1.844, + "step": 208 + }, + { + "epoch": 0.2638888888888889, + "grad_norm": 3.081913948059082, + "learning_rate": 4.096638655462185e-06, + "loss": 1.8198, + "step": 209 + }, + { + "epoch": 0.26515151515151514, + "grad_norm": 2.9697105884552, + "learning_rate": 4.11764705882353e-06, + "loss": 1.7885, + "step": 210 + }, + { + "epoch": 0.26515151515151514, + "eval_loss": 1.9749963283538818, + "eval_runtime": 36.962, + "eval_samples_per_second": 27.055, + "eval_steps_per_second": 1.136, + "step": 210 + }, + { + "epoch": 0.26641414141414144, + "grad_norm": 3.0732104778289795, + "learning_rate": 4.138655462184874e-06, + "loss": 1.8463, + "step": 211 + }, + { + "epoch": 0.2676767676767677, + "grad_norm": 3.4714958667755127, + "learning_rate": 4.159663865546219e-06, + "loss": 1.8537, + "step": 212 + }, + { + "epoch": 0.2689393939393939, + "grad_norm": 2.955864429473877, + "learning_rate": 4.180672268907564e-06, + "loss": 1.8566, + "step": 213 + }, + { + "epoch": 0.2702020202020202, + "grad_norm": 3.5314948558807373, + "learning_rate": 4.201680672268908e-06, + "loss": 1.7625, + "step": 214 + }, + { + "epoch": 0.27146464646464646, + "grad_norm": 2.8855810165405273, + "learning_rate": 4.222689075630252e-06, + "loss": 1.7968, + "step": 215 + }, + { + "epoch": 0.2727272727272727, + "grad_norm": 3.7866363525390625, + "learning_rate": 4.243697478991597e-06, + "loss": 1.8341, + "step": 216 + }, + { + "epoch": 0.273989898989899, + "grad_norm": 3.0352272987365723, + "learning_rate": 4.264705882352942e-06, + "loss": 1.8276, + "step": 217 + }, + { + "epoch": 0.27525252525252525, + "grad_norm": 3.4880247116088867, + "learning_rate": 4.2857142857142855e-06, + "loss": 1.8194, + "step": 218 + }, + { + "epoch": 0.2765151515151515, + "grad_norm": 3.6807212829589844, + "learning_rate": 4.306722689075631e-06, + "loss": 1.8495, + "step": 219 + }, + { + "epoch": 0.2777777777777778, + "grad_norm": 3.029017448425293, + "learning_rate": 4.327731092436976e-06, + "loss": 1.8125, + "step": 220 + }, + { + "epoch": 0.2777777777777778, + "eval_loss": 1.9633830785751343, + "eval_runtime": 37.1259, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 220 + }, + { + "epoch": 0.27904040404040403, + "grad_norm": 3.4820573329925537, + "learning_rate": 4.3487394957983194e-06, + "loss": 1.8159, + "step": 221 + }, + { + "epoch": 0.2803030303030303, + "grad_norm": 3.2057833671569824, + "learning_rate": 4.369747899159664e-06, + "loss": 1.8215, + "step": 222 + }, + { + "epoch": 0.2815656565656566, + "grad_norm": 3.477302312850952, + "learning_rate": 4.390756302521009e-06, + "loss": 1.7837, + "step": 223 + }, + { + "epoch": 0.2828282828282828, + "grad_norm": 2.8349595069885254, + "learning_rate": 4.411764705882353e-06, + "loss": 1.8555, + "step": 224 + }, + { + "epoch": 0.2840909090909091, + "grad_norm": 3.616478443145752, + "learning_rate": 4.432773109243698e-06, + "loss": 1.8067, + "step": 225 + }, + { + "epoch": 0.28535353535353536, + "grad_norm": 3.111969470977783, + "learning_rate": 4.453781512605043e-06, + "loss": 1.8322, + "step": 226 + }, + { + "epoch": 0.2866161616161616, + "grad_norm": 3.248695135116577, + "learning_rate": 4.474789915966387e-06, + "loss": 1.8004, + "step": 227 + }, + { + "epoch": 0.2878787878787879, + "grad_norm": 3.017082691192627, + "learning_rate": 4.495798319327731e-06, + "loss": 1.776, + "step": 228 + }, + { + "epoch": 0.28914141414141414, + "grad_norm": 3.6790099143981934, + "learning_rate": 4.516806722689076e-06, + "loss": 1.8176, + "step": 229 + }, + { + "epoch": 0.2904040404040404, + "grad_norm": 3.1513001918792725, + "learning_rate": 4.5378151260504205e-06, + "loss": 1.8016, + "step": 230 + }, + { + "epoch": 0.2904040404040404, + "eval_loss": 1.949769377708435, + "eval_runtime": 37.2573, + "eval_samples_per_second": 26.84, + "eval_steps_per_second": 1.127, + "step": 230 + }, + { + "epoch": 0.2916666666666667, + "grad_norm": 3.0559096336364746, + "learning_rate": 4.558823529411765e-06, + "loss": 1.7915, + "step": 231 + }, + { + "epoch": 0.29292929292929293, + "grad_norm": 3.0974984169006348, + "learning_rate": 4.57983193277311e-06, + "loss": 1.8277, + "step": 232 + }, + { + "epoch": 0.29419191919191917, + "grad_norm": 2.8514139652252197, + "learning_rate": 4.600840336134454e-06, + "loss": 1.7616, + "step": 233 + }, + { + "epoch": 0.29545454545454547, + "grad_norm": 3.355198383331299, + "learning_rate": 4.621848739495799e-06, + "loss": 1.841, + "step": 234 + }, + { + "epoch": 0.2967171717171717, + "grad_norm": 3.1467435359954834, + "learning_rate": 4.642857142857144e-06, + "loss": 1.7393, + "step": 235 + }, + { + "epoch": 0.29797979797979796, + "grad_norm": 4.065819263458252, + "learning_rate": 4.663865546218488e-06, + "loss": 1.7508, + "step": 236 + }, + { + "epoch": 0.29924242424242425, + "grad_norm": 3.2602460384368896, + "learning_rate": 4.684873949579832e-06, + "loss": 1.7415, + "step": 237 + }, + { + "epoch": 0.3005050505050505, + "grad_norm": 3.6884894371032715, + "learning_rate": 4.705882352941177e-06, + "loss": 1.7608, + "step": 238 + }, + { + "epoch": 0.30176767676767674, + "grad_norm": 3.9843380451202393, + "learning_rate": 4.7268907563025216e-06, + "loss": 1.8519, + "step": 239 + }, + { + "epoch": 0.30303030303030304, + "grad_norm": 3.2823405265808105, + "learning_rate": 4.747899159663865e-06, + "loss": 1.7689, + "step": 240 + }, + { + "epoch": 0.30303030303030304, + "eval_loss": 1.939724326133728, + "eval_runtime": 37.465, + "eval_samples_per_second": 26.692, + "eval_steps_per_second": 1.121, + "step": 240 + }, + { + "epoch": 0.3042929292929293, + "grad_norm": 3.4200189113616943, + "learning_rate": 4.768907563025211e-06, + "loss": 1.7944, + "step": 241 + }, + { + "epoch": 0.3055555555555556, + "grad_norm": 3.7541189193725586, + "learning_rate": 4.7899159663865555e-06, + "loss": 1.8151, + "step": 242 + }, + { + "epoch": 0.3068181818181818, + "grad_norm": 3.9103832244873047, + "learning_rate": 4.810924369747899e-06, + "loss": 1.7921, + "step": 243 + }, + { + "epoch": 0.30808080808080807, + "grad_norm": 3.2769975662231445, + "learning_rate": 4.831932773109244e-06, + "loss": 1.8408, + "step": 244 + }, + { + "epoch": 0.30934343434343436, + "grad_norm": 3.414168357849121, + "learning_rate": 4.852941176470589e-06, + "loss": 1.7937, + "step": 245 + }, + { + "epoch": 0.3106060606060606, + "grad_norm": 3.219625473022461, + "learning_rate": 4.873949579831933e-06, + "loss": 1.796, + "step": 246 + }, + { + "epoch": 0.31186868686868685, + "grad_norm": 3.2573912143707275, + "learning_rate": 4.894957983193277e-06, + "loss": 1.8052, + "step": 247 + }, + { + "epoch": 0.31313131313131315, + "grad_norm": 3.4201040267944336, + "learning_rate": 4.915966386554622e-06, + "loss": 1.8529, + "step": 248 + }, + { + "epoch": 0.3143939393939394, + "grad_norm": 3.7448058128356934, + "learning_rate": 4.936974789915967e-06, + "loss": 1.8183, + "step": 249 + }, + { + "epoch": 0.31565656565656564, + "grad_norm": 3.4108386039733887, + "learning_rate": 4.957983193277311e-06, + "loss": 1.7926, + "step": 250 + }, + { + "epoch": 0.31565656565656564, + "eval_loss": 1.9319047927856445, + "eval_runtime": 37.4745, + "eval_samples_per_second": 26.685, + "eval_steps_per_second": 1.121, + "step": 250 + }, + { + "epoch": 0.31691919191919193, + "grad_norm": 3.50185227394104, + "learning_rate": 4.978991596638656e-06, + "loss": 1.792, + "step": 251 + }, + { + "epoch": 0.3181818181818182, + "grad_norm": 3.225581407546997, + "learning_rate": 5e-06, + "loss": 1.7286, + "step": 252 + }, + { + "epoch": 0.3194444444444444, + "grad_norm": 3.3070669174194336, + "learning_rate": 4.997661365762395e-06, + "loss": 1.7957, + "step": 253 + }, + { + "epoch": 0.3207070707070707, + "grad_norm": 3.498079776763916, + "learning_rate": 4.99532273152479e-06, + "loss": 1.823, + "step": 254 + }, + { + "epoch": 0.32196969696969696, + "grad_norm": 3.316534996032715, + "learning_rate": 4.9929840972871845e-06, + "loss": 1.8249, + "step": 255 + }, + { + "epoch": 0.32323232323232326, + "grad_norm": 3.1641833782196045, + "learning_rate": 4.990645463049579e-06, + "loss": 1.7658, + "step": 256 + }, + { + "epoch": 0.3244949494949495, + "grad_norm": 3.0794270038604736, + "learning_rate": 4.988306828811974e-06, + "loss": 1.8146, + "step": 257 + }, + { + "epoch": 0.32575757575757575, + "grad_norm": 3.2410383224487305, + "learning_rate": 4.985968194574369e-06, + "loss": 1.7894, + "step": 258 + }, + { + "epoch": 0.32702020202020204, + "grad_norm": 3.0899274349212646, + "learning_rate": 4.983629560336764e-06, + "loss": 1.7381, + "step": 259 + }, + { + "epoch": 0.3282828282828283, + "grad_norm": 3.0132803916931152, + "learning_rate": 4.981290926099159e-06, + "loss": 1.7713, + "step": 260 + }, + { + "epoch": 0.3282828282828283, + "eval_loss": 1.9123589992523193, + "eval_runtime": 37.3201, + "eval_samples_per_second": 26.795, + "eval_steps_per_second": 1.125, + "step": 260 + }, + { + "epoch": 0.32954545454545453, + "grad_norm": 2.92564058303833, + "learning_rate": 4.9789522918615536e-06, + "loss": 1.7659, + "step": 261 + }, + { + "epoch": 0.33080808080808083, + "grad_norm": 3.1402878761291504, + "learning_rate": 4.976613657623948e-06, + "loss": 1.8439, + "step": 262 + }, + { + "epoch": 0.33207070707070707, + "grad_norm": 3.170881509780884, + "learning_rate": 4.974275023386343e-06, + "loss": 1.7605, + "step": 263 + }, + { + "epoch": 0.3333333333333333, + "grad_norm": 3.014850378036499, + "learning_rate": 4.971936389148738e-06, + "loss": 1.7899, + "step": 264 + }, + { + "epoch": 0.3345959595959596, + "grad_norm": 3.169818878173828, + "learning_rate": 4.969597754911132e-06, + "loss": 1.8263, + "step": 265 + }, + { + "epoch": 0.33585858585858586, + "grad_norm": 2.8709678649902344, + "learning_rate": 4.967259120673527e-06, + "loss": 1.7834, + "step": 266 + }, + { + "epoch": 0.3371212121212121, + "grad_norm": 2.7527096271514893, + "learning_rate": 4.964920486435922e-06, + "loss": 1.8106, + "step": 267 + }, + { + "epoch": 0.3383838383838384, + "grad_norm": 2.9960227012634277, + "learning_rate": 4.962581852198316e-06, + "loss": 1.7452, + "step": 268 + }, + { + "epoch": 0.33964646464646464, + "grad_norm": 2.979163885116577, + "learning_rate": 4.9602432179607115e-06, + "loss": 1.6943, + "step": 269 + }, + { + "epoch": 0.3409090909090909, + "grad_norm": 3.2396063804626465, + "learning_rate": 4.957904583723106e-06, + "loss": 1.7798, + "step": 270 + }, + { + "epoch": 0.3409090909090909, + "eval_loss": 1.9126806259155273, + "eval_runtime": 37.123, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 270 + }, + { + "epoch": 0.3421717171717172, + "grad_norm": 2.7868010997772217, + "learning_rate": 4.955565949485501e-06, + "loss": 1.8053, + "step": 271 + }, + { + "epoch": 0.3434343434343434, + "grad_norm": 2.8895349502563477, + "learning_rate": 4.9532273152478956e-06, + "loss": 1.7672, + "step": 272 + }, + { + "epoch": 0.3446969696969697, + "grad_norm": 3.0839710235595703, + "learning_rate": 4.95088868101029e-06, + "loss": 1.7414, + "step": 273 + }, + { + "epoch": 0.34595959595959597, + "grad_norm": 2.81974458694458, + "learning_rate": 4.948550046772685e-06, + "loss": 1.7527, + "step": 274 + }, + { + "epoch": 0.3472222222222222, + "grad_norm": 3.7469775676727295, + "learning_rate": 4.9462114125350805e-06, + "loss": 1.7689, + "step": 275 + }, + { + "epoch": 0.3484848484848485, + "grad_norm": 3.434573173522949, + "learning_rate": 4.943872778297475e-06, + "loss": 1.739, + "step": 276 + }, + { + "epoch": 0.34974747474747475, + "grad_norm": 2.985041379928589, + "learning_rate": 4.941534144059869e-06, + "loss": 1.7894, + "step": 277 + }, + { + "epoch": 0.351010101010101, + "grad_norm": 3.441204309463501, + "learning_rate": 4.939195509822265e-06, + "loss": 1.7804, + "step": 278 + }, + { + "epoch": 0.3522727272727273, + "grad_norm": 2.8371102809906006, + "learning_rate": 4.936856875584659e-06, + "loss": 1.7415, + "step": 279 + }, + { + "epoch": 0.35353535353535354, + "grad_norm": 3.662773847579956, + "learning_rate": 4.9345182413470535e-06, + "loss": 1.7894, + "step": 280 + }, + { + "epoch": 0.35353535353535354, + "eval_loss": 1.915844440460205, + "eval_runtime": 37.0941, + "eval_samples_per_second": 26.958, + "eval_steps_per_second": 1.132, + "step": 280 + }, + { + "epoch": 0.3547979797979798, + "grad_norm": 3.3254036903381348, + "learning_rate": 4.932179607109449e-06, + "loss": 1.7238, + "step": 281 + }, + { + "epoch": 0.3560606060606061, + "grad_norm": 2.9872665405273438, + "learning_rate": 4.929840972871843e-06, + "loss": 1.769, + "step": 282 + }, + { + "epoch": 0.3573232323232323, + "grad_norm": 3.0163521766662598, + "learning_rate": 4.9275023386342376e-06, + "loss": 1.7702, + "step": 283 + }, + { + "epoch": 0.35858585858585856, + "grad_norm": 2.9687387943267822, + "learning_rate": 4.925163704396633e-06, + "loss": 1.7927, + "step": 284 + }, + { + "epoch": 0.35984848484848486, + "grad_norm": 2.7064011096954346, + "learning_rate": 4.922825070159027e-06, + "loss": 1.7638, + "step": 285 + }, + { + "epoch": 0.3611111111111111, + "grad_norm": 3.0043134689331055, + "learning_rate": 4.9204864359214225e-06, + "loss": 1.8041, + "step": 286 + }, + { + "epoch": 0.36237373737373735, + "grad_norm": 3.037100315093994, + "learning_rate": 4.918147801683817e-06, + "loss": 1.7579, + "step": 287 + }, + { + "epoch": 0.36363636363636365, + "grad_norm": 2.7528676986694336, + "learning_rate": 4.915809167446211e-06, + "loss": 1.7629, + "step": 288 + }, + { + "epoch": 0.3648989898989899, + "grad_norm": 3.075815439224243, + "learning_rate": 4.913470533208607e-06, + "loss": 1.7966, + "step": 289 + }, + { + "epoch": 0.3661616161616162, + "grad_norm": 2.964785575866699, + "learning_rate": 4.911131898971001e-06, + "loss": 1.801, + "step": 290 + }, + { + "epoch": 0.3661616161616162, + "eval_loss": 1.8839410543441772, + "eval_runtime": 37.0844, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 290 + }, + { + "epoch": 0.36742424242424243, + "grad_norm": 2.668747663497925, + "learning_rate": 4.908793264733396e-06, + "loss": 1.7703, + "step": 291 + }, + { + "epoch": 0.3686868686868687, + "grad_norm": 2.814751148223877, + "learning_rate": 4.906454630495791e-06, + "loss": 1.7075, + "step": 292 + }, + { + "epoch": 0.369949494949495, + "grad_norm": 3.005922794342041, + "learning_rate": 4.904115996258186e-06, + "loss": 1.7418, + "step": 293 + }, + { + "epoch": 0.3712121212121212, + "grad_norm": 3.050936460494995, + "learning_rate": 4.90177736202058e-06, + "loss": 1.793, + "step": 294 + }, + { + "epoch": 0.37247474747474746, + "grad_norm": 2.822004556655884, + "learning_rate": 4.899438727782975e-06, + "loss": 1.8303, + "step": 295 + }, + { + "epoch": 0.37373737373737376, + "grad_norm": 2.950254201889038, + "learning_rate": 4.89710009354537e-06, + "loss": 1.808, + "step": 296 + }, + { + "epoch": 0.375, + "grad_norm": 2.8140666484832764, + "learning_rate": 4.8947614593077645e-06, + "loss": 1.7262, + "step": 297 + }, + { + "epoch": 0.37626262626262624, + "grad_norm": 2.899620532989502, + "learning_rate": 4.89242282507016e-06, + "loss": 1.7157, + "step": 298 + }, + { + "epoch": 0.37752525252525254, + "grad_norm": 2.805753469467163, + "learning_rate": 4.890084190832554e-06, + "loss": 1.7294, + "step": 299 + }, + { + "epoch": 0.3787878787878788, + "grad_norm": 3.110246181488037, + "learning_rate": 4.887745556594949e-06, + "loss": 1.7415, + "step": 300 + }, + { + "epoch": 0.3787878787878788, + "eval_loss": 1.892285943031311, + "eval_runtime": 37.0732, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 300 + }, + { + "epoch": 0.380050505050505, + "grad_norm": 2.9197490215301514, + "learning_rate": 4.885406922357344e-06, + "loss": 1.7132, + "step": 301 + }, + { + "epoch": 0.3813131313131313, + "grad_norm": 3.3165123462677, + "learning_rate": 4.883068288119738e-06, + "loss": 1.777, + "step": 302 + }, + { + "epoch": 0.38257575757575757, + "grad_norm": 2.8409054279327393, + "learning_rate": 4.880729653882133e-06, + "loss": 1.7092, + "step": 303 + }, + { + "epoch": 0.3838383838383838, + "grad_norm": 2.6800901889801025, + "learning_rate": 4.878391019644528e-06, + "loss": 1.7622, + "step": 304 + }, + { + "epoch": 0.3851010101010101, + "grad_norm": 3.2847490310668945, + "learning_rate": 4.876052385406922e-06, + "loss": 1.729, + "step": 305 + }, + { + "epoch": 0.38636363636363635, + "grad_norm": 2.786057949066162, + "learning_rate": 4.873713751169318e-06, + "loss": 1.7168, + "step": 306 + }, + { + "epoch": 0.38762626262626265, + "grad_norm": 3.236050605773926, + "learning_rate": 4.871375116931712e-06, + "loss": 1.6809, + "step": 307 + }, + { + "epoch": 0.3888888888888889, + "grad_norm": 2.851910352706909, + "learning_rate": 4.869036482694107e-06, + "loss": 1.7336, + "step": 308 + }, + { + "epoch": 0.39015151515151514, + "grad_norm": 2.892421007156372, + "learning_rate": 4.866697848456502e-06, + "loss": 1.7451, + "step": 309 + }, + { + "epoch": 0.39141414141414144, + "grad_norm": 2.8175048828125, + "learning_rate": 4.864359214218897e-06, + "loss": 1.7734, + "step": 310 + }, + { + "epoch": 0.39141414141414144, + "eval_loss": 1.8777843713760376, + "eval_runtime": 36.9321, + "eval_samples_per_second": 27.077, + "eval_steps_per_second": 1.137, + "step": 310 + }, + { + "epoch": 0.3926767676767677, + "grad_norm": 3.005329132080078, + "learning_rate": 4.8620205799812915e-06, + "loss": 1.7476, + "step": 311 + }, + { + "epoch": 0.3939393939393939, + "grad_norm": 2.849017858505249, + "learning_rate": 4.859681945743686e-06, + "loss": 1.7854, + "step": 312 + }, + { + "epoch": 0.3952020202020202, + "grad_norm": 2.801048517227173, + "learning_rate": 4.857343311506081e-06, + "loss": 1.7186, + "step": 313 + }, + { + "epoch": 0.39646464646464646, + "grad_norm": 2.5772831439971924, + "learning_rate": 4.855004677268476e-06, + "loss": 1.7279, + "step": 314 + }, + { + "epoch": 0.3977272727272727, + "grad_norm": 2.7969934940338135, + "learning_rate": 4.85266604303087e-06, + "loss": 1.7475, + "step": 315 + }, + { + "epoch": 0.398989898989899, + "grad_norm": 2.7067770957946777, + "learning_rate": 4.850327408793265e-06, + "loss": 1.7055, + "step": 316 + }, + { + "epoch": 0.40025252525252525, + "grad_norm": 2.70928955078125, + "learning_rate": 4.84798877455566e-06, + "loss": 1.7203, + "step": 317 + }, + { + "epoch": 0.4015151515151515, + "grad_norm": 2.7704474925994873, + "learning_rate": 4.845650140318054e-06, + "loss": 1.8058, + "step": 318 + }, + { + "epoch": 0.4027777777777778, + "grad_norm": 2.7680182456970215, + "learning_rate": 4.843311506080449e-06, + "loss": 1.7162, + "step": 319 + }, + { + "epoch": 0.40404040404040403, + "grad_norm": 3.0953330993652344, + "learning_rate": 4.840972871842844e-06, + "loss": 1.7617, + "step": 320 + }, + { + "epoch": 0.40404040404040403, + "eval_loss": 1.8680965900421143, + "eval_runtime": 37.0382, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 320 + }, + { + "epoch": 0.4053030303030303, + "grad_norm": 2.8340392112731934, + "learning_rate": 4.838634237605239e-06, + "loss": 1.7767, + "step": 321 + }, + { + "epoch": 0.4065656565656566, + "grad_norm": 3.057861804962158, + "learning_rate": 4.836295603367634e-06, + "loss": 1.8043, + "step": 322 + }, + { + "epoch": 0.4078282828282828, + "grad_norm": 3.0123419761657715, + "learning_rate": 4.833956969130029e-06, + "loss": 1.7561, + "step": 323 + }, + { + "epoch": 0.4090909090909091, + "grad_norm": 3.0873124599456787, + "learning_rate": 4.831618334892423e-06, + "loss": 1.7581, + "step": 324 + }, + { + "epoch": 0.41035353535353536, + "grad_norm": 2.6600539684295654, + "learning_rate": 4.8292797006548184e-06, + "loss": 1.7005, + "step": 325 + }, + { + "epoch": 0.4116161616161616, + "grad_norm": 2.905280590057373, + "learning_rate": 4.826941066417213e-06, + "loss": 1.7592, + "step": 326 + }, + { + "epoch": 0.4128787878787879, + "grad_norm": 3.0642189979553223, + "learning_rate": 4.824602432179607e-06, + "loss": 1.6456, + "step": 327 + }, + { + "epoch": 0.41414141414141414, + "grad_norm": 2.606531858444214, + "learning_rate": 4.8222637979420025e-06, + "loss": 1.6731, + "step": 328 + }, + { + "epoch": 0.4154040404040404, + "grad_norm": 2.9918158054351807, + "learning_rate": 4.819925163704397e-06, + "loss": 1.6854, + "step": 329 + }, + { + "epoch": 0.4166666666666667, + "grad_norm": 2.721940040588379, + "learning_rate": 4.817586529466791e-06, + "loss": 1.7346, + "step": 330 + }, + { + "epoch": 0.4166666666666667, + "eval_loss": 1.863736629486084, + "eval_runtime": 37.0887, + "eval_samples_per_second": 26.962, + "eval_steps_per_second": 1.132, + "step": 330 + }, + { + "epoch": 0.41792929292929293, + "grad_norm": 3.1076390743255615, + "learning_rate": 4.815247895229187e-06, + "loss": 1.6286, + "step": 331 + }, + { + "epoch": 0.41919191919191917, + "grad_norm": 3.1235783100128174, + "learning_rate": 4.812909260991581e-06, + "loss": 1.7404, + "step": 332 + }, + { + "epoch": 0.42045454545454547, + "grad_norm": 3.1211330890655518, + "learning_rate": 4.8105706267539755e-06, + "loss": 1.7143, + "step": 333 + }, + { + "epoch": 0.4217171717171717, + "grad_norm": 3.072317600250244, + "learning_rate": 4.808231992516371e-06, + "loss": 1.7514, + "step": 334 + }, + { + "epoch": 0.42297979797979796, + "grad_norm": 2.880671739578247, + "learning_rate": 4.805893358278765e-06, + "loss": 1.7018, + "step": 335 + }, + { + "epoch": 0.42424242424242425, + "grad_norm": 2.700063943862915, + "learning_rate": 4.8035547240411604e-06, + "loss": 1.7089, + "step": 336 + }, + { + "epoch": 0.4255050505050505, + "grad_norm": 2.9038445949554443, + "learning_rate": 4.801216089803556e-06, + "loss": 1.7483, + "step": 337 + }, + { + "epoch": 0.42676767676767674, + "grad_norm": 2.7822766304016113, + "learning_rate": 4.79887745556595e-06, + "loss": 1.6927, + "step": 338 + }, + { + "epoch": 0.42803030303030304, + "grad_norm": 2.957554340362549, + "learning_rate": 4.7965388213283445e-06, + "loss": 1.7162, + "step": 339 + }, + { + "epoch": 0.4292929292929293, + "grad_norm": 3.1333892345428467, + "learning_rate": 4.79420018709074e-06, + "loss": 1.731, + "step": 340 + }, + { + "epoch": 0.4292929292929293, + "eval_loss": 1.852413535118103, + "eval_runtime": 37.1809, + "eval_samples_per_second": 26.896, + "eval_steps_per_second": 1.13, + "step": 340 + }, + { + "epoch": 0.4305555555555556, + "grad_norm": 2.9801084995269775, + "learning_rate": 4.791861552853134e-06, + "loss": 1.6892, + "step": 341 + }, + { + "epoch": 0.4318181818181818, + "grad_norm": 2.5077271461486816, + "learning_rate": 4.789522918615529e-06, + "loss": 1.637, + "step": 342 + }, + { + "epoch": 0.43308080808080807, + "grad_norm": 2.8654391765594482, + "learning_rate": 4.787184284377924e-06, + "loss": 1.6915, + "step": 343 + }, + { + "epoch": 0.43434343434343436, + "grad_norm": 2.6204264163970947, + "learning_rate": 4.784845650140318e-06, + "loss": 1.6722, + "step": 344 + }, + { + "epoch": 0.4356060606060606, + "grad_norm": 2.633991003036499, + "learning_rate": 4.782507015902713e-06, + "loss": 1.6797, + "step": 345 + }, + { + "epoch": 0.43686868686868685, + "grad_norm": 2.7203924655914307, + "learning_rate": 4.780168381665108e-06, + "loss": 1.7018, + "step": 346 + }, + { + "epoch": 0.43813131313131315, + "grad_norm": 2.6789209842681885, + "learning_rate": 4.7778297474275024e-06, + "loss": 1.7153, + "step": 347 + }, + { + "epoch": 0.4393939393939394, + "grad_norm": 2.6285433769226074, + "learning_rate": 4.775491113189898e-06, + "loss": 1.7082, + "step": 348 + }, + { + "epoch": 0.44065656565656564, + "grad_norm": 2.623793840408325, + "learning_rate": 4.773152478952292e-06, + "loss": 1.7094, + "step": 349 + }, + { + "epoch": 0.44191919191919193, + "grad_norm": 2.647160291671753, + "learning_rate": 4.7708138447146865e-06, + "loss": 1.6528, + "step": 350 + }, + { + "epoch": 0.44191919191919193, + "eval_loss": 1.8506718873977661, + "eval_runtime": 37.4232, + "eval_samples_per_second": 26.721, + "eval_steps_per_second": 1.122, + "step": 350 + }, + { + "epoch": 0.4431818181818182, + "grad_norm": 2.664553642272949, + "learning_rate": 4.768475210477082e-06, + "loss": 1.7008, + "step": 351 + }, + { + "epoch": 0.4444444444444444, + "grad_norm": 2.8486251831054688, + "learning_rate": 4.766136576239477e-06, + "loss": 1.6811, + "step": 352 + }, + { + "epoch": 0.4457070707070707, + "grad_norm": 2.7514266967773438, + "learning_rate": 4.7637979420018715e-06, + "loss": 1.6969, + "step": 353 + }, + { + "epoch": 0.44696969696969696, + "grad_norm": 2.848083257675171, + "learning_rate": 4.761459307764266e-06, + "loss": 1.7058, + "step": 354 + }, + { + "epoch": 0.44823232323232326, + "grad_norm": 2.759218454360962, + "learning_rate": 4.759120673526661e-06, + "loss": 1.6109, + "step": 355 + }, + { + "epoch": 0.4494949494949495, + "grad_norm": 2.905318021774292, + "learning_rate": 4.756782039289056e-06, + "loss": 1.7689, + "step": 356 + }, + { + "epoch": 0.45075757575757575, + "grad_norm": 2.845646381378174, + "learning_rate": 4.75444340505145e-06, + "loss": 1.6687, + "step": 357 + }, + { + "epoch": 0.45202020202020204, + "grad_norm": 2.8609306812286377, + "learning_rate": 4.752104770813845e-06, + "loss": 1.6332, + "step": 358 + }, + { + "epoch": 0.4532828282828283, + "grad_norm": 3.016479969024658, + "learning_rate": 4.74976613657624e-06, + "loss": 1.6692, + "step": 359 + }, + { + "epoch": 0.45454545454545453, + "grad_norm": 3.2790887355804443, + "learning_rate": 4.747427502338635e-06, + "loss": 1.704, + "step": 360 + }, + { + "epoch": 0.45454545454545453, + "eval_loss": 1.848663330078125, + "eval_runtime": 37.4817, + "eval_samples_per_second": 26.68, + "eval_steps_per_second": 1.121, + "step": 360 + }, + { + "epoch": 0.45580808080808083, + "grad_norm": 3.1137545108795166, + "learning_rate": 4.745088868101029e-06, + "loss": 1.6797, + "step": 361 + }, + { + "epoch": 0.45707070707070707, + "grad_norm": 2.7598798274993896, + "learning_rate": 4.742750233863424e-06, + "loss": 1.6477, + "step": 362 + }, + { + "epoch": 0.4583333333333333, + "grad_norm": 3.4851579666137695, + "learning_rate": 4.740411599625819e-06, + "loss": 1.6481, + "step": 363 + }, + { + "epoch": 0.4595959595959596, + "grad_norm": 2.9285919666290283, + "learning_rate": 4.7380729653882135e-06, + "loss": 1.7094, + "step": 364 + }, + { + "epoch": 0.46085858585858586, + "grad_norm": 2.901458501815796, + "learning_rate": 4.735734331150608e-06, + "loss": 1.6906, + "step": 365 + }, + { + "epoch": 0.4621212121212121, + "grad_norm": 2.8737902641296387, + "learning_rate": 4.733395696913003e-06, + "loss": 1.6819, + "step": 366 + }, + { + "epoch": 0.4633838383838384, + "grad_norm": 3.2480952739715576, + "learning_rate": 4.7310570626753984e-06, + "loss": 1.6827, + "step": 367 + }, + { + "epoch": 0.46464646464646464, + "grad_norm": 2.8714661598205566, + "learning_rate": 4.728718428437793e-06, + "loss": 1.7349, + "step": 368 + }, + { + "epoch": 0.4659090909090909, + "grad_norm": 2.8410098552703857, + "learning_rate": 4.726379794200187e-06, + "loss": 1.7214, + "step": 369 + }, + { + "epoch": 0.4671717171717172, + "grad_norm": 2.744807720184326, + "learning_rate": 4.7240411599625826e-06, + "loss": 1.6742, + "step": 370 + }, + { + "epoch": 0.4671717171717172, + "eval_loss": 1.842191457748413, + "eval_runtime": 37.454, + "eval_samples_per_second": 26.699, + "eval_steps_per_second": 1.121, + "step": 370 + }, + { + "epoch": 0.4684343434343434, + "grad_norm": 2.8135552406311035, + "learning_rate": 4.721702525724977e-06, + "loss": 1.6957, + "step": 371 + }, + { + "epoch": 0.4696969696969697, + "grad_norm": 2.856764078140259, + "learning_rate": 4.719363891487372e-06, + "loss": 1.6802, + "step": 372 + }, + { + "epoch": 0.47095959595959597, + "grad_norm": 3.0707688331604004, + "learning_rate": 4.717025257249767e-06, + "loss": 1.7383, + "step": 373 + }, + { + "epoch": 0.4722222222222222, + "grad_norm": 2.8832876682281494, + "learning_rate": 4.714686623012161e-06, + "loss": 1.6309, + "step": 374 + }, + { + "epoch": 0.4734848484848485, + "grad_norm": 2.6840622425079346, + "learning_rate": 4.712347988774556e-06, + "loss": 1.6624, + "step": 375 + }, + { + "epoch": 0.47474747474747475, + "grad_norm": 3.2920548915863037, + "learning_rate": 4.710009354536951e-06, + "loss": 1.6887, + "step": 376 + }, + { + "epoch": 0.476010101010101, + "grad_norm": 3.0317189693450928, + "learning_rate": 4.707670720299345e-06, + "loss": 1.6514, + "step": 377 + }, + { + "epoch": 0.4772727272727273, + "grad_norm": 2.7284793853759766, + "learning_rate": 4.7053320860617404e-06, + "loss": 1.6595, + "step": 378 + }, + { + "epoch": 0.47853535353535354, + "grad_norm": 2.86665415763855, + "learning_rate": 4.702993451824135e-06, + "loss": 1.7219, + "step": 379 + }, + { + "epoch": 0.4797979797979798, + "grad_norm": 3.041191577911377, + "learning_rate": 4.700654817586529e-06, + "loss": 1.6959, + "step": 380 + }, + { + "epoch": 0.4797979797979798, + "eval_loss": 1.8414148092269897, + "eval_runtime": 37.2537, + "eval_samples_per_second": 26.843, + "eval_steps_per_second": 1.127, + "step": 380 + }, + { + "epoch": 0.4810606060606061, + "grad_norm": 3.1729421615600586, + "learning_rate": 4.6983161833489246e-06, + "loss": 1.6538, + "step": 381 + }, + { + "epoch": 0.4823232323232323, + "grad_norm": 2.8562991619110107, + "learning_rate": 4.69597754911132e-06, + "loss": 1.6667, + "step": 382 + }, + { + "epoch": 0.48358585858585856, + "grad_norm": 3.0348711013793945, + "learning_rate": 4.693638914873714e-06, + "loss": 1.6888, + "step": 383 + }, + { + "epoch": 0.48484848484848486, + "grad_norm": 2.9156477451324463, + "learning_rate": 4.6913002806361095e-06, + "loss": 1.628, + "step": 384 + }, + { + "epoch": 0.4861111111111111, + "grad_norm": 2.832689046859741, + "learning_rate": 4.688961646398504e-06, + "loss": 1.664, + "step": 385 + }, + { + "epoch": 0.48737373737373735, + "grad_norm": 2.9858946800231934, + "learning_rate": 4.686623012160898e-06, + "loss": 1.6696, + "step": 386 + }, + { + "epoch": 0.48863636363636365, + "grad_norm": 2.732468843460083, + "learning_rate": 4.684284377923294e-06, + "loss": 1.6444, + "step": 387 + }, + { + "epoch": 0.4898989898989899, + "grad_norm": 2.7801132202148438, + "learning_rate": 4.681945743685688e-06, + "loss": 1.7084, + "step": 388 + }, + { + "epoch": 0.4911616161616162, + "grad_norm": 2.646303415298462, + "learning_rate": 4.6796071094480824e-06, + "loss": 1.6963, + "step": 389 + }, + { + "epoch": 0.49242424242424243, + "grad_norm": 2.8750250339508057, + "learning_rate": 4.677268475210478e-06, + "loss": 1.6814, + "step": 390 + }, + { + "epoch": 0.49242424242424243, + "eval_loss": 1.8283894062042236, + "eval_runtime": 37.2058, + "eval_samples_per_second": 26.877, + "eval_steps_per_second": 1.129, + "step": 390 + }, + { + "epoch": 0.4936868686868687, + "grad_norm": 2.5373270511627197, + "learning_rate": 4.674929840972872e-06, + "loss": 1.5956, + "step": 391 + }, + { + "epoch": 0.494949494949495, + "grad_norm": 2.837170124053955, + "learning_rate": 4.6725912067352666e-06, + "loss": 1.6237, + "step": 392 + }, + { + "epoch": 0.4962121212121212, + "grad_norm": 3.199463367462158, + "learning_rate": 4.670252572497662e-06, + "loss": 1.7087, + "step": 393 + }, + { + "epoch": 0.49747474747474746, + "grad_norm": 2.9850947856903076, + "learning_rate": 4.667913938260056e-06, + "loss": 1.5913, + "step": 394 + }, + { + "epoch": 0.49873737373737376, + "grad_norm": 2.954148292541504, + "learning_rate": 4.665575304022451e-06, + "loss": 1.6272, + "step": 395 + }, + { + "epoch": 0.5, + "grad_norm": 2.8625757694244385, + "learning_rate": 4.663236669784846e-06, + "loss": 1.6623, + "step": 396 + }, + { + "epoch": 0.5012626262626263, + "grad_norm": 3.196354627609253, + "learning_rate": 4.66089803554724e-06, + "loss": 1.7249, + "step": 397 + }, + { + "epoch": 0.5025252525252525, + "grad_norm": 2.7257323265075684, + "learning_rate": 4.658559401309636e-06, + "loss": 1.665, + "step": 398 + }, + { + "epoch": 0.5037878787878788, + "grad_norm": 3.0976016521453857, + "learning_rate": 4.656220767072031e-06, + "loss": 1.6763, + "step": 399 + }, + { + "epoch": 0.5050505050505051, + "grad_norm": 3.0398128032684326, + "learning_rate": 4.653882132834425e-06, + "loss": 1.6375, + "step": 400 + }, + { + "epoch": 0.5050505050505051, + "eval_loss": 1.8324880599975586, + "eval_runtime": 37.1754, + "eval_samples_per_second": 26.9, + "eval_steps_per_second": 1.13, + "step": 400 + }, + { + "epoch": 0.5063131313131313, + "grad_norm": 3.314958095550537, + "learning_rate": 4.65154349859682e-06, + "loss": 1.6699, + "step": 401 + }, + { + "epoch": 0.5075757575757576, + "grad_norm": 2.8478870391845703, + "learning_rate": 4.649204864359215e-06, + "loss": 1.6495, + "step": 402 + }, + { + "epoch": 0.5088383838383839, + "grad_norm": 2.8476510047912598, + "learning_rate": 4.646866230121609e-06, + "loss": 1.633, + "step": 403 + }, + { + "epoch": 0.51010101010101, + "grad_norm": 2.916667938232422, + "learning_rate": 4.644527595884004e-06, + "loss": 1.6092, + "step": 404 + }, + { + "epoch": 0.5113636363636364, + "grad_norm": 2.977226495742798, + "learning_rate": 4.642188961646399e-06, + "loss": 1.6588, + "step": 405 + }, + { + "epoch": 0.5126262626262627, + "grad_norm": 3.0540225505828857, + "learning_rate": 4.6398503274087935e-06, + "loss": 1.6529, + "step": 406 + }, + { + "epoch": 0.5138888888888888, + "grad_norm": 2.902035713195801, + "learning_rate": 4.637511693171188e-06, + "loss": 1.5983, + "step": 407 + }, + { + "epoch": 0.5151515151515151, + "grad_norm": 2.9009487628936768, + "learning_rate": 4.635173058933583e-06, + "loss": 1.6389, + "step": 408 + }, + { + "epoch": 0.5164141414141414, + "grad_norm": 2.904982805252075, + "learning_rate": 4.632834424695978e-06, + "loss": 1.6971, + "step": 409 + }, + { + "epoch": 0.5176767676767676, + "grad_norm": 2.8340721130371094, + "learning_rate": 4.630495790458373e-06, + "loss": 1.6336, + "step": 410 + }, + { + "epoch": 0.5176767676767676, + "eval_loss": 1.8218064308166504, + "eval_runtime": 36.9963, + "eval_samples_per_second": 27.03, + "eval_steps_per_second": 1.135, + "step": 410 + }, + { + "epoch": 0.5189393939393939, + "grad_norm": 2.9507815837860107, + "learning_rate": 4.628157156220767e-06, + "loss": 1.6795, + "step": 411 + }, + { + "epoch": 0.5202020202020202, + "grad_norm": 2.868793249130249, + "learning_rate": 4.625818521983162e-06, + "loss": 1.6494, + "step": 412 + }, + { + "epoch": 0.5214646464646465, + "grad_norm": 2.786395788192749, + "learning_rate": 4.623479887745557e-06, + "loss": 1.7156, + "step": 413 + }, + { + "epoch": 0.5227272727272727, + "grad_norm": 2.7456934452056885, + "learning_rate": 4.621141253507952e-06, + "loss": 1.6528, + "step": 414 + }, + { + "epoch": 0.523989898989899, + "grad_norm": 3.1211695671081543, + "learning_rate": 4.618802619270347e-06, + "loss": 1.6891, + "step": 415 + }, + { + "epoch": 0.5252525252525253, + "grad_norm": 3.2319254875183105, + "learning_rate": 4.616463985032741e-06, + "loss": 1.6452, + "step": 416 + }, + { + "epoch": 0.5265151515151515, + "grad_norm": 2.6718053817749023, + "learning_rate": 4.614125350795136e-06, + "loss": 1.6442, + "step": 417 + }, + { + "epoch": 0.5277777777777778, + "grad_norm": 2.985020637512207, + "learning_rate": 4.611786716557531e-06, + "loss": 1.6935, + "step": 418 + }, + { + "epoch": 0.5290404040404041, + "grad_norm": 2.798720598220825, + "learning_rate": 4.609448082319925e-06, + "loss": 1.6303, + "step": 419 + }, + { + "epoch": 0.5303030303030303, + "grad_norm": 2.860403299331665, + "learning_rate": 4.6071094480823205e-06, + "loss": 1.6445, + "step": 420 + }, + { + "epoch": 0.5303030303030303, + "eval_loss": 1.8173401355743408, + "eval_runtime": 37.1334, + "eval_samples_per_second": 26.93, + "eval_steps_per_second": 1.131, + "step": 420 + }, + { + "epoch": 0.5315656565656566, + "grad_norm": 2.883209705352783, + "learning_rate": 4.604770813844715e-06, + "loss": 1.6182, + "step": 421 + }, + { + "epoch": 0.5328282828282829, + "grad_norm": 3.0330183506011963, + "learning_rate": 4.60243217960711e-06, + "loss": 1.6208, + "step": 422 + }, + { + "epoch": 0.5340909090909091, + "grad_norm": 3.224184274673462, + "learning_rate": 4.6000935453695046e-06, + "loss": 1.6687, + "step": 423 + }, + { + "epoch": 0.5353535353535354, + "grad_norm": 3.0345349311828613, + "learning_rate": 4.597754911131899e-06, + "loss": 1.6654, + "step": 424 + }, + { + "epoch": 0.5366161616161617, + "grad_norm": 3.1425750255584717, + "learning_rate": 4.595416276894294e-06, + "loss": 1.6569, + "step": 425 + }, + { + "epoch": 0.5378787878787878, + "grad_norm": 2.6387979984283447, + "learning_rate": 4.593077642656689e-06, + "loss": 1.6868, + "step": 426 + }, + { + "epoch": 0.5391414141414141, + "grad_norm": 2.959446430206299, + "learning_rate": 4.590739008419083e-06, + "loss": 1.621, + "step": 427 + }, + { + "epoch": 0.5404040404040404, + "grad_norm": 3.009673833847046, + "learning_rate": 4.588400374181478e-06, + "loss": 1.6414, + "step": 428 + }, + { + "epoch": 0.5416666666666666, + "grad_norm": 2.827531576156616, + "learning_rate": 4.586061739943874e-06, + "loss": 1.6561, + "step": 429 + }, + { + "epoch": 0.5429292929292929, + "grad_norm": 2.8770735263824463, + "learning_rate": 4.583723105706268e-06, + "loss": 1.6685, + "step": 430 + }, + { + "epoch": 0.5429292929292929, + "eval_loss": 1.8234716653823853, + "eval_runtime": 37.1647, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 430 + }, + { + "epoch": 0.5441919191919192, + "grad_norm": 3.1215782165527344, + "learning_rate": 4.5813844714686625e-06, + "loss": 1.6548, + "step": 431 + }, + { + "epoch": 0.5454545454545454, + "grad_norm": 2.6587517261505127, + "learning_rate": 4.579045837231058e-06, + "loss": 1.649, + "step": 432 + }, + { + "epoch": 0.5467171717171717, + "grad_norm": 2.7059030532836914, + "learning_rate": 4.576707202993452e-06, + "loss": 1.6293, + "step": 433 + }, + { + "epoch": 0.547979797979798, + "grad_norm": 2.7390925884246826, + "learning_rate": 4.574368568755847e-06, + "loss": 1.6957, + "step": 434 + }, + { + "epoch": 0.5492424242424242, + "grad_norm": 2.9457757472991943, + "learning_rate": 4.572029934518242e-06, + "loss": 1.6714, + "step": 435 + }, + { + "epoch": 0.5505050505050505, + "grad_norm": 2.6871347427368164, + "learning_rate": 4.569691300280636e-06, + "loss": 1.6648, + "step": 436 + }, + { + "epoch": 0.5517676767676768, + "grad_norm": 3.1708056926727295, + "learning_rate": 4.5673526660430315e-06, + "loss": 1.6415, + "step": 437 + }, + { + "epoch": 0.553030303030303, + "grad_norm": 2.8767378330230713, + "learning_rate": 4.565014031805426e-06, + "loss": 1.6475, + "step": 438 + }, + { + "epoch": 0.5542929292929293, + "grad_norm": 3.083566188812256, + "learning_rate": 4.56267539756782e-06, + "loss": 1.632, + "step": 439 + }, + { + "epoch": 0.5555555555555556, + "grad_norm": 2.584460735321045, + "learning_rate": 4.560336763330216e-06, + "loss": 1.6469, + "step": 440 + }, + { + "epoch": 0.5555555555555556, + "eval_loss": 1.8140095472335815, + "eval_runtime": 37.1986, + "eval_samples_per_second": 26.883, + "eval_steps_per_second": 1.129, + "step": 440 + }, + { + "epoch": 0.5568181818181818, + "grad_norm": 2.5885088443756104, + "learning_rate": 4.55799812909261e-06, + "loss": 1.6058, + "step": 441 + }, + { + "epoch": 0.5580808080808081, + "grad_norm": 2.8467836380004883, + "learning_rate": 4.5556594948550045e-06, + "loss": 1.6959, + "step": 442 + }, + { + "epoch": 0.5593434343434344, + "grad_norm": 2.764326333999634, + "learning_rate": 4.5533208606174e-06, + "loss": 1.6565, + "step": 443 + }, + { + "epoch": 0.5606060606060606, + "grad_norm": 2.88089919090271, + "learning_rate": 4.550982226379795e-06, + "loss": 1.6651, + "step": 444 + }, + { + "epoch": 0.5618686868686869, + "grad_norm": 2.6239469051361084, + "learning_rate": 4.548643592142189e-06, + "loss": 1.6255, + "step": 445 + }, + { + "epoch": 0.5631313131313131, + "grad_norm": 2.7748732566833496, + "learning_rate": 4.546304957904585e-06, + "loss": 1.6587, + "step": 446 + }, + { + "epoch": 0.5643939393939394, + "grad_norm": 2.7075202465057373, + "learning_rate": 4.543966323666979e-06, + "loss": 1.6448, + "step": 447 + }, + { + "epoch": 0.5656565656565656, + "grad_norm": 2.9291634559631348, + "learning_rate": 4.5416276894293735e-06, + "loss": 1.6564, + "step": 448 + }, + { + "epoch": 0.5669191919191919, + "grad_norm": 3.0039470195770264, + "learning_rate": 4.539289055191769e-06, + "loss": 1.623, + "step": 449 + }, + { + "epoch": 0.5681818181818182, + "grad_norm": 3.085277557373047, + "learning_rate": 4.536950420954163e-06, + "loss": 1.6176, + "step": 450 + }, + { + "epoch": 0.5681818181818182, + "eval_loss": 1.8022958040237427, + "eval_runtime": 37.2204, + "eval_samples_per_second": 26.867, + "eval_steps_per_second": 1.128, + "step": 450 + }, + { + "epoch": 0.5694444444444444, + "grad_norm": 2.93575119972229, + "learning_rate": 4.534611786716558e-06, + "loss": 1.615, + "step": 451 + }, + { + "epoch": 0.5707070707070707, + "grad_norm": 2.847141742706299, + "learning_rate": 4.532273152478953e-06, + "loss": 1.6474, + "step": 452 + }, + { + "epoch": 0.571969696969697, + "grad_norm": 3.0299603939056396, + "learning_rate": 4.529934518241347e-06, + "loss": 1.6424, + "step": 453 + }, + { + "epoch": 0.5732323232323232, + "grad_norm": 3.0307669639587402, + "learning_rate": 4.527595884003742e-06, + "loss": 1.6473, + "step": 454 + }, + { + "epoch": 0.5744949494949495, + "grad_norm": 2.864269971847534, + "learning_rate": 4.525257249766137e-06, + "loss": 1.6631, + "step": 455 + }, + { + "epoch": 0.5757575757575758, + "grad_norm": 2.606637716293335, + "learning_rate": 4.522918615528531e-06, + "loss": 1.6118, + "step": 456 + }, + { + "epoch": 0.577020202020202, + "grad_norm": 3.127891778945923, + "learning_rate": 4.520579981290926e-06, + "loss": 1.6286, + "step": 457 + }, + { + "epoch": 0.5782828282828283, + "grad_norm": 2.8809893131256104, + "learning_rate": 4.518241347053321e-06, + "loss": 1.6158, + "step": 458 + }, + { + "epoch": 0.5795454545454546, + "grad_norm": 3.0662436485290527, + "learning_rate": 4.515902712815716e-06, + "loss": 1.7336, + "step": 459 + }, + { + "epoch": 0.5808080808080808, + "grad_norm": 3.0795772075653076, + "learning_rate": 4.513564078578111e-06, + "loss": 1.6974, + "step": 460 + }, + { + "epoch": 0.5808080808080808, + "eval_loss": 1.7981404066085815, + "eval_runtime": 37.299, + "eval_samples_per_second": 26.81, + "eval_steps_per_second": 1.126, + "step": 460 + }, + { + "epoch": 0.5820707070707071, + "grad_norm": 3.091688871383667, + "learning_rate": 4.511225444340506e-06, + "loss": 1.6803, + "step": 461 + }, + { + "epoch": 0.5833333333333334, + "grad_norm": 2.5597071647644043, + "learning_rate": 4.5088868101029005e-06, + "loss": 1.6104, + "step": 462 + }, + { + "epoch": 0.5845959595959596, + "grad_norm": 2.849806070327759, + "learning_rate": 4.506548175865295e-06, + "loss": 1.6147, + "step": 463 + }, + { + "epoch": 0.5858585858585859, + "grad_norm": 2.9791204929351807, + "learning_rate": 4.50420954162769e-06, + "loss": 1.6376, + "step": 464 + }, + { + "epoch": 0.5871212121212122, + "grad_norm": 2.8582305908203125, + "learning_rate": 4.501870907390085e-06, + "loss": 1.6831, + "step": 465 + }, + { + "epoch": 0.5883838383838383, + "grad_norm": 2.8445184230804443, + "learning_rate": 4.499532273152479e-06, + "loss": 1.6463, + "step": 466 + }, + { + "epoch": 0.5896464646464646, + "grad_norm": 2.858168125152588, + "learning_rate": 4.497193638914874e-06, + "loss": 1.6662, + "step": 467 + }, + { + "epoch": 0.5909090909090909, + "grad_norm": 2.829702854156494, + "learning_rate": 4.494855004677269e-06, + "loss": 1.6058, + "step": 468 + }, + { + "epoch": 0.5921717171717171, + "grad_norm": 2.9292726516723633, + "learning_rate": 4.492516370439663e-06, + "loss": 1.6341, + "step": 469 + }, + { + "epoch": 0.5934343434343434, + "grad_norm": 2.9006292819976807, + "learning_rate": 4.490177736202058e-06, + "loss": 1.5726, + "step": 470 + }, + { + "epoch": 0.5934343434343434, + "eval_loss": 1.7953916788101196, + "eval_runtime": 37.5058, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 470 + }, + { + "epoch": 0.5946969696969697, + "grad_norm": 2.8116230964660645, + "learning_rate": 4.487839101964453e-06, + "loss": 1.6083, + "step": 471 + }, + { + "epoch": 0.5959595959595959, + "grad_norm": 2.540607452392578, + "learning_rate": 4.485500467726848e-06, + "loss": 1.6208, + "step": 472 + }, + { + "epoch": 0.5972222222222222, + "grad_norm": 2.7339096069335938, + "learning_rate": 4.4831618334892425e-06, + "loss": 1.6839, + "step": 473 + }, + { + "epoch": 0.5984848484848485, + "grad_norm": 2.9094762802124023, + "learning_rate": 4.480823199251638e-06, + "loss": 1.6962, + "step": 474 + }, + { + "epoch": 0.5997474747474747, + "grad_norm": 2.7733027935028076, + "learning_rate": 4.478484565014032e-06, + "loss": 1.6252, + "step": 475 + }, + { + "epoch": 0.601010101010101, + "grad_norm": 2.5709428787231445, + "learning_rate": 4.4761459307764274e-06, + "loss": 1.5896, + "step": 476 + }, + { + "epoch": 0.6022727272727273, + "grad_norm": 2.8358278274536133, + "learning_rate": 4.473807296538822e-06, + "loss": 1.6247, + "step": 477 + }, + { + "epoch": 0.6035353535353535, + "grad_norm": 2.7504897117614746, + "learning_rate": 4.471468662301216e-06, + "loss": 1.7138, + "step": 478 + }, + { + "epoch": 0.6047979797979798, + "grad_norm": 2.6965620517730713, + "learning_rate": 4.4691300280636115e-06, + "loss": 1.6054, + "step": 479 + }, + { + "epoch": 0.6060606060606061, + "grad_norm": 2.9401891231536865, + "learning_rate": 4.466791393826006e-06, + "loss": 1.6172, + "step": 480 + }, + { + "epoch": 0.6060606060606061, + "eval_loss": 1.7946635484695435, + "eval_runtime": 37.5366, + "eval_samples_per_second": 26.641, + "eval_steps_per_second": 1.119, + "step": 480 + }, + { + "epoch": 0.6073232323232324, + "grad_norm": 2.71535587310791, + "learning_rate": 4.4644527595884e-06, + "loss": 1.61, + "step": 481 + }, + { + "epoch": 0.6085858585858586, + "grad_norm": 2.7910654544830322, + "learning_rate": 4.462114125350796e-06, + "loss": 1.6193, + "step": 482 + }, + { + "epoch": 0.6098484848484849, + "grad_norm": 2.809241533279419, + "learning_rate": 4.45977549111319e-06, + "loss": 1.602, + "step": 483 + }, + { + "epoch": 0.6111111111111112, + "grad_norm": 2.7152137756347656, + "learning_rate": 4.457436856875585e-06, + "loss": 1.6516, + "step": 484 + }, + { + "epoch": 0.6123737373737373, + "grad_norm": 2.7653465270996094, + "learning_rate": 4.45509822263798e-06, + "loss": 1.5904, + "step": 485 + }, + { + "epoch": 0.6136363636363636, + "grad_norm": 2.729876756668091, + "learning_rate": 4.452759588400374e-06, + "loss": 1.6259, + "step": 486 + }, + { + "epoch": 0.61489898989899, + "grad_norm": 2.6290180683135986, + "learning_rate": 4.4504209541627694e-06, + "loss": 1.5796, + "step": 487 + }, + { + "epoch": 0.6161616161616161, + "grad_norm": 2.939884662628174, + "learning_rate": 4.448082319925164e-06, + "loss": 1.6774, + "step": 488 + }, + { + "epoch": 0.6174242424242424, + "grad_norm": 2.7554702758789062, + "learning_rate": 4.445743685687559e-06, + "loss": 1.6226, + "step": 489 + }, + { + "epoch": 0.6186868686868687, + "grad_norm": 2.890458106994629, + "learning_rate": 4.4434050514499535e-06, + "loss": 1.592, + "step": 490 + }, + { + "epoch": 0.6186868686868687, + "eval_loss": 1.7868437767028809, + "eval_runtime": 37.5267, + "eval_samples_per_second": 26.648, + "eval_steps_per_second": 1.119, + "step": 490 + }, + { + "epoch": 0.6199494949494949, + "grad_norm": 2.757178783416748, + "learning_rate": 4.441066417212349e-06, + "loss": 1.6297, + "step": 491 + }, + { + "epoch": 0.6212121212121212, + "grad_norm": 2.7179954051971436, + "learning_rate": 4.438727782974743e-06, + "loss": 1.5685, + "step": 492 + }, + { + "epoch": 0.6224747474747475, + "grad_norm": 2.824514389038086, + "learning_rate": 4.436389148737138e-06, + "loss": 1.6625, + "step": 493 + }, + { + "epoch": 0.6237373737373737, + "grad_norm": 3.0062615871429443, + "learning_rate": 4.434050514499533e-06, + "loss": 1.5889, + "step": 494 + }, + { + "epoch": 0.625, + "grad_norm": 2.962176561355591, + "learning_rate": 4.431711880261927e-06, + "loss": 1.6813, + "step": 495 + }, + { + "epoch": 0.6262626262626263, + "grad_norm": 2.7531204223632812, + "learning_rate": 4.429373246024323e-06, + "loss": 1.6152, + "step": 496 + }, + { + "epoch": 0.6275252525252525, + "grad_norm": 2.764021873474121, + "learning_rate": 4.427034611786717e-06, + "loss": 1.6164, + "step": 497 + }, + { + "epoch": 0.6287878787878788, + "grad_norm": 2.836826801300049, + "learning_rate": 4.4246959775491114e-06, + "loss": 1.6192, + "step": 498 + }, + { + "epoch": 0.6300505050505051, + "grad_norm": 2.871220827102661, + "learning_rate": 4.422357343311507e-06, + "loss": 1.6096, + "step": 499 + }, + { + "epoch": 0.6313131313131313, + "grad_norm": 2.8129167556762695, + "learning_rate": 4.420018709073901e-06, + "loss": 1.6674, + "step": 500 + }, + { + "epoch": 0.6313131313131313, + "eval_loss": 1.785156011581421, + "eval_runtime": 37.2971, + "eval_samples_per_second": 26.812, + "eval_steps_per_second": 1.126, + "step": 500 + }, + { + "epoch": 0.6325757575757576, + "grad_norm": 2.775108814239502, + "learning_rate": 4.4176800748362955e-06, + "loss": 1.5851, + "step": 501 + }, + { + "epoch": 0.6338383838383839, + "grad_norm": 2.7775518894195557, + "learning_rate": 4.415341440598691e-06, + "loss": 1.6839, + "step": 502 + }, + { + "epoch": 0.63510101010101, + "grad_norm": 2.7997918128967285, + "learning_rate": 4.413002806361085e-06, + "loss": 1.5849, + "step": 503 + }, + { + "epoch": 0.6363636363636364, + "grad_norm": 2.91103196144104, + "learning_rate": 4.4106641721234805e-06, + "loss": 1.6369, + "step": 504 + }, + { + "epoch": 0.6376262626262627, + "grad_norm": 2.6800882816314697, + "learning_rate": 4.408325537885875e-06, + "loss": 1.6049, + "step": 505 + }, + { + "epoch": 0.6388888888888888, + "grad_norm": 2.7451021671295166, + "learning_rate": 4.40598690364827e-06, + "loss": 1.658, + "step": 506 + }, + { + "epoch": 0.6401515151515151, + "grad_norm": 2.794813632965088, + "learning_rate": 4.403648269410665e-06, + "loss": 1.6015, + "step": 507 + }, + { + "epoch": 0.6414141414141414, + "grad_norm": 2.769373655319214, + "learning_rate": 4.401309635173059e-06, + "loss": 1.6217, + "step": 508 + }, + { + "epoch": 0.6426767676767676, + "grad_norm": 2.9081523418426514, + "learning_rate": 4.398971000935454e-06, + "loss": 1.6235, + "step": 509 + }, + { + "epoch": 0.6439393939393939, + "grad_norm": 2.707618236541748, + "learning_rate": 4.396632366697849e-06, + "loss": 1.6003, + "step": 510 + }, + { + "epoch": 0.6439393939393939, + "eval_loss": 1.7876425981521606, + "eval_runtime": 36.9832, + "eval_samples_per_second": 27.039, + "eval_steps_per_second": 1.136, + "step": 510 + }, + { + "epoch": 0.6452020202020202, + "grad_norm": 2.7923099994659424, + "learning_rate": 4.394293732460244e-06, + "loss": 1.5647, + "step": 511 + }, + { + "epoch": 0.6464646464646465, + "grad_norm": 2.8449175357818604, + "learning_rate": 4.391955098222638e-06, + "loss": 1.6456, + "step": 512 + }, + { + "epoch": 0.6477272727272727, + "grad_norm": 2.755856990814209, + "learning_rate": 4.389616463985033e-06, + "loss": 1.6962, + "step": 513 + }, + { + "epoch": 0.648989898989899, + "grad_norm": 2.5578861236572266, + "learning_rate": 4.387277829747428e-06, + "loss": 1.6097, + "step": 514 + }, + { + "epoch": 0.6502525252525253, + "grad_norm": 2.8804502487182617, + "learning_rate": 4.3849391955098225e-06, + "loss": 1.6212, + "step": 515 + }, + { + "epoch": 0.6515151515151515, + "grad_norm": 2.684215545654297, + "learning_rate": 4.382600561272217e-06, + "loss": 1.5529, + "step": 516 + }, + { + "epoch": 0.6527777777777778, + "grad_norm": 2.9128026962280273, + "learning_rate": 4.380261927034612e-06, + "loss": 1.6424, + "step": 517 + }, + { + "epoch": 0.6540404040404041, + "grad_norm": 2.733415126800537, + "learning_rate": 4.377923292797007e-06, + "loss": 1.6201, + "step": 518 + }, + { + "epoch": 0.6553030303030303, + "grad_norm": 2.7211062908172607, + "learning_rate": 4.375584658559401e-06, + "loss": 1.5833, + "step": 519 + }, + { + "epoch": 0.6565656565656566, + "grad_norm": 2.7308316230773926, + "learning_rate": 4.373246024321796e-06, + "loss": 1.6288, + "step": 520 + }, + { + "epoch": 0.6565656565656566, + "eval_loss": 1.7787725925445557, + "eval_runtime": 37.1089, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 520 + }, + { + "epoch": 0.6578282828282829, + "grad_norm": 2.760362148284912, + "learning_rate": 4.3709073900841915e-06, + "loss": 1.5814, + "step": 521 + }, + { + "epoch": 0.6590909090909091, + "grad_norm": 2.7625153064727783, + "learning_rate": 4.368568755846586e-06, + "loss": 1.6239, + "step": 522 + }, + { + "epoch": 0.6603535353535354, + "grad_norm": 2.715710401535034, + "learning_rate": 4.366230121608981e-06, + "loss": 1.6069, + "step": 523 + }, + { + "epoch": 0.6616161616161617, + "grad_norm": 2.7753846645355225, + "learning_rate": 4.363891487371376e-06, + "loss": 1.6178, + "step": 524 + }, + { + "epoch": 0.6628787878787878, + "grad_norm": 2.650716543197632, + "learning_rate": 4.36155285313377e-06, + "loss": 1.6534, + "step": 525 + }, + { + "epoch": 0.6641414141414141, + "grad_norm": 2.573969602584839, + "learning_rate": 4.359214218896165e-06, + "loss": 1.6721, + "step": 526 + }, + { + "epoch": 0.6654040404040404, + "grad_norm": 2.9017715454101562, + "learning_rate": 4.35687558465856e-06, + "loss": 1.6273, + "step": 527 + }, + { + "epoch": 0.6666666666666666, + "grad_norm": 2.691643714904785, + "learning_rate": 4.354536950420954e-06, + "loss": 1.6199, + "step": 528 + }, + { + "epoch": 0.6679292929292929, + "grad_norm": 2.818662166595459, + "learning_rate": 4.3521983161833494e-06, + "loss": 1.5758, + "step": 529 + }, + { + "epoch": 0.6691919191919192, + "grad_norm": 3.017738103866577, + "learning_rate": 4.349859681945744e-06, + "loss": 1.6437, + "step": 530 + }, + { + "epoch": 0.6691919191919192, + "eval_loss": 1.7835458517074585, + "eval_runtime": 37.1359, + "eval_samples_per_second": 26.928, + "eval_steps_per_second": 1.131, + "step": 530 + }, + { + "epoch": 0.6704545454545454, + "grad_norm": 2.8027284145355225, + "learning_rate": 4.347521047708138e-06, + "loss": 1.6688, + "step": 531 + }, + { + "epoch": 0.6717171717171717, + "grad_norm": 2.6642649173736572, + "learning_rate": 4.3451824134705335e-06, + "loss": 1.6327, + "step": 532 + }, + { + "epoch": 0.672979797979798, + "grad_norm": 2.7406396865844727, + "learning_rate": 4.342843779232928e-06, + "loss": 1.6297, + "step": 533 + }, + { + "epoch": 0.6742424242424242, + "grad_norm": 2.758538007736206, + "learning_rate": 4.340505144995323e-06, + "loss": 1.5276, + "step": 534 + }, + { + "epoch": 0.6755050505050505, + "grad_norm": 2.815817356109619, + "learning_rate": 4.3381665107577185e-06, + "loss": 1.6192, + "step": 535 + }, + { + "epoch": 0.6767676767676768, + "grad_norm": 2.8416755199432373, + "learning_rate": 4.335827876520113e-06, + "loss": 1.5966, + "step": 536 + }, + { + "epoch": 0.678030303030303, + "grad_norm": 3.014401435852051, + "learning_rate": 4.333489242282507e-06, + "loss": 1.6104, + "step": 537 + }, + { + "epoch": 0.6792929292929293, + "grad_norm": 2.731356143951416, + "learning_rate": 4.331150608044903e-06, + "loss": 1.6177, + "step": 538 + }, + { + "epoch": 0.6805555555555556, + "grad_norm": 2.633310317993164, + "learning_rate": 4.328811973807297e-06, + "loss": 1.5004, + "step": 539 + }, + { + "epoch": 0.6818181818181818, + "grad_norm": 2.977476119995117, + "learning_rate": 4.3264733395696914e-06, + "loss": 1.6171, + "step": 540 + }, + { + "epoch": 0.6818181818181818, + "eval_loss": 1.7790141105651855, + "eval_runtime": 37.12, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 540 + }, + { + "epoch": 0.6830808080808081, + "grad_norm": 2.769422769546509, + "learning_rate": 4.324134705332087e-06, + "loss": 1.5591, + "step": 541 + }, + { + "epoch": 0.6843434343434344, + "grad_norm": 3.088426351547241, + "learning_rate": 4.321796071094481e-06, + "loss": 1.5603, + "step": 542 + }, + { + "epoch": 0.6856060606060606, + "grad_norm": 2.6784212589263916, + "learning_rate": 4.3194574368568755e-06, + "loss": 1.6061, + "step": 543 + }, + { + "epoch": 0.6868686868686869, + "grad_norm": 3.209821939468384, + "learning_rate": 4.317118802619271e-06, + "loss": 1.585, + "step": 544 + }, + { + "epoch": 0.6881313131313131, + "grad_norm": 2.940791368484497, + "learning_rate": 4.314780168381665e-06, + "loss": 1.6092, + "step": 545 + }, + { + "epoch": 0.6893939393939394, + "grad_norm": 2.679884672164917, + "learning_rate": 4.3124415341440605e-06, + "loss": 1.6207, + "step": 546 + }, + { + "epoch": 0.6906565656565656, + "grad_norm": 2.853809118270874, + "learning_rate": 4.310102899906455e-06, + "loss": 1.5674, + "step": 547 + }, + { + "epoch": 0.6919191919191919, + "grad_norm": 2.693838119506836, + "learning_rate": 4.307764265668849e-06, + "loss": 1.5886, + "step": 548 + }, + { + "epoch": 0.6931818181818182, + "grad_norm": 2.871680736541748, + "learning_rate": 4.305425631431245e-06, + "loss": 1.6309, + "step": 549 + }, + { + "epoch": 0.6944444444444444, + "grad_norm": 2.9908292293548584, + "learning_rate": 4.30308699719364e-06, + "loss": 1.5577, + "step": 550 + }, + { + "epoch": 0.6944444444444444, + "eval_loss": 1.7790372371673584, + "eval_runtime": 37.114, + "eval_samples_per_second": 26.944, + "eval_steps_per_second": 1.132, + "step": 550 + }, + { + "epoch": 0.6957070707070707, + "grad_norm": 2.883481740951538, + "learning_rate": 4.300748362956034e-06, + "loss": 1.5749, + "step": 551 + }, + { + "epoch": 0.696969696969697, + "grad_norm": 2.7523458003997803, + "learning_rate": 4.298409728718429e-06, + "loss": 1.624, + "step": 552 + }, + { + "epoch": 0.6982323232323232, + "grad_norm": 2.69509220123291, + "learning_rate": 4.296071094480824e-06, + "loss": 1.5856, + "step": 553 + }, + { + "epoch": 0.6994949494949495, + "grad_norm": 2.6340222358703613, + "learning_rate": 4.293732460243218e-06, + "loss": 1.5857, + "step": 554 + }, + { + "epoch": 0.7007575757575758, + "grad_norm": 2.4998607635498047, + "learning_rate": 4.291393826005613e-06, + "loss": 1.6038, + "step": 555 + }, + { + "epoch": 0.702020202020202, + "grad_norm": 2.7913858890533447, + "learning_rate": 4.289055191768008e-06, + "loss": 1.6411, + "step": 556 + }, + { + "epoch": 0.7032828282828283, + "grad_norm": 2.7149479389190674, + "learning_rate": 4.2867165575304025e-06, + "loss": 1.592, + "step": 557 + }, + { + "epoch": 0.7045454545454546, + "grad_norm": 2.7187132835388184, + "learning_rate": 4.284377923292797e-06, + "loss": 1.5413, + "step": 558 + }, + { + "epoch": 0.7058080808080808, + "grad_norm": 2.8913674354553223, + "learning_rate": 4.282039289055192e-06, + "loss": 1.6105, + "step": 559 + }, + { + "epoch": 0.7070707070707071, + "grad_norm": 2.8470897674560547, + "learning_rate": 4.279700654817587e-06, + "loss": 1.6481, + "step": 560 + }, + { + "epoch": 0.7070707070707071, + "eval_loss": 1.7711386680603027, + "eval_runtime": 37.1243, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 560 + }, + { + "epoch": 0.7083333333333334, + "grad_norm": 2.7525687217712402, + "learning_rate": 4.277362020579982e-06, + "loss": 1.5749, + "step": 561 + }, + { + "epoch": 0.7095959595959596, + "grad_norm": 2.6935207843780518, + "learning_rate": 4.275023386342376e-06, + "loss": 1.637, + "step": 562 + }, + { + "epoch": 0.7108585858585859, + "grad_norm": 2.6935207843780518, + "learning_rate": 4.275023386342376e-06, + "loss": 1.5903, + "step": 563 + }, + { + "epoch": 0.7121212121212122, + "grad_norm": 2.992992401123047, + "learning_rate": 4.272684752104771e-06, + "loss": 1.6654, + "step": 564 + }, + { + "epoch": 0.7133838383838383, + "grad_norm": 2.7314348220825195, + "learning_rate": 4.270346117867166e-06, + "loss": 1.6775, + "step": 565 + }, + { + "epoch": 0.7146464646464646, + "grad_norm": 2.6952013969421387, + "learning_rate": 4.26800748362956e-06, + "loss": 1.5936, + "step": 566 + }, + { + "epoch": 0.7159090909090909, + "grad_norm": 2.8244829177856445, + "learning_rate": 4.265668849391956e-06, + "loss": 1.6228, + "step": 567 + }, + { + "epoch": 0.7171717171717171, + "grad_norm": 2.9110143184661865, + "learning_rate": 4.26333021515435e-06, + "loss": 1.6016, + "step": 568 + }, + { + "epoch": 0.7184343434343434, + "grad_norm": 2.7953436374664307, + "learning_rate": 4.260991580916745e-06, + "loss": 1.611, + "step": 569 + }, + { + "epoch": 0.7196969696969697, + "grad_norm": 2.854447364807129, + "learning_rate": 4.25865294667914e-06, + "loss": 1.6365, + "step": 570 + }, + { + "epoch": 0.7196969696969697, + "eval_loss": 1.7711069583892822, + "eval_runtime": 37.1374, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 570 + }, + { + "epoch": 0.7209595959595959, + "grad_norm": 2.7885239124298096, + "learning_rate": 4.256314312441534e-06, + "loss": 1.6106, + "step": 571 + }, + { + "epoch": 0.7222222222222222, + "grad_norm": 2.8012993335723877, + "learning_rate": 4.2539756782039295e-06, + "loss": 1.6081, + "step": 572 + }, + { + "epoch": 0.7234848484848485, + "grad_norm": 2.9223906993865967, + "learning_rate": 4.251637043966324e-06, + "loss": 1.5836, + "step": 573 + }, + { + "epoch": 0.7247474747474747, + "grad_norm": 2.885830879211426, + "learning_rate": 4.249298409728719e-06, + "loss": 1.5616, + "step": 574 + }, + { + "epoch": 0.726010101010101, + "grad_norm": 2.687372922897339, + "learning_rate": 4.2469597754911136e-06, + "loss": 1.5873, + "step": 575 + }, + { + "epoch": 0.7272727272727273, + "grad_norm": 3.0156168937683105, + "learning_rate": 4.244621141253508e-06, + "loss": 1.5763, + "step": 576 + }, + { + "epoch": 0.7285353535353535, + "grad_norm": 3.061842918395996, + "learning_rate": 4.242282507015903e-06, + "loss": 1.6148, + "step": 577 + }, + { + "epoch": 0.7297979797979798, + "grad_norm": 2.7467715740203857, + "learning_rate": 4.239943872778298e-06, + "loss": 1.577, + "step": 578 + }, + { + "epoch": 0.7310606060606061, + "grad_norm": 2.669501543045044, + "learning_rate": 4.237605238540692e-06, + "loss": 1.5653, + "step": 579 + }, + { + "epoch": 0.7323232323232324, + "grad_norm": 2.874349355697632, + "learning_rate": 4.235266604303087e-06, + "loss": 1.6523, + "step": 580 + }, + { + "epoch": 0.7323232323232324, + "eval_loss": 1.7680890560150146, + "eval_runtime": 37.2931, + "eval_samples_per_second": 26.815, + "eval_steps_per_second": 1.126, + "step": 580 + }, + { + "epoch": 0.7335858585858586, + "grad_norm": 2.7468504905700684, + "learning_rate": 4.232927970065482e-06, + "loss": 1.5935, + "step": 581 + }, + { + "epoch": 0.7348484848484849, + "grad_norm": 2.936697244644165, + "learning_rate": 4.230589335827877e-06, + "loss": 1.6335, + "step": 582 + }, + { + "epoch": 0.7361111111111112, + "grad_norm": 2.6878974437713623, + "learning_rate": 4.2282507015902715e-06, + "loss": 1.5699, + "step": 583 + }, + { + "epoch": 0.7373737373737373, + "grad_norm": 3.3578853607177734, + "learning_rate": 4.225912067352667e-06, + "loss": 1.5979, + "step": 584 + }, + { + "epoch": 0.7386363636363636, + "grad_norm": 2.8441176414489746, + "learning_rate": 4.223573433115061e-06, + "loss": 1.5736, + "step": 585 + }, + { + "epoch": 0.73989898989899, + "grad_norm": 2.7186129093170166, + "learning_rate": 4.221234798877456e-06, + "loss": 1.636, + "step": 586 + }, + { + "epoch": 0.7411616161616161, + "grad_norm": 2.940002918243408, + "learning_rate": 4.218896164639851e-06, + "loss": 1.5721, + "step": 587 + }, + { + "epoch": 0.7424242424242424, + "grad_norm": 2.676284074783325, + "learning_rate": 4.216557530402245e-06, + "loss": 1.5972, + "step": 588 + }, + { + "epoch": 0.7436868686868687, + "grad_norm": 2.9176571369171143, + "learning_rate": 4.2142188961646405e-06, + "loss": 1.6045, + "step": 589 + }, + { + "epoch": 0.7449494949494949, + "grad_norm": 2.7716288566589355, + "learning_rate": 4.211880261927035e-06, + "loss": 1.5988, + "step": 590 + }, + { + "epoch": 0.7449494949494949, + "eval_loss": 1.775631070137024, + "eval_runtime": 37.4113, + "eval_samples_per_second": 26.73, + "eval_steps_per_second": 1.123, + "step": 590 + }, + { + "epoch": 0.7462121212121212, + "grad_norm": 2.7201945781707764, + "learning_rate": 4.209541627689429e-06, + "loss": 1.5767, + "step": 591 + }, + { + "epoch": 0.7474747474747475, + "grad_norm": 2.838362455368042, + "learning_rate": 4.207202993451825e-06, + "loss": 1.5793, + "step": 592 + }, + { + "epoch": 0.7487373737373737, + "grad_norm": 2.6511716842651367, + "learning_rate": 4.204864359214219e-06, + "loss": 1.5776, + "step": 593 + }, + { + "epoch": 0.75, + "grad_norm": 2.7062957286834717, + "learning_rate": 4.2025257249766135e-06, + "loss": 1.5593, + "step": 594 + }, + { + "epoch": 0.7512626262626263, + "grad_norm": 2.808570384979248, + "learning_rate": 4.200187090739009e-06, + "loss": 1.6249, + "step": 595 + }, + { + "epoch": 0.7525252525252525, + "grad_norm": 2.600278377532959, + "learning_rate": 4.197848456501403e-06, + "loss": 1.6171, + "step": 596 + }, + { + "epoch": 0.7537878787878788, + "grad_norm": 2.9669382572174072, + "learning_rate": 4.195509822263798e-06, + "loss": 1.608, + "step": 597 + }, + { + "epoch": 0.7550505050505051, + "grad_norm": 2.6576340198516846, + "learning_rate": 4.193171188026194e-06, + "loss": 1.5986, + "step": 598 + }, + { + "epoch": 0.7563131313131313, + "grad_norm": 2.787060022354126, + "learning_rate": 4.190832553788588e-06, + "loss": 1.548, + "step": 599 + }, + { + "epoch": 0.7575757575757576, + "grad_norm": 2.854971408843994, + "learning_rate": 4.1884939195509825e-06, + "loss": 1.6421, + "step": 600 + }, + { + "epoch": 0.7575757575757576, + "eval_loss": 1.7721046209335327, + "eval_runtime": 37.4471, + "eval_samples_per_second": 26.704, + "eval_steps_per_second": 1.122, + "step": 600 + }, + { + "epoch": 0.7588383838383839, + "grad_norm": 2.7417967319488525, + "learning_rate": 4.186155285313378e-06, + "loss": 1.6399, + "step": 601 + }, + { + "epoch": 0.76010101010101, + "grad_norm": 3.0036182403564453, + "learning_rate": 4.183816651075772e-06, + "loss": 1.5953, + "step": 602 + }, + { + "epoch": 0.7613636363636364, + "grad_norm": 2.502861261367798, + "learning_rate": 4.181478016838167e-06, + "loss": 1.5284, + "step": 603 + }, + { + "epoch": 0.7626262626262627, + "grad_norm": 2.7262094020843506, + "learning_rate": 4.179139382600562e-06, + "loss": 1.619, + "step": 604 + }, + { + "epoch": 0.7638888888888888, + "grad_norm": 2.8748724460601807, + "learning_rate": 4.176800748362956e-06, + "loss": 1.5625, + "step": 605 + }, + { + "epoch": 0.7651515151515151, + "grad_norm": 2.7272109985351562, + "learning_rate": 4.174462114125351e-06, + "loss": 1.5935, + "step": 606 + }, + { + "epoch": 0.7664141414141414, + "grad_norm": 2.7930479049682617, + "learning_rate": 4.172123479887746e-06, + "loss": 1.61, + "step": 607 + }, + { + "epoch": 0.7676767676767676, + "grad_norm": 2.952056407928467, + "learning_rate": 4.16978484565014e-06, + "loss": 1.6369, + "step": 608 + }, + { + "epoch": 0.7689393939393939, + "grad_norm": 2.979564905166626, + "learning_rate": 4.167446211412535e-06, + "loss": 1.5748, + "step": 609 + }, + { + "epoch": 0.7702020202020202, + "grad_norm": 2.7945773601531982, + "learning_rate": 4.16510757717493e-06, + "loss": 1.6096, + "step": 610 + }, + { + "epoch": 0.7702020202020202, + "eval_loss": 1.7568180561065674, + "eval_runtime": 36.969, + "eval_samples_per_second": 27.05, + "eval_steps_per_second": 1.136, + "step": 610 + }, + { + "epoch": 0.7714646464646465, + "grad_norm": 2.9217135906219482, + "learning_rate": 4.1627689429373245e-06, + "loss": 1.5902, + "step": 611 + }, + { + "epoch": 0.7727272727272727, + "grad_norm": 2.8184452056884766, + "learning_rate": 4.16043030869972e-06, + "loss": 1.5775, + "step": 612 + }, + { + "epoch": 0.773989898989899, + "grad_norm": 2.7121212482452393, + "learning_rate": 4.158091674462115e-06, + "loss": 1.5778, + "step": 613 + }, + { + "epoch": 0.7752525252525253, + "grad_norm": 2.7600955963134766, + "learning_rate": 4.1557530402245095e-06, + "loss": 1.6145, + "step": 614 + }, + { + "epoch": 0.7765151515151515, + "grad_norm": 2.6388845443725586, + "learning_rate": 4.153414405986904e-06, + "loss": 1.6475, + "step": 615 + }, + { + "epoch": 0.7777777777777778, + "grad_norm": 2.684379816055298, + "learning_rate": 4.151075771749299e-06, + "loss": 1.6187, + "step": 616 + }, + { + "epoch": 0.7790404040404041, + "grad_norm": 2.6907055377960205, + "learning_rate": 4.1487371375116936e-06, + "loss": 1.6108, + "step": 617 + }, + { + "epoch": 0.7803030303030303, + "grad_norm": 2.8458995819091797, + "learning_rate": 4.146398503274088e-06, + "loss": 1.5596, + "step": 618 + }, + { + "epoch": 0.7815656565656566, + "grad_norm": 2.68843674659729, + "learning_rate": 4.144059869036483e-06, + "loss": 1.5912, + "step": 619 + }, + { + "epoch": 0.7828282828282829, + "grad_norm": 2.7620413303375244, + "learning_rate": 4.141721234798878e-06, + "loss": 1.627, + "step": 620 + }, + { + "epoch": 0.7828282828282829, + "eval_loss": 1.757076621055603, + "eval_runtime": 37.0489, + "eval_samples_per_second": 26.991, + "eval_steps_per_second": 1.134, + "step": 620 + }, + { + "epoch": 0.7840909090909091, + "grad_norm": 2.7084949016571045, + "learning_rate": 4.139382600561272e-06, + "loss": 1.5249, + "step": 621 + }, + { + "epoch": 0.7853535353535354, + "grad_norm": 3.0410406589508057, + "learning_rate": 4.137043966323667e-06, + "loss": 1.6458, + "step": 622 + }, + { + "epoch": 0.7866161616161617, + "grad_norm": 2.8445630073547363, + "learning_rate": 4.134705332086062e-06, + "loss": 1.6404, + "step": 623 + }, + { + "epoch": 0.7878787878787878, + "grad_norm": 3.0764353275299072, + "learning_rate": 4.132366697848457e-06, + "loss": 1.5041, + "step": 624 + }, + { + "epoch": 0.7891414141414141, + "grad_norm": 2.626544237136841, + "learning_rate": 4.1300280636108515e-06, + "loss": 1.6145, + "step": 625 + }, + { + "epoch": 0.7904040404040404, + "grad_norm": 2.9735796451568604, + "learning_rate": 4.127689429373246e-06, + "loss": 1.5672, + "step": 626 + }, + { + "epoch": 0.7916666666666666, + "grad_norm": 2.7798562049865723, + "learning_rate": 4.125350795135641e-06, + "loss": 1.6272, + "step": 627 + }, + { + "epoch": 0.7929292929292929, + "grad_norm": 2.7438931465148926, + "learning_rate": 4.123012160898036e-06, + "loss": 1.5318, + "step": 628 + }, + { + "epoch": 0.7941919191919192, + "grad_norm": 3.103142499923706, + "learning_rate": 4.120673526660431e-06, + "loss": 1.6, + "step": 629 + }, + { + "epoch": 0.7954545454545454, + "grad_norm": 2.9056239128112793, + "learning_rate": 4.118334892422825e-06, + "loss": 1.6581, + "step": 630 + }, + { + "epoch": 0.7954545454545454, + "eval_loss": 1.7552870512008667, + "eval_runtime": 37.1223, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 630 + }, + { + "epoch": 0.7967171717171717, + "grad_norm": 2.7320070266723633, + "learning_rate": 4.1159962581852205e-06, + "loss": 1.5808, + "step": 631 + }, + { + "epoch": 0.797979797979798, + "grad_norm": 2.9518637657165527, + "learning_rate": 4.113657623947615e-06, + "loss": 1.6034, + "step": 632 + }, + { + "epoch": 0.7992424242424242, + "grad_norm": 2.7135164737701416, + "learning_rate": 4.111318989710009e-06, + "loss": 1.5684, + "step": 633 + }, + { + "epoch": 0.8005050505050505, + "grad_norm": 2.948230266571045, + "learning_rate": 4.108980355472405e-06, + "loss": 1.6378, + "step": 634 + }, + { + "epoch": 0.8017676767676768, + "grad_norm": 2.816206216812134, + "learning_rate": 4.106641721234799e-06, + "loss": 1.6118, + "step": 635 + }, + { + "epoch": 0.803030303030303, + "grad_norm": 2.8032639026641846, + "learning_rate": 4.104303086997194e-06, + "loss": 1.5627, + "step": 636 + }, + { + "epoch": 0.8042929292929293, + "grad_norm": 3.032348871231079, + "learning_rate": 4.101964452759589e-06, + "loss": 1.6192, + "step": 637 + }, + { + "epoch": 0.8055555555555556, + "grad_norm": 2.9512710571289062, + "learning_rate": 4.099625818521983e-06, + "loss": 1.5655, + "step": 638 + }, + { + "epoch": 0.8068181818181818, + "grad_norm": 2.738973379135132, + "learning_rate": 4.097287184284378e-06, + "loss": 1.5686, + "step": 639 + }, + { + "epoch": 0.8080808080808081, + "grad_norm": 2.6510915756225586, + "learning_rate": 4.094948550046773e-06, + "loss": 1.5889, + "step": 640 + }, + { + "epoch": 0.8080808080808081, + "eval_loss": 1.762879490852356, + "eval_runtime": 37.1457, + "eval_samples_per_second": 26.921, + "eval_steps_per_second": 1.131, + "step": 640 + }, + { + "epoch": 0.8093434343434344, + "grad_norm": 2.6933884620666504, + "learning_rate": 4.092609915809167e-06, + "loss": 1.5436, + "step": 641 + }, + { + "epoch": 0.8106060606060606, + "grad_norm": 2.703784227371216, + "learning_rate": 4.0902712815715625e-06, + "loss": 1.5869, + "step": 642 + }, + { + "epoch": 0.8118686868686869, + "grad_norm": 2.7742321491241455, + "learning_rate": 4.087932647333958e-06, + "loss": 1.5779, + "step": 643 + }, + { + "epoch": 0.8131313131313131, + "grad_norm": 2.836832284927368, + "learning_rate": 4.085594013096352e-06, + "loss": 1.5511, + "step": 644 + }, + { + "epoch": 0.8143939393939394, + "grad_norm": 2.9212028980255127, + "learning_rate": 4.083255378858747e-06, + "loss": 1.5604, + "step": 645 + }, + { + "epoch": 0.8156565656565656, + "grad_norm": 2.8608527183532715, + "learning_rate": 4.080916744621142e-06, + "loss": 1.5975, + "step": 646 + }, + { + "epoch": 0.8169191919191919, + "grad_norm": 2.8246703147888184, + "learning_rate": 4.078578110383536e-06, + "loss": 1.5962, + "step": 647 + }, + { + "epoch": 0.8181818181818182, + "grad_norm": 3.0648651123046875, + "learning_rate": 4.076239476145932e-06, + "loss": 1.6267, + "step": 648 + }, + { + "epoch": 0.8194444444444444, + "grad_norm": 2.849240303039551, + "learning_rate": 4.073900841908326e-06, + "loss": 1.6271, + "step": 649 + }, + { + "epoch": 0.8207070707070707, + "grad_norm": 2.8817555904388428, + "learning_rate": 4.07156220767072e-06, + "loss": 1.5916, + "step": 650 + }, + { + "epoch": 0.8207070707070707, + "eval_loss": 1.7502237558364868, + "eval_runtime": 37.1497, + "eval_samples_per_second": 26.918, + "eval_steps_per_second": 1.131, + "step": 650 + }, + { + "epoch": 0.821969696969697, + "grad_norm": 2.5747952461242676, + "learning_rate": 4.069223573433116e-06, + "loss": 1.5661, + "step": 651 + }, + { + "epoch": 0.8232323232323232, + "grad_norm": 2.7721974849700928, + "learning_rate": 4.06688493919551e-06, + "loss": 1.5903, + "step": 652 + }, + { + "epoch": 0.8244949494949495, + "grad_norm": 2.9606690406799316, + "learning_rate": 4.0645463049579045e-06, + "loss": 1.5431, + "step": 653 + }, + { + "epoch": 0.8257575757575758, + "grad_norm": 2.941169261932373, + "learning_rate": 4.0622076707203e-06, + "loss": 1.5711, + "step": 654 + }, + { + "epoch": 0.827020202020202, + "grad_norm": 2.8408572673797607, + "learning_rate": 4.059869036482694e-06, + "loss": 1.5975, + "step": 655 + }, + { + "epoch": 0.8282828282828283, + "grad_norm": 2.619875431060791, + "learning_rate": 4.057530402245089e-06, + "loss": 1.5194, + "step": 656 + }, + { + "epoch": 0.8295454545454546, + "grad_norm": 2.910982847213745, + "learning_rate": 4.055191768007484e-06, + "loss": 1.6318, + "step": 657 + }, + { + "epoch": 0.8308080808080808, + "grad_norm": 2.747422456741333, + "learning_rate": 4.052853133769879e-06, + "loss": 1.6205, + "step": 658 + }, + { + "epoch": 0.8320707070707071, + "grad_norm": 2.6889796257019043, + "learning_rate": 4.050514499532274e-06, + "loss": 1.5537, + "step": 659 + }, + { + "epoch": 0.8333333333333334, + "grad_norm": 2.9849202632904053, + "learning_rate": 4.048175865294669e-06, + "loss": 1.5438, + "step": 660 + }, + { + "epoch": 0.8333333333333334, + "eval_loss": 1.7484922409057617, + "eval_runtime": 37.161, + "eval_samples_per_second": 26.91, + "eval_steps_per_second": 1.13, + "step": 660 + }, + { + "epoch": 0.8345959595959596, + "grad_norm": 2.816930055618286, + "learning_rate": 4.045837231057063e-06, + "loss": 1.555, + "step": 661 + }, + { + "epoch": 0.8358585858585859, + "grad_norm": 2.990126371383667, + "learning_rate": 4.043498596819458e-06, + "loss": 1.6263, + "step": 662 + }, + { + "epoch": 0.8371212121212122, + "grad_norm": 2.898263692855835, + "learning_rate": 4.041159962581853e-06, + "loss": 1.5522, + "step": 663 + }, + { + "epoch": 0.8383838383838383, + "grad_norm": 2.7824151515960693, + "learning_rate": 4.038821328344247e-06, + "loss": 1.5366, + "step": 664 + }, + { + "epoch": 0.8396464646464646, + "grad_norm": 2.9943954944610596, + "learning_rate": 4.036482694106642e-06, + "loss": 1.6214, + "step": 665 + }, + { + "epoch": 0.8409090909090909, + "grad_norm": 2.628308057785034, + "learning_rate": 4.034144059869037e-06, + "loss": 1.5797, + "step": 666 + }, + { + "epoch": 0.8421717171717171, + "grad_norm": 2.7565176486968994, + "learning_rate": 4.0318054256314315e-06, + "loss": 1.5567, + "step": 667 + }, + { + "epoch": 0.8434343434343434, + "grad_norm": 2.688354969024658, + "learning_rate": 4.029466791393826e-06, + "loss": 1.5516, + "step": 668 + }, + { + "epoch": 0.8446969696969697, + "grad_norm": 2.899620771408081, + "learning_rate": 4.027128157156221e-06, + "loss": 1.6014, + "step": 669 + }, + { + "epoch": 0.8459595959595959, + "grad_norm": 2.7894718647003174, + "learning_rate": 4.024789522918616e-06, + "loss": 1.5688, + "step": 670 + }, + { + "epoch": 0.8459595959595959, + "eval_loss": 1.7486555576324463, + "eval_runtime": 37.1824, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 670 + }, + { + "epoch": 0.8472222222222222, + "grad_norm": 2.7441444396972656, + "learning_rate": 4.02245088868101e-06, + "loss": 1.5949, + "step": 671 + }, + { + "epoch": 0.8484848484848485, + "grad_norm": 2.7088451385498047, + "learning_rate": 4.020112254443405e-06, + "loss": 1.5878, + "step": 672 + }, + { + "epoch": 0.8497474747474747, + "grad_norm": 2.756920576095581, + "learning_rate": 4.0177736202058005e-06, + "loss": 1.6356, + "step": 673 + }, + { + "epoch": 0.851010101010101, + "grad_norm": 2.5605697631835938, + "learning_rate": 4.015434985968195e-06, + "loss": 1.5586, + "step": 674 + }, + { + "epoch": 0.8522727272727273, + "grad_norm": 2.946573257446289, + "learning_rate": 4.01309635173059e-06, + "loss": 1.5537, + "step": 675 + }, + { + "epoch": 0.8535353535353535, + "grad_norm": 2.65500545501709, + "learning_rate": 4.010757717492985e-06, + "loss": 1.5875, + "step": 676 + }, + { + "epoch": 0.8547979797979798, + "grad_norm": 2.6632118225097656, + "learning_rate": 4.008419083255379e-06, + "loss": 1.5329, + "step": 677 + }, + { + "epoch": 0.8560606060606061, + "grad_norm": 2.8909032344818115, + "learning_rate": 4.006080449017774e-06, + "loss": 1.6396, + "step": 678 + }, + { + "epoch": 0.8573232323232324, + "grad_norm": 2.8780951499938965, + "learning_rate": 4.003741814780169e-06, + "loss": 1.6189, + "step": 679 + }, + { + "epoch": 0.8585858585858586, + "grad_norm": 2.7757010459899902, + "learning_rate": 4.001403180542563e-06, + "loss": 1.5968, + "step": 680 + }, + { + "epoch": 0.8585858585858586, + "eval_loss": 1.7423925399780273, + "eval_runtime": 37.1537, + "eval_samples_per_second": 26.915, + "eval_steps_per_second": 1.13, + "step": 680 + }, + { + "epoch": 0.8598484848484849, + "grad_norm": 2.6610794067382812, + "learning_rate": 3.9990645463049584e-06, + "loss": 1.5258, + "step": 681 + }, + { + "epoch": 0.8611111111111112, + "grad_norm": 2.9292140007019043, + "learning_rate": 3.996725912067353e-06, + "loss": 1.5669, + "step": 682 + }, + { + "epoch": 0.8623737373737373, + "grad_norm": 2.7885360717773438, + "learning_rate": 3.994387277829747e-06, + "loss": 1.5737, + "step": 683 + }, + { + "epoch": 0.8636363636363636, + "grad_norm": 2.7511544227600098, + "learning_rate": 3.9920486435921425e-06, + "loss": 1.5715, + "step": 684 + }, + { + "epoch": 0.86489898989899, + "grad_norm": 2.8561315536499023, + "learning_rate": 3.989710009354537e-06, + "loss": 1.5965, + "step": 685 + }, + { + "epoch": 0.8661616161616161, + "grad_norm": 2.9127237796783447, + "learning_rate": 3.987371375116932e-06, + "loss": 1.5407, + "step": 686 + }, + { + "epoch": 0.8674242424242424, + "grad_norm": 2.735577344894409, + "learning_rate": 3.985032740879327e-06, + "loss": 1.5449, + "step": 687 + }, + { + "epoch": 0.8686868686868687, + "grad_norm": 2.6626689434051514, + "learning_rate": 3.982694106641721e-06, + "loss": 1.5173, + "step": 688 + }, + { + "epoch": 0.8699494949494949, + "grad_norm": 2.901592969894409, + "learning_rate": 3.980355472404116e-06, + "loss": 1.5757, + "step": 689 + }, + { + "epoch": 0.8712121212121212, + "grad_norm": 2.8801231384277344, + "learning_rate": 3.978016838166512e-06, + "loss": 1.5633, + "step": 690 + }, + { + "epoch": 0.8712121212121212, + "eval_loss": 1.7453432083129883, + "eval_runtime": 37.4368, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 690 + }, + { + "epoch": 0.8724747474747475, + "grad_norm": 2.7463626861572266, + "learning_rate": 3.975678203928906e-06, + "loss": 1.5786, + "step": 691 + }, + { + "epoch": 0.8737373737373737, + "grad_norm": 2.659820318222046, + "learning_rate": 3.9733395696913004e-06, + "loss": 1.5923, + "step": 692 + }, + { + "epoch": 0.875, + "grad_norm": 2.759030342102051, + "learning_rate": 3.971000935453696e-06, + "loss": 1.6217, + "step": 693 + }, + { + "epoch": 0.8762626262626263, + "grad_norm": 2.689967632293701, + "learning_rate": 3.96866230121609e-06, + "loss": 1.5533, + "step": 694 + }, + { + "epoch": 0.8775252525252525, + "grad_norm": 3.1230740547180176, + "learning_rate": 3.9663236669784845e-06, + "loss": 1.6538, + "step": 695 + }, + { + "epoch": 0.8787878787878788, + "grad_norm": 2.657022476196289, + "learning_rate": 3.96398503274088e-06, + "loss": 1.5432, + "step": 696 + }, + { + "epoch": 0.8800505050505051, + "grad_norm": 2.919379711151123, + "learning_rate": 3.961646398503274e-06, + "loss": 1.5867, + "step": 697 + }, + { + "epoch": 0.8813131313131313, + "grad_norm": 2.607135057449341, + "learning_rate": 3.9593077642656695e-06, + "loss": 1.5079, + "step": 698 + }, + { + "epoch": 0.8825757575757576, + "grad_norm": 2.9978976249694824, + "learning_rate": 3.956969130028064e-06, + "loss": 1.5905, + "step": 699 + }, + { + "epoch": 0.8838383838383839, + "grad_norm": 2.8117148876190186, + "learning_rate": 3.954630495790458e-06, + "loss": 1.5588, + "step": 700 + }, + { + "epoch": 0.8838383838383839, + "eval_loss": 1.7376645803451538, + "eval_runtime": 37.5045, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 700 + }, + { + "epoch": 0.88510101010101, + "grad_norm": 2.7135095596313477, + "learning_rate": 3.952291861552854e-06, + "loss": 1.6013, + "step": 701 + }, + { + "epoch": 0.8863636363636364, + "grad_norm": 2.812061071395874, + "learning_rate": 3.949953227315248e-06, + "loss": 1.5518, + "step": 702 + }, + { + "epoch": 0.8876262626262627, + "grad_norm": 3.0532422065734863, + "learning_rate": 3.9476145930776424e-06, + "loss": 1.5579, + "step": 703 + }, + { + "epoch": 0.8888888888888888, + "grad_norm": 2.8306567668914795, + "learning_rate": 3.945275958840038e-06, + "loss": 1.5478, + "step": 704 + }, + { + "epoch": 0.8901515151515151, + "grad_norm": 2.737921953201294, + "learning_rate": 3.942937324602433e-06, + "loss": 1.4958, + "step": 705 + }, + { + "epoch": 0.8914141414141414, + "grad_norm": 2.902211904525757, + "learning_rate": 3.940598690364827e-06, + "loss": 1.5715, + "step": 706 + }, + { + "epoch": 0.8926767676767676, + "grad_norm": 2.725818395614624, + "learning_rate": 3.938260056127222e-06, + "loss": 1.5502, + "step": 707 + }, + { + "epoch": 0.8939393939393939, + "grad_norm": 2.9530506134033203, + "learning_rate": 3.935921421889617e-06, + "loss": 1.564, + "step": 708 + }, + { + "epoch": 0.8952020202020202, + "grad_norm": 2.798231363296509, + "learning_rate": 3.9335827876520115e-06, + "loss": 1.5445, + "step": 709 + }, + { + "epoch": 0.8964646464646465, + "grad_norm": 3.0233471393585205, + "learning_rate": 3.931244153414407e-06, + "loss": 1.6193, + "step": 710 + }, + { + "epoch": 0.8964646464646465, + "eval_loss": 1.742077350616455, + "eval_runtime": 37.0255, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 710 + }, + { + "epoch": 0.8977272727272727, + "grad_norm": 2.84309458732605, + "learning_rate": 3.928905519176801e-06, + "loss": 1.5379, + "step": 711 + }, + { + "epoch": 0.898989898989899, + "grad_norm": 2.7412185668945312, + "learning_rate": 3.926566884939196e-06, + "loss": 1.5837, + "step": 712 + }, + { + "epoch": 0.9002525252525253, + "grad_norm": 3.1094822883605957, + "learning_rate": 3.924228250701591e-06, + "loss": 1.5621, + "step": 713 + }, + { + "epoch": 0.9015151515151515, + "grad_norm": 2.776059627532959, + "learning_rate": 3.921889616463985e-06, + "loss": 1.5714, + "step": 714 + }, + { + "epoch": 0.9027777777777778, + "grad_norm": 2.7650489807128906, + "learning_rate": 3.91955098222638e-06, + "loss": 1.5418, + "step": 715 + }, + { + "epoch": 0.9040404040404041, + "grad_norm": 2.8586974143981934, + "learning_rate": 3.917212347988775e-06, + "loss": 1.5891, + "step": 716 + }, + { + "epoch": 0.9053030303030303, + "grad_norm": 2.74332594871521, + "learning_rate": 3.914873713751169e-06, + "loss": 1.5492, + "step": 717 + }, + { + "epoch": 0.9065656565656566, + "grad_norm": 2.931682586669922, + "learning_rate": 3.912535079513564e-06, + "loss": 1.54, + "step": 718 + }, + { + "epoch": 0.9078282828282829, + "grad_norm": 2.678812026977539, + "learning_rate": 3.910196445275959e-06, + "loss": 1.5677, + "step": 719 + }, + { + "epoch": 0.9090909090909091, + "grad_norm": 2.7282893657684326, + "learning_rate": 3.907857811038354e-06, + "loss": 1.5413, + "step": 720 + }, + { + "epoch": 0.9090909090909091, + "eval_loss": 1.7409753799438477, + "eval_runtime": 37.1883, + "eval_samples_per_second": 26.89, + "eval_steps_per_second": 1.129, + "step": 720 + }, + { + "epoch": 0.9103535353535354, + "grad_norm": 3.073997735977173, + "learning_rate": 3.905519176800749e-06, + "loss": 1.5734, + "step": 721 + }, + { + "epoch": 0.9116161616161617, + "grad_norm": 2.6249935626983643, + "learning_rate": 3.903180542563144e-06, + "loss": 1.5063, + "step": 722 + }, + { + "epoch": 0.9128787878787878, + "grad_norm": 2.7758615016937256, + "learning_rate": 3.9008419083255384e-06, + "loss": 1.5112, + "step": 723 + }, + { + "epoch": 0.9141414141414141, + "grad_norm": 2.8954596519470215, + "learning_rate": 3.898503274087933e-06, + "loss": 1.5508, + "step": 724 + }, + { + "epoch": 0.9154040404040404, + "grad_norm": 2.79073166847229, + "learning_rate": 3.896164639850328e-06, + "loss": 1.5773, + "step": 725 + }, + { + "epoch": 0.9166666666666666, + "grad_norm": 2.687027931213379, + "learning_rate": 3.8938260056127226e-06, + "loss": 1.5243, + "step": 726 + }, + { + "epoch": 0.9179292929292929, + "grad_norm": 2.5704851150512695, + "learning_rate": 3.891487371375117e-06, + "loss": 1.5075, + "step": 727 + }, + { + "epoch": 0.9191919191919192, + "grad_norm": 2.8684048652648926, + "learning_rate": 3.889148737137512e-06, + "loss": 1.6471, + "step": 728 + }, + { + "epoch": 0.9204545454545454, + "grad_norm": 2.8995590209960938, + "learning_rate": 3.886810102899907e-06, + "loss": 1.5429, + "step": 729 + }, + { + "epoch": 0.9217171717171717, + "grad_norm": 2.7575254440307617, + "learning_rate": 3.884471468662301e-06, + "loss": 1.5241, + "step": 730 + }, + { + "epoch": 0.9217171717171717, + "eval_loss": 1.7294232845306396, + "eval_runtime": 37.1941, + "eval_samples_per_second": 26.886, + "eval_steps_per_second": 1.129, + "step": 730 + }, + { + "epoch": 0.922979797979798, + "grad_norm": 2.7057321071624756, + "learning_rate": 3.882132834424696e-06, + "loss": 1.5498, + "step": 731 + }, + { + "epoch": 0.9242424242424242, + "grad_norm": 2.7679436206817627, + "learning_rate": 3.879794200187091e-06, + "loss": 1.563, + "step": 732 + }, + { + "epoch": 0.9255050505050505, + "grad_norm": 2.7258687019348145, + "learning_rate": 3.877455565949485e-06, + "loss": 1.5016, + "step": 733 + }, + { + "epoch": 0.9267676767676768, + "grad_norm": 2.8474221229553223, + "learning_rate": 3.8751169317118804e-06, + "loss": 1.5318, + "step": 734 + }, + { + "epoch": 0.928030303030303, + "grad_norm": 2.7718396186828613, + "learning_rate": 3.872778297474276e-06, + "loss": 1.5919, + "step": 735 + }, + { + "epoch": 0.9292929292929293, + "grad_norm": 2.7785022258758545, + "learning_rate": 3.87043966323667e-06, + "loss": 1.5399, + "step": 736 + }, + { + "epoch": 0.9305555555555556, + "grad_norm": 2.742476463317871, + "learning_rate": 3.868101028999065e-06, + "loss": 1.5428, + "step": 737 + }, + { + "epoch": 0.9318181818181818, + "grad_norm": 2.7050726413726807, + "learning_rate": 3.86576239476146e-06, + "loss": 1.5604, + "step": 738 + }, + { + "epoch": 0.9330808080808081, + "grad_norm": 2.6755242347717285, + "learning_rate": 3.863423760523854e-06, + "loss": 1.5409, + "step": 739 + }, + { + "epoch": 0.9343434343434344, + "grad_norm": 2.6947829723358154, + "learning_rate": 3.8610851262862495e-06, + "loss": 1.5537, + "step": 740 + }, + { + "epoch": 0.9343434343434344, + "eval_loss": 1.7396037578582764, + "eval_runtime": 37.1912, + "eval_samples_per_second": 26.888, + "eval_steps_per_second": 1.129, + "step": 740 + }, + { + "epoch": 0.9356060606060606, + "grad_norm": 2.729816436767578, + "learning_rate": 3.858746492048644e-06, + "loss": 1.5581, + "step": 741 + }, + { + "epoch": 0.9368686868686869, + "grad_norm": 2.6951887607574463, + "learning_rate": 3.856407857811038e-06, + "loss": 1.5929, + "step": 742 + }, + { + "epoch": 0.9381313131313131, + "grad_norm": 2.6601064205169678, + "learning_rate": 3.854069223573434e-06, + "loss": 1.4881, + "step": 743 + }, + { + "epoch": 0.9393939393939394, + "grad_norm": 2.623898983001709, + "learning_rate": 3.851730589335828e-06, + "loss": 1.5499, + "step": 744 + }, + { + "epoch": 0.9406565656565656, + "grad_norm": 2.703871250152588, + "learning_rate": 3.8493919550982224e-06, + "loss": 1.5585, + "step": 745 + }, + { + "epoch": 0.9419191919191919, + "grad_norm": 2.674893379211426, + "learning_rate": 3.847053320860618e-06, + "loss": 1.5586, + "step": 746 + }, + { + "epoch": 0.9431818181818182, + "grad_norm": 2.8353004455566406, + "learning_rate": 3.844714686623012e-06, + "loss": 1.5937, + "step": 747 + }, + { + "epoch": 0.9444444444444444, + "grad_norm": 2.68009614944458, + "learning_rate": 3.842376052385407e-06, + "loss": 1.5192, + "step": 748 + }, + { + "epoch": 0.9457070707070707, + "grad_norm": 2.6847407817840576, + "learning_rate": 3.840037418147802e-06, + "loss": 1.512, + "step": 749 + }, + { + "epoch": 0.946969696969697, + "grad_norm": 2.624290943145752, + "learning_rate": 3.837698783910197e-06, + "loss": 1.5348, + "step": 750 + }, + { + "epoch": 0.946969696969697, + "eval_loss": 1.738332748413086, + "eval_runtime": 37.1782, + "eval_samples_per_second": 26.898, + "eval_steps_per_second": 1.13, + "step": 750 + }, + { + "epoch": 0.9482323232323232, + "grad_norm": 2.7467777729034424, + "learning_rate": 3.8353601496725915e-06, + "loss": 1.5603, + "step": 751 + }, + { + "epoch": 0.9494949494949495, + "grad_norm": 2.851987600326538, + "learning_rate": 3.833021515434987e-06, + "loss": 1.5974, + "step": 752 + }, + { + "epoch": 0.9507575757575758, + "grad_norm": 2.6642227172851562, + "learning_rate": 3.830682881197381e-06, + "loss": 1.5138, + "step": 753 + }, + { + "epoch": 0.952020202020202, + "grad_norm": 2.7351491451263428, + "learning_rate": 3.828344246959776e-06, + "loss": 1.5656, + "step": 754 + }, + { + "epoch": 0.9532828282828283, + "grad_norm": 2.593748092651367, + "learning_rate": 3.826005612722171e-06, + "loss": 1.5395, + "step": 755 + }, + { + "epoch": 0.9545454545454546, + "grad_norm": 2.639180898666382, + "learning_rate": 3.823666978484565e-06, + "loss": 1.5349, + "step": 756 + }, + { + "epoch": 0.9558080808080808, + "grad_norm": 2.805148124694824, + "learning_rate": 3.82132834424696e-06, + "loss": 1.5805, + "step": 757 + }, + { + "epoch": 0.9570707070707071, + "grad_norm": 2.5619425773620605, + "learning_rate": 3.818989710009355e-06, + "loss": 1.5389, + "step": 758 + }, + { + "epoch": 0.9583333333333334, + "grad_norm": 2.7224011421203613, + "learning_rate": 3.816651075771749e-06, + "loss": 1.5316, + "step": 759 + }, + { + "epoch": 0.9595959595959596, + "grad_norm": 2.4667656421661377, + "learning_rate": 3.8143124415341442e-06, + "loss": 1.4942, + "step": 760 + }, + { + "epoch": 0.9595959595959596, + "eval_loss": 1.7205606698989868, + "eval_runtime": 37.1654, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 760 + }, + { + "epoch": 0.9608585858585859, + "grad_norm": 2.456643581390381, + "learning_rate": 3.811973807296539e-06, + "loss": 1.538, + "step": 761 + }, + { + "epoch": 0.9621212121212122, + "grad_norm": 2.696720838546753, + "learning_rate": 3.8096351730589335e-06, + "loss": 1.5462, + "step": 762 + }, + { + "epoch": 0.9633838383838383, + "grad_norm": 2.7314131259918213, + "learning_rate": 3.8072965388213284e-06, + "loss": 1.5931, + "step": 763 + }, + { + "epoch": 0.9646464646464646, + "grad_norm": 2.616532325744629, + "learning_rate": 3.804957904583723e-06, + "loss": 1.5469, + "step": 764 + }, + { + "epoch": 0.9659090909090909, + "grad_norm": 2.7471885681152344, + "learning_rate": 3.8026192703461185e-06, + "loss": 1.5431, + "step": 765 + }, + { + "epoch": 0.9671717171717171, + "grad_norm": 2.7478253841400146, + "learning_rate": 3.8002806361085133e-06, + "loss": 1.5452, + "step": 766 + }, + { + "epoch": 0.9684343434343434, + "grad_norm": 2.543790340423584, + "learning_rate": 3.7979420018709077e-06, + "loss": 1.6091, + "step": 767 + }, + { + "epoch": 0.9696969696969697, + "grad_norm": 2.6169509887695312, + "learning_rate": 3.7956033676333026e-06, + "loss": 1.5557, + "step": 768 + }, + { + "epoch": 0.9709595959595959, + "grad_norm": 2.4768588542938232, + "learning_rate": 3.7932647333956974e-06, + "loss": 1.5247, + "step": 769 + }, + { + "epoch": 0.9722222222222222, + "grad_norm": 2.654648780822754, + "learning_rate": 3.7909260991580923e-06, + "loss": 1.5281, + "step": 770 + }, + { + "epoch": 0.9722222222222222, + "eval_loss": 1.7289395332336426, + "eval_runtime": 37.1374, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 770 + }, + { + "epoch": 0.9734848484848485, + "grad_norm": 2.893073081970215, + "learning_rate": 3.7885874649204867e-06, + "loss": 1.5749, + "step": 771 + }, + { + "epoch": 0.9747474747474747, + "grad_norm": 2.696341037750244, + "learning_rate": 3.7862488306828815e-06, + "loss": 1.5025, + "step": 772 + }, + { + "epoch": 0.976010101010101, + "grad_norm": 2.678222417831421, + "learning_rate": 3.7839101964452764e-06, + "loss": 1.538, + "step": 773 + }, + { + "epoch": 0.9772727272727273, + "grad_norm": 2.7772746086120605, + "learning_rate": 3.7815715622076708e-06, + "loss": 1.5648, + "step": 774 + }, + { + "epoch": 0.9785353535353535, + "grad_norm": 3.076620101928711, + "learning_rate": 3.7792329279700656e-06, + "loss": 1.6114, + "step": 775 + }, + { + "epoch": 0.9797979797979798, + "grad_norm": 2.700056791305542, + "learning_rate": 3.7768942937324605e-06, + "loss": 1.5881, + "step": 776 + }, + { + "epoch": 0.9810606060606061, + "grad_norm": 2.601419687271118, + "learning_rate": 3.7745556594948553e-06, + "loss": 1.5319, + "step": 777 + }, + { + "epoch": 0.9823232323232324, + "grad_norm": 2.6664950847625732, + "learning_rate": 3.7722170252572497e-06, + "loss": 1.5497, + "step": 778 + }, + { + "epoch": 0.9835858585858586, + "grad_norm": 2.7374589443206787, + "learning_rate": 3.7698783910196446e-06, + "loss": 1.5488, + "step": 779 + }, + { + "epoch": 0.9848484848484849, + "grad_norm": 2.797938823699951, + "learning_rate": 3.76753975678204e-06, + "loss": 1.5376, + "step": 780 + }, + { + "epoch": 0.9848484848484849, + "eval_loss": 1.7334709167480469, + "eval_runtime": 37.1372, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 780 + }, + { + "epoch": 0.9861111111111112, + "grad_norm": 2.881591796875, + "learning_rate": 3.7652011225444347e-06, + "loss": 1.5405, + "step": 781 + }, + { + "epoch": 0.9873737373737373, + "grad_norm": 2.6965644359588623, + "learning_rate": 3.7628624883068295e-06, + "loss": 1.5066, + "step": 782 + }, + { + "epoch": 0.9886363636363636, + "grad_norm": 2.7286899089813232, + "learning_rate": 3.760523854069224e-06, + "loss": 1.5277, + "step": 783 + }, + { + "epoch": 0.98989898989899, + "grad_norm": 2.795644998550415, + "learning_rate": 3.7581852198316188e-06, + "loss": 1.5335, + "step": 784 + }, + { + "epoch": 0.9911616161616161, + "grad_norm": 2.8160440921783447, + "learning_rate": 3.7558465855940136e-06, + "loss": 1.585, + "step": 785 + }, + { + "epoch": 0.9924242424242424, + "grad_norm": 2.7984936237335205, + "learning_rate": 3.753507951356408e-06, + "loss": 1.5569, + "step": 786 + }, + { + "epoch": 0.9936868686868687, + "grad_norm": 2.7596278190612793, + "learning_rate": 3.751169317118803e-06, + "loss": 1.4838, + "step": 787 + }, + { + "epoch": 0.9949494949494949, + "grad_norm": 2.666383743286133, + "learning_rate": 3.7488306828811977e-06, + "loss": 1.5656, + "step": 788 + }, + { + "epoch": 0.9962121212121212, + "grad_norm": 2.8139331340789795, + "learning_rate": 3.7464920486435926e-06, + "loss": 1.5753, + "step": 789 + }, + { + "epoch": 0.9974747474747475, + "grad_norm": 2.7714269161224365, + "learning_rate": 3.744153414405987e-06, + "loss": 1.5187, + "step": 790 + }, + { + "epoch": 0.9974747474747475, + "eval_loss": 1.7257990837097168, + "eval_runtime": 37.1065, + "eval_samples_per_second": 26.949, + "eval_steps_per_second": 1.132, + "step": 790 + }, + { + "epoch": 0.9987373737373737, + "grad_norm": 2.697606325149536, + "learning_rate": 3.741814780168382e-06, + "loss": 1.56, + "step": 791 + }, + { + "epoch": 1.0, + "grad_norm": 2.7685229778289795, + "learning_rate": 3.7394761459307767e-06, + "loss": 1.5705, + "step": 792 + }, + { + "epoch": 1.0012626262626263, + "grad_norm": 2.778200626373291, + "learning_rate": 3.737137511693171e-06, + "loss": 1.5288, + "step": 793 + }, + { + "epoch": 1.0025252525252526, + "grad_norm": 2.7211337089538574, + "learning_rate": 3.734798877455566e-06, + "loss": 1.5348, + "step": 794 + }, + { + "epoch": 1.003787878787879, + "grad_norm": 2.690739154815674, + "learning_rate": 3.7324602432179608e-06, + "loss": 1.4734, + "step": 795 + }, + { + "epoch": 1.005050505050505, + "grad_norm": 2.6809420585632324, + "learning_rate": 3.730121608980356e-06, + "loss": 1.4852, + "step": 796 + }, + { + "epoch": 1.0063131313131313, + "grad_norm": 2.8550117015838623, + "learning_rate": 3.727782974742751e-06, + "loss": 1.4562, + "step": 797 + }, + { + "epoch": 1.0075757575757576, + "grad_norm": 2.713935613632202, + "learning_rate": 3.7254443405051453e-06, + "loss": 1.5005, + "step": 798 + }, + { + "epoch": 1.0088383838383839, + "grad_norm": 2.668363332748413, + "learning_rate": 3.72310570626754e-06, + "loss": 1.485, + "step": 799 + }, + { + "epoch": 1.0101010101010102, + "grad_norm": 2.840769052505493, + "learning_rate": 3.720767072029935e-06, + "loss": 1.4531, + "step": 800 + }, + { + "epoch": 1.0101010101010102, + "eval_loss": 1.725937008857727, + "eval_runtime": 37.2471, + "eval_samples_per_second": 26.848, + "eval_steps_per_second": 1.128, + "step": 800 + }, + { + "epoch": 1.0113636363636365, + "grad_norm": 2.7345798015594482, + "learning_rate": 3.71842843779233e-06, + "loss": 1.5091, + "step": 801 + }, + { + "epoch": 1.0126262626262625, + "grad_norm": 2.9052624702453613, + "learning_rate": 3.7160898035547243e-06, + "loss": 1.5039, + "step": 802 + }, + { + "epoch": 1.0138888888888888, + "grad_norm": 2.893073558807373, + "learning_rate": 3.713751169317119e-06, + "loss": 1.4613, + "step": 803 + }, + { + "epoch": 1.0151515151515151, + "grad_norm": 2.8505427837371826, + "learning_rate": 3.711412535079514e-06, + "loss": 1.4932, + "step": 804 + }, + { + "epoch": 1.0164141414141414, + "grad_norm": 2.9065170288085938, + "learning_rate": 3.7090739008419084e-06, + "loss": 1.5082, + "step": 805 + }, + { + "epoch": 1.0176767676767677, + "grad_norm": 2.888068675994873, + "learning_rate": 3.706735266604303e-06, + "loss": 1.5675, + "step": 806 + }, + { + "epoch": 1.018939393939394, + "grad_norm": 2.7325260639190674, + "learning_rate": 3.704396632366698e-06, + "loss": 1.4649, + "step": 807 + }, + { + "epoch": 1.02020202020202, + "grad_norm": 2.8861136436462402, + "learning_rate": 3.702057998129093e-06, + "loss": 1.4517, + "step": 808 + }, + { + "epoch": 1.0214646464646464, + "grad_norm": 2.8837692737579346, + "learning_rate": 3.6997193638914873e-06, + "loss": 1.5198, + "step": 809 + }, + { + "epoch": 1.0227272727272727, + "grad_norm": 2.790330410003662, + "learning_rate": 3.697380729653882e-06, + "loss": 1.5066, + "step": 810 + }, + { + "epoch": 1.0227272727272727, + "eval_loss": 1.7269084453582764, + "eval_runtime": 36.89, + "eval_samples_per_second": 27.108, + "eval_steps_per_second": 1.139, + "step": 810 + }, + { + "epoch": 1.023989898989899, + "grad_norm": 3.03643536567688, + "learning_rate": 3.6950420954162774e-06, + "loss": 1.4928, + "step": 811 + }, + { + "epoch": 1.0252525252525253, + "grad_norm": 2.841055154800415, + "learning_rate": 3.6927034611786723e-06, + "loss": 1.5273, + "step": 812 + }, + { + "epoch": 1.0265151515151516, + "grad_norm": 2.84134578704834, + "learning_rate": 3.690364826941067e-06, + "loss": 1.5443, + "step": 813 + }, + { + "epoch": 1.0277777777777777, + "grad_norm": 2.837758779525757, + "learning_rate": 3.6880261927034615e-06, + "loss": 1.5586, + "step": 814 + }, + { + "epoch": 1.029040404040404, + "grad_norm": 2.731412410736084, + "learning_rate": 3.6856875584658564e-06, + "loss": 1.4465, + "step": 815 + }, + { + "epoch": 1.0303030303030303, + "grad_norm": 2.8979599475860596, + "learning_rate": 3.683348924228251e-06, + "loss": 1.5663, + "step": 816 + }, + { + "epoch": 1.0315656565656566, + "grad_norm": 2.7740373611450195, + "learning_rate": 3.6810102899906456e-06, + "loss": 1.4578, + "step": 817 + }, + { + "epoch": 1.0328282828282829, + "grad_norm": 2.8984005451202393, + "learning_rate": 3.6786716557530405e-06, + "loss": 1.493, + "step": 818 + }, + { + "epoch": 1.0340909090909092, + "grad_norm": 2.880680561065674, + "learning_rate": 3.6763330215154353e-06, + "loss": 1.4858, + "step": 819 + }, + { + "epoch": 1.0353535353535352, + "grad_norm": 3.0220961570739746, + "learning_rate": 3.67399438727783e-06, + "loss": 1.5009, + "step": 820 + }, + { + "epoch": 1.0353535353535352, + "eval_loss": 1.7245674133300781, + "eval_runtime": 37.0235, + "eval_samples_per_second": 27.01, + "eval_steps_per_second": 1.134, + "step": 820 + }, + { + "epoch": 1.0366161616161615, + "grad_norm": 2.7905962467193604, + "learning_rate": 3.6716557530402246e-06, + "loss": 1.4728, + "step": 821 + }, + { + "epoch": 1.0378787878787878, + "grad_norm": 2.856935501098633, + "learning_rate": 3.6693171188026194e-06, + "loss": 1.521, + "step": 822 + }, + { + "epoch": 1.0391414141414141, + "grad_norm": 2.956584930419922, + "learning_rate": 3.6669784845650143e-06, + "loss": 1.4725, + "step": 823 + }, + { + "epoch": 1.0404040404040404, + "grad_norm": 3.130960702896118, + "learning_rate": 3.6646398503274087e-06, + "loss": 1.6127, + "step": 824 + }, + { + "epoch": 1.0416666666666667, + "grad_norm": 2.8855528831481934, + "learning_rate": 3.6623012160898035e-06, + "loss": 1.5078, + "step": 825 + }, + { + "epoch": 1.0429292929292928, + "grad_norm": 2.8030972480773926, + "learning_rate": 3.659962581852199e-06, + "loss": 1.4635, + "step": 826 + }, + { + "epoch": 1.0441919191919191, + "grad_norm": 3.061593770980835, + "learning_rate": 3.6576239476145936e-06, + "loss": 1.4944, + "step": 827 + }, + { + "epoch": 1.0454545454545454, + "grad_norm": 2.9983572959899902, + "learning_rate": 3.6552853133769885e-06, + "loss": 1.505, + "step": 828 + }, + { + "epoch": 1.0467171717171717, + "grad_norm": 2.7681095600128174, + "learning_rate": 3.652946679139383e-06, + "loss": 1.4774, + "step": 829 + }, + { + "epoch": 1.047979797979798, + "grad_norm": 3.0530946254730225, + "learning_rate": 3.6506080449017777e-06, + "loss": 1.4676, + "step": 830 + }, + { + "epoch": 1.047979797979798, + "eval_loss": 1.7242811918258667, + "eval_runtime": 37.0752, + "eval_samples_per_second": 26.972, + "eval_steps_per_second": 1.133, + "step": 830 + }, + { + "epoch": 1.0492424242424243, + "grad_norm": 2.8383126258850098, + "learning_rate": 3.6482694106641726e-06, + "loss": 1.486, + "step": 831 + }, + { + "epoch": 1.0505050505050506, + "grad_norm": 3.095928907394409, + "learning_rate": 3.6459307764265674e-06, + "loss": 1.5331, + "step": 832 + }, + { + "epoch": 1.0517676767676767, + "grad_norm": 2.9225456714630127, + "learning_rate": 3.643592142188962e-06, + "loss": 1.556, + "step": 833 + }, + { + "epoch": 1.053030303030303, + "grad_norm": 2.9037210941314697, + "learning_rate": 3.6412535079513567e-06, + "loss": 1.4834, + "step": 834 + }, + { + "epoch": 1.0542929292929293, + "grad_norm": 2.915022373199463, + "learning_rate": 3.6389148737137515e-06, + "loss": 1.5346, + "step": 835 + }, + { + "epoch": 1.0555555555555556, + "grad_norm": 3.250411033630371, + "learning_rate": 3.636576239476146e-06, + "loss": 1.4885, + "step": 836 + }, + { + "epoch": 1.0568181818181819, + "grad_norm": 2.759061336517334, + "learning_rate": 3.634237605238541e-06, + "loss": 1.456, + "step": 837 + }, + { + "epoch": 1.0580808080808082, + "grad_norm": 2.955981731414795, + "learning_rate": 3.6318989710009356e-06, + "loss": 1.4889, + "step": 838 + }, + { + "epoch": 1.0593434343434343, + "grad_norm": 3.243338108062744, + "learning_rate": 3.6295603367633305e-06, + "loss": 1.5207, + "step": 839 + }, + { + "epoch": 1.0606060606060606, + "grad_norm": 3.075840950012207, + "learning_rate": 3.627221702525725e-06, + "loss": 1.4906, + "step": 840 + }, + { + "epoch": 1.0606060606060606, + "eval_loss": 1.7290290594100952, + "eval_runtime": 37.097, + "eval_samples_per_second": 26.956, + "eval_steps_per_second": 1.132, + "step": 840 + }, + { + "epoch": 1.0618686868686869, + "grad_norm": 3.149125099182129, + "learning_rate": 3.62488306828812e-06, + "loss": 1.4755, + "step": 841 + }, + { + "epoch": 1.0631313131313131, + "grad_norm": 3.099318504333496, + "learning_rate": 3.622544434050515e-06, + "loss": 1.534, + "step": 842 + }, + { + "epoch": 1.0643939393939394, + "grad_norm": 2.8942148685455322, + "learning_rate": 3.62020579981291e-06, + "loss": 1.4844, + "step": 843 + }, + { + "epoch": 1.0656565656565657, + "grad_norm": 3.142664670944214, + "learning_rate": 3.6178671655753047e-06, + "loss": 1.5179, + "step": 844 + }, + { + "epoch": 1.0669191919191918, + "grad_norm": 3.106552839279175, + "learning_rate": 3.615528531337699e-06, + "loss": 1.4667, + "step": 845 + }, + { + "epoch": 1.0681818181818181, + "grad_norm": 2.8507936000823975, + "learning_rate": 3.613189897100094e-06, + "loss": 1.4463, + "step": 846 + }, + { + "epoch": 1.0694444444444444, + "grad_norm": 3.2198901176452637, + "learning_rate": 3.610851262862489e-06, + "loss": 1.5146, + "step": 847 + }, + { + "epoch": 1.0707070707070707, + "grad_norm": 3.146388292312622, + "learning_rate": 3.6085126286248832e-06, + "loss": 1.4927, + "step": 848 + }, + { + "epoch": 1.071969696969697, + "grad_norm": 2.8304760456085205, + "learning_rate": 3.606173994387278e-06, + "loss": 1.4971, + "step": 849 + }, + { + "epoch": 1.0732323232323233, + "grad_norm": 3.1714529991149902, + "learning_rate": 3.603835360149673e-06, + "loss": 1.5418, + "step": 850 + }, + { + "epoch": 1.0732323232323233, + "eval_loss": 1.725188970565796, + "eval_runtime": 37.117, + "eval_samples_per_second": 26.942, + "eval_steps_per_second": 1.132, + "step": 850 + }, + { + "epoch": 1.0744949494949494, + "grad_norm": 3.06162428855896, + "learning_rate": 3.6014967259120677e-06, + "loss": 1.4959, + "step": 851 + }, + { + "epoch": 1.0757575757575757, + "grad_norm": 2.812206268310547, + "learning_rate": 3.599158091674462e-06, + "loss": 1.5249, + "step": 852 + }, + { + "epoch": 1.077020202020202, + "grad_norm": 3.236210584640503, + "learning_rate": 3.596819457436857e-06, + "loss": 1.5166, + "step": 853 + }, + { + "epoch": 1.0782828282828283, + "grad_norm": 2.8397765159606934, + "learning_rate": 3.594480823199252e-06, + "loss": 1.5043, + "step": 854 + }, + { + "epoch": 1.0795454545454546, + "grad_norm": 2.9215919971466064, + "learning_rate": 3.5921421889616463e-06, + "loss": 1.5235, + "step": 855 + }, + { + "epoch": 1.0808080808080809, + "grad_norm": 2.8219058513641357, + "learning_rate": 3.589803554724041e-06, + "loss": 1.5091, + "step": 856 + }, + { + "epoch": 1.0820707070707072, + "grad_norm": 2.9344675540924072, + "learning_rate": 3.5874649204864364e-06, + "loss": 1.5355, + "step": 857 + }, + { + "epoch": 1.0833333333333333, + "grad_norm": 2.949094533920288, + "learning_rate": 3.5851262862488312e-06, + "loss": 1.5273, + "step": 858 + }, + { + "epoch": 1.0845959595959596, + "grad_norm": 2.9801411628723145, + "learning_rate": 3.582787652011226e-06, + "loss": 1.5078, + "step": 859 + }, + { + "epoch": 1.0858585858585859, + "grad_norm": 2.802182912826538, + "learning_rate": 3.5804490177736205e-06, + "loss": 1.5015, + "step": 860 + }, + { + "epoch": 1.0858585858585859, + "eval_loss": 1.7209502458572388, + "eval_runtime": 37.1238, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 860 + }, + { + "epoch": 1.0871212121212122, + "grad_norm": 2.857457160949707, + "learning_rate": 3.5781103835360153e-06, + "loss": 1.5057, + "step": 861 + }, + { + "epoch": 1.0883838383838385, + "grad_norm": 2.794478178024292, + "learning_rate": 3.57577174929841e-06, + "loss": 1.4594, + "step": 862 + }, + { + "epoch": 1.0896464646464645, + "grad_norm": 3.0028598308563232, + "learning_rate": 3.573433115060805e-06, + "loss": 1.4666, + "step": 863 + }, + { + "epoch": 1.0909090909090908, + "grad_norm": 2.7701406478881836, + "learning_rate": 3.5710944808231994e-06, + "loss": 1.4596, + "step": 864 + }, + { + "epoch": 1.0921717171717171, + "grad_norm": 2.9705960750579834, + "learning_rate": 3.5687558465855943e-06, + "loss": 1.5092, + "step": 865 + }, + { + "epoch": 1.0934343434343434, + "grad_norm": 2.9820821285247803, + "learning_rate": 3.566417212347989e-06, + "loss": 1.5246, + "step": 866 + }, + { + "epoch": 1.0946969696969697, + "grad_norm": 2.988356113433838, + "learning_rate": 3.5640785781103835e-06, + "loss": 1.5035, + "step": 867 + }, + { + "epoch": 1.095959595959596, + "grad_norm": 2.901599884033203, + "learning_rate": 3.5617399438727784e-06, + "loss": 1.5097, + "step": 868 + }, + { + "epoch": 1.0972222222222223, + "grad_norm": 2.9128799438476562, + "learning_rate": 3.5594013096351732e-06, + "loss": 1.5424, + "step": 869 + }, + { + "epoch": 1.0984848484848484, + "grad_norm": 2.719289779663086, + "learning_rate": 3.557062675397568e-06, + "loss": 1.4313, + "step": 870 + }, + { + "epoch": 1.0984848484848484, + "eval_loss": 1.7159698009490967, + "eval_runtime": 37.122, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 870 + }, + { + "epoch": 1.0997474747474747, + "grad_norm": 3.0080976486206055, + "learning_rate": 3.5547240411599625e-06, + "loss": 1.4483, + "step": 871 + }, + { + "epoch": 1.101010101010101, + "grad_norm": 2.856523036956787, + "learning_rate": 3.5523854069223578e-06, + "loss": 1.5463, + "step": 872 + }, + { + "epoch": 1.1022727272727273, + "grad_norm": 2.8585214614868164, + "learning_rate": 3.5500467726847526e-06, + "loss": 1.5391, + "step": 873 + }, + { + "epoch": 1.1035353535353536, + "grad_norm": 2.8440098762512207, + "learning_rate": 3.5477081384471474e-06, + "loss": 1.499, + "step": 874 + }, + { + "epoch": 1.10479797979798, + "grad_norm": 2.8114542961120605, + "learning_rate": 3.545369504209542e-06, + "loss": 1.4691, + "step": 875 + }, + { + "epoch": 1.106060606060606, + "grad_norm": 2.980273962020874, + "learning_rate": 3.5430308699719367e-06, + "loss": 1.4541, + "step": 876 + }, + { + "epoch": 1.1073232323232323, + "grad_norm": 3.0053417682647705, + "learning_rate": 3.5406922357343315e-06, + "loss": 1.4895, + "step": 877 + }, + { + "epoch": 1.1085858585858586, + "grad_norm": 2.924046754837036, + "learning_rate": 3.5383536014967264e-06, + "loss": 1.5186, + "step": 878 + }, + { + "epoch": 1.1098484848484849, + "grad_norm": 2.969461679458618, + "learning_rate": 3.536014967259121e-06, + "loss": 1.5213, + "step": 879 + }, + { + "epoch": 1.1111111111111112, + "grad_norm": 2.8936445713043213, + "learning_rate": 3.5336763330215156e-06, + "loss": 1.5585, + "step": 880 + }, + { + "epoch": 1.1111111111111112, + "eval_loss": 1.7150688171386719, + "eval_runtime": 37.1105, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 880 + }, + { + "epoch": 1.1123737373737375, + "grad_norm": 2.992034435272217, + "learning_rate": 3.5313376987839105e-06, + "loss": 1.5264, + "step": 881 + }, + { + "epoch": 1.1136363636363635, + "grad_norm": 3.0453898906707764, + "learning_rate": 3.5289990645463053e-06, + "loss": 1.5482, + "step": 882 + }, + { + "epoch": 1.1148989898989898, + "grad_norm": 2.7480897903442383, + "learning_rate": 3.5266604303086998e-06, + "loss": 1.4851, + "step": 883 + }, + { + "epoch": 1.1161616161616161, + "grad_norm": 2.9408013820648193, + "learning_rate": 3.5243217960710946e-06, + "loss": 1.536, + "step": 884 + }, + { + "epoch": 1.1174242424242424, + "grad_norm": 2.854438543319702, + "learning_rate": 3.5219831618334894e-06, + "loss": 1.5814, + "step": 885 + }, + { + "epoch": 1.1186868686868687, + "grad_norm": 2.9060397148132324, + "learning_rate": 3.519644527595884e-06, + "loss": 1.5019, + "step": 886 + }, + { + "epoch": 1.119949494949495, + "grad_norm": 2.8741681575775146, + "learning_rate": 3.517305893358279e-06, + "loss": 1.4713, + "step": 887 + }, + { + "epoch": 1.121212121212121, + "grad_norm": 2.972510576248169, + "learning_rate": 3.514967259120674e-06, + "loss": 1.5104, + "step": 888 + }, + { + "epoch": 1.1224747474747474, + "grad_norm": 2.8113558292388916, + "learning_rate": 3.512628624883069e-06, + "loss": 1.4973, + "step": 889 + }, + { + "epoch": 1.1237373737373737, + "grad_norm": 2.9687416553497314, + "learning_rate": 3.5102899906454637e-06, + "loss": 1.4774, + "step": 890 + }, + { + "epoch": 1.1237373737373737, + "eval_loss": 1.7169743776321411, + "eval_runtime": 37.0911, + "eval_samples_per_second": 26.961, + "eval_steps_per_second": 1.132, + "step": 890 + }, + { + "epoch": 1.125, + "grad_norm": 2.911409854888916, + "learning_rate": 3.507951356407858e-06, + "loss": 1.4536, + "step": 891 + }, + { + "epoch": 1.1262626262626263, + "grad_norm": 2.8209526538848877, + "learning_rate": 3.505612722170253e-06, + "loss": 1.5357, + "step": 892 + }, + { + "epoch": 1.1275252525252526, + "grad_norm": 2.811971426010132, + "learning_rate": 3.5032740879326478e-06, + "loss": 1.4889, + "step": 893 + }, + { + "epoch": 1.128787878787879, + "grad_norm": 2.969878673553467, + "learning_rate": 3.5009354536950426e-06, + "loss": 1.4758, + "step": 894 + }, + { + "epoch": 1.130050505050505, + "grad_norm": 2.858635425567627, + "learning_rate": 3.498596819457437e-06, + "loss": 1.5355, + "step": 895 + }, + { + "epoch": 1.1313131313131313, + "grad_norm": 3.018164873123169, + "learning_rate": 3.496258185219832e-06, + "loss": 1.5421, + "step": 896 + }, + { + "epoch": 1.1325757575757576, + "grad_norm": 2.846277952194214, + "learning_rate": 3.4939195509822267e-06, + "loss": 1.4936, + "step": 897 + }, + { + "epoch": 1.1338383838383839, + "grad_norm": 2.965245246887207, + "learning_rate": 3.491580916744621e-06, + "loss": 1.4903, + "step": 898 + }, + { + "epoch": 1.1351010101010102, + "grad_norm": 2.8500895500183105, + "learning_rate": 3.489242282507016e-06, + "loss": 1.5152, + "step": 899 + }, + { + "epoch": 1.1363636363636362, + "grad_norm": 2.8242156505584717, + "learning_rate": 3.486903648269411e-06, + "loss": 1.5531, + "step": 900 + }, + { + "epoch": 1.1363636363636362, + "eval_loss": 1.7180876731872559, + "eval_runtime": 37.4145, + "eval_samples_per_second": 26.728, + "eval_steps_per_second": 1.123, + "step": 900 + }, + { + "epoch": 1.1376262626262625, + "grad_norm": 2.8846635818481445, + "learning_rate": 3.4845650140318057e-06, + "loss": 1.5314, + "step": 901 + }, + { + "epoch": 1.1388888888888888, + "grad_norm": 2.9443984031677246, + "learning_rate": 3.4822263797942e-06, + "loss": 1.4699, + "step": 902 + }, + { + "epoch": 1.1401515151515151, + "grad_norm": 2.7435739040374756, + "learning_rate": 3.4798877455565953e-06, + "loss": 1.4949, + "step": 903 + }, + { + "epoch": 1.1414141414141414, + "grad_norm": 2.979102373123169, + "learning_rate": 3.47754911131899e-06, + "loss": 1.5689, + "step": 904 + }, + { + "epoch": 1.1426767676767677, + "grad_norm": 2.9814586639404297, + "learning_rate": 3.475210477081385e-06, + "loss": 1.5183, + "step": 905 + }, + { + "epoch": 1.143939393939394, + "grad_norm": 2.8380966186523438, + "learning_rate": 3.4728718428437794e-06, + "loss": 1.5177, + "step": 906 + }, + { + "epoch": 1.14520202020202, + "grad_norm": 2.8478410243988037, + "learning_rate": 3.4705332086061743e-06, + "loss": 1.4787, + "step": 907 + }, + { + "epoch": 1.1464646464646464, + "grad_norm": 2.8247780799865723, + "learning_rate": 3.468194574368569e-06, + "loss": 1.425, + "step": 908 + }, + { + "epoch": 1.1477272727272727, + "grad_norm": 2.869649887084961, + "learning_rate": 3.465855940130964e-06, + "loss": 1.4619, + "step": 909 + }, + { + "epoch": 1.148989898989899, + "grad_norm": 2.9439175128936768, + "learning_rate": 3.4635173058933584e-06, + "loss": 1.5194, + "step": 910 + }, + { + "epoch": 1.148989898989899, + "eval_loss": 1.7227450609207153, + "eval_runtime": 36.9723, + "eval_samples_per_second": 27.047, + "eval_steps_per_second": 1.136, + "step": 910 + }, + { + "epoch": 1.1502525252525253, + "grad_norm": 3.0219292640686035, + "learning_rate": 3.4611786716557532e-06, + "loss": 1.5878, + "step": 911 + }, + { + "epoch": 1.1515151515151516, + "grad_norm": 2.988475799560547, + "learning_rate": 3.458840037418148e-06, + "loss": 1.5349, + "step": 912 + }, + { + "epoch": 1.1527777777777777, + "grad_norm": 2.708740234375, + "learning_rate": 3.456501403180543e-06, + "loss": 1.4637, + "step": 913 + }, + { + "epoch": 1.154040404040404, + "grad_norm": 3.0844545364379883, + "learning_rate": 3.4541627689429373e-06, + "loss": 1.546, + "step": 914 + }, + { + "epoch": 1.1553030303030303, + "grad_norm": 2.800487518310547, + "learning_rate": 3.451824134705332e-06, + "loss": 1.5374, + "step": 915 + }, + { + "epoch": 1.1565656565656566, + "grad_norm": 2.8701722621917725, + "learning_rate": 3.449485500467727e-06, + "loss": 1.4549, + "step": 916 + }, + { + "epoch": 1.1578282828282829, + "grad_norm": 3.036418914794922, + "learning_rate": 3.4471468662301214e-06, + "loss": 1.5253, + "step": 917 + }, + { + "epoch": 1.1590909090909092, + "grad_norm": 3.0300986766815186, + "learning_rate": 3.4448082319925167e-06, + "loss": 1.4727, + "step": 918 + }, + { + "epoch": 1.1603535353535355, + "grad_norm": 2.8652496337890625, + "learning_rate": 3.4424695977549116e-06, + "loss": 1.4842, + "step": 919 + }, + { + "epoch": 1.1616161616161615, + "grad_norm": 2.9140491485595703, + "learning_rate": 3.4401309635173064e-06, + "loss": 1.4968, + "step": 920 + }, + { + "epoch": 1.1616161616161615, + "eval_loss": 1.7184255123138428, + "eval_runtime": 37.1186, + "eval_samples_per_second": 26.941, + "eval_steps_per_second": 1.132, + "step": 920 + }, + { + "epoch": 1.1628787878787878, + "grad_norm": 2.9708127975463867, + "learning_rate": 3.4377923292797012e-06, + "loss": 1.4817, + "step": 921 + }, + { + "epoch": 1.1641414141414141, + "grad_norm": 2.954814910888672, + "learning_rate": 3.4354536950420957e-06, + "loss": 1.4937, + "step": 922 + }, + { + "epoch": 1.1654040404040404, + "grad_norm": 2.803945779800415, + "learning_rate": 3.4331150608044905e-06, + "loss": 1.527, + "step": 923 + }, + { + "epoch": 1.1666666666666667, + "grad_norm": 2.863250255584717, + "learning_rate": 3.4307764265668853e-06, + "loss": 1.4922, + "step": 924 + }, + { + "epoch": 1.1679292929292928, + "grad_norm": 2.8030712604522705, + "learning_rate": 3.4284377923292798e-06, + "loss": 1.4427, + "step": 925 + }, + { + "epoch": 1.1691919191919191, + "grad_norm": 2.944110155105591, + "learning_rate": 3.4260991580916746e-06, + "loss": 1.4435, + "step": 926 + }, + { + "epoch": 1.1704545454545454, + "grad_norm": 2.9672467708587646, + "learning_rate": 3.4237605238540695e-06, + "loss": 1.5005, + "step": 927 + }, + { + "epoch": 1.1717171717171717, + "grad_norm": 2.920205593109131, + "learning_rate": 3.4214218896164643e-06, + "loss": 1.5303, + "step": 928 + }, + { + "epoch": 1.172979797979798, + "grad_norm": 3.1062121391296387, + "learning_rate": 3.4190832553788587e-06, + "loss": 1.5395, + "step": 929 + }, + { + "epoch": 1.1742424242424243, + "grad_norm": 3.1106555461883545, + "learning_rate": 3.4167446211412536e-06, + "loss": 1.5038, + "step": 930 + }, + { + "epoch": 1.1742424242424243, + "eval_loss": 1.7127314805984497, + "eval_runtime": 37.1193, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 930 + }, + { + "epoch": 1.1755050505050506, + "grad_norm": 2.7898054122924805, + "learning_rate": 3.4144059869036484e-06, + "loss": 1.5096, + "step": 931 + }, + { + "epoch": 1.1767676767676767, + "grad_norm": 2.881694793701172, + "learning_rate": 3.4120673526660432e-06, + "loss": 1.4556, + "step": 932 + }, + { + "epoch": 1.178030303030303, + "grad_norm": 3.131551742553711, + "learning_rate": 3.4097287184284385e-06, + "loss": 1.5487, + "step": 933 + }, + { + "epoch": 1.1792929292929293, + "grad_norm": 2.9734153747558594, + "learning_rate": 3.407390084190833e-06, + "loss": 1.439, + "step": 934 + }, + { + "epoch": 1.1805555555555556, + "grad_norm": 3.009657144546509, + "learning_rate": 3.4050514499532278e-06, + "loss": 1.462, + "step": 935 + }, + { + "epoch": 1.1818181818181819, + "grad_norm": 2.8378758430480957, + "learning_rate": 3.4027128157156226e-06, + "loss": 1.5083, + "step": 936 + }, + { + "epoch": 1.1830808080808082, + "grad_norm": 2.923969268798828, + "learning_rate": 3.400374181478017e-06, + "loss": 1.5844, + "step": 937 + }, + { + "epoch": 1.1843434343434343, + "grad_norm": 3.2384307384490967, + "learning_rate": 3.398035547240412e-06, + "loss": 1.5175, + "step": 938 + }, + { + "epoch": 1.1856060606060606, + "grad_norm": 3.024385929107666, + "learning_rate": 3.3956969130028067e-06, + "loss": 1.4756, + "step": 939 + }, + { + "epoch": 1.1868686868686869, + "grad_norm": 3.088745594024658, + "learning_rate": 3.3933582787652016e-06, + "loss": 1.4914, + "step": 940 + }, + { + "epoch": 1.1868686868686869, + "eval_loss": 1.705490231513977, + "eval_runtime": 37.1291, + "eval_samples_per_second": 26.933, + "eval_steps_per_second": 1.131, + "step": 940 + }, + { + "epoch": 1.1881313131313131, + "grad_norm": 2.898893117904663, + "learning_rate": 3.391019644527596e-06, + "loss": 1.5002, + "step": 941 + }, + { + "epoch": 1.1893939393939394, + "grad_norm": 2.800103187561035, + "learning_rate": 3.388681010289991e-06, + "loss": 1.4842, + "step": 942 + }, + { + "epoch": 1.1906565656565657, + "grad_norm": 3.0287230014801025, + "learning_rate": 3.3863423760523857e-06, + "loss": 1.5454, + "step": 943 + }, + { + "epoch": 1.1919191919191918, + "grad_norm": 2.9379546642303467, + "learning_rate": 3.3840037418147805e-06, + "loss": 1.5427, + "step": 944 + }, + { + "epoch": 1.1931818181818181, + "grad_norm": 2.879699468612671, + "learning_rate": 3.381665107577175e-06, + "loss": 1.4804, + "step": 945 + }, + { + "epoch": 1.1944444444444444, + "grad_norm": 2.8708608150482178, + "learning_rate": 3.3793264733395698e-06, + "loss": 1.479, + "step": 946 + }, + { + "epoch": 1.1957070707070707, + "grad_norm": 2.8513312339782715, + "learning_rate": 3.3769878391019646e-06, + "loss": 1.5343, + "step": 947 + }, + { + "epoch": 1.196969696969697, + "grad_norm": 2.8715100288391113, + "learning_rate": 3.37464920486436e-06, + "loss": 1.5114, + "step": 948 + }, + { + "epoch": 1.1982323232323233, + "grad_norm": 2.819213628768921, + "learning_rate": 3.3723105706267543e-06, + "loss": 1.4873, + "step": 949 + }, + { + "epoch": 1.1994949494949494, + "grad_norm": 2.946578025817871, + "learning_rate": 3.369971936389149e-06, + "loss": 1.5378, + "step": 950 + }, + { + "epoch": 1.1994949494949494, + "eval_loss": 1.7094330787658691, + "eval_runtime": 37.1453, + "eval_samples_per_second": 26.921, + "eval_steps_per_second": 1.131, + "step": 950 + }, + { + "epoch": 1.2007575757575757, + "grad_norm": 2.8170437812805176, + "learning_rate": 3.367633302151544e-06, + "loss": 1.4416, + "step": 951 + }, + { + "epoch": 1.202020202020202, + "grad_norm": 2.9654042720794678, + "learning_rate": 3.365294667913939e-06, + "loss": 1.5115, + "step": 952 + }, + { + "epoch": 1.2032828282828283, + "grad_norm": 2.9138922691345215, + "learning_rate": 3.3629560336763333e-06, + "loss": 1.5106, + "step": 953 + }, + { + "epoch": 1.2045454545454546, + "grad_norm": 2.989961862564087, + "learning_rate": 3.360617399438728e-06, + "loss": 1.5278, + "step": 954 + }, + { + "epoch": 1.2058080808080809, + "grad_norm": 2.9620680809020996, + "learning_rate": 3.358278765201123e-06, + "loss": 1.4949, + "step": 955 + }, + { + "epoch": 1.2070707070707072, + "grad_norm": 2.945505380630493, + "learning_rate": 3.3559401309635174e-06, + "loss": 1.5018, + "step": 956 + }, + { + "epoch": 1.2083333333333333, + "grad_norm": 2.806464433670044, + "learning_rate": 3.353601496725912e-06, + "loss": 1.4951, + "step": 957 + }, + { + "epoch": 1.2095959595959596, + "grad_norm": 2.7847559452056885, + "learning_rate": 3.351262862488307e-06, + "loss": 1.4259, + "step": 958 + }, + { + "epoch": 1.2108585858585859, + "grad_norm": 2.760003089904785, + "learning_rate": 3.348924228250702e-06, + "loss": 1.4757, + "step": 959 + }, + { + "epoch": 1.2121212121212122, + "grad_norm": 3.037522792816162, + "learning_rate": 3.3465855940130963e-06, + "loss": 1.5462, + "step": 960 + }, + { + "epoch": 1.2121212121212122, + "eval_loss": 1.7071956396102905, + "eval_runtime": 37.1351, + "eval_samples_per_second": 26.929, + "eval_steps_per_second": 1.131, + "step": 960 + }, + { + "epoch": 1.2133838383838385, + "grad_norm": 2.7755539417266846, + "learning_rate": 3.344246959775491e-06, + "loss": 1.453, + "step": 961 + }, + { + "epoch": 1.2146464646464645, + "grad_norm": 2.910369634628296, + "learning_rate": 3.341908325537886e-06, + "loss": 1.4914, + "step": 962 + }, + { + "epoch": 1.2159090909090908, + "grad_norm": 2.8999392986297607, + "learning_rate": 3.339569691300281e-06, + "loss": 1.5389, + "step": 963 + }, + { + "epoch": 1.2171717171717171, + "grad_norm": 2.891995906829834, + "learning_rate": 3.337231057062676e-06, + "loss": 1.4915, + "step": 964 + }, + { + "epoch": 1.2184343434343434, + "grad_norm": 2.814624309539795, + "learning_rate": 3.3348924228250705e-06, + "loss": 1.4936, + "step": 965 + }, + { + "epoch": 1.2196969696969697, + "grad_norm": 2.971163034439087, + "learning_rate": 3.3325537885874654e-06, + "loss": 1.4813, + "step": 966 + }, + { + "epoch": 1.220959595959596, + "grad_norm": 2.934155225753784, + "learning_rate": 3.33021515434986e-06, + "loss": 1.5162, + "step": 967 + }, + { + "epoch": 1.2222222222222223, + "grad_norm": 3.007847309112549, + "learning_rate": 3.3278765201122546e-06, + "loss": 1.4741, + "step": 968 + }, + { + "epoch": 1.2234848484848484, + "grad_norm": 2.977207660675049, + "learning_rate": 3.3255378858746495e-06, + "loss": 1.5186, + "step": 969 + }, + { + "epoch": 1.2247474747474747, + "grad_norm": 2.926567316055298, + "learning_rate": 3.3231992516370443e-06, + "loss": 1.5111, + "step": 970 + }, + { + "epoch": 1.2247474747474747, + "eval_loss": 1.7068836688995361, + "eval_runtime": 37.1419, + "eval_samples_per_second": 26.924, + "eval_steps_per_second": 1.131, + "step": 970 + }, + { + "epoch": 1.226010101010101, + "grad_norm": 3.063671350479126, + "learning_rate": 3.320860617399439e-06, + "loss": 1.4376, + "step": 971 + }, + { + "epoch": 1.2272727272727273, + "grad_norm": 2.7263271808624268, + "learning_rate": 3.3185219831618336e-06, + "loss": 1.4555, + "step": 972 + }, + { + "epoch": 1.2285353535353536, + "grad_norm": 2.912658929824829, + "learning_rate": 3.3161833489242284e-06, + "loss": 1.5116, + "step": 973 + }, + { + "epoch": 1.22979797979798, + "grad_norm": 3.2216858863830566, + "learning_rate": 3.3138447146866233e-06, + "loss": 1.5124, + "step": 974 + }, + { + "epoch": 1.231060606060606, + "grad_norm": 2.9138946533203125, + "learning_rate": 3.311506080449018e-06, + "loss": 1.4917, + "step": 975 + }, + { + "epoch": 1.2323232323232323, + "grad_norm": 3.0519120693206787, + "learning_rate": 3.3091674462114125e-06, + "loss": 1.4986, + "step": 976 + }, + { + "epoch": 1.2335858585858586, + "grad_norm": 2.8186402320861816, + "learning_rate": 3.3068288119738074e-06, + "loss": 1.4835, + "step": 977 + }, + { + "epoch": 1.2348484848484849, + "grad_norm": 2.799189329147339, + "learning_rate": 3.304490177736202e-06, + "loss": 1.5069, + "step": 978 + }, + { + "epoch": 1.2361111111111112, + "grad_norm": 3.010010004043579, + "learning_rate": 3.3021515434985975e-06, + "loss": 1.4276, + "step": 979 + }, + { + "epoch": 1.2373737373737375, + "grad_norm": 3.153592109680176, + "learning_rate": 3.299812909260992e-06, + "loss": 1.5459, + "step": 980 + }, + { + "epoch": 1.2373737373737375, + "eval_loss": 1.7070198059082031, + "eval_runtime": 37.1397, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 980 + }, + { + "epoch": 1.2386363636363638, + "grad_norm": 2.7818424701690674, + "learning_rate": 3.2974742750233867e-06, + "loss": 1.5045, + "step": 981 + }, + { + "epoch": 1.2398989898989898, + "grad_norm": 2.9211390018463135, + "learning_rate": 3.2951356407857816e-06, + "loss": 1.4966, + "step": 982 + }, + { + "epoch": 1.2411616161616161, + "grad_norm": 2.979743242263794, + "learning_rate": 3.2927970065481764e-06, + "loss": 1.4754, + "step": 983 + }, + { + "epoch": 1.2424242424242424, + "grad_norm": 2.8476738929748535, + "learning_rate": 3.290458372310571e-06, + "loss": 1.4619, + "step": 984 + }, + { + "epoch": 1.2436868686868687, + "grad_norm": 2.8304896354675293, + "learning_rate": 3.2881197380729657e-06, + "loss": 1.4192, + "step": 985 + }, + { + "epoch": 1.244949494949495, + "grad_norm": 3.3150880336761475, + "learning_rate": 3.2857811038353605e-06, + "loss": 1.515, + "step": 986 + }, + { + "epoch": 1.246212121212121, + "grad_norm": 3.049398422241211, + "learning_rate": 3.283442469597755e-06, + "loss": 1.5382, + "step": 987 + }, + { + "epoch": 1.2474747474747474, + "grad_norm": 3.0910539627075195, + "learning_rate": 3.2811038353601498e-06, + "loss": 1.5458, + "step": 988 + }, + { + "epoch": 1.2487373737373737, + "grad_norm": 3.002788782119751, + "learning_rate": 3.2787652011225446e-06, + "loss": 1.4668, + "step": 989 + }, + { + "epoch": 1.25, + "grad_norm": 2.9342103004455566, + "learning_rate": 3.2764265668849395e-06, + "loss": 1.5187, + "step": 990 + }, + { + "epoch": 1.25, + "eval_loss": 1.7082287073135376, + "eval_runtime": 37.1399, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 990 + }, + { + "epoch": 1.2512626262626263, + "grad_norm": 2.7596492767333984, + "learning_rate": 3.274087932647334e-06, + "loss": 1.5191, + "step": 991 + }, + { + "epoch": 1.2525252525252526, + "grad_norm": 2.949368953704834, + "learning_rate": 3.2717492984097287e-06, + "loss": 1.4776, + "step": 992 + }, + { + "epoch": 1.253787878787879, + "grad_norm": 2.7099363803863525, + "learning_rate": 3.2694106641721236e-06, + "loss": 1.455, + "step": 993 + }, + { + "epoch": 1.255050505050505, + "grad_norm": 2.8631885051727295, + "learning_rate": 3.267072029934519e-06, + "loss": 1.4394, + "step": 994 + }, + { + "epoch": 1.2563131313131313, + "grad_norm": 3.09389591217041, + "learning_rate": 3.2647333956969137e-06, + "loss": 1.4958, + "step": 995 + }, + { + "epoch": 1.2575757575757576, + "grad_norm": 3.108271360397339, + "learning_rate": 3.262394761459308e-06, + "loss": 1.5108, + "step": 996 + }, + { + "epoch": 1.2588383838383839, + "grad_norm": 3.079176902770996, + "learning_rate": 3.260056127221703e-06, + "loss": 1.4749, + "step": 997 + }, + { + "epoch": 1.2601010101010102, + "grad_norm": 3.0458056926727295, + "learning_rate": 3.257717492984098e-06, + "loss": 1.4911, + "step": 998 + }, + { + "epoch": 1.2613636363636362, + "grad_norm": 2.952151298522949, + "learning_rate": 3.2553788587464922e-06, + "loss": 1.4842, + "step": 999 + }, + { + "epoch": 1.2626262626262625, + "grad_norm": 2.7918665409088135, + "learning_rate": 3.253040224508887e-06, + "loss": 1.4684, + "step": 1000 + }, + { + "epoch": 1.2626262626262625, + "eval_loss": 1.7100834846496582, + "eval_runtime": 37.3594, + "eval_samples_per_second": 26.767, + "eval_steps_per_second": 1.124, + "step": 1000 + }, + { + "epoch": 1.2638888888888888, + "grad_norm": 2.798064947128296, + "learning_rate": 3.250701590271282e-06, + "loss": 1.4295, + "step": 1001 + }, + { + "epoch": 1.2651515151515151, + "grad_norm": 2.852234363555908, + "learning_rate": 3.2483629560336767e-06, + "loss": 1.5039, + "step": 1002 + }, + { + "epoch": 1.2664141414141414, + "grad_norm": 2.762645721435547, + "learning_rate": 3.246024321796071e-06, + "loss": 1.4365, + "step": 1003 + }, + { + "epoch": 1.2676767676767677, + "grad_norm": 2.944326162338257, + "learning_rate": 3.243685687558466e-06, + "loss": 1.4849, + "step": 1004 + }, + { + "epoch": 1.268939393939394, + "grad_norm": 2.9298877716064453, + "learning_rate": 3.241347053320861e-06, + "loss": 1.4397, + "step": 1005 + }, + { + "epoch": 1.2702020202020203, + "grad_norm": 3.1259615421295166, + "learning_rate": 3.2390084190832553e-06, + "loss": 1.4819, + "step": 1006 + }, + { + "epoch": 1.2714646464646464, + "grad_norm": 3.114792823791504, + "learning_rate": 3.23666978484565e-06, + "loss": 1.4728, + "step": 1007 + }, + { + "epoch": 1.2727272727272727, + "grad_norm": 2.9800360202789307, + "learning_rate": 3.234331150608045e-06, + "loss": 1.4426, + "step": 1008 + }, + { + "epoch": 1.273989898989899, + "grad_norm": 2.8487257957458496, + "learning_rate": 3.2319925163704402e-06, + "loss": 1.461, + "step": 1009 + }, + { + "epoch": 1.2752525252525253, + "grad_norm": 2.800760507583618, + "learning_rate": 3.229653882132835e-06, + "loss": 1.4764, + "step": 1010 + }, + { + "epoch": 1.2752525252525253, + "eval_loss": 1.701302409172058, + "eval_runtime": 36.9359, + "eval_samples_per_second": 27.074, + "eval_steps_per_second": 1.137, + "step": 1010 + }, + { + "epoch": 1.2765151515151514, + "grad_norm": 2.851012706756592, + "learning_rate": 3.2273152478952295e-06, + "loss": 1.5124, + "step": 1011 + }, + { + "epoch": 1.2777777777777777, + "grad_norm": 2.9498190879821777, + "learning_rate": 3.2249766136576243e-06, + "loss": 1.4789, + "step": 1012 + }, + { + "epoch": 1.279040404040404, + "grad_norm": 2.80710768699646, + "learning_rate": 3.222637979420019e-06, + "loss": 1.5592, + "step": 1013 + }, + { + "epoch": 1.2803030303030303, + "grad_norm": 2.9525952339172363, + "learning_rate": 3.220299345182414e-06, + "loss": 1.4941, + "step": 1014 + }, + { + "epoch": 1.2815656565656566, + "grad_norm": 2.9770407676696777, + "learning_rate": 3.2179607109448084e-06, + "loss": 1.4437, + "step": 1015 + }, + { + "epoch": 1.2828282828282829, + "grad_norm": 3.2553279399871826, + "learning_rate": 3.2156220767072033e-06, + "loss": 1.5248, + "step": 1016 + }, + { + "epoch": 1.2840909090909092, + "grad_norm": 3.0591554641723633, + "learning_rate": 3.213283442469598e-06, + "loss": 1.5704, + "step": 1017 + }, + { + "epoch": 1.2853535353535355, + "grad_norm": 2.776968002319336, + "learning_rate": 3.2109448082319925e-06, + "loss": 1.4239, + "step": 1018 + }, + { + "epoch": 1.2866161616161615, + "grad_norm": 2.8728222846984863, + "learning_rate": 3.2086061739943874e-06, + "loss": 1.5115, + "step": 1019 + }, + { + "epoch": 1.2878787878787878, + "grad_norm": 2.907308340072632, + "learning_rate": 3.2062675397567822e-06, + "loss": 1.4764, + "step": 1020 + }, + { + "epoch": 1.2878787878787878, + "eval_loss": 1.7016223669052124, + "eval_runtime": 37.0728, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 1020 + }, + { + "epoch": 1.2891414141414141, + "grad_norm": 3.0686309337615967, + "learning_rate": 3.203928905519177e-06, + "loss": 1.5175, + "step": 1021 + }, + { + "epoch": 1.2904040404040404, + "grad_norm": 2.8636863231658936, + "learning_rate": 3.2015902712815715e-06, + "loss": 1.5025, + "step": 1022 + }, + { + "epoch": 1.2916666666666667, + "grad_norm": 2.8539068698883057, + "learning_rate": 3.1992516370439663e-06, + "loss": 1.4774, + "step": 1023 + }, + { + "epoch": 1.2929292929292928, + "grad_norm": 2.9830520153045654, + "learning_rate": 3.196913002806361e-06, + "loss": 1.533, + "step": 1024 + }, + { + "epoch": 1.2941919191919191, + "grad_norm": 2.8806374073028564, + "learning_rate": 3.1945743685687564e-06, + "loss": 1.5298, + "step": 1025 + }, + { + "epoch": 1.2954545454545454, + "grad_norm": 2.932870626449585, + "learning_rate": 3.1922357343311513e-06, + "loss": 1.5028, + "step": 1026 + }, + { + "epoch": 1.2967171717171717, + "grad_norm": 2.889601945877075, + "learning_rate": 3.1898971000935457e-06, + "loss": 1.4534, + "step": 1027 + }, + { + "epoch": 1.297979797979798, + "grad_norm": 2.930546522140503, + "learning_rate": 3.1875584658559405e-06, + "loss": 1.5047, + "step": 1028 + }, + { + "epoch": 1.2992424242424243, + "grad_norm": 3.0944340229034424, + "learning_rate": 3.1852198316183354e-06, + "loss": 1.5198, + "step": 1029 + }, + { + "epoch": 1.3005050505050506, + "grad_norm": 2.8930089473724365, + "learning_rate": 3.18288119738073e-06, + "loss": 1.4655, + "step": 1030 + }, + { + "epoch": 1.3005050505050506, + "eval_loss": 1.7008370161056519, + "eval_runtime": 37.1263, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 1030 + }, + { + "epoch": 1.3017676767676767, + "grad_norm": 2.888303518295288, + "learning_rate": 3.1805425631431246e-06, + "loss": 1.4249, + "step": 1031 + }, + { + "epoch": 1.303030303030303, + "grad_norm": 3.0144314765930176, + "learning_rate": 3.1782039289055195e-06, + "loss": 1.4963, + "step": 1032 + }, + { + "epoch": 1.3042929292929293, + "grad_norm": 3.0118296146392822, + "learning_rate": 3.1758652946679143e-06, + "loss": 1.5119, + "step": 1033 + }, + { + "epoch": 1.3055555555555556, + "grad_norm": 2.8148205280303955, + "learning_rate": 3.1735266604303087e-06, + "loss": 1.4921, + "step": 1034 + }, + { + "epoch": 1.3068181818181819, + "grad_norm": 2.811166286468506, + "learning_rate": 3.1711880261927036e-06, + "loss": 1.4689, + "step": 1035 + }, + { + "epoch": 1.308080808080808, + "grad_norm": 2.924088478088379, + "learning_rate": 3.1688493919550984e-06, + "loss": 1.5682, + "step": 1036 + }, + { + "epoch": 1.3093434343434343, + "grad_norm": 2.905988931655884, + "learning_rate": 3.166510757717493e-06, + "loss": 1.5104, + "step": 1037 + }, + { + "epoch": 1.3106060606060606, + "grad_norm": 2.972963571548462, + "learning_rate": 3.1641721234798877e-06, + "loss": 1.4543, + "step": 1038 + }, + { + "epoch": 1.3118686868686869, + "grad_norm": 2.8034229278564453, + "learning_rate": 3.1618334892422825e-06, + "loss": 1.45, + "step": 1039 + }, + { + "epoch": 1.3131313131313131, + "grad_norm": 2.8977267742156982, + "learning_rate": 3.159494855004678e-06, + "loss": 1.4531, + "step": 1040 + }, + { + "epoch": 1.3131313131313131, + "eval_loss": 1.7027597427368164, + "eval_runtime": 37.0629, + "eval_samples_per_second": 26.981, + "eval_steps_per_second": 1.133, + "step": 1040 + }, + { + "epoch": 1.3143939393939394, + "grad_norm": 3.106973886489868, + "learning_rate": 3.1571562207670726e-06, + "loss": 1.4723, + "step": 1041 + }, + { + "epoch": 1.3156565656565657, + "grad_norm": 2.7628490924835205, + "learning_rate": 3.154817586529467e-06, + "loss": 1.4902, + "step": 1042 + }, + { + "epoch": 1.316919191919192, + "grad_norm": 3.119751453399658, + "learning_rate": 3.152478952291862e-06, + "loss": 1.5326, + "step": 1043 + }, + { + "epoch": 1.3181818181818181, + "grad_norm": 3.1470835208892822, + "learning_rate": 3.1501403180542568e-06, + "loss": 1.4626, + "step": 1044 + }, + { + "epoch": 1.3194444444444444, + "grad_norm": 2.866586208343506, + "learning_rate": 3.1478016838166516e-06, + "loss": 1.5006, + "step": 1045 + }, + { + "epoch": 1.3207070707070707, + "grad_norm": 2.878464698791504, + "learning_rate": 3.145463049579046e-06, + "loss": 1.4727, + "step": 1046 + }, + { + "epoch": 1.321969696969697, + "grad_norm": 2.8567826747894287, + "learning_rate": 3.143124415341441e-06, + "loss": 1.4202, + "step": 1047 + }, + { + "epoch": 1.3232323232323233, + "grad_norm": 2.8718442916870117, + "learning_rate": 3.1407857811038357e-06, + "loss": 1.473, + "step": 1048 + }, + { + "epoch": 1.3244949494949494, + "grad_norm": 3.003040313720703, + "learning_rate": 3.13844714686623e-06, + "loss": 1.4784, + "step": 1049 + }, + { + "epoch": 1.3257575757575757, + "grad_norm": 3.080288887023926, + "learning_rate": 3.136108512628625e-06, + "loss": 1.4514, + "step": 1050 + }, + { + "epoch": 1.3257575757575757, + "eval_loss": 1.704396367073059, + "eval_runtime": 37.0757, + "eval_samples_per_second": 26.972, + "eval_steps_per_second": 1.133, + "step": 1050 + }, + { + "epoch": 1.327020202020202, + "grad_norm": 2.8699166774749756, + "learning_rate": 3.13376987839102e-06, + "loss": 1.4966, + "step": 1051 + }, + { + "epoch": 1.3282828282828283, + "grad_norm": 2.979238271713257, + "learning_rate": 3.1314312441534146e-06, + "loss": 1.4422, + "step": 1052 + }, + { + "epoch": 1.3295454545454546, + "grad_norm": 2.8612024784088135, + "learning_rate": 3.129092609915809e-06, + "loss": 1.4982, + "step": 1053 + }, + { + "epoch": 1.3308080808080809, + "grad_norm": 2.7759032249450684, + "learning_rate": 3.126753975678204e-06, + "loss": 1.413, + "step": 1054 + }, + { + "epoch": 1.3320707070707072, + "grad_norm": 2.9772675037384033, + "learning_rate": 3.124415341440599e-06, + "loss": 1.4789, + "step": 1055 + }, + { + "epoch": 1.3333333333333333, + "grad_norm": 2.9382526874542236, + "learning_rate": 3.122076707202994e-06, + "loss": 1.4547, + "step": 1056 + }, + { + "epoch": 1.3345959595959596, + "grad_norm": 2.85050630569458, + "learning_rate": 3.119738072965389e-06, + "loss": 1.4193, + "step": 1057 + }, + { + "epoch": 1.3358585858585859, + "grad_norm": 3.1148695945739746, + "learning_rate": 3.1173994387277833e-06, + "loss": 1.4527, + "step": 1058 + }, + { + "epoch": 1.3371212121212122, + "grad_norm": 3.3906028270721436, + "learning_rate": 3.115060804490178e-06, + "loss": 1.5273, + "step": 1059 + }, + { + "epoch": 1.3383838383838385, + "grad_norm": 2.9665908813476562, + "learning_rate": 3.112722170252573e-06, + "loss": 1.5237, + "step": 1060 + }, + { + "epoch": 1.3383838383838385, + "eval_loss": 1.7004541158676147, + "eval_runtime": 37.081, + "eval_samples_per_second": 26.968, + "eval_steps_per_second": 1.133, + "step": 1060 + }, + { + "epoch": 1.3396464646464645, + "grad_norm": 2.9809224605560303, + "learning_rate": 3.1103835360149674e-06, + "loss": 1.498, + "step": 1061 + }, + { + "epoch": 1.3409090909090908, + "grad_norm": 2.95265793800354, + "learning_rate": 3.1080449017773622e-06, + "loss": 1.511, + "step": 1062 + }, + { + "epoch": 1.3421717171717171, + "grad_norm": 3.0814359188079834, + "learning_rate": 3.105706267539757e-06, + "loss": 1.4448, + "step": 1063 + }, + { + "epoch": 1.3434343434343434, + "grad_norm": 2.8267405033111572, + "learning_rate": 3.103367633302152e-06, + "loss": 1.4648, + "step": 1064 + }, + { + "epoch": 1.3446969696969697, + "grad_norm": 3.08182954788208, + "learning_rate": 3.1010289990645463e-06, + "loss": 1.5245, + "step": 1065 + }, + { + "epoch": 1.345959595959596, + "grad_norm": 3.183209180831909, + "learning_rate": 3.098690364826941e-06, + "loss": 1.4388, + "step": 1066 + }, + { + "epoch": 1.3472222222222223, + "grad_norm": 3.2121284008026123, + "learning_rate": 3.096351730589336e-06, + "loss": 1.4328, + "step": 1067 + }, + { + "epoch": 1.3484848484848486, + "grad_norm": 3.093921661376953, + "learning_rate": 3.0940130963517304e-06, + "loss": 1.5176, + "step": 1068 + }, + { + "epoch": 1.3497474747474747, + "grad_norm": 2.977710008621216, + "learning_rate": 3.0916744621141253e-06, + "loss": 1.4153, + "step": 1069 + }, + { + "epoch": 1.351010101010101, + "grad_norm": 3.119001865386963, + "learning_rate": 3.08933582787652e-06, + "loss": 1.4879, + "step": 1070 + }, + { + "epoch": 1.351010101010101, + "eval_loss": 1.7006300687789917, + "eval_runtime": 37.0959, + "eval_samples_per_second": 26.957, + "eval_steps_per_second": 1.132, + "step": 1070 + }, + { + "epoch": 1.3522727272727273, + "grad_norm": 3.026405096054077, + "learning_rate": 3.0869971936389154e-06, + "loss": 1.483, + "step": 1071 + }, + { + "epoch": 1.3535353535353536, + "grad_norm": 2.770728826522827, + "learning_rate": 3.0846585594013102e-06, + "loss": 1.5175, + "step": 1072 + }, + { + "epoch": 1.3547979797979797, + "grad_norm": 2.7930657863616943, + "learning_rate": 3.0823199251637047e-06, + "loss": 1.4833, + "step": 1073 + }, + { + "epoch": 1.356060606060606, + "grad_norm": 2.9128804206848145, + "learning_rate": 3.0799812909260995e-06, + "loss": 1.5016, + "step": 1074 + }, + { + "epoch": 1.3573232323232323, + "grad_norm": 3.0432205200195312, + "learning_rate": 3.0776426566884943e-06, + "loss": 1.5211, + "step": 1075 + }, + { + "epoch": 1.3585858585858586, + "grad_norm": 2.9197583198547363, + "learning_rate": 3.075304022450889e-06, + "loss": 1.5065, + "step": 1076 + }, + { + "epoch": 1.3598484848484849, + "grad_norm": 2.944986581802368, + "learning_rate": 3.0729653882132836e-06, + "loss": 1.5015, + "step": 1077 + }, + { + "epoch": 1.3611111111111112, + "grad_norm": 2.7553489208221436, + "learning_rate": 3.0706267539756784e-06, + "loss": 1.4692, + "step": 1078 + }, + { + "epoch": 1.3623737373737375, + "grad_norm": 2.7607858180999756, + "learning_rate": 3.0682881197380733e-06, + "loss": 1.4494, + "step": 1079 + }, + { + "epoch": 1.3636363636363638, + "grad_norm": 2.7896173000335693, + "learning_rate": 3.0659494855004677e-06, + "loss": 1.4625, + "step": 1080 + }, + { + "epoch": 1.3636363636363638, + "eval_loss": 1.6998732089996338, + "eval_runtime": 37.1079, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 1080 + }, + { + "epoch": 1.3648989898989898, + "grad_norm": 3.020059823989868, + "learning_rate": 3.0636108512628626e-06, + "loss": 1.5331, + "step": 1081 + }, + { + "epoch": 1.3661616161616161, + "grad_norm": 2.805593729019165, + "learning_rate": 3.0612722170252574e-06, + "loss": 1.4867, + "step": 1082 + }, + { + "epoch": 1.3674242424242424, + "grad_norm": 2.965871810913086, + "learning_rate": 3.0589335827876522e-06, + "loss": 1.4494, + "step": 1083 + }, + { + "epoch": 1.3686868686868687, + "grad_norm": 2.8010783195495605, + "learning_rate": 3.0565949485500467e-06, + "loss": 1.4484, + "step": 1084 + }, + { + "epoch": 1.369949494949495, + "grad_norm": 2.985258102416992, + "learning_rate": 3.0542563143124415e-06, + "loss": 1.5029, + "step": 1085 + }, + { + "epoch": 1.371212121212121, + "grad_norm": 2.7528188228607178, + "learning_rate": 3.0519176800748368e-06, + "loss": 1.4695, + "step": 1086 + }, + { + "epoch": 1.3724747474747474, + "grad_norm": 2.9123780727386475, + "learning_rate": 3.0495790458372316e-06, + "loss": 1.5527, + "step": 1087 + }, + { + "epoch": 1.3737373737373737, + "grad_norm": 2.8589160442352295, + "learning_rate": 3.0472404115996265e-06, + "loss": 1.5104, + "step": 1088 + }, + { + "epoch": 1.375, + "grad_norm": 2.853546142578125, + "learning_rate": 3.044901777362021e-06, + "loss": 1.4498, + "step": 1089 + }, + { + "epoch": 1.3762626262626263, + "grad_norm": 2.7989797592163086, + "learning_rate": 3.0425631431244157e-06, + "loss": 1.4494, + "step": 1090 + }, + { + "epoch": 1.3762626262626263, + "eval_loss": 1.6923459768295288, + "eval_runtime": 37.2917, + "eval_samples_per_second": 26.816, + "eval_steps_per_second": 1.126, + "step": 1090 + }, + { + "epoch": 1.3775252525252526, + "grad_norm": 2.8750412464141846, + "learning_rate": 3.0402245088868106e-06, + "loss": 1.5026, + "step": 1091 + }, + { + "epoch": 1.378787878787879, + "grad_norm": 2.8185842037200928, + "learning_rate": 3.037885874649205e-06, + "loss": 1.4859, + "step": 1092 + }, + { + "epoch": 1.380050505050505, + "grad_norm": 2.736332893371582, + "learning_rate": 3.0355472404116e-06, + "loss": 1.4944, + "step": 1093 + }, + { + "epoch": 1.3813131313131313, + "grad_norm": 2.93930983543396, + "learning_rate": 3.0332086061739947e-06, + "loss": 1.4975, + "step": 1094 + }, + { + "epoch": 1.3825757575757576, + "grad_norm": 2.964146852493286, + "learning_rate": 3.0308699719363895e-06, + "loss": 1.4861, + "step": 1095 + }, + { + "epoch": 1.3838383838383839, + "grad_norm": 2.8321545124053955, + "learning_rate": 3.028531337698784e-06, + "loss": 1.4535, + "step": 1096 + }, + { + "epoch": 1.3851010101010102, + "grad_norm": 2.921372413635254, + "learning_rate": 3.0261927034611788e-06, + "loss": 1.5242, + "step": 1097 + }, + { + "epoch": 1.3863636363636362, + "grad_norm": 2.7309911251068115, + "learning_rate": 3.0238540692235736e-06, + "loss": 1.4105, + "step": 1098 + }, + { + "epoch": 1.3876262626262625, + "grad_norm": 2.768899917602539, + "learning_rate": 3.021515434985968e-06, + "loss": 1.4183, + "step": 1099 + }, + { + "epoch": 1.3888888888888888, + "grad_norm": 2.883399248123169, + "learning_rate": 3.019176800748363e-06, + "loss": 1.5307, + "step": 1100 + }, + { + "epoch": 1.3888888888888888, + "eval_loss": 1.6948082447052002, + "eval_runtime": 37.488, + "eval_samples_per_second": 26.675, + "eval_steps_per_second": 1.12, + "step": 1100 + }, + { + "epoch": 1.3901515151515151, + "grad_norm": 2.8610544204711914, + "learning_rate": 3.016838166510758e-06, + "loss": 1.5473, + "step": 1101 + }, + { + "epoch": 1.3914141414141414, + "grad_norm": 2.762694835662842, + "learning_rate": 3.014499532273153e-06, + "loss": 1.4179, + "step": 1102 + }, + { + "epoch": 1.3926767676767677, + "grad_norm": 2.975828170776367, + "learning_rate": 3.012160898035548e-06, + "loss": 1.4959, + "step": 1103 + }, + { + "epoch": 1.393939393939394, + "grad_norm": 3.0633926391601562, + "learning_rate": 3.0098222637979422e-06, + "loss": 1.4979, + "step": 1104 + }, + { + "epoch": 1.3952020202020203, + "grad_norm": 2.972900867462158, + "learning_rate": 3.007483629560337e-06, + "loss": 1.5311, + "step": 1105 + }, + { + "epoch": 1.3964646464646464, + "grad_norm": 2.936033248901367, + "learning_rate": 3.005144995322732e-06, + "loss": 1.4856, + "step": 1106 + }, + { + "epoch": 1.3977272727272727, + "grad_norm": 3.0375452041625977, + "learning_rate": 3.0028063610851268e-06, + "loss": 1.5001, + "step": 1107 + }, + { + "epoch": 1.398989898989899, + "grad_norm": 2.998283624649048, + "learning_rate": 3.000467726847521e-06, + "loss": 1.5219, + "step": 1108 + }, + { + "epoch": 1.4002525252525253, + "grad_norm": 2.831354856491089, + "learning_rate": 2.998129092609916e-06, + "loss": 1.4498, + "step": 1109 + }, + { + "epoch": 1.4015151515151514, + "grad_norm": 2.975390911102295, + "learning_rate": 2.995790458372311e-06, + "loss": 1.4655, + "step": 1110 + }, + { + "epoch": 1.4015151515151514, + "eval_loss": 1.6940916776657104, + "eval_runtime": 36.9469, + "eval_samples_per_second": 27.066, + "eval_steps_per_second": 1.137, + "step": 1110 + }, + { + "epoch": 1.4027777777777777, + "grad_norm": 2.9645442962646484, + "learning_rate": 2.9934518241347053e-06, + "loss": 1.4568, + "step": 1111 + }, + { + "epoch": 1.404040404040404, + "grad_norm": 2.873746633529663, + "learning_rate": 2.9911131898971e-06, + "loss": 1.4483, + "step": 1112 + }, + { + "epoch": 1.4053030303030303, + "grad_norm": 3.1041767597198486, + "learning_rate": 2.988774555659495e-06, + "loss": 1.4914, + "step": 1113 + }, + { + "epoch": 1.4065656565656566, + "grad_norm": 3.0122971534729004, + "learning_rate": 2.98643592142189e-06, + "loss": 1.4494, + "step": 1114 + }, + { + "epoch": 1.4078282828282829, + "grad_norm": 2.842233180999756, + "learning_rate": 2.9840972871842842e-06, + "loss": 1.4274, + "step": 1115 + }, + { + "epoch": 1.4090909090909092, + "grad_norm": 3.058265447616577, + "learning_rate": 2.9817586529466795e-06, + "loss": 1.4637, + "step": 1116 + }, + { + "epoch": 1.4103535353535355, + "grad_norm": 3.06136155128479, + "learning_rate": 2.9794200187090744e-06, + "loss": 1.4791, + "step": 1117 + }, + { + "epoch": 1.4116161616161615, + "grad_norm": 2.8931167125701904, + "learning_rate": 2.977081384471469e-06, + "loss": 1.4476, + "step": 1118 + }, + { + "epoch": 1.4128787878787878, + "grad_norm": 2.8025107383728027, + "learning_rate": 2.974742750233864e-06, + "loss": 1.4891, + "step": 1119 + }, + { + "epoch": 1.4141414141414141, + "grad_norm": 3.1213603019714355, + "learning_rate": 2.9724041159962585e-06, + "loss": 1.5513, + "step": 1120 + }, + { + "epoch": 1.4141414141414141, + "eval_loss": 1.6972154378890991, + "eval_runtime": 37.0863, + "eval_samples_per_second": 26.964, + "eval_steps_per_second": 1.132, + "step": 1120 + }, + { + "epoch": 1.4154040404040404, + "grad_norm": 3.126791477203369, + "learning_rate": 2.9700654817586533e-06, + "loss": 1.4952, + "step": 1121 + }, + { + "epoch": 1.4166666666666667, + "grad_norm": 2.959385395050049, + "learning_rate": 2.967726847521048e-06, + "loss": 1.5098, + "step": 1122 + }, + { + "epoch": 1.4179292929292928, + "grad_norm": 2.907566547393799, + "learning_rate": 2.9653882132834426e-06, + "loss": 1.4959, + "step": 1123 + }, + { + "epoch": 1.4191919191919191, + "grad_norm": 2.9181833267211914, + "learning_rate": 2.9630495790458374e-06, + "loss": 1.4607, + "step": 1124 + }, + { + "epoch": 1.4204545454545454, + "grad_norm": 2.8703668117523193, + "learning_rate": 2.9607109448082323e-06, + "loss": 1.5215, + "step": 1125 + }, + { + "epoch": 1.4217171717171717, + "grad_norm": 3.166255235671997, + "learning_rate": 2.958372310570627e-06, + "loss": 1.4451, + "step": 1126 + }, + { + "epoch": 1.422979797979798, + "grad_norm": 2.863943576812744, + "learning_rate": 2.9560336763330215e-06, + "loss": 1.4481, + "step": 1127 + }, + { + "epoch": 1.4242424242424243, + "grad_norm": 2.975482702255249, + "learning_rate": 2.9536950420954164e-06, + "loss": 1.5118, + "step": 1128 + }, + { + "epoch": 1.4255050505050506, + "grad_norm": 3.077357769012451, + "learning_rate": 2.951356407857811e-06, + "loss": 1.4678, + "step": 1129 + }, + { + "epoch": 1.4267676767676767, + "grad_norm": 3.0375332832336426, + "learning_rate": 2.9490177736202056e-06, + "loss": 1.5036, + "step": 1130 + }, + { + "epoch": 1.4267676767676767, + "eval_loss": 1.6973206996917725, + "eval_runtime": 37.1448, + "eval_samples_per_second": 26.922, + "eval_steps_per_second": 1.131, + "step": 1130 + }, + { + "epoch": 1.428030303030303, + "grad_norm": 3.022068500518799, + "learning_rate": 2.9466791393826005e-06, + "loss": 1.442, + "step": 1131 + }, + { + "epoch": 1.4292929292929293, + "grad_norm": 2.9089183807373047, + "learning_rate": 2.9443405051449957e-06, + "loss": 1.4627, + "step": 1132 + }, + { + "epoch": 1.4305555555555556, + "grad_norm": 2.840149402618408, + "learning_rate": 2.9420018709073906e-06, + "loss": 1.4745, + "step": 1133 + }, + { + "epoch": 1.4318181818181819, + "grad_norm": 2.938650608062744, + "learning_rate": 2.9396632366697854e-06, + "loss": 1.4795, + "step": 1134 + }, + { + "epoch": 1.433080808080808, + "grad_norm": 2.9602437019348145, + "learning_rate": 2.93732460243218e-06, + "loss": 1.4228, + "step": 1135 + }, + { + "epoch": 1.4343434343434343, + "grad_norm": 2.929257869720459, + "learning_rate": 2.9349859681945747e-06, + "loss": 1.5016, + "step": 1136 + }, + { + "epoch": 1.4356060606060606, + "grad_norm": 2.8120505809783936, + "learning_rate": 2.9326473339569695e-06, + "loss": 1.474, + "step": 1137 + }, + { + "epoch": 1.4368686868686869, + "grad_norm": 3.0380513668060303, + "learning_rate": 2.9303086997193644e-06, + "loss": 1.4714, + "step": 1138 + }, + { + "epoch": 1.4381313131313131, + "grad_norm": 2.91106915473938, + "learning_rate": 2.9279700654817588e-06, + "loss": 1.4558, + "step": 1139 + }, + { + "epoch": 1.4393939393939394, + "grad_norm": 2.9970154762268066, + "learning_rate": 2.9256314312441536e-06, + "loss": 1.4895, + "step": 1140 + }, + { + "epoch": 1.4393939393939394, + "eval_loss": 1.6916719675064087, + "eval_runtime": 37.1538, + "eval_samples_per_second": 26.915, + "eval_steps_per_second": 1.13, + "step": 1140 + }, + { + "epoch": 1.4406565656565657, + "grad_norm": 3.041616201400757, + "learning_rate": 2.9232927970065485e-06, + "loss": 1.5177, + "step": 1141 + }, + { + "epoch": 1.441919191919192, + "grad_norm": 2.875385046005249, + "learning_rate": 2.920954162768943e-06, + "loss": 1.4697, + "step": 1142 + }, + { + "epoch": 1.4431818181818181, + "grad_norm": 2.7117881774902344, + "learning_rate": 2.9186155285313377e-06, + "loss": 1.4809, + "step": 1143 + }, + { + "epoch": 1.4444444444444444, + "grad_norm": 2.912123203277588, + "learning_rate": 2.9162768942937326e-06, + "loss": 1.4032, + "step": 1144 + }, + { + "epoch": 1.4457070707070707, + "grad_norm": 2.798440933227539, + "learning_rate": 2.9139382600561274e-06, + "loss": 1.4442, + "step": 1145 + }, + { + "epoch": 1.446969696969697, + "grad_norm": 2.9299135208129883, + "learning_rate": 2.911599625818522e-06, + "loss": 1.4412, + "step": 1146 + }, + { + "epoch": 1.4482323232323233, + "grad_norm": 2.988434076309204, + "learning_rate": 2.909260991580917e-06, + "loss": 1.4473, + "step": 1147 + }, + { + "epoch": 1.4494949494949494, + "grad_norm": 3.0697057247161865, + "learning_rate": 2.906922357343312e-06, + "loss": 1.5151, + "step": 1148 + }, + { + "epoch": 1.4507575757575757, + "grad_norm": 2.9009311199188232, + "learning_rate": 2.9045837231057068e-06, + "loss": 1.4434, + "step": 1149 + }, + { + "epoch": 1.452020202020202, + "grad_norm": 2.9949986934661865, + "learning_rate": 2.902245088868101e-06, + "loss": 1.5436, + "step": 1150 + }, + { + "epoch": 1.452020202020202, + "eval_loss": 1.6880143880844116, + "eval_runtime": 37.1212, + "eval_samples_per_second": 26.939, + "eval_steps_per_second": 1.131, + "step": 1150 + }, + { + "epoch": 1.4532828282828283, + "grad_norm": 2.7827775478363037, + "learning_rate": 2.899906454630496e-06, + "loss": 1.4284, + "step": 1151 + }, + { + "epoch": 1.4545454545454546, + "grad_norm": 2.8879382610321045, + "learning_rate": 2.897567820392891e-06, + "loss": 1.4772, + "step": 1152 + }, + { + "epoch": 1.4558080808080809, + "grad_norm": 3.0119130611419678, + "learning_rate": 2.8952291861552857e-06, + "loss": 1.4546, + "step": 1153 + }, + { + "epoch": 1.4570707070707072, + "grad_norm": 2.9002952575683594, + "learning_rate": 2.89289055191768e-06, + "loss": 1.4898, + "step": 1154 + }, + { + "epoch": 1.4583333333333333, + "grad_norm": 2.8384580612182617, + "learning_rate": 2.890551917680075e-06, + "loss": 1.4685, + "step": 1155 + }, + { + "epoch": 1.4595959595959596, + "grad_norm": 2.9149580001831055, + "learning_rate": 2.88821328344247e-06, + "loss": 1.5133, + "step": 1156 + }, + { + "epoch": 1.4608585858585859, + "grad_norm": 3.041875123977661, + "learning_rate": 2.8858746492048647e-06, + "loss": 1.5038, + "step": 1157 + }, + { + "epoch": 1.4621212121212122, + "grad_norm": 2.9623730182647705, + "learning_rate": 2.883536014967259e-06, + "loss": 1.4776, + "step": 1158 + }, + { + "epoch": 1.4633838383838385, + "grad_norm": 2.935295343399048, + "learning_rate": 2.881197380729654e-06, + "loss": 1.4166, + "step": 1159 + }, + { + "epoch": 1.4646464646464645, + "grad_norm": 2.9556751251220703, + "learning_rate": 2.8788587464920488e-06, + "loss": 1.5132, + "step": 1160 + }, + { + "epoch": 1.4646464646464645, + "eval_loss": 1.69142746925354, + "eval_runtime": 37.1425, + "eval_samples_per_second": 26.923, + "eval_steps_per_second": 1.131, + "step": 1160 + }, + { + "epoch": 1.4659090909090908, + "grad_norm": 2.992587089538574, + "learning_rate": 2.876520112254443e-06, + "loss": 1.4695, + "step": 1161 + }, + { + "epoch": 1.4671717171717171, + "grad_norm": 3.005990982055664, + "learning_rate": 2.8741814780168385e-06, + "loss": 1.5122, + "step": 1162 + }, + { + "epoch": 1.4684343434343434, + "grad_norm": 2.9341115951538086, + "learning_rate": 2.8718428437792333e-06, + "loss": 1.4916, + "step": 1163 + }, + { + "epoch": 1.4696969696969697, + "grad_norm": 2.9665164947509766, + "learning_rate": 2.869504209541628e-06, + "loss": 1.4827, + "step": 1164 + }, + { + "epoch": 1.470959595959596, + "grad_norm": 2.9972221851348877, + "learning_rate": 2.867165575304023e-06, + "loss": 1.4845, + "step": 1165 + }, + { + "epoch": 1.4722222222222223, + "grad_norm": 2.981387138366699, + "learning_rate": 2.8648269410664174e-06, + "loss": 1.4374, + "step": 1166 + }, + { + "epoch": 1.4734848484848486, + "grad_norm": 3.04587459564209, + "learning_rate": 2.8624883068288123e-06, + "loss": 1.431, + "step": 1167 + }, + { + "epoch": 1.4747474747474747, + "grad_norm": 3.194389820098877, + "learning_rate": 2.860149672591207e-06, + "loss": 1.4847, + "step": 1168 + }, + { + "epoch": 1.476010101010101, + "grad_norm": 2.949721574783325, + "learning_rate": 2.857811038353602e-06, + "loss": 1.4596, + "step": 1169 + }, + { + "epoch": 1.4772727272727273, + "grad_norm": 3.0032384395599365, + "learning_rate": 2.8554724041159964e-06, + "loss": 1.4821, + "step": 1170 + }, + { + "epoch": 1.4772727272727273, + "eval_loss": 1.6863130331039429, + "eval_runtime": 37.1386, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 1170 + }, + { + "epoch": 1.4785353535353536, + "grad_norm": 2.943141460418701, + "learning_rate": 2.853133769878391e-06, + "loss": 1.4998, + "step": 1171 + }, + { + "epoch": 1.4797979797979797, + "grad_norm": 3.0163350105285645, + "learning_rate": 2.850795135640786e-06, + "loss": 1.4565, + "step": 1172 + }, + { + "epoch": 1.481060606060606, + "grad_norm": 2.9713706970214844, + "learning_rate": 2.8484565014031805e-06, + "loss": 1.4663, + "step": 1173 + }, + { + "epoch": 1.4823232323232323, + "grad_norm": 2.9195785522460938, + "learning_rate": 2.8461178671655753e-06, + "loss": 1.4594, + "step": 1174 + }, + { + "epoch": 1.4835858585858586, + "grad_norm": 3.127619504928589, + "learning_rate": 2.84377923292797e-06, + "loss": 1.4742, + "step": 1175 + }, + { + "epoch": 1.4848484848484849, + "grad_norm": 3.0264782905578613, + "learning_rate": 2.841440598690365e-06, + "loss": 1.4366, + "step": 1176 + }, + { + "epoch": 1.4861111111111112, + "grad_norm": 3.0541677474975586, + "learning_rate": 2.8391019644527603e-06, + "loss": 1.4629, + "step": 1177 + }, + { + "epoch": 1.4873737373737375, + "grad_norm": 2.8967537879943848, + "learning_rate": 2.8367633302151547e-06, + "loss": 1.4281, + "step": 1178 + }, + { + "epoch": 1.4886363636363638, + "grad_norm": 2.975773811340332, + "learning_rate": 2.8344246959775495e-06, + "loss": 1.4272, + "step": 1179 + }, + { + "epoch": 1.4898989898989898, + "grad_norm": 2.9768712520599365, + "learning_rate": 2.8320860617399444e-06, + "loss": 1.453, + "step": 1180 + }, + { + "epoch": 1.4898989898989898, + "eval_loss": 1.6810754537582397, + "eval_runtime": 37.1659, + "eval_samples_per_second": 26.906, + "eval_steps_per_second": 1.13, + "step": 1180 + }, + { + "epoch": 1.4911616161616161, + "grad_norm": 2.8897762298583984, + "learning_rate": 2.829747427502339e-06, + "loss": 1.4577, + "step": 1181 + }, + { + "epoch": 1.4924242424242424, + "grad_norm": 2.848259925842285, + "learning_rate": 2.8274087932647336e-06, + "loss": 1.4323, + "step": 1182 + }, + { + "epoch": 1.4936868686868687, + "grad_norm": 2.9275119304656982, + "learning_rate": 2.8250701590271285e-06, + "loss": 1.4712, + "step": 1183 + }, + { + "epoch": 1.494949494949495, + "grad_norm": 2.8643906116485596, + "learning_rate": 2.8227315247895233e-06, + "loss": 1.4375, + "step": 1184 + }, + { + "epoch": 1.496212121212121, + "grad_norm": 3.0250704288482666, + "learning_rate": 2.8203928905519177e-06, + "loss": 1.5482, + "step": 1185 + }, + { + "epoch": 1.4974747474747474, + "grad_norm": 3.1019556522369385, + "learning_rate": 2.8180542563143126e-06, + "loss": 1.4618, + "step": 1186 + }, + { + "epoch": 1.4987373737373737, + "grad_norm": 3.1480300426483154, + "learning_rate": 2.8157156220767074e-06, + "loss": 1.4152, + "step": 1187 + }, + { + "epoch": 1.5, + "grad_norm": 2.82671856880188, + "learning_rate": 2.8133769878391023e-06, + "loss": 1.4692, + "step": 1188 + }, + { + "epoch": 1.5012626262626263, + "grad_norm": 3.2128899097442627, + "learning_rate": 2.8110383536014967e-06, + "loss": 1.4087, + "step": 1189 + }, + { + "epoch": 1.5025252525252526, + "grad_norm": 3.1939210891723633, + "learning_rate": 2.8086997193638915e-06, + "loss": 1.4919, + "step": 1190 + }, + { + "epoch": 1.5025252525252526, + "eval_loss": 1.6861544847488403, + "eval_runtime": 37.3328, + "eval_samples_per_second": 26.786, + "eval_steps_per_second": 1.125, + "step": 1190 + }, + { + "epoch": 1.503787878787879, + "grad_norm": 3.022059679031372, + "learning_rate": 2.8063610851262864e-06, + "loss": 1.4865, + "step": 1191 + }, + { + "epoch": 1.5050505050505052, + "grad_norm": 3.0068860054016113, + "learning_rate": 2.804022450888681e-06, + "loss": 1.478, + "step": 1192 + }, + { + "epoch": 1.5063131313131313, + "grad_norm": 3.2898383140563965, + "learning_rate": 2.801683816651076e-06, + "loss": 1.4773, + "step": 1193 + }, + { + "epoch": 1.5075757575757576, + "grad_norm": 3.090339422225952, + "learning_rate": 2.799345182413471e-06, + "loss": 1.4012, + "step": 1194 + }, + { + "epoch": 1.5088383838383839, + "grad_norm": 2.977215528488159, + "learning_rate": 2.7970065481758657e-06, + "loss": 1.4651, + "step": 1195 + }, + { + "epoch": 1.51010101010101, + "grad_norm": 3.1349570751190186, + "learning_rate": 2.7946679139382606e-06, + "loss": 1.5572, + "step": 1196 + }, + { + "epoch": 1.5113636363636362, + "grad_norm": 2.8715298175811768, + "learning_rate": 2.792329279700655e-06, + "loss": 1.4558, + "step": 1197 + }, + { + "epoch": 1.5126262626262625, + "grad_norm": 2.9848744869232178, + "learning_rate": 2.78999064546305e-06, + "loss": 1.4383, + "step": 1198 + }, + { + "epoch": 1.5138888888888888, + "grad_norm": 3.0774083137512207, + "learning_rate": 2.7876520112254447e-06, + "loss": 1.4607, + "step": 1199 + }, + { + "epoch": 1.5151515151515151, + "grad_norm": 2.9721343517303467, + "learning_rate": 2.785313376987839e-06, + "loss": 1.4341, + "step": 1200 + }, + { + "epoch": 1.5151515151515151, + "eval_loss": 1.6859924793243408, + "eval_runtime": 37.4589, + "eval_samples_per_second": 26.696, + "eval_steps_per_second": 1.121, + "step": 1200 + }, + { + "epoch": 1.5164141414141414, + "grad_norm": 3.041095733642578, + "learning_rate": 2.782974742750234e-06, + "loss": 1.4736, + "step": 1201 + }, + { + "epoch": 1.5176767676767677, + "grad_norm": 2.8400142192840576, + "learning_rate": 2.780636108512629e-06, + "loss": 1.4169, + "step": 1202 + }, + { + "epoch": 1.518939393939394, + "grad_norm": 2.785592555999756, + "learning_rate": 2.7782974742750236e-06, + "loss": 1.4553, + "step": 1203 + }, + { + "epoch": 1.5202020202020203, + "grad_norm": 2.75209641456604, + "learning_rate": 2.775958840037418e-06, + "loss": 1.4267, + "step": 1204 + }, + { + "epoch": 1.5214646464646466, + "grad_norm": 2.927558183670044, + "learning_rate": 2.773620205799813e-06, + "loss": 1.4377, + "step": 1205 + }, + { + "epoch": 1.5227272727272727, + "grad_norm": 2.9335508346557617, + "learning_rate": 2.7712815715622077e-06, + "loss": 1.4295, + "step": 1206 + }, + { + "epoch": 1.523989898989899, + "grad_norm": 2.7812745571136475, + "learning_rate": 2.7689429373246026e-06, + "loss": 1.4029, + "step": 1207 + }, + { + "epoch": 1.5252525252525253, + "grad_norm": 3.230062961578369, + "learning_rate": 2.766604303086998e-06, + "loss": 1.539, + "step": 1208 + }, + { + "epoch": 1.5265151515151514, + "grad_norm": 3.1244332790374756, + "learning_rate": 2.7642656688493923e-06, + "loss": 1.4829, + "step": 1209 + }, + { + "epoch": 1.5277777777777777, + "grad_norm": 3.2391650676727295, + "learning_rate": 2.761927034611787e-06, + "loss": 1.4677, + "step": 1210 + }, + { + "epoch": 1.5277777777777777, + "eval_loss": 1.6849907636642456, + "eval_runtime": 36.8638, + "eval_samples_per_second": 27.127, + "eval_steps_per_second": 1.139, + "step": 1210 + }, + { + "epoch": 1.529040404040404, + "grad_norm": 3.1884987354278564, + "learning_rate": 2.759588400374182e-06, + "loss": 1.4185, + "step": 1211 + }, + { + "epoch": 1.5303030303030303, + "grad_norm": 2.959118366241455, + "learning_rate": 2.7572497661365764e-06, + "loss": 1.4536, + "step": 1212 + }, + { + "epoch": 1.5315656565656566, + "grad_norm": 2.882380247116089, + "learning_rate": 2.7549111318989712e-06, + "loss": 1.4872, + "step": 1213 + }, + { + "epoch": 1.5328282828282829, + "grad_norm": 3.0574429035186768, + "learning_rate": 2.752572497661366e-06, + "loss": 1.4343, + "step": 1214 + }, + { + "epoch": 1.5340909090909092, + "grad_norm": 3.1645233631134033, + "learning_rate": 2.750233863423761e-06, + "loss": 1.5101, + "step": 1215 + }, + { + "epoch": 1.5353535353535355, + "grad_norm": 2.83034086227417, + "learning_rate": 2.7478952291861553e-06, + "loss": 1.45, + "step": 1216 + }, + { + "epoch": 1.5366161616161618, + "grad_norm": 2.964780330657959, + "learning_rate": 2.74555659494855e-06, + "loss": 1.453, + "step": 1217 + }, + { + "epoch": 1.5378787878787878, + "grad_norm": 2.9896774291992188, + "learning_rate": 2.743217960710945e-06, + "loss": 1.4648, + "step": 1218 + }, + { + "epoch": 1.5391414141414141, + "grad_norm": 3.013664484024048, + "learning_rate": 2.74087932647334e-06, + "loss": 1.4828, + "step": 1219 + }, + { + "epoch": 1.5404040404040404, + "grad_norm": 3.0011773109436035, + "learning_rate": 2.7385406922357343e-06, + "loss": 1.4705, + "step": 1220 + }, + { + "epoch": 1.5404040404040404, + "eval_loss": 1.6872731447219849, + "eval_runtime": 36.9864, + "eval_samples_per_second": 27.037, + "eval_steps_per_second": 1.136, + "step": 1220 + }, + { + "epoch": 1.5416666666666665, + "grad_norm": 3.239349842071533, + "learning_rate": 2.736202057998129e-06, + "loss": 1.484, + "step": 1221 + }, + { + "epoch": 1.5429292929292928, + "grad_norm": 2.9267003536224365, + "learning_rate": 2.733863423760524e-06, + "loss": 1.4587, + "step": 1222 + }, + { + "epoch": 1.5441919191919191, + "grad_norm": 2.890094518661499, + "learning_rate": 2.7315247895229192e-06, + "loss": 1.4456, + "step": 1223 + }, + { + "epoch": 1.5454545454545454, + "grad_norm": 3.0287511348724365, + "learning_rate": 2.7291861552853136e-06, + "loss": 1.5059, + "step": 1224 + }, + { + "epoch": 1.5467171717171717, + "grad_norm": 3.041609525680542, + "learning_rate": 2.7268475210477085e-06, + "loss": 1.4341, + "step": 1225 + }, + { + "epoch": 1.547979797979798, + "grad_norm": 2.8771862983703613, + "learning_rate": 2.7245088868101033e-06, + "loss": 1.4308, + "step": 1226 + }, + { + "epoch": 1.5492424242424243, + "grad_norm": 3.1173794269561768, + "learning_rate": 2.722170252572498e-06, + "loss": 1.514, + "step": 1227 + }, + { + "epoch": 1.5505050505050506, + "grad_norm": 2.838275671005249, + "learning_rate": 2.7198316183348926e-06, + "loss": 1.4798, + "step": 1228 + }, + { + "epoch": 1.551767676767677, + "grad_norm": 2.9917166233062744, + "learning_rate": 2.7174929840972874e-06, + "loss": 1.5256, + "step": 1229 + }, + { + "epoch": 1.553030303030303, + "grad_norm": 3.071840763092041, + "learning_rate": 2.7151543498596823e-06, + "loss": 1.5187, + "step": 1230 + }, + { + "epoch": 1.553030303030303, + "eval_loss": 1.6850661039352417, + "eval_runtime": 37.0094, + "eval_samples_per_second": 27.02, + "eval_steps_per_second": 1.135, + "step": 1230 + }, + { + "epoch": 1.5542929292929293, + "grad_norm": 3.253169536590576, + "learning_rate": 2.7128157156220767e-06, + "loss": 1.5522, + "step": 1231 + }, + { + "epoch": 1.5555555555555556, + "grad_norm": 3.037151336669922, + "learning_rate": 2.7104770813844715e-06, + "loss": 1.5125, + "step": 1232 + }, + { + "epoch": 1.5568181818181817, + "grad_norm": 2.8486485481262207, + "learning_rate": 2.7081384471468664e-06, + "loss": 1.4169, + "step": 1233 + }, + { + "epoch": 1.558080808080808, + "grad_norm": 2.953824996948242, + "learning_rate": 2.7057998129092612e-06, + "loss": 1.4505, + "step": 1234 + }, + { + "epoch": 1.5593434343434343, + "grad_norm": 3.012303590774536, + "learning_rate": 2.7034611786716557e-06, + "loss": 1.4816, + "step": 1235 + }, + { + "epoch": 1.5606060606060606, + "grad_norm": 3.1798768043518066, + "learning_rate": 2.7011225444340505e-06, + "loss": 1.5116, + "step": 1236 + }, + { + "epoch": 1.5618686868686869, + "grad_norm": 3.11370587348938, + "learning_rate": 2.6987839101964453e-06, + "loss": 1.5035, + "step": 1237 + }, + { + "epoch": 1.5631313131313131, + "grad_norm": 3.0563342571258545, + "learning_rate": 2.69644527595884e-06, + "loss": 1.4387, + "step": 1238 + }, + { + "epoch": 1.5643939393939394, + "grad_norm": 2.9884026050567627, + "learning_rate": 2.6941066417212354e-06, + "loss": 1.428, + "step": 1239 + }, + { + "epoch": 1.5656565656565657, + "grad_norm": 2.955773115158081, + "learning_rate": 2.69176800748363e-06, + "loss": 1.4208, + "step": 1240 + }, + { + "epoch": 1.5656565656565657, + "eval_loss": 1.685523509979248, + "eval_runtime": 37.0248, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 1240 + }, + { + "epoch": 1.566919191919192, + "grad_norm": 2.9858005046844482, + "learning_rate": 2.6894293732460247e-06, + "loss": 1.4706, + "step": 1241 + }, + { + "epoch": 1.5681818181818183, + "grad_norm": 3.122340440750122, + "learning_rate": 2.6870907390084196e-06, + "loss": 1.4892, + "step": 1242 + }, + { + "epoch": 1.5694444444444444, + "grad_norm": 2.9828262329101562, + "learning_rate": 2.684752104770814e-06, + "loss": 1.4255, + "step": 1243 + }, + { + "epoch": 1.5707070707070707, + "grad_norm": 2.916653633117676, + "learning_rate": 2.682413470533209e-06, + "loss": 1.4241, + "step": 1244 + }, + { + "epoch": 1.571969696969697, + "grad_norm": 3.1615540981292725, + "learning_rate": 2.6800748362956037e-06, + "loss": 1.4719, + "step": 1245 + }, + { + "epoch": 1.573232323232323, + "grad_norm": 3.154130697250366, + "learning_rate": 2.6777362020579985e-06, + "loss": 1.4731, + "step": 1246 + }, + { + "epoch": 1.5744949494949494, + "grad_norm": 3.040848731994629, + "learning_rate": 2.675397567820393e-06, + "loss": 1.5149, + "step": 1247 + }, + { + "epoch": 1.5757575757575757, + "grad_norm": 2.9808318614959717, + "learning_rate": 2.6730589335827878e-06, + "loss": 1.431, + "step": 1248 + }, + { + "epoch": 1.577020202020202, + "grad_norm": 2.8538382053375244, + "learning_rate": 2.6707202993451826e-06, + "loss": 1.4458, + "step": 1249 + }, + { + "epoch": 1.5782828282828283, + "grad_norm": 3.113945960998535, + "learning_rate": 2.6683816651075774e-06, + "loss": 1.4873, + "step": 1250 + }, + { + "epoch": 1.5782828282828283, + "eval_loss": 1.6823841333389282, + "eval_runtime": 37.0384, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 1250 + }, + { + "epoch": 1.5795454545454546, + "grad_norm": 3.0044517517089844, + "learning_rate": 2.666043030869972e-06, + "loss": 1.4904, + "step": 1251 + }, + { + "epoch": 1.5808080808080809, + "grad_norm": 2.845820665359497, + "learning_rate": 2.6637043966323667e-06, + "loss": 1.4174, + "step": 1252 + }, + { + "epoch": 1.5820707070707072, + "grad_norm": 2.9123542308807373, + "learning_rate": 2.6613657623947616e-06, + "loss": 1.4896, + "step": 1253 + }, + { + "epoch": 1.5833333333333335, + "grad_norm": 3.1593222618103027, + "learning_rate": 2.659027128157157e-06, + "loss": 1.5208, + "step": 1254 + }, + { + "epoch": 1.5845959595959596, + "grad_norm": 3.0782854557037354, + "learning_rate": 2.6566884939195512e-06, + "loss": 1.4911, + "step": 1255 + }, + { + "epoch": 1.5858585858585859, + "grad_norm": 2.77309513092041, + "learning_rate": 2.654349859681946e-06, + "loss": 1.4475, + "step": 1256 + }, + { + "epoch": 1.5871212121212122, + "grad_norm": 2.9154622554779053, + "learning_rate": 2.652011225444341e-06, + "loss": 1.4324, + "step": 1257 + }, + { + "epoch": 1.5883838383838382, + "grad_norm": 3.107395887374878, + "learning_rate": 2.6496725912067358e-06, + "loss": 1.5126, + "step": 1258 + }, + { + "epoch": 1.5896464646464645, + "grad_norm": 2.825025796890259, + "learning_rate": 2.64733395696913e-06, + "loss": 1.4723, + "step": 1259 + }, + { + "epoch": 1.5909090909090908, + "grad_norm": 2.833296060562134, + "learning_rate": 2.644995322731525e-06, + "loss": 1.4516, + "step": 1260 + }, + { + "epoch": 1.5909090909090908, + "eval_loss": 1.6950432062149048, + "eval_runtime": 37.0343, + "eval_samples_per_second": 27.002, + "eval_steps_per_second": 1.134, + "step": 1260 + }, + { + "epoch": 1.5921717171717171, + "grad_norm": 2.9508421421051025, + "learning_rate": 2.64265668849392e-06, + "loss": 1.4002, + "step": 1261 + }, + { + "epoch": 1.5934343434343434, + "grad_norm": 2.8778491020202637, + "learning_rate": 2.6403180542563143e-06, + "loss": 1.4258, + "step": 1262 + }, + { + "epoch": 1.5946969696969697, + "grad_norm": 2.81261944770813, + "learning_rate": 2.637979420018709e-06, + "loss": 1.435, + "step": 1263 + }, + { + "epoch": 1.595959595959596, + "grad_norm": 3.1811511516571045, + "learning_rate": 2.635640785781104e-06, + "loss": 1.4792, + "step": 1264 + }, + { + "epoch": 1.5972222222222223, + "grad_norm": 3.0168304443359375, + "learning_rate": 2.633302151543499e-06, + "loss": 1.5394, + "step": 1265 + }, + { + "epoch": 1.5984848484848486, + "grad_norm": 2.9531311988830566, + "learning_rate": 2.6309635173058932e-06, + "loss": 1.4775, + "step": 1266 + }, + { + "epoch": 1.5997474747474747, + "grad_norm": 2.835839033126831, + "learning_rate": 2.628624883068288e-06, + "loss": 1.478, + "step": 1267 + }, + { + "epoch": 1.601010101010101, + "grad_norm": 2.9014039039611816, + "learning_rate": 2.626286248830683e-06, + "loss": 1.4926, + "step": 1268 + }, + { + "epoch": 1.6022727272727273, + "grad_norm": 2.89461088180542, + "learning_rate": 2.623947614593078e-06, + "loss": 1.448, + "step": 1269 + }, + { + "epoch": 1.6035353535353534, + "grad_norm": 2.8005120754241943, + "learning_rate": 2.621608980355473e-06, + "loss": 1.4733, + "step": 1270 + }, + { + "epoch": 1.6035353535353534, + "eval_loss": 1.6889415979385376, + "eval_runtime": 37.0345, + "eval_samples_per_second": 27.002, + "eval_steps_per_second": 1.134, + "step": 1270 + }, + { + "epoch": 1.6047979797979797, + "grad_norm": 2.9576175212860107, + "learning_rate": 2.6192703461178675e-06, + "loss": 1.5116, + "step": 1271 + }, + { + "epoch": 1.606060606060606, + "grad_norm": 2.926638603210449, + "learning_rate": 2.6169317118802623e-06, + "loss": 1.4744, + "step": 1272 + }, + { + "epoch": 1.6073232323232323, + "grad_norm": 2.911587715148926, + "learning_rate": 2.614593077642657e-06, + "loss": 1.4695, + "step": 1273 + }, + { + "epoch": 1.6085858585858586, + "grad_norm": 2.8878750801086426, + "learning_rate": 2.6122544434050516e-06, + "loss": 1.4574, + "step": 1274 + }, + { + "epoch": 1.6098484848484849, + "grad_norm": 2.8489067554473877, + "learning_rate": 2.6099158091674464e-06, + "loss": 1.4599, + "step": 1275 + }, + { + "epoch": 1.6111111111111112, + "grad_norm": 2.8223719596862793, + "learning_rate": 2.6075771749298412e-06, + "loss": 1.4431, + "step": 1276 + }, + { + "epoch": 1.6123737373737375, + "grad_norm": 2.869910955429077, + "learning_rate": 2.605238540692236e-06, + "loss": 1.48, + "step": 1277 + }, + { + "epoch": 1.6136363636363638, + "grad_norm": 2.8435168266296387, + "learning_rate": 2.6028999064546305e-06, + "loss": 1.4471, + "step": 1278 + }, + { + "epoch": 1.61489898989899, + "grad_norm": 2.9384665489196777, + "learning_rate": 2.6005612722170253e-06, + "loss": 1.5176, + "step": 1279 + }, + { + "epoch": 1.6161616161616161, + "grad_norm": 3.021785020828247, + "learning_rate": 2.59822263797942e-06, + "loss": 1.4789, + "step": 1280 + }, + { + "epoch": 1.6161616161616161, + "eval_loss": 1.6864516735076904, + "eval_runtime": 37.0249, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 1280 + }, + { + "epoch": 1.6174242424242424, + "grad_norm": 3.007183313369751, + "learning_rate": 2.5958840037418146e-06, + "loss": 1.4686, + "step": 1281 + }, + { + "epoch": 1.6186868686868687, + "grad_norm": 3.115910053253174, + "learning_rate": 2.5935453695042095e-06, + "loss": 1.5382, + "step": 1282 + }, + { + "epoch": 1.6199494949494948, + "grad_norm": 2.98108172416687, + "learning_rate": 2.5912067352666043e-06, + "loss": 1.4569, + "step": 1283 + }, + { + "epoch": 1.621212121212121, + "grad_norm": 2.926832437515259, + "learning_rate": 2.5888681010289996e-06, + "loss": 1.4515, + "step": 1284 + }, + { + "epoch": 1.6224747474747474, + "grad_norm": 2.9045534133911133, + "learning_rate": 2.5865294667913944e-06, + "loss": 1.4259, + "step": 1285 + }, + { + "epoch": 1.6237373737373737, + "grad_norm": 3.0757768154144287, + "learning_rate": 2.584190832553789e-06, + "loss": 1.5436, + "step": 1286 + }, + { + "epoch": 1.625, + "grad_norm": 3.045257806777954, + "learning_rate": 2.5818521983161837e-06, + "loss": 1.4378, + "step": 1287 + }, + { + "epoch": 1.6262626262626263, + "grad_norm": 2.8131601810455322, + "learning_rate": 2.5795135640785785e-06, + "loss": 1.4224, + "step": 1288 + }, + { + "epoch": 1.6275252525252526, + "grad_norm": 2.953326940536499, + "learning_rate": 2.5771749298409734e-06, + "loss": 1.4137, + "step": 1289 + }, + { + "epoch": 1.628787878787879, + "grad_norm": 2.946232318878174, + "learning_rate": 2.5748362956033678e-06, + "loss": 1.4575, + "step": 1290 + }, + { + "epoch": 1.628787878787879, + "eval_loss": 1.6896334886550903, + "eval_runtime": 37.317, + "eval_samples_per_second": 26.797, + "eval_steps_per_second": 1.125, + "step": 1290 + }, + { + "epoch": 1.6300505050505052, + "grad_norm": 2.860867738723755, + "learning_rate": 2.5724976613657626e-06, + "loss": 1.4412, + "step": 1291 + }, + { + "epoch": 1.6313131313131313, + "grad_norm": 3.0417261123657227, + "learning_rate": 2.5701590271281575e-06, + "loss": 1.4919, + "step": 1292 + }, + { + "epoch": 1.6325757575757576, + "grad_norm": 3.21425199508667, + "learning_rate": 2.567820392890552e-06, + "loss": 1.5243, + "step": 1293 + }, + { + "epoch": 1.6338383838383839, + "grad_norm": 3.000302314758301, + "learning_rate": 2.5654817586529467e-06, + "loss": 1.4571, + "step": 1294 + }, + { + "epoch": 1.63510101010101, + "grad_norm": 2.88415789604187, + "learning_rate": 2.5631431244153416e-06, + "loss": 1.4527, + "step": 1295 + }, + { + "epoch": 1.6363636363636362, + "grad_norm": 2.7747156620025635, + "learning_rate": 2.5608044901777364e-06, + "loss": 1.3984, + "step": 1296 + }, + { + "epoch": 1.6376262626262625, + "grad_norm": 2.9298319816589355, + "learning_rate": 2.558465855940131e-06, + "loss": 1.4587, + "step": 1297 + }, + { + "epoch": 1.6388888888888888, + "grad_norm": 3.0199968814849854, + "learning_rate": 2.5561272217025257e-06, + "loss": 1.4539, + "step": 1298 + }, + { + "epoch": 1.6401515151515151, + "grad_norm": 3.09503436088562, + "learning_rate": 2.5537885874649205e-06, + "loss": 1.4551, + "step": 1299 + }, + { + "epoch": 1.6414141414141414, + "grad_norm": 2.972161293029785, + "learning_rate": 2.5514499532273158e-06, + "loss": 1.4123, + "step": 1300 + }, + { + "epoch": 1.6414141414141414, + "eval_loss": 1.6877377033233643, + "eval_runtime": 37.4088, + "eval_samples_per_second": 26.732, + "eval_steps_per_second": 1.123, + "step": 1300 + }, + { + "epoch": 1.6426767676767677, + "grad_norm": 3.081058979034424, + "learning_rate": 2.5491113189897106e-06, + "loss": 1.457, + "step": 1301 + }, + { + "epoch": 1.643939393939394, + "grad_norm": 3.040106773376465, + "learning_rate": 2.546772684752105e-06, + "loss": 1.4465, + "step": 1302 + }, + { + "epoch": 1.6452020202020203, + "grad_norm": 3.133812427520752, + "learning_rate": 2.5444340505145e-06, + "loss": 1.4382, + "step": 1303 + }, + { + "epoch": 1.6464646464646466, + "grad_norm": 2.9701125621795654, + "learning_rate": 2.5420954162768947e-06, + "loss": 1.457, + "step": 1304 + }, + { + "epoch": 1.6477272727272727, + "grad_norm": 2.9272050857543945, + "learning_rate": 2.539756782039289e-06, + "loss": 1.464, + "step": 1305 + }, + { + "epoch": 1.648989898989899, + "grad_norm": 3.048401117324829, + "learning_rate": 2.537418147801684e-06, + "loss": 1.4688, + "step": 1306 + }, + { + "epoch": 1.6502525252525253, + "grad_norm": 2.9103102684020996, + "learning_rate": 2.535079513564079e-06, + "loss": 1.5025, + "step": 1307 + }, + { + "epoch": 1.6515151515151514, + "grad_norm": 3.0841548442840576, + "learning_rate": 2.5327408793264737e-06, + "loss": 1.4938, + "step": 1308 + }, + { + "epoch": 1.6527777777777777, + "grad_norm": 3.022184371948242, + "learning_rate": 2.530402245088868e-06, + "loss": 1.4332, + "step": 1309 + }, + { + "epoch": 1.654040404040404, + "grad_norm": 2.881479024887085, + "learning_rate": 2.528063610851263e-06, + "loss": 1.4541, + "step": 1310 + }, + { + "epoch": 1.654040404040404, + "eval_loss": 1.6814675331115723, + "eval_runtime": 36.9446, + "eval_samples_per_second": 27.068, + "eval_steps_per_second": 1.137, + "step": 1310 + }, + { + "epoch": 1.6553030303030303, + "grad_norm": 2.906719446182251, + "learning_rate": 2.5257249766136578e-06, + "loss": 1.4968, + "step": 1311 + }, + { + "epoch": 1.6565656565656566, + "grad_norm": 3.0566461086273193, + "learning_rate": 2.523386342376052e-06, + "loss": 1.4513, + "step": 1312 + }, + { + "epoch": 1.6578282828282829, + "grad_norm": 2.895026206970215, + "learning_rate": 2.521047708138447e-06, + "loss": 1.4217, + "step": 1313 + }, + { + "epoch": 1.6590909090909092, + "grad_norm": 3.0006465911865234, + "learning_rate": 2.518709073900842e-06, + "loss": 1.4793, + "step": 1314 + }, + { + "epoch": 1.6603535353535355, + "grad_norm": 2.9853575229644775, + "learning_rate": 2.516370439663237e-06, + "loss": 1.4642, + "step": 1315 + }, + { + "epoch": 1.6616161616161618, + "grad_norm": 2.9729392528533936, + "learning_rate": 2.514031805425632e-06, + "loss": 1.4644, + "step": 1316 + }, + { + "epoch": 1.6628787878787878, + "grad_norm": 3.0997586250305176, + "learning_rate": 2.5116931711880264e-06, + "loss": 1.4392, + "step": 1317 + }, + { + "epoch": 1.6641414141414141, + "grad_norm": 2.9621379375457764, + "learning_rate": 2.5093545369504213e-06, + "loss": 1.4387, + "step": 1318 + }, + { + "epoch": 1.6654040404040404, + "grad_norm": 3.0571911334991455, + "learning_rate": 2.507015902712816e-06, + "loss": 1.4638, + "step": 1319 + }, + { + "epoch": 1.6666666666666665, + "grad_norm": 2.9580321311950684, + "learning_rate": 2.504677268475211e-06, + "loss": 1.4985, + "step": 1320 + }, + { + "epoch": 1.6666666666666665, + "eval_loss": 1.6790977716445923, + "eval_runtime": 37.0242, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 1320 + }, + { + "epoch": 1.6679292929292928, + "grad_norm": 2.87432599067688, + "learning_rate": 2.5023386342376054e-06, + "loss": 1.4058, + "step": 1321 + }, + { + "epoch": 1.6691919191919191, + "grad_norm": 2.925546407699585, + "learning_rate": 2.5e-06, + "loss": 1.4986, + "step": 1322 + }, + { + "epoch": 1.6704545454545454, + "grad_norm": 2.837743043899536, + "learning_rate": 2.497661365762395e-06, + "loss": 1.455, + "step": 1323 + }, + { + "epoch": 1.6717171717171717, + "grad_norm": 3.111581563949585, + "learning_rate": 2.4953227315247895e-06, + "loss": 1.5005, + "step": 1324 + }, + { + "epoch": 1.672979797979798, + "grad_norm": 2.851076126098633, + "learning_rate": 2.4929840972871843e-06, + "loss": 1.4365, + "step": 1325 + }, + { + "epoch": 1.6742424242424243, + "grad_norm": 2.966965675354004, + "learning_rate": 2.4906454630495796e-06, + "loss": 1.4697, + "step": 1326 + }, + { + "epoch": 1.6755050505050506, + "grad_norm": 3.015193223953247, + "learning_rate": 2.488306828811974e-06, + "loss": 1.4486, + "step": 1327 + }, + { + "epoch": 1.676767676767677, + "grad_norm": 3.017681121826172, + "learning_rate": 2.485968194574369e-06, + "loss": 1.4017, + "step": 1328 + }, + { + "epoch": 1.678030303030303, + "grad_norm": 3.071981906890869, + "learning_rate": 2.4836295603367637e-06, + "loss": 1.4572, + "step": 1329 + }, + { + "epoch": 1.6792929292929293, + "grad_norm": 3.051741123199463, + "learning_rate": 2.481290926099158e-06, + "loss": 1.4498, + "step": 1330 + }, + { + "epoch": 1.6792929292929293, + "eval_loss": 1.6771777868270874, + "eval_runtime": 37.0747, + "eval_samples_per_second": 26.973, + "eval_steps_per_second": 1.133, + "step": 1330 + }, + { + "epoch": 1.6805555555555556, + "grad_norm": 2.9561071395874023, + "learning_rate": 2.478952291861553e-06, + "loss": 1.4428, + "step": 1331 + }, + { + "epoch": 1.6818181818181817, + "grad_norm": 2.992584466934204, + "learning_rate": 2.4766136576239478e-06, + "loss": 1.461, + "step": 1332 + }, + { + "epoch": 1.683080808080808, + "grad_norm": 3.0473921298980713, + "learning_rate": 2.4742750233863426e-06, + "loss": 1.489, + "step": 1333 + }, + { + "epoch": 1.6843434343434343, + "grad_norm": 2.8721725940704346, + "learning_rate": 2.4719363891487375e-06, + "loss": 1.487, + "step": 1334 + }, + { + "epoch": 1.6856060606060606, + "grad_norm": 2.8650765419006348, + "learning_rate": 2.4695977549111323e-06, + "loss": 1.4504, + "step": 1335 + }, + { + "epoch": 1.6868686868686869, + "grad_norm": 2.998887538909912, + "learning_rate": 2.4672591206735267e-06, + "loss": 1.4586, + "step": 1336 + }, + { + "epoch": 1.6881313131313131, + "grad_norm": 2.915188789367676, + "learning_rate": 2.4649204864359216e-06, + "loss": 1.4333, + "step": 1337 + }, + { + "epoch": 1.6893939393939394, + "grad_norm": 2.9660091400146484, + "learning_rate": 2.4625818521983164e-06, + "loss": 1.4711, + "step": 1338 + }, + { + "epoch": 1.6906565656565657, + "grad_norm": 2.916274070739746, + "learning_rate": 2.4602432179607113e-06, + "loss": 1.4642, + "step": 1339 + }, + { + "epoch": 1.691919191919192, + "grad_norm": 3.04154896736145, + "learning_rate": 2.4579045837231057e-06, + "loss": 1.4233, + "step": 1340 + }, + { + "epoch": 1.691919191919192, + "eval_loss": 1.6764302253723145, + "eval_runtime": 37.0686, + "eval_samples_per_second": 26.977, + "eval_steps_per_second": 1.133, + "step": 1340 + }, + { + "epoch": 1.6931818181818183, + "grad_norm": 2.9976465702056885, + "learning_rate": 2.4555659494855005e-06, + "loss": 1.4775, + "step": 1341 + }, + { + "epoch": 1.6944444444444444, + "grad_norm": 2.9140090942382812, + "learning_rate": 2.4532273152478954e-06, + "loss": 1.4532, + "step": 1342 + }, + { + "epoch": 1.6957070707070707, + "grad_norm": 3.0224454402923584, + "learning_rate": 2.45088868101029e-06, + "loss": 1.5091, + "step": 1343 + }, + { + "epoch": 1.696969696969697, + "grad_norm": 2.8979439735412598, + "learning_rate": 2.448550046772685e-06, + "loss": 1.4634, + "step": 1344 + }, + { + "epoch": 1.698232323232323, + "grad_norm": 2.913433313369751, + "learning_rate": 2.44621141253508e-06, + "loss": 1.4483, + "step": 1345 + }, + { + "epoch": 1.6994949494949494, + "grad_norm": 2.813455820083618, + "learning_rate": 2.4438727782974743e-06, + "loss": 1.446, + "step": 1346 + }, + { + "epoch": 1.7007575757575757, + "grad_norm": 2.988722324371338, + "learning_rate": 2.441534144059869e-06, + "loss": 1.4908, + "step": 1347 + }, + { + "epoch": 1.702020202020202, + "grad_norm": 2.850446939468384, + "learning_rate": 2.439195509822264e-06, + "loss": 1.4221, + "step": 1348 + }, + { + "epoch": 1.7032828282828283, + "grad_norm": 2.8405508995056152, + "learning_rate": 2.436856875584659e-06, + "loss": 1.4496, + "step": 1349 + }, + { + "epoch": 1.7045454545454546, + "grad_norm": 3.0188651084899902, + "learning_rate": 2.4345182413470537e-06, + "loss": 1.489, + "step": 1350 + }, + { + "epoch": 1.7045454545454546, + "eval_loss": 1.6780083179473877, + "eval_runtime": 37.0724, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 1350 + }, + { + "epoch": 1.7058080808080809, + "grad_norm": 2.905088424682617, + "learning_rate": 2.4321796071094485e-06, + "loss": 1.4283, + "step": 1351 + }, + { + "epoch": 1.7070707070707072, + "grad_norm": 2.878998279571533, + "learning_rate": 2.429840972871843e-06, + "loss": 1.4357, + "step": 1352 + }, + { + "epoch": 1.7083333333333335, + "grad_norm": 3.1707355976104736, + "learning_rate": 2.427502338634238e-06, + "loss": 1.496, + "step": 1353 + }, + { + "epoch": 1.7095959595959596, + "grad_norm": 2.9610493183135986, + "learning_rate": 2.4251637043966326e-06, + "loss": 1.4905, + "step": 1354 + }, + { + "epoch": 1.7108585858585859, + "grad_norm": 2.8372201919555664, + "learning_rate": 2.422825070159027e-06, + "loss": 1.4563, + "step": 1355 + }, + { + "epoch": 1.7121212121212122, + "grad_norm": 2.8741753101348877, + "learning_rate": 2.420486435921422e-06, + "loss": 1.4525, + "step": 1356 + }, + { + "epoch": 1.7133838383838382, + "grad_norm": 2.973726749420166, + "learning_rate": 2.418147801683817e-06, + "loss": 1.4878, + "step": 1357 + }, + { + "epoch": 1.7146464646464645, + "grad_norm": 3.05717396736145, + "learning_rate": 2.4158091674462116e-06, + "loss": 1.5402, + "step": 1358 + }, + { + "epoch": 1.7159090909090908, + "grad_norm": 2.883772373199463, + "learning_rate": 2.4134705332086064e-06, + "loss": 1.4329, + "step": 1359 + }, + { + "epoch": 1.7171717171717171, + "grad_norm": 2.8246653079986572, + "learning_rate": 2.4111318989710013e-06, + "loss": 1.4821, + "step": 1360 + }, + { + "epoch": 1.7171717171717171, + "eval_loss": 1.6737746000289917, + "eval_runtime": 37.0666, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 1360 + }, + { + "epoch": 1.7184343434343434, + "grad_norm": 2.905977964401245, + "learning_rate": 2.4087932647333957e-06, + "loss": 1.4378, + "step": 1361 + }, + { + "epoch": 1.7196969696969697, + "grad_norm": 2.9268195629119873, + "learning_rate": 2.4064546304957905e-06, + "loss": 1.4765, + "step": 1362 + }, + { + "epoch": 1.720959595959596, + "grad_norm": 3.1369476318359375, + "learning_rate": 2.4041159962581854e-06, + "loss": 1.4532, + "step": 1363 + }, + { + "epoch": 1.7222222222222223, + "grad_norm": 3.04546856880188, + "learning_rate": 2.4017773620205802e-06, + "loss": 1.4921, + "step": 1364 + }, + { + "epoch": 1.7234848484848486, + "grad_norm": 2.872852325439453, + "learning_rate": 2.399438727782975e-06, + "loss": 1.4323, + "step": 1365 + }, + { + "epoch": 1.7247474747474747, + "grad_norm": 2.961681604385376, + "learning_rate": 2.39710009354537e-06, + "loss": 1.5181, + "step": 1366 + }, + { + "epoch": 1.726010101010101, + "grad_norm": 2.997729778289795, + "learning_rate": 2.3947614593077643e-06, + "loss": 1.4817, + "step": 1367 + }, + { + "epoch": 1.7272727272727273, + "grad_norm": 2.9097466468811035, + "learning_rate": 2.392422825070159e-06, + "loss": 1.4478, + "step": 1368 + }, + { + "epoch": 1.7285353535353534, + "grad_norm": 2.82433819770813, + "learning_rate": 2.390084190832554e-06, + "loss": 1.4065, + "step": 1369 + }, + { + "epoch": 1.7297979797979797, + "grad_norm": 2.940910577774048, + "learning_rate": 2.387745556594949e-06, + "loss": 1.4735, + "step": 1370 + }, + { + "epoch": 1.7297979797979797, + "eval_loss": 1.6717814207077026, + "eval_runtime": 37.0937, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 1370 + }, + { + "epoch": 1.731060606060606, + "grad_norm": 2.79101300239563, + "learning_rate": 2.3854069223573433e-06, + "loss": 1.4545, + "step": 1371 + }, + { + "epoch": 1.7323232323232323, + "grad_norm": 2.916954517364502, + "learning_rate": 2.3830682881197385e-06, + "loss": 1.4293, + "step": 1372 + }, + { + "epoch": 1.7335858585858586, + "grad_norm": 2.9609909057617188, + "learning_rate": 2.380729653882133e-06, + "loss": 1.4578, + "step": 1373 + }, + { + "epoch": 1.7348484848484849, + "grad_norm": 2.935925245285034, + "learning_rate": 2.378391019644528e-06, + "loss": 1.4577, + "step": 1374 + }, + { + "epoch": 1.7361111111111112, + "grad_norm": 2.9803526401519775, + "learning_rate": 2.3760523854069226e-06, + "loss": 1.4449, + "step": 1375 + }, + { + "epoch": 1.7373737373737375, + "grad_norm": 3.0292913913726807, + "learning_rate": 2.3737137511693175e-06, + "loss": 1.4308, + "step": 1376 + }, + { + "epoch": 1.7386363636363638, + "grad_norm": 3.0247879028320312, + "learning_rate": 2.371375116931712e-06, + "loss": 1.5406, + "step": 1377 + }, + { + "epoch": 1.73989898989899, + "grad_norm": 2.86775541305542, + "learning_rate": 2.3690364826941067e-06, + "loss": 1.4324, + "step": 1378 + }, + { + "epoch": 1.7411616161616161, + "grad_norm": 2.8593552112579346, + "learning_rate": 2.3666978484565016e-06, + "loss": 1.4526, + "step": 1379 + }, + { + "epoch": 1.7424242424242424, + "grad_norm": 2.8735320568084717, + "learning_rate": 2.3643592142188964e-06, + "loss": 1.4097, + "step": 1380 + }, + { + "epoch": 1.7424242424242424, + "eval_loss": 1.675827145576477, + "eval_runtime": 37.1206, + "eval_samples_per_second": 26.939, + "eval_steps_per_second": 1.131, + "step": 1380 + }, + { + "epoch": 1.7436868686868687, + "grad_norm": 2.948153495788574, + "learning_rate": 2.3620205799812913e-06, + "loss": 1.4234, + "step": 1381 + }, + { + "epoch": 1.7449494949494948, + "grad_norm": 2.9344282150268555, + "learning_rate": 2.359681945743686e-06, + "loss": 1.4653, + "step": 1382 + }, + { + "epoch": 1.746212121212121, + "grad_norm": 2.9309439659118652, + "learning_rate": 2.3573433115060805e-06, + "loss": 1.4676, + "step": 1383 + }, + { + "epoch": 1.7474747474747474, + "grad_norm": 3.0762970447540283, + "learning_rate": 2.3550046772684754e-06, + "loss": 1.4407, + "step": 1384 + }, + { + "epoch": 1.7487373737373737, + "grad_norm": 3.149824857711792, + "learning_rate": 2.3526660430308702e-06, + "loss": 1.4629, + "step": 1385 + }, + { + "epoch": 1.75, + "grad_norm": 3.109571695327759, + "learning_rate": 2.3503274087932646e-06, + "loss": 1.494, + "step": 1386 + }, + { + "epoch": 1.7512626262626263, + "grad_norm": 2.8359758853912354, + "learning_rate": 2.34798877455566e-06, + "loss": 1.4056, + "step": 1387 + }, + { + "epoch": 1.7525252525252526, + "grad_norm": 2.8523781299591064, + "learning_rate": 2.3456501403180548e-06, + "loss": 1.4389, + "step": 1388 + }, + { + "epoch": 1.753787878787879, + "grad_norm": 3.0218093395233154, + "learning_rate": 2.343311506080449e-06, + "loss": 1.4587, + "step": 1389 + }, + { + "epoch": 1.7550505050505052, + "grad_norm": 3.0975496768951416, + "learning_rate": 2.340972871842844e-06, + "loss": 1.5133, + "step": 1390 + }, + { + "epoch": 1.7550505050505052, + "eval_loss": 1.671931266784668, + "eval_runtime": 37.2749, + "eval_samples_per_second": 26.828, + "eval_steps_per_second": 1.127, + "step": 1390 + }, + { + "epoch": 1.7563131313131313, + "grad_norm": 3.0177836418151855, + "learning_rate": 2.338634237605239e-06, + "loss": 1.4716, + "step": 1391 + }, + { + "epoch": 1.7575757575757576, + "grad_norm": 3.049105167388916, + "learning_rate": 2.3362956033676333e-06, + "loss": 1.4858, + "step": 1392 + }, + { + "epoch": 1.7588383838383839, + "grad_norm": 3.0062026977539062, + "learning_rate": 2.333956969130028e-06, + "loss": 1.4239, + "step": 1393 + }, + { + "epoch": 1.76010101010101, + "grad_norm": 3.071772813796997, + "learning_rate": 2.331618334892423e-06, + "loss": 1.4654, + "step": 1394 + }, + { + "epoch": 1.7613636363636362, + "grad_norm": 3.140510320663452, + "learning_rate": 2.329279700654818e-06, + "loss": 1.4854, + "step": 1395 + }, + { + "epoch": 1.7626262626262625, + "grad_norm": 2.9903907775878906, + "learning_rate": 2.3269410664172126e-06, + "loss": 1.4406, + "step": 1396 + }, + { + "epoch": 1.7638888888888888, + "grad_norm": 2.839891195297241, + "learning_rate": 2.3246024321796075e-06, + "loss": 1.4498, + "step": 1397 + }, + { + "epoch": 1.7651515151515151, + "grad_norm": 3.0252437591552734, + "learning_rate": 2.322263797942002e-06, + "loss": 1.472, + "step": 1398 + }, + { + "epoch": 1.7664141414141414, + "grad_norm": 3.0214762687683105, + "learning_rate": 2.3199251637043968e-06, + "loss": 1.4031, + "step": 1399 + }, + { + "epoch": 1.7676767676767677, + "grad_norm": 3.190088987350464, + "learning_rate": 2.3175865294667916e-06, + "loss": 1.4497, + "step": 1400 + }, + { + "epoch": 1.7676767676767677, + "eval_loss": 1.673120379447937, + "eval_runtime": 37.4414, + "eval_samples_per_second": 26.708, + "eval_steps_per_second": 1.122, + "step": 1400 + }, + { + "epoch": 1.768939393939394, + "grad_norm": 2.9620201587677, + "learning_rate": 2.3152478952291864e-06, + "loss": 1.4149, + "step": 1401 + }, + { + "epoch": 1.7702020202020203, + "grad_norm": 3.112372875213623, + "learning_rate": 2.312909260991581e-06, + "loss": 1.4569, + "step": 1402 + }, + { + "epoch": 1.7714646464646466, + "grad_norm": 3.138857364654541, + "learning_rate": 2.310570626753976e-06, + "loss": 1.4819, + "step": 1403 + }, + { + "epoch": 1.7727272727272727, + "grad_norm": 3.0957937240600586, + "learning_rate": 2.3082319925163705e-06, + "loss": 1.4374, + "step": 1404 + }, + { + "epoch": 1.773989898989899, + "grad_norm": 3.014039993286133, + "learning_rate": 2.3058933582787654e-06, + "loss": 1.4262, + "step": 1405 + }, + { + "epoch": 1.7752525252525253, + "grad_norm": 2.8127005100250244, + "learning_rate": 2.3035547240411602e-06, + "loss": 1.4468, + "step": 1406 + }, + { + "epoch": 1.7765151515151514, + "grad_norm": 2.9180614948272705, + "learning_rate": 2.301216089803555e-06, + "loss": 1.4259, + "step": 1407 + }, + { + "epoch": 1.7777777777777777, + "grad_norm": 3.0376124382019043, + "learning_rate": 2.2988774555659495e-06, + "loss": 1.4628, + "step": 1408 + }, + { + "epoch": 1.779040404040404, + "grad_norm": 3.031740665435791, + "learning_rate": 2.2965388213283443e-06, + "loss": 1.5058, + "step": 1409 + }, + { + "epoch": 1.7803030303030303, + "grad_norm": 2.9876646995544434, + "learning_rate": 2.294200187090739e-06, + "loss": 1.4605, + "step": 1410 + }, + { + "epoch": 1.7803030303030303, + "eval_loss": 1.668077826499939, + "eval_runtime": 36.9734, + "eval_samples_per_second": 27.046, + "eval_steps_per_second": 1.136, + "step": 1410 + }, + { + "epoch": 1.7815656565656566, + "grad_norm": 2.953744888305664, + "learning_rate": 2.291861552853134e-06, + "loss": 1.4257, + "step": 1411 + }, + { + "epoch": 1.7828282828282829, + "grad_norm": 3.049036741256714, + "learning_rate": 2.289522918615529e-06, + "loss": 1.4476, + "step": 1412 + }, + { + "epoch": 1.7840909090909092, + "grad_norm": 3.102116584777832, + "learning_rate": 2.2871842843779237e-06, + "loss": 1.4701, + "step": 1413 + }, + { + "epoch": 1.7853535353535355, + "grad_norm": 2.842970132827759, + "learning_rate": 2.284845650140318e-06, + "loss": 1.415, + "step": 1414 + }, + { + "epoch": 1.7866161616161618, + "grad_norm": 2.9386725425720215, + "learning_rate": 2.282507015902713e-06, + "loss": 1.4561, + "step": 1415 + }, + { + "epoch": 1.7878787878787878, + "grad_norm": 2.942713499069214, + "learning_rate": 2.280168381665108e-06, + "loss": 1.435, + "step": 1416 + }, + { + "epoch": 1.7891414141414141, + "grad_norm": 2.943105936050415, + "learning_rate": 2.2778297474275022e-06, + "loss": 1.4923, + "step": 1417 + }, + { + "epoch": 1.7904040404040404, + "grad_norm": 2.7094717025756836, + "learning_rate": 2.2754911131898975e-06, + "loss": 1.4457, + "step": 1418 + }, + { + "epoch": 1.7916666666666665, + "grad_norm": 2.7859625816345215, + "learning_rate": 2.2731524789522923e-06, + "loss": 1.3993, + "step": 1419 + }, + { + "epoch": 1.7929292929292928, + "grad_norm": 3.0923988819122314, + "learning_rate": 2.2708138447146868e-06, + "loss": 1.5054, + "step": 1420 + }, + { + "epoch": 1.7929292929292928, + "eval_loss": 1.6687692403793335, + "eval_runtime": 37.0567, + "eval_samples_per_second": 26.986, + "eval_steps_per_second": 1.133, + "step": 1420 + }, + { + "epoch": 1.7941919191919191, + "grad_norm": 2.849726676940918, + "learning_rate": 2.2684752104770816e-06, + "loss": 1.3978, + "step": 1421 + }, + { + "epoch": 1.7954545454545454, + "grad_norm": 3.028932809829712, + "learning_rate": 2.2661365762394764e-06, + "loss": 1.4786, + "step": 1422 + }, + { + "epoch": 1.7967171717171717, + "grad_norm": 2.973188638687134, + "learning_rate": 2.263797942001871e-06, + "loss": 1.4393, + "step": 1423 + }, + { + "epoch": 1.797979797979798, + "grad_norm": 2.884032964706421, + "learning_rate": 2.2614593077642657e-06, + "loss": 1.4197, + "step": 1424 + }, + { + "epoch": 1.7992424242424243, + "grad_norm": 2.9058549404144287, + "learning_rate": 2.2591206735266606e-06, + "loss": 1.4313, + "step": 1425 + }, + { + "epoch": 1.8005050505050506, + "grad_norm": 2.905841827392578, + "learning_rate": 2.2567820392890554e-06, + "loss": 1.439, + "step": 1426 + }, + { + "epoch": 1.801767676767677, + "grad_norm": 2.927028179168701, + "learning_rate": 2.2544434050514502e-06, + "loss": 1.4296, + "step": 1427 + }, + { + "epoch": 1.803030303030303, + "grad_norm": 2.965529203414917, + "learning_rate": 2.252104770813845e-06, + "loss": 1.4185, + "step": 1428 + }, + { + "epoch": 1.8042929292929293, + "grad_norm": 2.9494433403015137, + "learning_rate": 2.2497661365762395e-06, + "loss": 1.5087, + "step": 1429 + }, + { + "epoch": 1.8055555555555556, + "grad_norm": 3.0137765407562256, + "learning_rate": 2.2474275023386343e-06, + "loss": 1.4383, + "step": 1430 + }, + { + "epoch": 1.8055555555555556, + "eval_loss": 1.6728461980819702, + "eval_runtime": 37.1033, + "eval_samples_per_second": 26.952, + "eval_steps_per_second": 1.132, + "step": 1430 + }, + { + "epoch": 1.8068181818181817, + "grad_norm": 2.980062484741211, + "learning_rate": 2.245088868101029e-06, + "loss": 1.4333, + "step": 1431 + }, + { + "epoch": 1.808080808080808, + "grad_norm": 3.143759250640869, + "learning_rate": 2.242750233863424e-06, + "loss": 1.4676, + "step": 1432 + }, + { + "epoch": 1.8093434343434343, + "grad_norm": 3.0134518146514893, + "learning_rate": 2.240411599625819e-06, + "loss": 1.4223, + "step": 1433 + }, + { + "epoch": 1.8106060606060606, + "grad_norm": 3.000018358230591, + "learning_rate": 2.2380729653882137e-06, + "loss": 1.4537, + "step": 1434 + }, + { + "epoch": 1.8118686868686869, + "grad_norm": 2.9828288555145264, + "learning_rate": 2.235734331150608e-06, + "loss": 1.4379, + "step": 1435 + }, + { + "epoch": 1.8131313131313131, + "grad_norm": 2.965486764907837, + "learning_rate": 2.233395696913003e-06, + "loss": 1.4244, + "step": 1436 + }, + { + "epoch": 1.8143939393939394, + "grad_norm": 2.9787070751190186, + "learning_rate": 2.231057062675398e-06, + "loss": 1.4117, + "step": 1437 + }, + { + "epoch": 1.8156565656565657, + "grad_norm": 3.0893537998199463, + "learning_rate": 2.2287184284377927e-06, + "loss": 1.4809, + "step": 1438 + }, + { + "epoch": 1.816919191919192, + "grad_norm": 2.8568801879882812, + "learning_rate": 2.226379794200187e-06, + "loss": 1.4232, + "step": 1439 + }, + { + "epoch": 1.8181818181818183, + "grad_norm": 3.1005046367645264, + "learning_rate": 2.224041159962582e-06, + "loss": 1.482, + "step": 1440 + }, + { + "epoch": 1.8181818181818183, + "eval_loss": 1.6738661527633667, + "eval_runtime": 37.1286, + "eval_samples_per_second": 26.933, + "eval_steps_per_second": 1.131, + "step": 1440 + }, + { + "epoch": 1.8194444444444444, + "grad_norm": 2.9064700603485107, + "learning_rate": 2.2217025257249768e-06, + "loss": 1.4179, + "step": 1441 + }, + { + "epoch": 1.8207070707070707, + "grad_norm": 3.0111782550811768, + "learning_rate": 2.2193638914873716e-06, + "loss": 1.4084, + "step": 1442 + }, + { + "epoch": 1.821969696969697, + "grad_norm": 2.9269816875457764, + "learning_rate": 2.2170252572497665e-06, + "loss": 1.4361, + "step": 1443 + }, + { + "epoch": 1.823232323232323, + "grad_norm": 2.9985902309417725, + "learning_rate": 2.2146866230121613e-06, + "loss": 1.4345, + "step": 1444 + }, + { + "epoch": 1.8244949494949494, + "grad_norm": 3.1306066513061523, + "learning_rate": 2.2123479887745557e-06, + "loss": 1.4728, + "step": 1445 + }, + { + "epoch": 1.8257575757575757, + "grad_norm": 2.9876086711883545, + "learning_rate": 2.2100093545369506e-06, + "loss": 1.378, + "step": 1446 + }, + { + "epoch": 1.827020202020202, + "grad_norm": 3.018763780593872, + "learning_rate": 2.2076707202993454e-06, + "loss": 1.4506, + "step": 1447 + }, + { + "epoch": 1.8282828282828283, + "grad_norm": 2.960573673248291, + "learning_rate": 2.2053320860617402e-06, + "loss": 1.4674, + "step": 1448 + }, + { + "epoch": 1.8295454545454546, + "grad_norm": 2.849881887435913, + "learning_rate": 2.202993451824135e-06, + "loss": 1.4687, + "step": 1449 + }, + { + "epoch": 1.8308080808080809, + "grad_norm": 2.8798882961273193, + "learning_rate": 2.2006548175865295e-06, + "loss": 1.3868, + "step": 1450 + }, + { + "epoch": 1.8308080808080809, + "eval_loss": 1.6657519340515137, + "eval_runtime": 37.0993, + "eval_samples_per_second": 26.955, + "eval_steps_per_second": 1.132, + "step": 1450 + }, + { + "epoch": 1.8320707070707072, + "grad_norm": 2.937471628189087, + "learning_rate": 2.1983161833489243e-06, + "loss": 1.4085, + "step": 1451 + }, + { + "epoch": 1.8333333333333335, + "grad_norm": 3.0815699100494385, + "learning_rate": 2.195977549111319e-06, + "loss": 1.4055, + "step": 1452 + }, + { + "epoch": 1.8345959595959596, + "grad_norm": 2.927917718887329, + "learning_rate": 2.193638914873714e-06, + "loss": 1.4014, + "step": 1453 + }, + { + "epoch": 1.8358585858585859, + "grad_norm": 2.9576570987701416, + "learning_rate": 2.1913002806361085e-06, + "loss": 1.4394, + "step": 1454 + }, + { + "epoch": 1.8371212121212122, + "grad_norm": 2.9505863189697266, + "learning_rate": 2.1889616463985033e-06, + "loss": 1.4018, + "step": 1455 + }, + { + "epoch": 1.8383838383838382, + "grad_norm": 3.014965772628784, + "learning_rate": 2.186623012160898e-06, + "loss": 1.3735, + "step": 1456 + }, + { + "epoch": 1.8396464646464645, + "grad_norm": 3.121122121810913, + "learning_rate": 2.184284377923293e-06, + "loss": 1.4289, + "step": 1457 + }, + { + "epoch": 1.8409090909090908, + "grad_norm": 2.9710142612457275, + "learning_rate": 2.181945743685688e-06, + "loss": 1.4494, + "step": 1458 + }, + { + "epoch": 1.8421717171717171, + "grad_norm": 3.028554916381836, + "learning_rate": 2.1796071094480827e-06, + "loss": 1.5151, + "step": 1459 + }, + { + "epoch": 1.8434343434343434, + "grad_norm": 2.8973751068115234, + "learning_rate": 2.177268475210477e-06, + "loss": 1.4171, + "step": 1460 + }, + { + "epoch": 1.8434343434343434, + "eval_loss": 1.6670986413955688, + "eval_runtime": 37.0937, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 1460 + }, + { + "epoch": 1.8446969696969697, + "grad_norm": 3.046865701675415, + "learning_rate": 2.174929840972872e-06, + "loss": 1.4016, + "step": 1461 + }, + { + "epoch": 1.845959595959596, + "grad_norm": 3.0509603023529053, + "learning_rate": 2.1725912067352668e-06, + "loss": 1.4196, + "step": 1462 + }, + { + "epoch": 1.8472222222222223, + "grad_norm": 2.951892137527466, + "learning_rate": 2.1702525724976616e-06, + "loss": 1.4063, + "step": 1463 + }, + { + "epoch": 1.8484848484848486, + "grad_norm": 3.2126667499542236, + "learning_rate": 2.1679139382600565e-06, + "loss": 1.5106, + "step": 1464 + }, + { + "epoch": 1.8497474747474747, + "grad_norm": 2.965576648712158, + "learning_rate": 2.1655753040224513e-06, + "loss": 1.4096, + "step": 1465 + }, + { + "epoch": 1.851010101010101, + "grad_norm": 2.8640592098236084, + "learning_rate": 2.1632366697848457e-06, + "loss": 1.389, + "step": 1466 + }, + { + "epoch": 1.8522727272727273, + "grad_norm": 3.1476943492889404, + "learning_rate": 2.1608980355472406e-06, + "loss": 1.4793, + "step": 1467 + }, + { + "epoch": 1.8535353535353534, + "grad_norm": 2.814695119857788, + "learning_rate": 2.1585594013096354e-06, + "loss": 1.4265, + "step": 1468 + }, + { + "epoch": 1.8547979797979797, + "grad_norm": 2.9728431701660156, + "learning_rate": 2.1562207670720302e-06, + "loss": 1.386, + "step": 1469 + }, + { + "epoch": 1.856060606060606, + "grad_norm": 2.9062652587890625, + "learning_rate": 2.1538821328344247e-06, + "loss": 1.3986, + "step": 1470 + }, + { + "epoch": 1.856060606060606, + "eval_loss": 1.6698763370513916, + "eval_runtime": 37.1129, + "eval_samples_per_second": 26.945, + "eval_steps_per_second": 1.132, + "step": 1470 + }, + { + "epoch": 1.8573232323232323, + "grad_norm": 2.950857162475586, + "learning_rate": 2.15154349859682e-06, + "loss": 1.4193, + "step": 1471 + }, + { + "epoch": 1.8585858585858586, + "grad_norm": 3.017594337463379, + "learning_rate": 2.1492048643592144e-06, + "loss": 1.4318, + "step": 1472 + }, + { + "epoch": 1.8598484848484849, + "grad_norm": 2.991514205932617, + "learning_rate": 2.146866230121609e-06, + "loss": 1.4462, + "step": 1473 + }, + { + "epoch": 1.8611111111111112, + "grad_norm": 3.149475336074829, + "learning_rate": 2.144527595884004e-06, + "loss": 1.4486, + "step": 1474 + }, + { + "epoch": 1.8623737373737375, + "grad_norm": 2.948828935623169, + "learning_rate": 2.1421889616463985e-06, + "loss": 1.4522, + "step": 1475 + }, + { + "epoch": 1.8636363636363638, + "grad_norm": 3.0383694171905518, + "learning_rate": 2.1398503274087933e-06, + "loss": 1.45, + "step": 1476 + }, + { + "epoch": 1.86489898989899, + "grad_norm": 3.093449831008911, + "learning_rate": 2.137511693171188e-06, + "loss": 1.4266, + "step": 1477 + }, + { + "epoch": 1.8661616161616161, + "grad_norm": 3.017549753189087, + "learning_rate": 2.135173058933583e-06, + "loss": 1.4067, + "step": 1478 + }, + { + "epoch": 1.8674242424242424, + "grad_norm": 2.985044002532959, + "learning_rate": 2.132834424695978e-06, + "loss": 1.4904, + "step": 1479 + }, + { + "epoch": 1.8686868686868687, + "grad_norm": 2.9600346088409424, + "learning_rate": 2.1304957904583727e-06, + "loss": 1.4006, + "step": 1480 + }, + { + "epoch": 1.8686868686868687, + "eval_loss": 1.6648480892181396, + "eval_runtime": 37.1225, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 1480 + }, + { + "epoch": 1.8699494949494948, + "grad_norm": 2.8905625343322754, + "learning_rate": 2.128157156220767e-06, + "loss": 1.4096, + "step": 1481 + }, + { + "epoch": 1.871212121212121, + "grad_norm": 2.95342755317688, + "learning_rate": 2.125818521983162e-06, + "loss": 1.3888, + "step": 1482 + }, + { + "epoch": 1.8724747474747474, + "grad_norm": 2.9438536167144775, + "learning_rate": 2.1234798877455568e-06, + "loss": 1.413, + "step": 1483 + }, + { + "epoch": 1.8737373737373737, + "grad_norm": 3.1257364749908447, + "learning_rate": 2.1211412535079516e-06, + "loss": 1.4902, + "step": 1484 + }, + { + "epoch": 1.875, + "grad_norm": 3.1377382278442383, + "learning_rate": 2.118802619270346e-06, + "loss": 1.4815, + "step": 1485 + }, + { + "epoch": 1.8762626262626263, + "grad_norm": 3.122253656387329, + "learning_rate": 2.116463985032741e-06, + "loss": 1.4695, + "step": 1486 + }, + { + "epoch": 1.8775252525252526, + "grad_norm": 3.0128729343414307, + "learning_rate": 2.1141253507951357e-06, + "loss": 1.4673, + "step": 1487 + }, + { + "epoch": 1.878787878787879, + "grad_norm": 2.921835422515869, + "learning_rate": 2.1117867165575306e-06, + "loss": 1.4325, + "step": 1488 + }, + { + "epoch": 1.8800505050505052, + "grad_norm": 3.19289231300354, + "learning_rate": 2.1094480823199254e-06, + "loss": 1.4504, + "step": 1489 + }, + { + "epoch": 1.8813131313131313, + "grad_norm": 2.8579299449920654, + "learning_rate": 2.1071094480823203e-06, + "loss": 1.4042, + "step": 1490 + }, + { + "epoch": 1.8813131313131313, + "eval_loss": 1.6603271961212158, + "eval_runtime": 37.2997, + "eval_samples_per_second": 26.81, + "eval_steps_per_second": 1.126, + "step": 1490 + }, + { + "epoch": 1.8825757575757576, + "grad_norm": 2.9113969802856445, + "learning_rate": 2.1047708138447147e-06, + "loss": 1.4128, + "step": 1491 + }, + { + "epoch": 1.8838383838383839, + "grad_norm": 3.058133363723755, + "learning_rate": 2.1024321796071095e-06, + "loss": 1.4574, + "step": 1492 + }, + { + "epoch": 1.88510101010101, + "grad_norm": 3.0066583156585693, + "learning_rate": 2.1000935453695044e-06, + "loss": 1.4521, + "step": 1493 + }, + { + "epoch": 1.8863636363636362, + "grad_norm": 2.9154908657073975, + "learning_rate": 2.097754911131899e-06, + "loss": 1.4103, + "step": 1494 + }, + { + "epoch": 1.8876262626262625, + "grad_norm": 2.977062702178955, + "learning_rate": 2.095416276894294e-06, + "loss": 1.4003, + "step": 1495 + }, + { + "epoch": 1.8888888888888888, + "grad_norm": 2.9048168659210205, + "learning_rate": 2.093077642656689e-06, + "loss": 1.4216, + "step": 1496 + }, + { + "epoch": 1.8901515151515151, + "grad_norm": 3.008733034133911, + "learning_rate": 2.0907390084190833e-06, + "loss": 1.4516, + "step": 1497 + }, + { + "epoch": 1.8914141414141414, + "grad_norm": 3.0832529067993164, + "learning_rate": 2.088400374181478e-06, + "loss": 1.4248, + "step": 1498 + }, + { + "epoch": 1.8926767676767677, + "grad_norm": 2.848066568374634, + "learning_rate": 2.086061739943873e-06, + "loss": 1.4042, + "step": 1499 + }, + { + "epoch": 1.893939393939394, + "grad_norm": 2.846782684326172, + "learning_rate": 2.0837231057062674e-06, + "loss": 1.409, + "step": 1500 + }, + { + "epoch": 1.893939393939394, + "eval_loss": 1.6672260761260986, + "eval_runtime": 37.4695, + "eval_samples_per_second": 26.688, + "eval_steps_per_second": 1.121, + "step": 1500 + }, + { + "epoch": 1.8952020202020203, + "grad_norm": 3.026818037033081, + "learning_rate": 2.0813844714686623e-06, + "loss": 1.4842, + "step": 1501 + }, + { + "epoch": 1.8964646464646466, + "grad_norm": 2.977665662765503, + "learning_rate": 2.0790458372310575e-06, + "loss": 1.4635, + "step": 1502 + }, + { + "epoch": 1.8977272727272727, + "grad_norm": 2.9093048572540283, + "learning_rate": 2.076707202993452e-06, + "loss": 1.4303, + "step": 1503 + }, + { + "epoch": 1.898989898989899, + "grad_norm": 3.0777814388275146, + "learning_rate": 2.0743685687558468e-06, + "loss": 1.4581, + "step": 1504 + }, + { + "epoch": 1.9002525252525253, + "grad_norm": 3.0283970832824707, + "learning_rate": 2.0720299345182416e-06, + "loss": 1.4665, + "step": 1505 + }, + { + "epoch": 1.9015151515151514, + "grad_norm": 2.852412223815918, + "learning_rate": 2.069691300280636e-06, + "loss": 1.423, + "step": 1506 + }, + { + "epoch": 1.9027777777777777, + "grad_norm": 3.1163125038146973, + "learning_rate": 2.067352666043031e-06, + "loss": 1.4907, + "step": 1507 + }, + { + "epoch": 1.904040404040404, + "grad_norm": 2.9857897758483887, + "learning_rate": 2.0650140318054257e-06, + "loss": 1.4262, + "step": 1508 + }, + { + "epoch": 1.9053030303030303, + "grad_norm": 2.957388401031494, + "learning_rate": 2.0626753975678206e-06, + "loss": 1.4525, + "step": 1509 + }, + { + "epoch": 1.9065656565656566, + "grad_norm": 3.0107505321502686, + "learning_rate": 2.0603367633302154e-06, + "loss": 1.4461, + "step": 1510 + }, + { + "epoch": 1.9065656565656566, + "eval_loss": 1.6621215343475342, + "eval_runtime": 36.9336, + "eval_samples_per_second": 27.076, + "eval_steps_per_second": 1.137, + "step": 1510 + }, + { + "epoch": 1.9078282828282829, + "grad_norm": 3.112478494644165, + "learning_rate": 2.0579981290926103e-06, + "loss": 1.479, + "step": 1511 + }, + { + "epoch": 1.9090909090909092, + "grad_norm": 3.083494186401367, + "learning_rate": 2.0556594948550047e-06, + "loss": 1.4777, + "step": 1512 + }, + { + "epoch": 1.9103535353535355, + "grad_norm": 2.932652711868286, + "learning_rate": 2.0533208606173995e-06, + "loss": 1.3937, + "step": 1513 + }, + { + "epoch": 1.9116161616161618, + "grad_norm": 2.8349449634552, + "learning_rate": 2.0509822263797944e-06, + "loss": 1.4096, + "step": 1514 + }, + { + "epoch": 1.9128787878787878, + "grad_norm": 2.9760472774505615, + "learning_rate": 2.048643592142189e-06, + "loss": 1.4312, + "step": 1515 + }, + { + "epoch": 1.9141414141414141, + "grad_norm": 3.1012730598449707, + "learning_rate": 2.0463049579045836e-06, + "loss": 1.5189, + "step": 1516 + }, + { + "epoch": 1.9154040404040404, + "grad_norm": 2.9072659015655518, + "learning_rate": 2.043966323666979e-06, + "loss": 1.4051, + "step": 1517 + }, + { + "epoch": 1.9166666666666665, + "grad_norm": 2.840203285217285, + "learning_rate": 2.0416276894293733e-06, + "loss": 1.3899, + "step": 1518 + }, + { + "epoch": 1.9179292929292928, + "grad_norm": 3.130812406539917, + "learning_rate": 2.039289055191768e-06, + "loss": 1.4834, + "step": 1519 + }, + { + "epoch": 1.9191919191919191, + "grad_norm": 3.1507933139801025, + "learning_rate": 2.036950420954163e-06, + "loss": 1.402, + "step": 1520 + }, + { + "epoch": 1.9191919191919191, + "eval_loss": 1.6648352146148682, + "eval_runtime": 37.0844, + "eval_samples_per_second": 26.966, + "eval_steps_per_second": 1.133, + "step": 1520 + }, + { + "epoch": 1.9204545454545454, + "grad_norm": 3.099269151687622, + "learning_rate": 2.034611786716558e-06, + "loss": 1.4189, + "step": 1521 + }, + { + "epoch": 1.9217171717171717, + "grad_norm": 3.0046563148498535, + "learning_rate": 2.0322731524789523e-06, + "loss": 1.405, + "step": 1522 + }, + { + "epoch": 1.922979797979798, + "grad_norm": 3.229635715484619, + "learning_rate": 2.029934518241347e-06, + "loss": 1.5031, + "step": 1523 + }, + { + "epoch": 1.9242424242424243, + "grad_norm": 3.0336225032806396, + "learning_rate": 2.027595884003742e-06, + "loss": 1.4685, + "step": 1524 + }, + { + "epoch": 1.9255050505050506, + "grad_norm": 2.9998443126678467, + "learning_rate": 2.025257249766137e-06, + "loss": 1.4558, + "step": 1525 + }, + { + "epoch": 1.926767676767677, + "grad_norm": 3.1587722301483154, + "learning_rate": 2.0229186155285316e-06, + "loss": 1.4916, + "step": 1526 + }, + { + "epoch": 1.928030303030303, + "grad_norm": 3.2228925228118896, + "learning_rate": 2.0205799812909265e-06, + "loss": 1.4373, + "step": 1527 + }, + { + "epoch": 1.9292929292929293, + "grad_norm": 3.06015944480896, + "learning_rate": 2.018241347053321e-06, + "loss": 1.4147, + "step": 1528 + }, + { + "epoch": 1.9305555555555556, + "grad_norm": 3.02284574508667, + "learning_rate": 2.0159027128157157e-06, + "loss": 1.4574, + "step": 1529 + }, + { + "epoch": 1.9318181818181817, + "grad_norm": 2.9933414459228516, + "learning_rate": 2.0135640785781106e-06, + "loss": 1.4283, + "step": 1530 + }, + { + "epoch": 1.9318181818181817, + "eval_loss": 1.6711069345474243, + "eval_runtime": 37.1182, + "eval_samples_per_second": 26.941, + "eval_steps_per_second": 1.132, + "step": 1530 + }, + { + "epoch": 1.933080808080808, + "grad_norm": 3.325230121612549, + "learning_rate": 2.011225444340505e-06, + "loss": 1.4532, + "step": 1531 + }, + { + "epoch": 1.9343434343434343, + "grad_norm": 3.0051827430725098, + "learning_rate": 2.0088868101029003e-06, + "loss": 1.3965, + "step": 1532 + }, + { + "epoch": 1.9356060606060606, + "grad_norm": 3.1392757892608643, + "learning_rate": 2.006548175865295e-06, + "loss": 1.4678, + "step": 1533 + }, + { + "epoch": 1.9368686868686869, + "grad_norm": 3.0719001293182373, + "learning_rate": 2.0042095416276895e-06, + "loss": 1.4374, + "step": 1534 + }, + { + "epoch": 1.9381313131313131, + "grad_norm": 3.0406386852264404, + "learning_rate": 2.0018709073900844e-06, + "loss": 1.4199, + "step": 1535 + }, + { + "epoch": 1.9393939393939394, + "grad_norm": 2.9481236934661865, + "learning_rate": 1.9995322731524792e-06, + "loss": 1.4599, + "step": 1536 + }, + { + "epoch": 1.9406565656565657, + "grad_norm": 2.901155948638916, + "learning_rate": 1.9971936389148736e-06, + "loss": 1.381, + "step": 1537 + }, + { + "epoch": 1.941919191919192, + "grad_norm": 2.9623661041259766, + "learning_rate": 1.9948550046772685e-06, + "loss": 1.4769, + "step": 1538 + }, + { + "epoch": 1.9431818181818183, + "grad_norm": 2.8457608222961426, + "learning_rate": 1.9925163704396633e-06, + "loss": 1.4117, + "step": 1539 + }, + { + "epoch": 1.9444444444444444, + "grad_norm": 2.9782702922821045, + "learning_rate": 1.990177736202058e-06, + "loss": 1.4824, + "step": 1540 + }, + { + "epoch": 1.9444444444444444, + "eval_loss": 1.6663525104522705, + "eval_runtime": 37.1207, + "eval_samples_per_second": 26.939, + "eval_steps_per_second": 1.131, + "step": 1540 + }, + { + "epoch": 1.9457070707070707, + "grad_norm": 2.8265817165374756, + "learning_rate": 1.987839101964453e-06, + "loss": 1.4375, + "step": 1541 + }, + { + "epoch": 1.946969696969697, + "grad_norm": 2.9103686809539795, + "learning_rate": 1.985500467726848e-06, + "loss": 1.474, + "step": 1542 + }, + { + "epoch": 1.948232323232323, + "grad_norm": 2.837301254272461, + "learning_rate": 1.9831618334892423e-06, + "loss": 1.4171, + "step": 1543 + }, + { + "epoch": 1.9494949494949494, + "grad_norm": 3.0468413829803467, + "learning_rate": 1.980823199251637e-06, + "loss": 1.5095, + "step": 1544 + }, + { + "epoch": 1.9507575757575757, + "grad_norm": 2.968480348587036, + "learning_rate": 1.978484565014032e-06, + "loss": 1.4203, + "step": 1545 + }, + { + "epoch": 1.952020202020202, + "grad_norm": 2.9865822792053223, + "learning_rate": 1.976145930776427e-06, + "loss": 1.3916, + "step": 1546 + }, + { + "epoch": 1.9532828282828283, + "grad_norm": 2.9733808040618896, + "learning_rate": 1.9738072965388212e-06, + "loss": 1.4216, + "step": 1547 + }, + { + "epoch": 1.9545454545454546, + "grad_norm": 2.941718101501465, + "learning_rate": 1.9714686623012165e-06, + "loss": 1.426, + "step": 1548 + }, + { + "epoch": 1.9558080808080809, + "grad_norm": 3.0980522632598877, + "learning_rate": 1.969130028063611e-06, + "loss": 1.471, + "step": 1549 + }, + { + "epoch": 1.9570707070707072, + "grad_norm": 2.797102689743042, + "learning_rate": 1.9667913938260057e-06, + "loss": 1.3897, + "step": 1550 + }, + { + "epoch": 1.9570707070707072, + "eval_loss": 1.661893367767334, + "eval_runtime": 37.1157, + "eval_samples_per_second": 26.943, + "eval_steps_per_second": 1.132, + "step": 1550 + }, + { + "epoch": 1.9583333333333335, + "grad_norm": 2.9562623500823975, + "learning_rate": 1.9644527595884006e-06, + "loss": 1.4481, + "step": 1551 + }, + { + "epoch": 1.9595959595959596, + "grad_norm": 2.9586915969848633, + "learning_rate": 1.9621141253507954e-06, + "loss": 1.4068, + "step": 1552 + }, + { + "epoch": 1.9608585858585859, + "grad_norm": 3.061110496520996, + "learning_rate": 1.95977549111319e-06, + "loss": 1.4524, + "step": 1553 + }, + { + "epoch": 1.9621212121212122, + "grad_norm": 3.080491542816162, + "learning_rate": 1.9574368568755847e-06, + "loss": 1.4457, + "step": 1554 + }, + { + "epoch": 1.9633838383838382, + "grad_norm": 3.083254814147949, + "learning_rate": 1.9550982226379795e-06, + "loss": 1.4775, + "step": 1555 + }, + { + "epoch": 1.9646464646464645, + "grad_norm": 3.008596420288086, + "learning_rate": 1.9527595884003744e-06, + "loss": 1.425, + "step": 1556 + }, + { + "epoch": 1.9659090909090908, + "grad_norm": 2.999195098876953, + "learning_rate": 1.9504209541627692e-06, + "loss": 1.4114, + "step": 1557 + }, + { + "epoch": 1.9671717171717171, + "grad_norm": 3.055265188217163, + "learning_rate": 1.948082319925164e-06, + "loss": 1.4767, + "step": 1558 + }, + { + "epoch": 1.9684343434343434, + "grad_norm": 3.0922064781188965, + "learning_rate": 1.9457436856875585e-06, + "loss": 1.4548, + "step": 1559 + }, + { + "epoch": 1.9696969696969697, + "grad_norm": 2.8740789890289307, + "learning_rate": 1.9434050514499533e-06, + "loss": 1.4582, + "step": 1560 + }, + { + "epoch": 1.9696969696969697, + "eval_loss": 1.6608474254608154, + "eval_runtime": 37.1201, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 1560 + }, + { + "epoch": 1.970959595959596, + "grad_norm": 2.983570098876953, + "learning_rate": 1.941066417212348e-06, + "loss": 1.4602, + "step": 1561 + }, + { + "epoch": 1.9722222222222223, + "grad_norm": 3.132474184036255, + "learning_rate": 1.9387277829747426e-06, + "loss": 1.4689, + "step": 1562 + }, + { + "epoch": 1.9734848484848486, + "grad_norm": 2.9685440063476562, + "learning_rate": 1.936389148737138e-06, + "loss": 1.39, + "step": 1563 + }, + { + "epoch": 1.9747474747474747, + "grad_norm": 3.123037815093994, + "learning_rate": 1.9340505144995327e-06, + "loss": 1.4151, + "step": 1564 + }, + { + "epoch": 1.976010101010101, + "grad_norm": 3.135232925415039, + "learning_rate": 1.931711880261927e-06, + "loss": 1.4166, + "step": 1565 + }, + { + "epoch": 1.9772727272727273, + "grad_norm": 2.9131808280944824, + "learning_rate": 1.929373246024322e-06, + "loss": 1.3952, + "step": 1566 + }, + { + "epoch": 1.9785353535353534, + "grad_norm": 3.1187453269958496, + "learning_rate": 1.927034611786717e-06, + "loss": 1.4132, + "step": 1567 + }, + { + "epoch": 1.9797979797979797, + "grad_norm": 2.983447551727295, + "learning_rate": 1.9246959775491112e-06, + "loss": 1.4517, + "step": 1568 + }, + { + "epoch": 1.981060606060606, + "grad_norm": 2.9731497764587402, + "learning_rate": 1.922357343311506e-06, + "loss": 1.432, + "step": 1569 + }, + { + "epoch": 1.9823232323232323, + "grad_norm": 2.8526554107666016, + "learning_rate": 1.920018709073901e-06, + "loss": 1.4141, + "step": 1570 + }, + { + "epoch": 1.9823232323232323, + "eval_loss": 1.6579400300979614, + "eval_runtime": 37.1216, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 1570 + }, + { + "epoch": 1.9835858585858586, + "grad_norm": 2.900019407272339, + "learning_rate": 1.9176800748362958e-06, + "loss": 1.4008, + "step": 1571 + }, + { + "epoch": 1.9848484848484849, + "grad_norm": 2.96079683303833, + "learning_rate": 1.9153414405986906e-06, + "loss": 1.4586, + "step": 1572 + }, + { + "epoch": 1.9861111111111112, + "grad_norm": 3.1275715827941895, + "learning_rate": 1.9130028063610854e-06, + "loss": 1.4768, + "step": 1573 + }, + { + "epoch": 1.9873737373737375, + "grad_norm": 3.044041395187378, + "learning_rate": 1.91066417212348e-06, + "loss": 1.4428, + "step": 1574 + }, + { + "epoch": 1.9886363636363638, + "grad_norm": 3.0206832885742188, + "learning_rate": 1.9083255378858747e-06, + "loss": 1.4492, + "step": 1575 + }, + { + "epoch": 1.98989898989899, + "grad_norm": 2.935185194015503, + "learning_rate": 1.9059869036482695e-06, + "loss": 1.398, + "step": 1576 + }, + { + "epoch": 1.9911616161616161, + "grad_norm": 2.935185194015503, + "learning_rate": 1.9059869036482695e-06, + "loss": 1.4373, + "step": 1577 + }, + { + "epoch": 1.9924242424242424, + "grad_norm": 3.129822254180908, + "learning_rate": 1.9036482694106642e-06, + "loss": 1.4599, + "step": 1578 + }, + { + "epoch": 1.9936868686868687, + "grad_norm": 2.978015661239624, + "learning_rate": 1.9013096351730592e-06, + "loss": 1.409, + "step": 1579 + }, + { + "epoch": 1.9949494949494948, + "grad_norm": 2.980295181274414, + "learning_rate": 1.8989710009354539e-06, + "loss": 1.4251, + "step": 1580 + }, + { + "epoch": 1.9949494949494948, + "eval_loss": 1.661036491394043, + "eval_runtime": 37.1122, + "eval_samples_per_second": 26.945, + "eval_steps_per_second": 1.132, + "step": 1580 + }, + { + "epoch": 1.996212121212121, + "grad_norm": 2.878361225128174, + "learning_rate": 1.8966323666978487e-06, + "loss": 1.4289, + "step": 1581 + }, + { + "epoch": 1.9974747474747474, + "grad_norm": 2.967414379119873, + "learning_rate": 1.8942937324602433e-06, + "loss": 1.5151, + "step": 1582 + }, + { + "epoch": 1.9987373737373737, + "grad_norm": 2.9448516368865967, + "learning_rate": 1.8919550982226382e-06, + "loss": 1.4327, + "step": 1583 + }, + { + "epoch": 2.0, + "grad_norm": 2.9000067710876465, + "learning_rate": 1.8896164639850328e-06, + "loss": 1.4113, + "step": 1584 + }, + { + "epoch": 2.0012626262626263, + "grad_norm": 3.1630377769470215, + "learning_rate": 1.8872778297474277e-06, + "loss": 1.4179, + "step": 1585 + }, + { + "epoch": 2.0025252525252526, + "grad_norm": 3.2642927169799805, + "learning_rate": 1.8849391955098223e-06, + "loss": 1.3873, + "step": 1586 + }, + { + "epoch": 2.003787878787879, + "grad_norm": 3.1669039726257324, + "learning_rate": 1.8826005612722173e-06, + "loss": 1.3834, + "step": 1587 + }, + { + "epoch": 2.005050505050505, + "grad_norm": 3.145040988922119, + "learning_rate": 1.880261927034612e-06, + "loss": 1.3334, + "step": 1588 + }, + { + "epoch": 2.0063131313131315, + "grad_norm": 3.707200050354004, + "learning_rate": 1.8779232927970068e-06, + "loss": 1.3979, + "step": 1589 + }, + { + "epoch": 2.007575757575758, + "grad_norm": 3.936523199081421, + "learning_rate": 1.8755846585594014e-06, + "loss": 1.4221, + "step": 1590 + }, + { + "epoch": 2.007575757575758, + "eval_loss": 1.6722478866577148, + "eval_runtime": 37.2129, + "eval_samples_per_second": 26.872, + "eval_steps_per_second": 1.129, + "step": 1590 + }, + { + "epoch": 2.0088383838383836, + "grad_norm": 3.4221086502075195, + "learning_rate": 1.8732460243217963e-06, + "loss": 1.3894, + "step": 1591 + }, + { + "epoch": 2.01010101010101, + "grad_norm": 3.3115553855895996, + "learning_rate": 1.870907390084191e-06, + "loss": 1.4016, + "step": 1592 + }, + { + "epoch": 2.0113636363636362, + "grad_norm": 3.055779457092285, + "learning_rate": 1.8685687558465855e-06, + "loss": 1.3676, + "step": 1593 + }, + { + "epoch": 2.0126262626262625, + "grad_norm": 3.4615232944488525, + "learning_rate": 1.8662301216089804e-06, + "loss": 1.3624, + "step": 1594 + }, + { + "epoch": 2.013888888888889, + "grad_norm": 3.3298869132995605, + "learning_rate": 1.8638914873713754e-06, + "loss": 1.3939, + "step": 1595 + }, + { + "epoch": 2.015151515151515, + "grad_norm": 3.271921396255493, + "learning_rate": 1.86155285313377e-06, + "loss": 1.3994, + "step": 1596 + }, + { + "epoch": 2.0164141414141414, + "grad_norm": 3.272677421569824, + "learning_rate": 1.859214218896165e-06, + "loss": 1.3934, + "step": 1597 + }, + { + "epoch": 2.0176767676767677, + "grad_norm": 3.3485944271087646, + "learning_rate": 1.8568755846585596e-06, + "loss": 1.3626, + "step": 1598 + }, + { + "epoch": 2.018939393939394, + "grad_norm": 3.426981210708618, + "learning_rate": 1.8545369504209542e-06, + "loss": 1.4231, + "step": 1599 + }, + { + "epoch": 2.0202020202020203, + "grad_norm": 3.3859283924102783, + "learning_rate": 1.852198316183349e-06, + "loss": 1.3587, + "step": 1600 + }, + { + "epoch": 2.0202020202020203, + "eval_loss": 1.6732399463653564, + "eval_runtime": 37.3484, + "eval_samples_per_second": 26.775, + "eval_steps_per_second": 1.125, + "step": 1600 + }, + { + "epoch": 2.0214646464646466, + "grad_norm": 3.537806987762451, + "learning_rate": 1.8498596819457437e-06, + "loss": 1.3729, + "step": 1601 + }, + { + "epoch": 2.022727272727273, + "grad_norm": 3.5417566299438477, + "learning_rate": 1.8475210477081387e-06, + "loss": 1.4272, + "step": 1602 + }, + { + "epoch": 2.023989898989899, + "grad_norm": 3.287195920944214, + "learning_rate": 1.8451824134705336e-06, + "loss": 1.3647, + "step": 1603 + }, + { + "epoch": 2.025252525252525, + "grad_norm": 3.233365297317505, + "learning_rate": 1.8428437792329282e-06, + "loss": 1.4062, + "step": 1604 + }, + { + "epoch": 2.0265151515151514, + "grad_norm": 3.296578884124756, + "learning_rate": 1.8405051449953228e-06, + "loss": 1.3841, + "step": 1605 + }, + { + "epoch": 2.0277777777777777, + "grad_norm": 3.2697315216064453, + "learning_rate": 1.8381665107577177e-06, + "loss": 1.3957, + "step": 1606 + }, + { + "epoch": 2.029040404040404, + "grad_norm": 3.1117050647735596, + "learning_rate": 1.8358278765201123e-06, + "loss": 1.3643, + "step": 1607 + }, + { + "epoch": 2.0303030303030303, + "grad_norm": 3.14170241355896, + "learning_rate": 1.8334892422825071e-06, + "loss": 1.3483, + "step": 1608 + }, + { + "epoch": 2.0315656565656566, + "grad_norm": 3.2749483585357666, + "learning_rate": 1.8311506080449018e-06, + "loss": 1.4007, + "step": 1609 + }, + { + "epoch": 2.032828282828283, + "grad_norm": 3.266631603240967, + "learning_rate": 1.8288119738072968e-06, + "loss": 1.3943, + "step": 1610 + }, + { + "epoch": 2.032828282828283, + "eval_loss": 1.6679413318634033, + "eval_runtime": 36.8768, + "eval_samples_per_second": 27.117, + "eval_steps_per_second": 1.139, + "step": 1610 + }, + { + "epoch": 2.034090909090909, + "grad_norm": 3.535625696182251, + "learning_rate": 1.8264733395696914e-06, + "loss": 1.4027, + "step": 1611 + }, + { + "epoch": 2.0353535353535355, + "grad_norm": 3.2741401195526123, + "learning_rate": 1.8241347053320863e-06, + "loss": 1.4075, + "step": 1612 + }, + { + "epoch": 2.0366161616161618, + "grad_norm": 3.3183717727661133, + "learning_rate": 1.821796071094481e-06, + "loss": 1.4223, + "step": 1613 + }, + { + "epoch": 2.037878787878788, + "grad_norm": 3.323057174682617, + "learning_rate": 1.8194574368568758e-06, + "loss": 1.3595, + "step": 1614 + }, + { + "epoch": 2.0391414141414144, + "grad_norm": 3.202054738998413, + "learning_rate": 1.8171188026192704e-06, + "loss": 1.3399, + "step": 1615 + }, + { + "epoch": 2.04040404040404, + "grad_norm": 3.3273448944091797, + "learning_rate": 1.8147801683816652e-06, + "loss": 1.3848, + "step": 1616 + }, + { + "epoch": 2.0416666666666665, + "grad_norm": 3.396843671798706, + "learning_rate": 1.81244153414406e-06, + "loss": 1.3968, + "step": 1617 + }, + { + "epoch": 2.042929292929293, + "grad_norm": 3.2145836353302, + "learning_rate": 1.810102899906455e-06, + "loss": 1.3571, + "step": 1618 + }, + { + "epoch": 2.044191919191919, + "grad_norm": 3.156754732131958, + "learning_rate": 1.8077642656688496e-06, + "loss": 1.3439, + "step": 1619 + }, + { + "epoch": 2.0454545454545454, + "grad_norm": 3.4259471893310547, + "learning_rate": 1.8054256314312444e-06, + "loss": 1.4123, + "step": 1620 + }, + { + "epoch": 2.0454545454545454, + "eval_loss": 1.6703134775161743, + "eval_runtime": 37.002, + "eval_samples_per_second": 27.026, + "eval_steps_per_second": 1.135, + "step": 1620 + }, + { + "epoch": 2.0467171717171717, + "grad_norm": 3.3556408882141113, + "learning_rate": 1.803086997193639e-06, + "loss": 1.4252, + "step": 1621 + }, + { + "epoch": 2.047979797979798, + "grad_norm": 3.6319949626922607, + "learning_rate": 1.8007483629560339e-06, + "loss": 1.4634, + "step": 1622 + }, + { + "epoch": 2.0492424242424243, + "grad_norm": 3.4495089054107666, + "learning_rate": 1.7984097287184285e-06, + "loss": 1.4066, + "step": 1623 + }, + { + "epoch": 2.0505050505050506, + "grad_norm": 3.182297945022583, + "learning_rate": 1.7960710944808231e-06, + "loss": 1.3899, + "step": 1624 + }, + { + "epoch": 2.051767676767677, + "grad_norm": 3.097158193588257, + "learning_rate": 1.7937324602432182e-06, + "loss": 1.293, + "step": 1625 + }, + { + "epoch": 2.053030303030303, + "grad_norm": 3.429065704345703, + "learning_rate": 1.791393826005613e-06, + "loss": 1.3905, + "step": 1626 + }, + { + "epoch": 2.0542929292929295, + "grad_norm": 3.183366298675537, + "learning_rate": 1.7890551917680077e-06, + "loss": 1.3444, + "step": 1627 + }, + { + "epoch": 2.0555555555555554, + "grad_norm": 3.175612449645996, + "learning_rate": 1.7867165575304025e-06, + "loss": 1.3985, + "step": 1628 + }, + { + "epoch": 2.0568181818181817, + "grad_norm": 3.26491641998291, + "learning_rate": 1.7843779232927971e-06, + "loss": 1.3607, + "step": 1629 + }, + { + "epoch": 2.058080808080808, + "grad_norm": 3.2279412746429443, + "learning_rate": 1.7820392890551918e-06, + "loss": 1.3545, + "step": 1630 + }, + { + "epoch": 2.058080808080808, + "eval_loss": 1.6729198694229126, + "eval_runtime": 37.0369, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 1630 + }, + { + "epoch": 2.0593434343434343, + "grad_norm": 3.3255503177642822, + "learning_rate": 1.7797006548175866e-06, + "loss": 1.3938, + "step": 1631 + }, + { + "epoch": 2.0606060606060606, + "grad_norm": 3.2912404537200928, + "learning_rate": 1.7773620205799812e-06, + "loss": 1.366, + "step": 1632 + }, + { + "epoch": 2.061868686868687, + "grad_norm": 3.2314395904541016, + "learning_rate": 1.7750233863423763e-06, + "loss": 1.3298, + "step": 1633 + }, + { + "epoch": 2.063131313131313, + "grad_norm": 3.3304178714752197, + "learning_rate": 1.772684752104771e-06, + "loss": 1.3592, + "step": 1634 + }, + { + "epoch": 2.0643939393939394, + "grad_norm": 3.3113889694213867, + "learning_rate": 1.7703461178671658e-06, + "loss": 1.3865, + "step": 1635 + }, + { + "epoch": 2.0656565656565657, + "grad_norm": 3.260434150695801, + "learning_rate": 1.7680074836295604e-06, + "loss": 1.3306, + "step": 1636 + }, + { + "epoch": 2.066919191919192, + "grad_norm": 3.2890963554382324, + "learning_rate": 1.7656688493919552e-06, + "loss": 1.3533, + "step": 1637 + }, + { + "epoch": 2.0681818181818183, + "grad_norm": 3.3027193546295166, + "learning_rate": 1.7633302151543499e-06, + "loss": 1.3746, + "step": 1638 + }, + { + "epoch": 2.0694444444444446, + "grad_norm": 3.3163208961486816, + "learning_rate": 1.7609915809167447e-06, + "loss": 1.4113, + "step": 1639 + }, + { + "epoch": 2.0707070707070705, + "grad_norm": 3.300557851791382, + "learning_rate": 1.7586529466791396e-06, + "loss": 1.3669, + "step": 1640 + }, + { + "epoch": 2.0707070707070705, + "eval_loss": 1.6676570177078247, + "eval_runtime": 37.0589, + "eval_samples_per_second": 26.984, + "eval_steps_per_second": 1.133, + "step": 1640 + }, + { + "epoch": 2.071969696969697, + "grad_norm": 3.305338144302368, + "learning_rate": 1.7563143124415344e-06, + "loss": 1.3969, + "step": 1641 + }, + { + "epoch": 2.073232323232323, + "grad_norm": 3.2047572135925293, + "learning_rate": 1.753975678203929e-06, + "loss": 1.3856, + "step": 1642 + }, + { + "epoch": 2.0744949494949494, + "grad_norm": 3.575038433074951, + "learning_rate": 1.7516370439663239e-06, + "loss": 1.3947, + "step": 1643 + }, + { + "epoch": 2.0757575757575757, + "grad_norm": 3.44581937789917, + "learning_rate": 1.7492984097287185e-06, + "loss": 1.4037, + "step": 1644 + }, + { + "epoch": 2.077020202020202, + "grad_norm": 3.574326753616333, + "learning_rate": 1.7469597754911134e-06, + "loss": 1.4216, + "step": 1645 + }, + { + "epoch": 2.0782828282828283, + "grad_norm": 3.2762558460235596, + "learning_rate": 1.744621141253508e-06, + "loss": 1.3621, + "step": 1646 + }, + { + "epoch": 2.0795454545454546, + "grad_norm": 3.3749659061431885, + "learning_rate": 1.7422825070159028e-06, + "loss": 1.3762, + "step": 1647 + }, + { + "epoch": 2.080808080808081, + "grad_norm": 3.287447214126587, + "learning_rate": 1.7399438727782977e-06, + "loss": 1.3565, + "step": 1648 + }, + { + "epoch": 2.082070707070707, + "grad_norm": 3.5107264518737793, + "learning_rate": 1.7376052385406925e-06, + "loss": 1.392, + "step": 1649 + }, + { + "epoch": 2.0833333333333335, + "grad_norm": 3.319650173187256, + "learning_rate": 1.7352666043030871e-06, + "loss": 1.3835, + "step": 1650 + }, + { + "epoch": 2.0833333333333335, + "eval_loss": 1.6727185249328613, + "eval_runtime": 37.0658, + "eval_samples_per_second": 26.979, + "eval_steps_per_second": 1.133, + "step": 1650 + }, + { + "epoch": 2.08459595959596, + "grad_norm": 3.4447925090789795, + "learning_rate": 1.732927970065482e-06, + "loss": 1.4186, + "step": 1651 + }, + { + "epoch": 2.0858585858585856, + "grad_norm": 3.5944249629974365, + "learning_rate": 1.7305893358278766e-06, + "loss": 1.4461, + "step": 1652 + }, + { + "epoch": 2.087121212121212, + "grad_norm": 3.5437049865722656, + "learning_rate": 1.7282507015902715e-06, + "loss": 1.3343, + "step": 1653 + }, + { + "epoch": 2.0883838383838382, + "grad_norm": 3.361407518386841, + "learning_rate": 1.725912067352666e-06, + "loss": 1.3653, + "step": 1654 + }, + { + "epoch": 2.0896464646464645, + "grad_norm": 3.4728963375091553, + "learning_rate": 1.7235734331150607e-06, + "loss": 1.4297, + "step": 1655 + }, + { + "epoch": 2.090909090909091, + "grad_norm": 3.3794610500335693, + "learning_rate": 1.7212347988774558e-06, + "loss": 1.3941, + "step": 1656 + }, + { + "epoch": 2.092171717171717, + "grad_norm": 3.280857801437378, + "learning_rate": 1.7188961646398506e-06, + "loss": 1.3445, + "step": 1657 + }, + { + "epoch": 2.0934343434343434, + "grad_norm": 3.29036283493042, + "learning_rate": 1.7165575304022453e-06, + "loss": 1.3581, + "step": 1658 + }, + { + "epoch": 2.0946969696969697, + "grad_norm": 3.1731486320495605, + "learning_rate": 1.7142188961646399e-06, + "loss": 1.3873, + "step": 1659 + }, + { + "epoch": 2.095959595959596, + "grad_norm": 3.2283008098602295, + "learning_rate": 1.7118802619270347e-06, + "loss": 1.3462, + "step": 1660 + }, + { + "epoch": 2.095959595959596, + "eval_loss": 1.6726486682891846, + "eval_runtime": 37.0617, + "eval_samples_per_second": 26.982, + "eval_steps_per_second": 1.133, + "step": 1660 + }, + { + "epoch": 2.0972222222222223, + "grad_norm": 3.329171657562256, + "learning_rate": 1.7095416276894294e-06, + "loss": 1.3974, + "step": 1661 + }, + { + "epoch": 2.0984848484848486, + "grad_norm": 3.4113657474517822, + "learning_rate": 1.7072029934518242e-06, + "loss": 1.4237, + "step": 1662 + }, + { + "epoch": 2.099747474747475, + "grad_norm": 3.431184768676758, + "learning_rate": 1.7048643592142193e-06, + "loss": 1.38, + "step": 1663 + }, + { + "epoch": 2.101010101010101, + "grad_norm": 3.5471761226654053, + "learning_rate": 1.7025257249766139e-06, + "loss": 1.3762, + "step": 1664 + }, + { + "epoch": 2.102272727272727, + "grad_norm": 3.43166446685791, + "learning_rate": 1.7001870907390085e-06, + "loss": 1.4371, + "step": 1665 + }, + { + "epoch": 2.1035353535353534, + "grad_norm": 3.28086256980896, + "learning_rate": 1.6978484565014034e-06, + "loss": 1.4062, + "step": 1666 + }, + { + "epoch": 2.1047979797979797, + "grad_norm": 3.166083335876465, + "learning_rate": 1.695509822263798e-06, + "loss": 1.3535, + "step": 1667 + }, + { + "epoch": 2.106060606060606, + "grad_norm": 3.4707589149475098, + "learning_rate": 1.6931711880261928e-06, + "loss": 1.4237, + "step": 1668 + }, + { + "epoch": 2.1073232323232323, + "grad_norm": 3.5033648014068604, + "learning_rate": 1.6908325537885875e-06, + "loss": 1.4331, + "step": 1669 + }, + { + "epoch": 2.1085858585858586, + "grad_norm": 3.3915514945983887, + "learning_rate": 1.6884939195509823e-06, + "loss": 1.3783, + "step": 1670 + }, + { + "epoch": 2.1085858585858586, + "eval_loss": 1.6657648086547852, + "eval_runtime": 37.0742, + "eval_samples_per_second": 26.973, + "eval_steps_per_second": 1.133, + "step": 1670 + }, + { + "epoch": 2.109848484848485, + "grad_norm": 3.3420865535736084, + "learning_rate": 1.6861552853133772e-06, + "loss": 1.4016, + "step": 1671 + }, + { + "epoch": 2.111111111111111, + "grad_norm": 3.382840394973755, + "learning_rate": 1.683816651075772e-06, + "loss": 1.4209, + "step": 1672 + }, + { + "epoch": 2.1123737373737375, + "grad_norm": 3.3109424114227295, + "learning_rate": 1.6814780168381666e-06, + "loss": 1.3716, + "step": 1673 + }, + { + "epoch": 2.1136363636363638, + "grad_norm": 3.3781373500823975, + "learning_rate": 1.6791393826005615e-06, + "loss": 1.4106, + "step": 1674 + }, + { + "epoch": 2.11489898989899, + "grad_norm": 3.29980731010437, + "learning_rate": 1.676800748362956e-06, + "loss": 1.3628, + "step": 1675 + }, + { + "epoch": 2.1161616161616164, + "grad_norm": 3.38468599319458, + "learning_rate": 1.674462114125351e-06, + "loss": 1.3478, + "step": 1676 + }, + { + "epoch": 2.117424242424242, + "grad_norm": 3.4106788635253906, + "learning_rate": 1.6721234798877456e-06, + "loss": 1.391, + "step": 1677 + }, + { + "epoch": 2.1186868686868685, + "grad_norm": 3.412184476852417, + "learning_rate": 1.6697848456501404e-06, + "loss": 1.4245, + "step": 1678 + }, + { + "epoch": 2.119949494949495, + "grad_norm": 3.3659188747406006, + "learning_rate": 1.6674462114125353e-06, + "loss": 1.3358, + "step": 1679 + }, + { + "epoch": 2.121212121212121, + "grad_norm": 3.241384744644165, + "learning_rate": 1.66510757717493e-06, + "loss": 1.3855, + "step": 1680 + }, + { + "epoch": 2.121212121212121, + "eval_loss": 1.6706457138061523, + "eval_runtime": 37.0651, + "eval_samples_per_second": 26.98, + "eval_steps_per_second": 1.133, + "step": 1680 + }, + { + "epoch": 2.1224747474747474, + "grad_norm": 3.3625874519348145, + "learning_rate": 1.6627689429373247e-06, + "loss": 1.3955, + "step": 1681 + }, + { + "epoch": 2.1237373737373737, + "grad_norm": 3.378143787384033, + "learning_rate": 1.6604303086997196e-06, + "loss": 1.3635, + "step": 1682 + }, + { + "epoch": 2.125, + "grad_norm": 3.4872324466705322, + "learning_rate": 1.6580916744621142e-06, + "loss": 1.4037, + "step": 1683 + }, + { + "epoch": 2.1262626262626263, + "grad_norm": 3.4733831882476807, + "learning_rate": 1.655753040224509e-06, + "loss": 1.3846, + "step": 1684 + }, + { + "epoch": 2.1275252525252526, + "grad_norm": 3.4518625736236572, + "learning_rate": 1.6534144059869037e-06, + "loss": 1.3984, + "step": 1685 + }, + { + "epoch": 2.128787878787879, + "grad_norm": 3.3702666759490967, + "learning_rate": 1.6510757717492987e-06, + "loss": 1.3712, + "step": 1686 + }, + { + "epoch": 2.130050505050505, + "grad_norm": 3.357708215713501, + "learning_rate": 1.6487371375116934e-06, + "loss": 1.4204, + "step": 1687 + }, + { + "epoch": 2.1313131313131315, + "grad_norm": 3.3902523517608643, + "learning_rate": 1.6463985032740882e-06, + "loss": 1.439, + "step": 1688 + }, + { + "epoch": 2.132575757575758, + "grad_norm": 3.246450424194336, + "learning_rate": 1.6440598690364828e-06, + "loss": 1.4061, + "step": 1689 + }, + { + "epoch": 2.1338383838383836, + "grad_norm": 3.5247316360473633, + "learning_rate": 1.6417212347988775e-06, + "loss": 1.4148, + "step": 1690 + }, + { + "epoch": 2.1338383838383836, + "eval_loss": 1.6684828996658325, + "eval_runtime": 37.3778, + "eval_samples_per_second": 26.754, + "eval_steps_per_second": 1.124, + "step": 1690 + }, + { + "epoch": 2.13510101010101, + "grad_norm": 3.3932440280914307, + "learning_rate": 1.6393826005612723e-06, + "loss": 1.3756, + "step": 1691 + }, + { + "epoch": 2.1363636363636362, + "grad_norm": 3.5242092609405518, + "learning_rate": 1.637043966323667e-06, + "loss": 1.4009, + "step": 1692 + }, + { + "epoch": 2.1376262626262625, + "grad_norm": 3.464235544204712, + "learning_rate": 1.6347053320860618e-06, + "loss": 1.4129, + "step": 1693 + }, + { + "epoch": 2.138888888888889, + "grad_norm": 3.4409964084625244, + "learning_rate": 1.6323666978484568e-06, + "loss": 1.3969, + "step": 1694 + }, + { + "epoch": 2.140151515151515, + "grad_norm": 3.329281806945801, + "learning_rate": 1.6300280636108515e-06, + "loss": 1.367, + "step": 1695 + }, + { + "epoch": 2.1414141414141414, + "grad_norm": 3.28047251701355, + "learning_rate": 1.6276894293732461e-06, + "loss": 1.3994, + "step": 1696 + }, + { + "epoch": 2.1426767676767677, + "grad_norm": 3.1937954425811768, + "learning_rate": 1.625350795135641e-06, + "loss": 1.3626, + "step": 1697 + }, + { + "epoch": 2.143939393939394, + "grad_norm": 3.5084168910980225, + "learning_rate": 1.6230121608980356e-06, + "loss": 1.4067, + "step": 1698 + }, + { + "epoch": 2.1452020202020203, + "grad_norm": 3.3637681007385254, + "learning_rate": 1.6206735266604304e-06, + "loss": 1.3619, + "step": 1699 + }, + { + "epoch": 2.1464646464646466, + "grad_norm": 3.2855823040008545, + "learning_rate": 1.618334892422825e-06, + "loss": 1.4227, + "step": 1700 + }, + { + "epoch": 2.1464646464646466, + "eval_loss": 1.6681337356567383, + "eval_runtime": 37.4375, + "eval_samples_per_second": 26.711, + "eval_steps_per_second": 1.122, + "step": 1700 + } + ], + "logging_steps": 1, + "max_steps": 2376, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 5.299425995915264e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-1700/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/latest b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/latest new file mode 100644 index 0000000000000000000000000000000000000000..4b10acccf3e8395339ff8799cea202bbc54d7f7d --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/latest @@ -0,0 +1 @@ +global_step900 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..8991b49e9c2a43fc527dab9e09ad8171f0cc5943 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/tokenizer_config.json @@ -0,0 +1,2086 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128257": { + "content": "🅂", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + }, + "128258": { + "content": "🄿", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..5287b20560874fca907a132149d3a33ad4cdfc93 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/trainer_state.json @@ -0,0 +1,7053 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.1363636363636362, + "eval_steps": 10, + "global_step": 900, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0012626262626262627, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8571, + "step": 1 + }, + { + "epoch": 0.0025252525252525255, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7758, + "step": 2 + }, + { + "epoch": 0.003787878787878788, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8383, + "step": 3 + }, + { + "epoch": 0.005050505050505051, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.9088, + "step": 4 + }, + { + "epoch": 0.006313131313131313, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8748, + "step": 5 + }, + { + "epoch": 0.007575757575757576, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8575, + "step": 6 + }, + { + "epoch": 0.008838383838383838, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8271, + "step": 7 + }, + { + "epoch": 0.010101010101010102, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8336, + "step": 8 + }, + { + "epoch": 0.011363636363636364, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8037, + "step": 9 + }, + { + "epoch": 0.012626262626262626, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8323, + "step": 10 + }, + { + "epoch": 0.012626262626262626, + "eval_loss": 2.8736937046051025, + "eval_runtime": 37.7802, + "eval_samples_per_second": 26.469, + "eval_steps_per_second": 1.112, + "step": 10 + }, + { + "epoch": 0.013888888888888888, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8653, + "step": 11 + }, + { + "epoch": 0.015151515151515152, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8104, + "step": 12 + }, + { + "epoch": 0.016414141414141416, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8489, + "step": 13 + }, + { + "epoch": 0.017676767676767676, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8442, + "step": 14 + }, + { + "epoch": 0.01893939393939394, + "grad_norm": 5.356770992279053, + "learning_rate": 2.1008403361344538e-08, + "loss": 2.8207, + "step": 15 + }, + { + "epoch": 0.020202020202020204, + "grad_norm": 4.954462051391602, + "learning_rate": 4.2016806722689076e-08, + "loss": 2.8104, + "step": 16 + }, + { + "epoch": 0.021464646464646464, + "grad_norm": 5.248059272766113, + "learning_rate": 6.302521008403361e-08, + "loss": 2.8178, + "step": 17 + }, + { + "epoch": 0.022727272727272728, + "grad_norm": 5.277204513549805, + "learning_rate": 8.403361344537815e-08, + "loss": 2.8255, + "step": 18 + }, + { + "epoch": 0.023989898989898988, + "grad_norm": 5.344228267669678, + "learning_rate": 1.050420168067227e-07, + "loss": 2.8018, + "step": 19 + }, + { + "epoch": 0.025252525252525252, + "grad_norm": 5.004310131072998, + "learning_rate": 1.2605042016806723e-07, + "loss": 2.8337, + "step": 20 + }, + { + "epoch": 0.025252525252525252, + "eval_loss": 2.872413158416748, + "eval_runtime": 37.7426, + "eval_samples_per_second": 26.495, + "eval_steps_per_second": 1.113, + "step": 20 + }, + { + "epoch": 0.026515151515151516, + "grad_norm": 5.209287166595459, + "learning_rate": 1.4705882352941178e-07, + "loss": 2.8246, + "step": 21 + }, + { + "epoch": 0.027777777777777776, + "grad_norm": 5.369327545166016, + "learning_rate": 1.680672268907563e-07, + "loss": 2.8791, + "step": 22 + }, + { + "epoch": 0.02904040404040404, + "grad_norm": 4.646446228027344, + "learning_rate": 1.8907563025210085e-07, + "loss": 2.8425, + "step": 23 + }, + { + "epoch": 0.030303030303030304, + "grad_norm": 4.949256420135498, + "learning_rate": 2.100840336134454e-07, + "loss": 2.8629, + "step": 24 + }, + { + "epoch": 0.03156565656565657, + "grad_norm": 4.709571838378906, + "learning_rate": 2.3109243697478996e-07, + "loss": 2.8264, + "step": 25 + }, + { + "epoch": 0.03282828282828283, + "grad_norm": 4.519557952880859, + "learning_rate": 2.5210084033613445e-07, + "loss": 2.8143, + "step": 26 + }, + { + "epoch": 0.03409090909090909, + "grad_norm": 4.445594310760498, + "learning_rate": 2.73109243697479e-07, + "loss": 2.8031, + "step": 27 + }, + { + "epoch": 0.03535353535353535, + "grad_norm": 4.26489782333374, + "learning_rate": 2.9411764705882356e-07, + "loss": 2.7719, + "step": 28 + }, + { + "epoch": 0.036616161616161616, + "grad_norm": 4.153639793395996, + "learning_rate": 3.151260504201681e-07, + "loss": 2.8251, + "step": 29 + }, + { + "epoch": 0.03787878787878788, + "grad_norm": 3.40531325340271, + "learning_rate": 3.361344537815126e-07, + "loss": 2.849, + "step": 30 + }, + { + "epoch": 0.03787878787878788, + "eval_loss": 2.825382709503174, + "eval_runtime": 37.0715, + "eval_samples_per_second": 26.975, + "eval_steps_per_second": 1.133, + "step": 30 + }, + { + "epoch": 0.039141414141414144, + "grad_norm": 3.4759507179260254, + "learning_rate": 3.5714285714285716e-07, + "loss": 2.7653, + "step": 31 + }, + { + "epoch": 0.04040404040404041, + "grad_norm": 3.4147374629974365, + "learning_rate": 3.781512605042017e-07, + "loss": 2.7916, + "step": 32 + }, + { + "epoch": 0.041666666666666664, + "grad_norm": 3.2775769233703613, + "learning_rate": 3.991596638655462e-07, + "loss": 2.8072, + "step": 33 + }, + { + "epoch": 0.04292929292929293, + "grad_norm": 3.3541715145111084, + "learning_rate": 4.201680672268908e-07, + "loss": 2.7944, + "step": 34 + }, + { + "epoch": 0.04419191919191919, + "grad_norm": 2.991971969604492, + "learning_rate": 4.4117647058823536e-07, + "loss": 2.7827, + "step": 35 + }, + { + "epoch": 0.045454545454545456, + "grad_norm": 2.9127862453460693, + "learning_rate": 4.621848739495799e-07, + "loss": 2.7332, + "step": 36 + }, + { + "epoch": 0.04671717171717172, + "grad_norm": 2.839707136154175, + "learning_rate": 4.831932773109245e-07, + "loss": 2.7568, + "step": 37 + }, + { + "epoch": 0.047979797979797977, + "grad_norm": 2.9395432472229004, + "learning_rate": 5.042016806722689e-07, + "loss": 2.7445, + "step": 38 + }, + { + "epoch": 0.04924242424242424, + "grad_norm": 2.869896650314331, + "learning_rate": 5.252100840336135e-07, + "loss": 2.7157, + "step": 39 + }, + { + "epoch": 0.050505050505050504, + "grad_norm": 2.9064571857452393, + "learning_rate": 5.46218487394958e-07, + "loss": 2.7049, + "step": 40 + }, + { + "epoch": 0.050505050505050504, + "eval_loss": 2.7477543354034424, + "eval_runtime": 37.0669, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 40 + }, + { + "epoch": 0.05176767676767677, + "grad_norm": 2.6264421939849854, + "learning_rate": 5.672268907563026e-07, + "loss": 2.7651, + "step": 41 + }, + { + "epoch": 0.05303030303030303, + "grad_norm": 2.518627405166626, + "learning_rate": 5.882352941176471e-07, + "loss": 2.7241, + "step": 42 + }, + { + "epoch": 0.054292929292929296, + "grad_norm": 2.521034002304077, + "learning_rate": 6.092436974789917e-07, + "loss": 2.6691, + "step": 43 + }, + { + "epoch": 0.05555555555555555, + "grad_norm": 2.4679362773895264, + "learning_rate": 6.302521008403362e-07, + "loss": 2.743, + "step": 44 + }, + { + "epoch": 0.056818181818181816, + "grad_norm": 2.5817418098449707, + "learning_rate": 6.512605042016808e-07, + "loss": 2.7067, + "step": 45 + }, + { + "epoch": 0.05808080808080808, + "grad_norm": 2.3704075813293457, + "learning_rate": 6.722689075630252e-07, + "loss": 2.6495, + "step": 46 + }, + { + "epoch": 0.059343434343434344, + "grad_norm": 2.2321906089782715, + "learning_rate": 6.932773109243699e-07, + "loss": 2.6606, + "step": 47 + }, + { + "epoch": 0.06060606060606061, + "grad_norm": 1.985407829284668, + "learning_rate": 7.142857142857143e-07, + "loss": 2.662, + "step": 48 + }, + { + "epoch": 0.06186868686868687, + "grad_norm": 1.945688247680664, + "learning_rate": 7.352941176470589e-07, + "loss": 2.6319, + "step": 49 + }, + { + "epoch": 0.06313131313131314, + "grad_norm": 1.9392870664596558, + "learning_rate": 7.563025210084034e-07, + "loss": 2.6094, + "step": 50 + }, + { + "epoch": 0.06313131313131314, + "eval_loss": 2.6554172039031982, + "eval_runtime": 37.0569, + "eval_samples_per_second": 26.986, + "eval_steps_per_second": 1.133, + "step": 50 + }, + { + "epoch": 0.06439393939393939, + "grad_norm": 1.9975577592849731, + "learning_rate": 7.77310924369748e-07, + "loss": 2.6045, + "step": 51 + }, + { + "epoch": 0.06565656565656566, + "grad_norm": 1.9733693599700928, + "learning_rate": 7.983193277310924e-07, + "loss": 2.649, + "step": 52 + }, + { + "epoch": 0.06691919191919192, + "grad_norm": 1.7324917316436768, + "learning_rate": 8.193277310924371e-07, + "loss": 2.5689, + "step": 53 + }, + { + "epoch": 0.06818181818181818, + "grad_norm": 1.7553645372390747, + "learning_rate": 8.403361344537816e-07, + "loss": 2.644, + "step": 54 + }, + { + "epoch": 0.06944444444444445, + "grad_norm": 1.8163801431655884, + "learning_rate": 8.613445378151261e-07, + "loss": 2.4943, + "step": 55 + }, + { + "epoch": 0.0707070707070707, + "grad_norm": 1.7717015743255615, + "learning_rate": 8.823529411764707e-07, + "loss": 2.6093, + "step": 56 + }, + { + "epoch": 0.07196969696969698, + "grad_norm": 1.7013663053512573, + "learning_rate": 9.033613445378152e-07, + "loss": 2.6376, + "step": 57 + }, + { + "epoch": 0.07323232323232323, + "grad_norm": 1.752943992614746, + "learning_rate": 9.243697478991598e-07, + "loss": 2.5172, + "step": 58 + }, + { + "epoch": 0.07449494949494949, + "grad_norm": 1.5362415313720703, + "learning_rate": 9.453781512605043e-07, + "loss": 2.6129, + "step": 59 + }, + { + "epoch": 0.07575757575757576, + "grad_norm": 1.514885425567627, + "learning_rate": 9.66386554621849e-07, + "loss": 2.6064, + "step": 60 + }, + { + "epoch": 0.07575757575757576, + "eval_loss": 2.5934340953826904, + "eval_runtime": 37.0782, + "eval_samples_per_second": 26.97, + "eval_steps_per_second": 1.133, + "step": 60 + }, + { + "epoch": 0.07702020202020202, + "grad_norm": 1.5915579795837402, + "learning_rate": 9.873949579831934e-07, + "loss": 2.5491, + "step": 61 + }, + { + "epoch": 0.07828282828282829, + "grad_norm": 1.601773977279663, + "learning_rate": 1.0084033613445378e-06, + "loss": 2.5685, + "step": 62 + }, + { + "epoch": 0.07954545454545454, + "grad_norm": 1.8327804803848267, + "learning_rate": 1.0294117647058825e-06, + "loss": 2.5763, + "step": 63 + }, + { + "epoch": 0.08080808080808081, + "grad_norm": 1.5276741981506348, + "learning_rate": 1.050420168067227e-06, + "loss": 2.5494, + "step": 64 + }, + { + "epoch": 0.08207070707070707, + "grad_norm": 1.7065422534942627, + "learning_rate": 1.0714285714285714e-06, + "loss": 2.449, + "step": 65 + }, + { + "epoch": 0.08333333333333333, + "grad_norm": 2.0551910400390625, + "learning_rate": 1.092436974789916e-06, + "loss": 2.4371, + "step": 66 + }, + { + "epoch": 0.0845959595959596, + "grad_norm": 2.154855728149414, + "learning_rate": 1.1134453781512607e-06, + "loss": 2.5417, + "step": 67 + }, + { + "epoch": 0.08585858585858586, + "grad_norm": 1.7922390699386597, + "learning_rate": 1.1344537815126051e-06, + "loss": 2.4754, + "step": 68 + }, + { + "epoch": 0.08712121212121213, + "grad_norm": 1.745632290840149, + "learning_rate": 1.1554621848739498e-06, + "loss": 2.4669, + "step": 69 + }, + { + "epoch": 0.08838383838383838, + "grad_norm": 1.8907713890075684, + "learning_rate": 1.1764705882352942e-06, + "loss": 2.4943, + "step": 70 + }, + { + "epoch": 0.08838383838383838, + "eval_loss": 2.518237829208374, + "eval_runtime": 37.074, + "eval_samples_per_second": 26.973, + "eval_steps_per_second": 1.133, + "step": 70 + }, + { + "epoch": 0.08964646464646464, + "grad_norm": 1.8178164958953857, + "learning_rate": 1.1974789915966389e-06, + "loss": 2.4891, + "step": 71 + }, + { + "epoch": 0.09090909090909091, + "grad_norm": 1.970969796180725, + "learning_rate": 1.2184873949579833e-06, + "loss": 2.4797, + "step": 72 + }, + { + "epoch": 0.09217171717171717, + "grad_norm": 2.0211431980133057, + "learning_rate": 1.2394957983193278e-06, + "loss": 2.4716, + "step": 73 + }, + { + "epoch": 0.09343434343434344, + "grad_norm": 2.650787591934204, + "learning_rate": 1.2605042016806724e-06, + "loss": 2.4915, + "step": 74 + }, + { + "epoch": 0.0946969696969697, + "grad_norm": 2.235499143600464, + "learning_rate": 1.2815126050420169e-06, + "loss": 2.3917, + "step": 75 + }, + { + "epoch": 0.09595959595959595, + "grad_norm": 2.245999813079834, + "learning_rate": 1.3025210084033615e-06, + "loss": 2.4942, + "step": 76 + }, + { + "epoch": 0.09722222222222222, + "grad_norm": 2.4414420127868652, + "learning_rate": 1.323529411764706e-06, + "loss": 2.4004, + "step": 77 + }, + { + "epoch": 0.09848484848484848, + "grad_norm": 2.279937982559204, + "learning_rate": 1.3445378151260504e-06, + "loss": 2.4043, + "step": 78 + }, + { + "epoch": 0.09974747474747475, + "grad_norm": 2.3830068111419678, + "learning_rate": 1.3655462184873949e-06, + "loss": 2.4096, + "step": 79 + }, + { + "epoch": 0.10101010101010101, + "grad_norm": 2.0119733810424805, + "learning_rate": 1.3865546218487397e-06, + "loss": 2.4227, + "step": 80 + }, + { + "epoch": 0.10101010101010101, + "eval_loss": 2.4316186904907227, + "eval_runtime": 37.0837, + "eval_samples_per_second": 26.966, + "eval_steps_per_second": 1.133, + "step": 80 + }, + { + "epoch": 0.10227272727272728, + "grad_norm": 2.1273374557495117, + "learning_rate": 1.4075630252100842e-06, + "loss": 2.3541, + "step": 81 + }, + { + "epoch": 0.10353535353535354, + "grad_norm": 2.6852898597717285, + "learning_rate": 1.4285714285714286e-06, + "loss": 2.3321, + "step": 82 + }, + { + "epoch": 0.10479797979797979, + "grad_norm": 2.2622721195220947, + "learning_rate": 1.4495798319327733e-06, + "loss": 2.3166, + "step": 83 + }, + { + "epoch": 0.10606060606060606, + "grad_norm": 2.98258900642395, + "learning_rate": 1.4705882352941177e-06, + "loss": 2.3733, + "step": 84 + }, + { + "epoch": 0.10732323232323232, + "grad_norm": 2.5140459537506104, + "learning_rate": 1.4915966386554624e-06, + "loss": 2.314, + "step": 85 + }, + { + "epoch": 0.10858585858585859, + "grad_norm": 2.637608051300049, + "learning_rate": 1.5126050420168068e-06, + "loss": 2.3479, + "step": 86 + }, + { + "epoch": 0.10984848484848485, + "grad_norm": 2.399508476257324, + "learning_rate": 1.5336134453781515e-06, + "loss": 2.2932, + "step": 87 + }, + { + "epoch": 0.1111111111111111, + "grad_norm": 3.0835793018341064, + "learning_rate": 1.554621848739496e-06, + "loss": 2.2886, + "step": 88 + }, + { + "epoch": 0.11237373737373738, + "grad_norm": 3.288769483566284, + "learning_rate": 1.5756302521008404e-06, + "loss": 2.2914, + "step": 89 + }, + { + "epoch": 0.11363636363636363, + "grad_norm": 2.538361072540283, + "learning_rate": 1.5966386554621848e-06, + "loss": 2.3247, + "step": 90 + }, + { + "epoch": 0.11363636363636363, + "eval_loss": 2.3510072231292725, + "eval_runtime": 37.0847, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 90 + }, + { + "epoch": 0.1148989898989899, + "grad_norm": 2.8016724586486816, + "learning_rate": 1.6176470588235297e-06, + "loss": 2.281, + "step": 91 + }, + { + "epoch": 0.11616161616161616, + "grad_norm": 2.9251859188079834, + "learning_rate": 1.6386554621848741e-06, + "loss": 2.2715, + "step": 92 + }, + { + "epoch": 0.11742424242424243, + "grad_norm": 2.595750570297241, + "learning_rate": 1.6596638655462186e-06, + "loss": 2.2127, + "step": 93 + }, + { + "epoch": 0.11868686868686869, + "grad_norm": 2.6989378929138184, + "learning_rate": 1.6806722689075632e-06, + "loss": 2.2623, + "step": 94 + }, + { + "epoch": 0.11994949494949494, + "grad_norm": 2.554863929748535, + "learning_rate": 1.7016806722689077e-06, + "loss": 2.2649, + "step": 95 + }, + { + "epoch": 0.12121212121212122, + "grad_norm": 2.7013020515441895, + "learning_rate": 1.7226890756302521e-06, + "loss": 2.2428, + "step": 96 + }, + { + "epoch": 0.12247474747474747, + "grad_norm": 2.7415151596069336, + "learning_rate": 1.7436974789915968e-06, + "loss": 2.1882, + "step": 97 + }, + { + "epoch": 0.12373737373737374, + "grad_norm": 2.672806739807129, + "learning_rate": 1.7647058823529414e-06, + "loss": 2.2302, + "step": 98 + }, + { + "epoch": 0.125, + "grad_norm": 2.9508326053619385, + "learning_rate": 1.7857142857142859e-06, + "loss": 2.1894, + "step": 99 + }, + { + "epoch": 0.12626262626262627, + "grad_norm": 2.4668166637420654, + "learning_rate": 1.8067226890756303e-06, + "loss": 2.2043, + "step": 100 + }, + { + "epoch": 0.12626262626262627, + "eval_loss": 2.2788307666778564, + "eval_runtime": 37.1156, + "eval_samples_per_second": 26.943, + "eval_steps_per_second": 1.132, + "step": 100 + }, + { + "epoch": 0.1275252525252525, + "grad_norm": 2.8241467475891113, + "learning_rate": 1.8277310924369748e-06, + "loss": 2.2109, + "step": 101 + }, + { + "epoch": 0.12878787878787878, + "grad_norm": 2.4179229736328125, + "learning_rate": 1.8487394957983196e-06, + "loss": 2.2193, + "step": 102 + }, + { + "epoch": 0.13005050505050506, + "grad_norm": 2.714315891265869, + "learning_rate": 1.869747899159664e-06, + "loss": 2.2474, + "step": 103 + }, + { + "epoch": 0.13131313131313133, + "grad_norm": 2.8560359477996826, + "learning_rate": 1.8907563025210085e-06, + "loss": 2.2205, + "step": 104 + }, + { + "epoch": 0.13257575757575757, + "grad_norm": 2.7780051231384277, + "learning_rate": 1.9117647058823528e-06, + "loss": 2.1834, + "step": 105 + }, + { + "epoch": 0.13383838383838384, + "grad_norm": 3.3773467540740967, + "learning_rate": 1.932773109243698e-06, + "loss": 2.2063, + "step": 106 + }, + { + "epoch": 0.1351010101010101, + "grad_norm": 2.927551746368408, + "learning_rate": 1.953781512605042e-06, + "loss": 2.1592, + "step": 107 + }, + { + "epoch": 0.13636363636363635, + "grad_norm": 2.8094165325164795, + "learning_rate": 1.9747899159663867e-06, + "loss": 2.2619, + "step": 108 + }, + { + "epoch": 0.13762626262626262, + "grad_norm": 2.5312564373016357, + "learning_rate": 1.9957983193277314e-06, + "loss": 2.2323, + "step": 109 + }, + { + "epoch": 0.1388888888888889, + "grad_norm": 3.175746202468872, + "learning_rate": 2.0168067226890756e-06, + "loss": 2.1776, + "step": 110 + }, + { + "epoch": 0.1388888888888889, + "eval_loss": 2.2208564281463623, + "eval_runtime": 37.3885, + "eval_samples_per_second": 26.746, + "eval_steps_per_second": 1.123, + "step": 110 + }, + { + "epoch": 0.14015151515151514, + "grad_norm": 2.660886287689209, + "learning_rate": 2.0378151260504203e-06, + "loss": 2.1626, + "step": 111 + }, + { + "epoch": 0.1414141414141414, + "grad_norm": 2.6948511600494385, + "learning_rate": 2.058823529411765e-06, + "loss": 2.1731, + "step": 112 + }, + { + "epoch": 0.14267676767676768, + "grad_norm": 2.826976776123047, + "learning_rate": 2.0798319327731096e-06, + "loss": 2.1865, + "step": 113 + }, + { + "epoch": 0.14393939393939395, + "grad_norm": 3.210033655166626, + "learning_rate": 2.100840336134454e-06, + "loss": 2.1325, + "step": 114 + }, + { + "epoch": 0.1452020202020202, + "grad_norm": 2.509060859680176, + "learning_rate": 2.1218487394957985e-06, + "loss": 2.0926, + "step": 115 + }, + { + "epoch": 0.14646464646464646, + "grad_norm": 3.51330304145813, + "learning_rate": 2.1428571428571427e-06, + "loss": 2.1577, + "step": 116 + }, + { + "epoch": 0.14772727272727273, + "grad_norm": 2.468395471572876, + "learning_rate": 2.163865546218488e-06, + "loss": 2.1, + "step": 117 + }, + { + "epoch": 0.14898989898989898, + "grad_norm": 3.081529378890991, + "learning_rate": 2.184873949579832e-06, + "loss": 2.1082, + "step": 118 + }, + { + "epoch": 0.15025252525252525, + "grad_norm": 2.8856358528137207, + "learning_rate": 2.2058823529411767e-06, + "loss": 2.1288, + "step": 119 + }, + { + "epoch": 0.15151515151515152, + "grad_norm": 3.323089838027954, + "learning_rate": 2.2268907563025214e-06, + "loss": 2.0933, + "step": 120 + }, + { + "epoch": 0.15151515151515152, + "eval_loss": 2.1834487915039062, + "eval_runtime": 37.4344, + "eval_samples_per_second": 26.713, + "eval_steps_per_second": 1.122, + "step": 120 + }, + { + "epoch": 0.1527777777777778, + "grad_norm": 2.8871560096740723, + "learning_rate": 2.2478991596638656e-06, + "loss": 2.1519, + "step": 121 + }, + { + "epoch": 0.15404040404040403, + "grad_norm": 3.0718588829040527, + "learning_rate": 2.2689075630252102e-06, + "loss": 2.0321, + "step": 122 + }, + { + "epoch": 0.1553030303030303, + "grad_norm": 2.820739507675171, + "learning_rate": 2.289915966386555e-06, + "loss": 2.0944, + "step": 123 + }, + { + "epoch": 0.15656565656565657, + "grad_norm": 3.1259310245513916, + "learning_rate": 2.3109243697478996e-06, + "loss": 2.0141, + "step": 124 + }, + { + "epoch": 0.15782828282828282, + "grad_norm": 3.303004741668701, + "learning_rate": 2.331932773109244e-06, + "loss": 2.1095, + "step": 125 + }, + { + "epoch": 0.1590909090909091, + "grad_norm": 3.1838674545288086, + "learning_rate": 2.3529411764705885e-06, + "loss": 2.0727, + "step": 126 + }, + { + "epoch": 0.16035353535353536, + "grad_norm": 3.1013786792755127, + "learning_rate": 2.3739495798319327e-06, + "loss": 2.0963, + "step": 127 + }, + { + "epoch": 0.16161616161616163, + "grad_norm": 2.803283214569092, + "learning_rate": 2.3949579831932778e-06, + "loss": 2.0551, + "step": 128 + }, + { + "epoch": 0.16287878787878787, + "grad_norm": 3.1616814136505127, + "learning_rate": 2.415966386554622e-06, + "loss": 2.0485, + "step": 129 + }, + { + "epoch": 0.16414141414141414, + "grad_norm": 3.333782196044922, + "learning_rate": 2.4369747899159667e-06, + "loss": 2.041, + "step": 130 + }, + { + "epoch": 0.16414141414141414, + "eval_loss": 2.150329351425171, + "eval_runtime": 37.494, + "eval_samples_per_second": 26.671, + "eval_steps_per_second": 1.12, + "step": 130 + }, + { + "epoch": 0.16540404040404041, + "grad_norm": 2.9714694023132324, + "learning_rate": 2.457983193277311e-06, + "loss": 2.0686, + "step": 131 + }, + { + "epoch": 0.16666666666666666, + "grad_norm": 3.216895818710327, + "learning_rate": 2.4789915966386555e-06, + "loss": 2.003, + "step": 132 + }, + { + "epoch": 0.16792929292929293, + "grad_norm": 2.9636809825897217, + "learning_rate": 2.5e-06, + "loss": 2.0534, + "step": 133 + }, + { + "epoch": 0.1691919191919192, + "grad_norm": 3.6477177143096924, + "learning_rate": 2.521008403361345e-06, + "loss": 2.077, + "step": 134 + }, + { + "epoch": 0.17045454545454544, + "grad_norm": 2.7437169551849365, + "learning_rate": 2.542016806722689e-06, + "loss": 2.0501, + "step": 135 + }, + { + "epoch": 0.1717171717171717, + "grad_norm": 3.2992663383483887, + "learning_rate": 2.5630252100840338e-06, + "loss": 2.0284, + "step": 136 + }, + { + "epoch": 0.17297979797979798, + "grad_norm": 3.0735950469970703, + "learning_rate": 2.5840336134453784e-06, + "loss": 2.025, + "step": 137 + }, + { + "epoch": 0.17424242424242425, + "grad_norm": 3.588418960571289, + "learning_rate": 2.605042016806723e-06, + "loss": 2.002, + "step": 138 + }, + { + "epoch": 0.1755050505050505, + "grad_norm": 2.8772335052490234, + "learning_rate": 2.6260504201680677e-06, + "loss": 1.9873, + "step": 139 + }, + { + "epoch": 0.17676767676767677, + "grad_norm": 2.9078969955444336, + "learning_rate": 2.647058823529412e-06, + "loss": 2.011, + "step": 140 + }, + { + "epoch": 0.17676767676767677, + "eval_loss": 2.1225531101226807, + "eval_runtime": 37.4531, + "eval_samples_per_second": 26.7, + "eval_steps_per_second": 1.121, + "step": 140 + }, + { + "epoch": 0.17803030303030304, + "grad_norm": 3.556654214859009, + "learning_rate": 2.6680672268907566e-06, + "loss": 2.0401, + "step": 141 + }, + { + "epoch": 0.17929292929292928, + "grad_norm": 2.9375200271606445, + "learning_rate": 2.689075630252101e-06, + "loss": 2.0277, + "step": 142 + }, + { + "epoch": 0.18055555555555555, + "grad_norm": 3.4019014835357666, + "learning_rate": 2.7100840336134455e-06, + "loss": 1.975, + "step": 143 + }, + { + "epoch": 0.18181818181818182, + "grad_norm": 2.7653748989105225, + "learning_rate": 2.7310924369747897e-06, + "loss": 2.0102, + "step": 144 + }, + { + "epoch": 0.1830808080808081, + "grad_norm": 3.4416797161102295, + "learning_rate": 2.752100840336135e-06, + "loss": 2.0093, + "step": 145 + }, + { + "epoch": 0.18434343434343434, + "grad_norm": 3.2531726360321045, + "learning_rate": 2.7731092436974795e-06, + "loss": 1.9852, + "step": 146 + }, + { + "epoch": 0.1856060606060606, + "grad_norm": 3.171502113342285, + "learning_rate": 2.7941176470588237e-06, + "loss": 1.9165, + "step": 147 + }, + { + "epoch": 0.18686868686868688, + "grad_norm": 3.114550828933716, + "learning_rate": 2.8151260504201684e-06, + "loss": 1.9516, + "step": 148 + }, + { + "epoch": 0.18813131313131312, + "grad_norm": 3.3601152896881104, + "learning_rate": 2.8361344537815126e-06, + "loss": 2.0183, + "step": 149 + }, + { + "epoch": 0.1893939393939394, + "grad_norm": 3.059622049331665, + "learning_rate": 2.8571428571428573e-06, + "loss": 1.9682, + "step": 150 + }, + { + "epoch": 0.1893939393939394, + "eval_loss": 2.0962483882904053, + "eval_runtime": 37.2364, + "eval_samples_per_second": 26.855, + "eval_steps_per_second": 1.128, + "step": 150 + }, + { + "epoch": 0.19065656565656566, + "grad_norm": 2.8749427795410156, + "learning_rate": 2.8781512605042015e-06, + "loss": 2.0089, + "step": 151 + }, + { + "epoch": 0.1919191919191919, + "grad_norm": 3.067267894744873, + "learning_rate": 2.8991596638655466e-06, + "loss": 1.9661, + "step": 152 + }, + { + "epoch": 0.19318181818181818, + "grad_norm": 2.7126266956329346, + "learning_rate": 2.9201680672268912e-06, + "loss": 1.982, + "step": 153 + }, + { + "epoch": 0.19444444444444445, + "grad_norm": 3.1330652236938477, + "learning_rate": 2.9411764705882355e-06, + "loss": 1.9753, + "step": 154 + }, + { + "epoch": 0.19570707070707072, + "grad_norm": 3.158313274383545, + "learning_rate": 2.96218487394958e-06, + "loss": 1.9431, + "step": 155 + }, + { + "epoch": 0.19696969696969696, + "grad_norm": 3.30806303024292, + "learning_rate": 2.9831932773109248e-06, + "loss": 1.9812, + "step": 156 + }, + { + "epoch": 0.19823232323232323, + "grad_norm": 3.0344884395599365, + "learning_rate": 3.004201680672269e-06, + "loss": 1.9677, + "step": 157 + }, + { + "epoch": 0.1994949494949495, + "grad_norm": 2.9032909870147705, + "learning_rate": 3.0252100840336137e-06, + "loss": 1.9132, + "step": 158 + }, + { + "epoch": 0.20075757575757575, + "grad_norm": 3.459934711456299, + "learning_rate": 3.046218487394958e-06, + "loss": 1.9581, + "step": 159 + }, + { + "epoch": 0.20202020202020202, + "grad_norm": 2.9455478191375732, + "learning_rate": 3.067226890756303e-06, + "loss": 1.9509, + "step": 160 + }, + { + "epoch": 0.20202020202020202, + "eval_loss": 2.074032783508301, + "eval_runtime": 37.1824, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 160 + }, + { + "epoch": 0.2032828282828283, + "grad_norm": 3.2347211837768555, + "learning_rate": 3.0882352941176476e-06, + "loss": 1.9743, + "step": 161 + }, + { + "epoch": 0.20454545454545456, + "grad_norm": 3.3152947425842285, + "learning_rate": 3.109243697478992e-06, + "loss": 1.8971, + "step": 162 + }, + { + "epoch": 0.2058080808080808, + "grad_norm": 3.035548210144043, + "learning_rate": 3.1302521008403365e-06, + "loss": 1.8762, + "step": 163 + }, + { + "epoch": 0.20707070707070707, + "grad_norm": 3.1004066467285156, + "learning_rate": 3.1512605042016808e-06, + "loss": 1.9512, + "step": 164 + }, + { + "epoch": 0.20833333333333334, + "grad_norm": 3.2280282974243164, + "learning_rate": 3.1722689075630254e-06, + "loss": 1.8856, + "step": 165 + }, + { + "epoch": 0.20959595959595959, + "grad_norm": 3.021969795227051, + "learning_rate": 3.1932773109243696e-06, + "loss": 1.9127, + "step": 166 + }, + { + "epoch": 0.21085858585858586, + "grad_norm": 3.1647000312805176, + "learning_rate": 3.2142857142857147e-06, + "loss": 2.038, + "step": 167 + }, + { + "epoch": 0.21212121212121213, + "grad_norm": 3.0084872245788574, + "learning_rate": 3.2352941176470594e-06, + "loss": 1.9536, + "step": 168 + }, + { + "epoch": 0.21338383838383837, + "grad_norm": 2.951509952545166, + "learning_rate": 3.2563025210084036e-06, + "loss": 1.933, + "step": 169 + }, + { + "epoch": 0.21464646464646464, + "grad_norm": 3.472856283187866, + "learning_rate": 3.2773109243697483e-06, + "loss": 1.934, + "step": 170 + }, + { + "epoch": 0.21464646464646464, + "eval_loss": 2.0450048446655273, + "eval_runtime": 37.1669, + "eval_samples_per_second": 26.906, + "eval_steps_per_second": 1.13, + "step": 170 + }, + { + "epoch": 0.2159090909090909, + "grad_norm": 3.1104896068573, + "learning_rate": 3.2983193277310925e-06, + "loss": 1.9008, + "step": 171 + }, + { + "epoch": 0.21717171717171718, + "grad_norm": 2.847781181335449, + "learning_rate": 3.319327731092437e-06, + "loss": 1.9592, + "step": 172 + }, + { + "epoch": 0.21843434343434343, + "grad_norm": 3.384554624557495, + "learning_rate": 3.3403361344537814e-06, + "loss": 1.9087, + "step": 173 + }, + { + "epoch": 0.2196969696969697, + "grad_norm": 3.221564292907715, + "learning_rate": 3.3613445378151265e-06, + "loss": 1.9107, + "step": 174 + }, + { + "epoch": 0.22095959595959597, + "grad_norm": 3.30204439163208, + "learning_rate": 3.382352941176471e-06, + "loss": 1.8879, + "step": 175 + }, + { + "epoch": 0.2222222222222222, + "grad_norm": 3.549259662628174, + "learning_rate": 3.4033613445378154e-06, + "loss": 1.9592, + "step": 176 + }, + { + "epoch": 0.22348484848484848, + "grad_norm": 3.498199462890625, + "learning_rate": 3.42436974789916e-06, + "loss": 1.984, + "step": 177 + }, + { + "epoch": 0.22474747474747475, + "grad_norm": 3.576317548751831, + "learning_rate": 3.4453781512605043e-06, + "loss": 1.8897, + "step": 178 + }, + { + "epoch": 0.22601010101010102, + "grad_norm": 3.5337026119232178, + "learning_rate": 3.466386554621849e-06, + "loss": 1.9607, + "step": 179 + }, + { + "epoch": 0.22727272727272727, + "grad_norm": 3.342886209487915, + "learning_rate": 3.4873949579831936e-06, + "loss": 1.8797, + "step": 180 + }, + { + "epoch": 0.22727272727272727, + "eval_loss": 2.02158522605896, + "eval_runtime": 37.1382, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 180 + }, + { + "epoch": 0.22853535353535354, + "grad_norm": 2.941906213760376, + "learning_rate": 3.508403361344538e-06, + "loss": 1.9, + "step": 181 + }, + { + "epoch": 0.2297979797979798, + "grad_norm": 3.1913063526153564, + "learning_rate": 3.529411764705883e-06, + "loss": 1.8704, + "step": 182 + }, + { + "epoch": 0.23106060606060605, + "grad_norm": 3.0619900226593018, + "learning_rate": 3.550420168067227e-06, + "loss": 1.8851, + "step": 183 + }, + { + "epoch": 0.23232323232323232, + "grad_norm": 3.051288604736328, + "learning_rate": 3.5714285714285718e-06, + "loss": 1.9422, + "step": 184 + }, + { + "epoch": 0.2335858585858586, + "grad_norm": 2.999845266342163, + "learning_rate": 3.5924369747899164e-06, + "loss": 1.8757, + "step": 185 + }, + { + "epoch": 0.23484848484848486, + "grad_norm": 3.262355089187622, + "learning_rate": 3.6134453781512607e-06, + "loss": 1.9114, + "step": 186 + }, + { + "epoch": 0.2361111111111111, + "grad_norm": 3.0354340076446533, + "learning_rate": 3.6344537815126053e-06, + "loss": 1.8547, + "step": 187 + }, + { + "epoch": 0.23737373737373738, + "grad_norm": 2.9401369094848633, + "learning_rate": 3.6554621848739496e-06, + "loss": 1.8974, + "step": 188 + }, + { + "epoch": 0.23863636363636365, + "grad_norm": 3.0544443130493164, + "learning_rate": 3.6764705882352946e-06, + "loss": 1.8662, + "step": 189 + }, + { + "epoch": 0.2398989898989899, + "grad_norm": 2.7257750034332275, + "learning_rate": 3.6974789915966393e-06, + "loss": 1.8737, + "step": 190 + }, + { + "epoch": 0.2398989898989899, + "eval_loss": 2.0065128803253174, + "eval_runtime": 37.1296, + "eval_samples_per_second": 26.933, + "eval_steps_per_second": 1.131, + "step": 190 + }, + { + "epoch": 0.24116161616161616, + "grad_norm": 2.8592522144317627, + "learning_rate": 3.7184873949579835e-06, + "loss": 1.8984, + "step": 191 + }, + { + "epoch": 0.24242424242424243, + "grad_norm": 2.833958864212036, + "learning_rate": 3.739495798319328e-06, + "loss": 1.9084, + "step": 192 + }, + { + "epoch": 0.24368686868686867, + "grad_norm": 2.7842230796813965, + "learning_rate": 3.7605042016806724e-06, + "loss": 1.8843, + "step": 193 + }, + { + "epoch": 0.24494949494949494, + "grad_norm": 3.1309409141540527, + "learning_rate": 3.781512605042017e-06, + "loss": 1.8259, + "step": 194 + }, + { + "epoch": 0.24621212121212122, + "grad_norm": 3.120788335800171, + "learning_rate": 3.8025210084033613e-06, + "loss": 1.8951, + "step": 195 + }, + { + "epoch": 0.2474747474747475, + "grad_norm": 3.2228143215179443, + "learning_rate": 3.8235294117647055e-06, + "loss": 1.8608, + "step": 196 + }, + { + "epoch": 0.24873737373737373, + "grad_norm": 3.3362114429473877, + "learning_rate": 3.844537815126051e-06, + "loss": 1.9163, + "step": 197 + }, + { + "epoch": 0.25, + "grad_norm": 3.0973799228668213, + "learning_rate": 3.865546218487396e-06, + "loss": 1.911, + "step": 198 + }, + { + "epoch": 0.25126262626262624, + "grad_norm": 3.5567545890808105, + "learning_rate": 3.8865546218487395e-06, + "loss": 1.8795, + "step": 199 + }, + { + "epoch": 0.25252525252525254, + "grad_norm": 3.4758830070495605, + "learning_rate": 3.907563025210084e-06, + "loss": 1.8813, + "step": 200 + }, + { + "epoch": 0.25252525252525254, + "eval_loss": 1.9892102479934692, + "eval_runtime": 37.1311, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 200 + }, + { + "epoch": 0.2537878787878788, + "grad_norm": 3.5564475059509277, + "learning_rate": 3.928571428571429e-06, + "loss": 1.8469, + "step": 201 + }, + { + "epoch": 0.255050505050505, + "grad_norm": 3.7004899978637695, + "learning_rate": 3.9495798319327735e-06, + "loss": 1.873, + "step": 202 + }, + { + "epoch": 0.2563131313131313, + "grad_norm": 3.9285225868225098, + "learning_rate": 3.970588235294118e-06, + "loss": 1.8596, + "step": 203 + }, + { + "epoch": 0.25757575757575757, + "grad_norm": 3.1388907432556152, + "learning_rate": 3.991596638655463e-06, + "loss": 1.7991, + "step": 204 + }, + { + "epoch": 0.2588383838383838, + "grad_norm": 3.3797011375427246, + "learning_rate": 4.0126050420168075e-06, + "loss": 1.8682, + "step": 205 + }, + { + "epoch": 0.2601010101010101, + "grad_norm": 3.32183575630188, + "learning_rate": 4.033613445378151e-06, + "loss": 1.8396, + "step": 206 + }, + { + "epoch": 0.26136363636363635, + "grad_norm": 3.457291603088379, + "learning_rate": 4.054621848739496e-06, + "loss": 1.8314, + "step": 207 + }, + { + "epoch": 0.26262626262626265, + "grad_norm": 3.2906265258789062, + "learning_rate": 4.075630252100841e-06, + "loss": 1.844, + "step": 208 + }, + { + "epoch": 0.2638888888888889, + "grad_norm": 3.081913948059082, + "learning_rate": 4.096638655462185e-06, + "loss": 1.8198, + "step": 209 + }, + { + "epoch": 0.26515151515151514, + "grad_norm": 2.9697105884552, + "learning_rate": 4.11764705882353e-06, + "loss": 1.7885, + "step": 210 + }, + { + "epoch": 0.26515151515151514, + "eval_loss": 1.9749963283538818, + "eval_runtime": 36.962, + "eval_samples_per_second": 27.055, + "eval_steps_per_second": 1.136, + "step": 210 + }, + { + "epoch": 0.26641414141414144, + "grad_norm": 3.0732104778289795, + "learning_rate": 4.138655462184874e-06, + "loss": 1.8463, + "step": 211 + }, + { + "epoch": 0.2676767676767677, + "grad_norm": 3.4714958667755127, + "learning_rate": 4.159663865546219e-06, + "loss": 1.8537, + "step": 212 + }, + { + "epoch": 0.2689393939393939, + "grad_norm": 2.955864429473877, + "learning_rate": 4.180672268907564e-06, + "loss": 1.8566, + "step": 213 + }, + { + "epoch": 0.2702020202020202, + "grad_norm": 3.5314948558807373, + "learning_rate": 4.201680672268908e-06, + "loss": 1.7625, + "step": 214 + }, + { + "epoch": 0.27146464646464646, + "grad_norm": 2.8855810165405273, + "learning_rate": 4.222689075630252e-06, + "loss": 1.7968, + "step": 215 + }, + { + "epoch": 0.2727272727272727, + "grad_norm": 3.7866363525390625, + "learning_rate": 4.243697478991597e-06, + "loss": 1.8341, + "step": 216 + }, + { + "epoch": 0.273989898989899, + "grad_norm": 3.0352272987365723, + "learning_rate": 4.264705882352942e-06, + "loss": 1.8276, + "step": 217 + }, + { + "epoch": 0.27525252525252525, + "grad_norm": 3.4880247116088867, + "learning_rate": 4.2857142857142855e-06, + "loss": 1.8194, + "step": 218 + }, + { + "epoch": 0.2765151515151515, + "grad_norm": 3.6807212829589844, + "learning_rate": 4.306722689075631e-06, + "loss": 1.8495, + "step": 219 + }, + { + "epoch": 0.2777777777777778, + "grad_norm": 3.029017448425293, + "learning_rate": 4.327731092436976e-06, + "loss": 1.8125, + "step": 220 + }, + { + "epoch": 0.2777777777777778, + "eval_loss": 1.9633830785751343, + "eval_runtime": 37.1259, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 220 + }, + { + "epoch": 0.27904040404040403, + "grad_norm": 3.4820573329925537, + "learning_rate": 4.3487394957983194e-06, + "loss": 1.8159, + "step": 221 + }, + { + "epoch": 0.2803030303030303, + "grad_norm": 3.2057833671569824, + "learning_rate": 4.369747899159664e-06, + "loss": 1.8215, + "step": 222 + }, + { + "epoch": 0.2815656565656566, + "grad_norm": 3.477302312850952, + "learning_rate": 4.390756302521009e-06, + "loss": 1.7837, + "step": 223 + }, + { + "epoch": 0.2828282828282828, + "grad_norm": 2.8349595069885254, + "learning_rate": 4.411764705882353e-06, + "loss": 1.8555, + "step": 224 + }, + { + "epoch": 0.2840909090909091, + "grad_norm": 3.616478443145752, + "learning_rate": 4.432773109243698e-06, + "loss": 1.8067, + "step": 225 + }, + { + "epoch": 0.28535353535353536, + "grad_norm": 3.111969470977783, + "learning_rate": 4.453781512605043e-06, + "loss": 1.8322, + "step": 226 + }, + { + "epoch": 0.2866161616161616, + "grad_norm": 3.248695135116577, + "learning_rate": 4.474789915966387e-06, + "loss": 1.8004, + "step": 227 + }, + { + "epoch": 0.2878787878787879, + "grad_norm": 3.017082691192627, + "learning_rate": 4.495798319327731e-06, + "loss": 1.776, + "step": 228 + }, + { + "epoch": 0.28914141414141414, + "grad_norm": 3.6790099143981934, + "learning_rate": 4.516806722689076e-06, + "loss": 1.8176, + "step": 229 + }, + { + "epoch": 0.2904040404040404, + "grad_norm": 3.1513001918792725, + "learning_rate": 4.5378151260504205e-06, + "loss": 1.8016, + "step": 230 + }, + { + "epoch": 0.2904040404040404, + "eval_loss": 1.949769377708435, + "eval_runtime": 37.2573, + "eval_samples_per_second": 26.84, + "eval_steps_per_second": 1.127, + "step": 230 + }, + { + "epoch": 0.2916666666666667, + "grad_norm": 3.0559096336364746, + "learning_rate": 4.558823529411765e-06, + "loss": 1.7915, + "step": 231 + }, + { + "epoch": 0.29292929292929293, + "grad_norm": 3.0974984169006348, + "learning_rate": 4.57983193277311e-06, + "loss": 1.8277, + "step": 232 + }, + { + "epoch": 0.29419191919191917, + "grad_norm": 2.8514139652252197, + "learning_rate": 4.600840336134454e-06, + "loss": 1.7616, + "step": 233 + }, + { + "epoch": 0.29545454545454547, + "grad_norm": 3.355198383331299, + "learning_rate": 4.621848739495799e-06, + "loss": 1.841, + "step": 234 + }, + { + "epoch": 0.2967171717171717, + "grad_norm": 3.1467435359954834, + "learning_rate": 4.642857142857144e-06, + "loss": 1.7393, + "step": 235 + }, + { + "epoch": 0.29797979797979796, + "grad_norm": 4.065819263458252, + "learning_rate": 4.663865546218488e-06, + "loss": 1.7508, + "step": 236 + }, + { + "epoch": 0.29924242424242425, + "grad_norm": 3.2602460384368896, + "learning_rate": 4.684873949579832e-06, + "loss": 1.7415, + "step": 237 + }, + { + "epoch": 0.3005050505050505, + "grad_norm": 3.6884894371032715, + "learning_rate": 4.705882352941177e-06, + "loss": 1.7608, + "step": 238 + }, + { + "epoch": 0.30176767676767674, + "grad_norm": 3.9843380451202393, + "learning_rate": 4.7268907563025216e-06, + "loss": 1.8519, + "step": 239 + }, + { + "epoch": 0.30303030303030304, + "grad_norm": 3.2823405265808105, + "learning_rate": 4.747899159663865e-06, + "loss": 1.7689, + "step": 240 + }, + { + "epoch": 0.30303030303030304, + "eval_loss": 1.939724326133728, + "eval_runtime": 37.465, + "eval_samples_per_second": 26.692, + "eval_steps_per_second": 1.121, + "step": 240 + }, + { + "epoch": 0.3042929292929293, + "grad_norm": 3.4200189113616943, + "learning_rate": 4.768907563025211e-06, + "loss": 1.7944, + "step": 241 + }, + { + "epoch": 0.3055555555555556, + "grad_norm": 3.7541189193725586, + "learning_rate": 4.7899159663865555e-06, + "loss": 1.8151, + "step": 242 + }, + { + "epoch": 0.3068181818181818, + "grad_norm": 3.9103832244873047, + "learning_rate": 4.810924369747899e-06, + "loss": 1.7921, + "step": 243 + }, + { + "epoch": 0.30808080808080807, + "grad_norm": 3.2769975662231445, + "learning_rate": 4.831932773109244e-06, + "loss": 1.8408, + "step": 244 + }, + { + "epoch": 0.30934343434343436, + "grad_norm": 3.414168357849121, + "learning_rate": 4.852941176470589e-06, + "loss": 1.7937, + "step": 245 + }, + { + "epoch": 0.3106060606060606, + "grad_norm": 3.219625473022461, + "learning_rate": 4.873949579831933e-06, + "loss": 1.796, + "step": 246 + }, + { + "epoch": 0.31186868686868685, + "grad_norm": 3.2573912143707275, + "learning_rate": 4.894957983193277e-06, + "loss": 1.8052, + "step": 247 + }, + { + "epoch": 0.31313131313131315, + "grad_norm": 3.4201040267944336, + "learning_rate": 4.915966386554622e-06, + "loss": 1.8529, + "step": 248 + }, + { + "epoch": 0.3143939393939394, + "grad_norm": 3.7448058128356934, + "learning_rate": 4.936974789915967e-06, + "loss": 1.8183, + "step": 249 + }, + { + "epoch": 0.31565656565656564, + "grad_norm": 3.4108386039733887, + "learning_rate": 4.957983193277311e-06, + "loss": 1.7926, + "step": 250 + }, + { + "epoch": 0.31565656565656564, + "eval_loss": 1.9319047927856445, + "eval_runtime": 37.4745, + "eval_samples_per_second": 26.685, + "eval_steps_per_second": 1.121, + "step": 250 + }, + { + "epoch": 0.31691919191919193, + "grad_norm": 3.50185227394104, + "learning_rate": 4.978991596638656e-06, + "loss": 1.792, + "step": 251 + }, + { + "epoch": 0.3181818181818182, + "grad_norm": 3.225581407546997, + "learning_rate": 5e-06, + "loss": 1.7286, + "step": 252 + }, + { + "epoch": 0.3194444444444444, + "grad_norm": 3.3070669174194336, + "learning_rate": 4.997661365762395e-06, + "loss": 1.7957, + "step": 253 + }, + { + "epoch": 0.3207070707070707, + "grad_norm": 3.498079776763916, + "learning_rate": 4.99532273152479e-06, + "loss": 1.823, + "step": 254 + }, + { + "epoch": 0.32196969696969696, + "grad_norm": 3.316534996032715, + "learning_rate": 4.9929840972871845e-06, + "loss": 1.8249, + "step": 255 + }, + { + "epoch": 0.32323232323232326, + "grad_norm": 3.1641833782196045, + "learning_rate": 4.990645463049579e-06, + "loss": 1.7658, + "step": 256 + }, + { + "epoch": 0.3244949494949495, + "grad_norm": 3.0794270038604736, + "learning_rate": 4.988306828811974e-06, + "loss": 1.8146, + "step": 257 + }, + { + "epoch": 0.32575757575757575, + "grad_norm": 3.2410383224487305, + "learning_rate": 4.985968194574369e-06, + "loss": 1.7894, + "step": 258 + }, + { + "epoch": 0.32702020202020204, + "grad_norm": 3.0899274349212646, + "learning_rate": 4.983629560336764e-06, + "loss": 1.7381, + "step": 259 + }, + { + "epoch": 0.3282828282828283, + "grad_norm": 3.0132803916931152, + "learning_rate": 4.981290926099159e-06, + "loss": 1.7713, + "step": 260 + }, + { + "epoch": 0.3282828282828283, + "eval_loss": 1.9123589992523193, + "eval_runtime": 37.3201, + "eval_samples_per_second": 26.795, + "eval_steps_per_second": 1.125, + "step": 260 + }, + { + "epoch": 0.32954545454545453, + "grad_norm": 2.92564058303833, + "learning_rate": 4.9789522918615536e-06, + "loss": 1.7659, + "step": 261 + }, + { + "epoch": 0.33080808080808083, + "grad_norm": 3.1402878761291504, + "learning_rate": 4.976613657623948e-06, + "loss": 1.8439, + "step": 262 + }, + { + "epoch": 0.33207070707070707, + "grad_norm": 3.170881509780884, + "learning_rate": 4.974275023386343e-06, + "loss": 1.7605, + "step": 263 + }, + { + "epoch": 0.3333333333333333, + "grad_norm": 3.014850378036499, + "learning_rate": 4.971936389148738e-06, + "loss": 1.7899, + "step": 264 + }, + { + "epoch": 0.3345959595959596, + "grad_norm": 3.169818878173828, + "learning_rate": 4.969597754911132e-06, + "loss": 1.8263, + "step": 265 + }, + { + "epoch": 0.33585858585858586, + "grad_norm": 2.8709678649902344, + "learning_rate": 4.967259120673527e-06, + "loss": 1.7834, + "step": 266 + }, + { + "epoch": 0.3371212121212121, + "grad_norm": 2.7527096271514893, + "learning_rate": 4.964920486435922e-06, + "loss": 1.8106, + "step": 267 + }, + { + "epoch": 0.3383838383838384, + "grad_norm": 2.9960227012634277, + "learning_rate": 4.962581852198316e-06, + "loss": 1.7452, + "step": 268 + }, + { + "epoch": 0.33964646464646464, + "grad_norm": 2.979163885116577, + "learning_rate": 4.9602432179607115e-06, + "loss": 1.6943, + "step": 269 + }, + { + "epoch": 0.3409090909090909, + "grad_norm": 3.2396063804626465, + "learning_rate": 4.957904583723106e-06, + "loss": 1.7798, + "step": 270 + }, + { + "epoch": 0.3409090909090909, + "eval_loss": 1.9126806259155273, + "eval_runtime": 37.123, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 270 + }, + { + "epoch": 0.3421717171717172, + "grad_norm": 2.7868010997772217, + "learning_rate": 4.955565949485501e-06, + "loss": 1.8053, + "step": 271 + }, + { + "epoch": 0.3434343434343434, + "grad_norm": 2.8895349502563477, + "learning_rate": 4.9532273152478956e-06, + "loss": 1.7672, + "step": 272 + }, + { + "epoch": 0.3446969696969697, + "grad_norm": 3.0839710235595703, + "learning_rate": 4.95088868101029e-06, + "loss": 1.7414, + "step": 273 + }, + { + "epoch": 0.34595959595959597, + "grad_norm": 2.81974458694458, + "learning_rate": 4.948550046772685e-06, + "loss": 1.7527, + "step": 274 + }, + { + "epoch": 0.3472222222222222, + "grad_norm": 3.7469775676727295, + "learning_rate": 4.9462114125350805e-06, + "loss": 1.7689, + "step": 275 + }, + { + "epoch": 0.3484848484848485, + "grad_norm": 3.434573173522949, + "learning_rate": 4.943872778297475e-06, + "loss": 1.739, + "step": 276 + }, + { + "epoch": 0.34974747474747475, + "grad_norm": 2.985041379928589, + "learning_rate": 4.941534144059869e-06, + "loss": 1.7894, + "step": 277 + }, + { + "epoch": 0.351010101010101, + "grad_norm": 3.441204309463501, + "learning_rate": 4.939195509822265e-06, + "loss": 1.7804, + "step": 278 + }, + { + "epoch": 0.3522727272727273, + "grad_norm": 2.8371102809906006, + "learning_rate": 4.936856875584659e-06, + "loss": 1.7415, + "step": 279 + }, + { + "epoch": 0.35353535353535354, + "grad_norm": 3.662773847579956, + "learning_rate": 4.9345182413470535e-06, + "loss": 1.7894, + "step": 280 + }, + { + "epoch": 0.35353535353535354, + "eval_loss": 1.915844440460205, + "eval_runtime": 37.0941, + "eval_samples_per_second": 26.958, + "eval_steps_per_second": 1.132, + "step": 280 + }, + { + "epoch": 0.3547979797979798, + "grad_norm": 3.3254036903381348, + "learning_rate": 4.932179607109449e-06, + "loss": 1.7238, + "step": 281 + }, + { + "epoch": 0.3560606060606061, + "grad_norm": 2.9872665405273438, + "learning_rate": 4.929840972871843e-06, + "loss": 1.769, + "step": 282 + }, + { + "epoch": 0.3573232323232323, + "grad_norm": 3.0163521766662598, + "learning_rate": 4.9275023386342376e-06, + "loss": 1.7702, + "step": 283 + }, + { + "epoch": 0.35858585858585856, + "grad_norm": 2.9687387943267822, + "learning_rate": 4.925163704396633e-06, + "loss": 1.7927, + "step": 284 + }, + { + "epoch": 0.35984848484848486, + "grad_norm": 2.7064011096954346, + "learning_rate": 4.922825070159027e-06, + "loss": 1.7638, + "step": 285 + }, + { + "epoch": 0.3611111111111111, + "grad_norm": 3.0043134689331055, + "learning_rate": 4.9204864359214225e-06, + "loss": 1.8041, + "step": 286 + }, + { + "epoch": 0.36237373737373735, + "grad_norm": 3.037100315093994, + "learning_rate": 4.918147801683817e-06, + "loss": 1.7579, + "step": 287 + }, + { + "epoch": 0.36363636363636365, + "grad_norm": 2.7528676986694336, + "learning_rate": 4.915809167446211e-06, + "loss": 1.7629, + "step": 288 + }, + { + "epoch": 0.3648989898989899, + "grad_norm": 3.075815439224243, + "learning_rate": 4.913470533208607e-06, + "loss": 1.7966, + "step": 289 + }, + { + "epoch": 0.3661616161616162, + "grad_norm": 2.964785575866699, + "learning_rate": 4.911131898971001e-06, + "loss": 1.801, + "step": 290 + }, + { + "epoch": 0.3661616161616162, + "eval_loss": 1.8839410543441772, + "eval_runtime": 37.0844, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 290 + }, + { + "epoch": 0.36742424242424243, + "grad_norm": 2.668747663497925, + "learning_rate": 4.908793264733396e-06, + "loss": 1.7703, + "step": 291 + }, + { + "epoch": 0.3686868686868687, + "grad_norm": 2.814751148223877, + "learning_rate": 4.906454630495791e-06, + "loss": 1.7075, + "step": 292 + }, + { + "epoch": 0.369949494949495, + "grad_norm": 3.005922794342041, + "learning_rate": 4.904115996258186e-06, + "loss": 1.7418, + "step": 293 + }, + { + "epoch": 0.3712121212121212, + "grad_norm": 3.050936460494995, + "learning_rate": 4.90177736202058e-06, + "loss": 1.793, + "step": 294 + }, + { + "epoch": 0.37247474747474746, + "grad_norm": 2.822004556655884, + "learning_rate": 4.899438727782975e-06, + "loss": 1.8303, + "step": 295 + }, + { + "epoch": 0.37373737373737376, + "grad_norm": 2.950254201889038, + "learning_rate": 4.89710009354537e-06, + "loss": 1.808, + "step": 296 + }, + { + "epoch": 0.375, + "grad_norm": 2.8140666484832764, + "learning_rate": 4.8947614593077645e-06, + "loss": 1.7262, + "step": 297 + }, + { + "epoch": 0.37626262626262624, + "grad_norm": 2.899620532989502, + "learning_rate": 4.89242282507016e-06, + "loss": 1.7157, + "step": 298 + }, + { + "epoch": 0.37752525252525254, + "grad_norm": 2.805753469467163, + "learning_rate": 4.890084190832554e-06, + "loss": 1.7294, + "step": 299 + }, + { + "epoch": 0.3787878787878788, + "grad_norm": 3.110246181488037, + "learning_rate": 4.887745556594949e-06, + "loss": 1.7415, + "step": 300 + }, + { + "epoch": 0.3787878787878788, + "eval_loss": 1.892285943031311, + "eval_runtime": 37.0732, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 300 + }, + { + "epoch": 0.380050505050505, + "grad_norm": 2.9197490215301514, + "learning_rate": 4.885406922357344e-06, + "loss": 1.7132, + "step": 301 + }, + { + "epoch": 0.3813131313131313, + "grad_norm": 3.3165123462677, + "learning_rate": 4.883068288119738e-06, + "loss": 1.777, + "step": 302 + }, + { + "epoch": 0.38257575757575757, + "grad_norm": 2.8409054279327393, + "learning_rate": 4.880729653882133e-06, + "loss": 1.7092, + "step": 303 + }, + { + "epoch": 0.3838383838383838, + "grad_norm": 2.6800901889801025, + "learning_rate": 4.878391019644528e-06, + "loss": 1.7622, + "step": 304 + }, + { + "epoch": 0.3851010101010101, + "grad_norm": 3.2847490310668945, + "learning_rate": 4.876052385406922e-06, + "loss": 1.729, + "step": 305 + }, + { + "epoch": 0.38636363636363635, + "grad_norm": 2.786057949066162, + "learning_rate": 4.873713751169318e-06, + "loss": 1.7168, + "step": 306 + }, + { + "epoch": 0.38762626262626265, + "grad_norm": 3.236050605773926, + "learning_rate": 4.871375116931712e-06, + "loss": 1.6809, + "step": 307 + }, + { + "epoch": 0.3888888888888889, + "grad_norm": 2.851910352706909, + "learning_rate": 4.869036482694107e-06, + "loss": 1.7336, + "step": 308 + }, + { + "epoch": 0.39015151515151514, + "grad_norm": 2.892421007156372, + "learning_rate": 4.866697848456502e-06, + "loss": 1.7451, + "step": 309 + }, + { + "epoch": 0.39141414141414144, + "grad_norm": 2.8175048828125, + "learning_rate": 4.864359214218897e-06, + "loss": 1.7734, + "step": 310 + }, + { + "epoch": 0.39141414141414144, + "eval_loss": 1.8777843713760376, + "eval_runtime": 36.9321, + "eval_samples_per_second": 27.077, + "eval_steps_per_second": 1.137, + "step": 310 + }, + { + "epoch": 0.3926767676767677, + "grad_norm": 3.005329132080078, + "learning_rate": 4.8620205799812915e-06, + "loss": 1.7476, + "step": 311 + }, + { + "epoch": 0.3939393939393939, + "grad_norm": 2.849017858505249, + "learning_rate": 4.859681945743686e-06, + "loss": 1.7854, + "step": 312 + }, + { + "epoch": 0.3952020202020202, + "grad_norm": 2.801048517227173, + "learning_rate": 4.857343311506081e-06, + "loss": 1.7186, + "step": 313 + }, + { + "epoch": 0.39646464646464646, + "grad_norm": 2.5772831439971924, + "learning_rate": 4.855004677268476e-06, + "loss": 1.7279, + "step": 314 + }, + { + "epoch": 0.3977272727272727, + "grad_norm": 2.7969934940338135, + "learning_rate": 4.85266604303087e-06, + "loss": 1.7475, + "step": 315 + }, + { + "epoch": 0.398989898989899, + "grad_norm": 2.7067770957946777, + "learning_rate": 4.850327408793265e-06, + "loss": 1.7055, + "step": 316 + }, + { + "epoch": 0.40025252525252525, + "grad_norm": 2.70928955078125, + "learning_rate": 4.84798877455566e-06, + "loss": 1.7203, + "step": 317 + }, + { + "epoch": 0.4015151515151515, + "grad_norm": 2.7704474925994873, + "learning_rate": 4.845650140318054e-06, + "loss": 1.8058, + "step": 318 + }, + { + "epoch": 0.4027777777777778, + "grad_norm": 2.7680182456970215, + "learning_rate": 4.843311506080449e-06, + "loss": 1.7162, + "step": 319 + }, + { + "epoch": 0.40404040404040403, + "grad_norm": 3.0953330993652344, + "learning_rate": 4.840972871842844e-06, + "loss": 1.7617, + "step": 320 + }, + { + "epoch": 0.40404040404040403, + "eval_loss": 1.8680965900421143, + "eval_runtime": 37.0382, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 320 + }, + { + "epoch": 0.4053030303030303, + "grad_norm": 2.8340392112731934, + "learning_rate": 4.838634237605239e-06, + "loss": 1.7767, + "step": 321 + }, + { + "epoch": 0.4065656565656566, + "grad_norm": 3.057861804962158, + "learning_rate": 4.836295603367634e-06, + "loss": 1.8043, + "step": 322 + }, + { + "epoch": 0.4078282828282828, + "grad_norm": 3.0123419761657715, + "learning_rate": 4.833956969130029e-06, + "loss": 1.7561, + "step": 323 + }, + { + "epoch": 0.4090909090909091, + "grad_norm": 3.0873124599456787, + "learning_rate": 4.831618334892423e-06, + "loss": 1.7581, + "step": 324 + }, + { + "epoch": 0.41035353535353536, + "grad_norm": 2.6600539684295654, + "learning_rate": 4.8292797006548184e-06, + "loss": 1.7005, + "step": 325 + }, + { + "epoch": 0.4116161616161616, + "grad_norm": 2.905280590057373, + "learning_rate": 4.826941066417213e-06, + "loss": 1.7592, + "step": 326 + }, + { + "epoch": 0.4128787878787879, + "grad_norm": 3.0642189979553223, + "learning_rate": 4.824602432179607e-06, + "loss": 1.6456, + "step": 327 + }, + { + "epoch": 0.41414141414141414, + "grad_norm": 2.606531858444214, + "learning_rate": 4.8222637979420025e-06, + "loss": 1.6731, + "step": 328 + }, + { + "epoch": 0.4154040404040404, + "grad_norm": 2.9918158054351807, + "learning_rate": 4.819925163704397e-06, + "loss": 1.6854, + "step": 329 + }, + { + "epoch": 0.4166666666666667, + "grad_norm": 2.721940040588379, + "learning_rate": 4.817586529466791e-06, + "loss": 1.7346, + "step": 330 + }, + { + "epoch": 0.4166666666666667, + "eval_loss": 1.863736629486084, + "eval_runtime": 37.0887, + "eval_samples_per_second": 26.962, + "eval_steps_per_second": 1.132, + "step": 330 + }, + { + "epoch": 0.41792929292929293, + "grad_norm": 3.1076390743255615, + "learning_rate": 4.815247895229187e-06, + "loss": 1.6286, + "step": 331 + }, + { + "epoch": 0.41919191919191917, + "grad_norm": 3.1235783100128174, + "learning_rate": 4.812909260991581e-06, + "loss": 1.7404, + "step": 332 + }, + { + "epoch": 0.42045454545454547, + "grad_norm": 3.1211330890655518, + "learning_rate": 4.8105706267539755e-06, + "loss": 1.7143, + "step": 333 + }, + { + "epoch": 0.4217171717171717, + "grad_norm": 3.072317600250244, + "learning_rate": 4.808231992516371e-06, + "loss": 1.7514, + "step": 334 + }, + { + "epoch": 0.42297979797979796, + "grad_norm": 2.880671739578247, + "learning_rate": 4.805893358278765e-06, + "loss": 1.7018, + "step": 335 + }, + { + "epoch": 0.42424242424242425, + "grad_norm": 2.700063943862915, + "learning_rate": 4.8035547240411604e-06, + "loss": 1.7089, + "step": 336 + }, + { + "epoch": 0.4255050505050505, + "grad_norm": 2.9038445949554443, + "learning_rate": 4.801216089803556e-06, + "loss": 1.7483, + "step": 337 + }, + { + "epoch": 0.42676767676767674, + "grad_norm": 2.7822766304016113, + "learning_rate": 4.79887745556595e-06, + "loss": 1.6927, + "step": 338 + }, + { + "epoch": 0.42803030303030304, + "grad_norm": 2.957554340362549, + "learning_rate": 4.7965388213283445e-06, + "loss": 1.7162, + "step": 339 + }, + { + "epoch": 0.4292929292929293, + "grad_norm": 3.1333892345428467, + "learning_rate": 4.79420018709074e-06, + "loss": 1.731, + "step": 340 + }, + { + "epoch": 0.4292929292929293, + "eval_loss": 1.852413535118103, + "eval_runtime": 37.1809, + "eval_samples_per_second": 26.896, + "eval_steps_per_second": 1.13, + "step": 340 + }, + { + "epoch": 0.4305555555555556, + "grad_norm": 2.9801084995269775, + "learning_rate": 4.791861552853134e-06, + "loss": 1.6892, + "step": 341 + }, + { + "epoch": 0.4318181818181818, + "grad_norm": 2.5077271461486816, + "learning_rate": 4.789522918615529e-06, + "loss": 1.637, + "step": 342 + }, + { + "epoch": 0.43308080808080807, + "grad_norm": 2.8654391765594482, + "learning_rate": 4.787184284377924e-06, + "loss": 1.6915, + "step": 343 + }, + { + "epoch": 0.43434343434343436, + "grad_norm": 2.6204264163970947, + "learning_rate": 4.784845650140318e-06, + "loss": 1.6722, + "step": 344 + }, + { + "epoch": 0.4356060606060606, + "grad_norm": 2.633991003036499, + "learning_rate": 4.782507015902713e-06, + "loss": 1.6797, + "step": 345 + }, + { + "epoch": 0.43686868686868685, + "grad_norm": 2.7203924655914307, + "learning_rate": 4.780168381665108e-06, + "loss": 1.7018, + "step": 346 + }, + { + "epoch": 0.43813131313131315, + "grad_norm": 2.6789209842681885, + "learning_rate": 4.7778297474275024e-06, + "loss": 1.7153, + "step": 347 + }, + { + "epoch": 0.4393939393939394, + "grad_norm": 2.6285433769226074, + "learning_rate": 4.775491113189898e-06, + "loss": 1.7082, + "step": 348 + }, + { + "epoch": 0.44065656565656564, + "grad_norm": 2.623793840408325, + "learning_rate": 4.773152478952292e-06, + "loss": 1.7094, + "step": 349 + }, + { + "epoch": 0.44191919191919193, + "grad_norm": 2.647160291671753, + "learning_rate": 4.7708138447146865e-06, + "loss": 1.6528, + "step": 350 + }, + { + "epoch": 0.44191919191919193, + "eval_loss": 1.8506718873977661, + "eval_runtime": 37.4232, + "eval_samples_per_second": 26.721, + "eval_steps_per_second": 1.122, + "step": 350 + }, + { + "epoch": 0.4431818181818182, + "grad_norm": 2.664553642272949, + "learning_rate": 4.768475210477082e-06, + "loss": 1.7008, + "step": 351 + }, + { + "epoch": 0.4444444444444444, + "grad_norm": 2.8486251831054688, + "learning_rate": 4.766136576239477e-06, + "loss": 1.6811, + "step": 352 + }, + { + "epoch": 0.4457070707070707, + "grad_norm": 2.7514266967773438, + "learning_rate": 4.7637979420018715e-06, + "loss": 1.6969, + "step": 353 + }, + { + "epoch": 0.44696969696969696, + "grad_norm": 2.848083257675171, + "learning_rate": 4.761459307764266e-06, + "loss": 1.7058, + "step": 354 + }, + { + "epoch": 0.44823232323232326, + "grad_norm": 2.759218454360962, + "learning_rate": 4.759120673526661e-06, + "loss": 1.6109, + "step": 355 + }, + { + "epoch": 0.4494949494949495, + "grad_norm": 2.905318021774292, + "learning_rate": 4.756782039289056e-06, + "loss": 1.7689, + "step": 356 + }, + { + "epoch": 0.45075757575757575, + "grad_norm": 2.845646381378174, + "learning_rate": 4.75444340505145e-06, + "loss": 1.6687, + "step": 357 + }, + { + "epoch": 0.45202020202020204, + "grad_norm": 2.8609306812286377, + "learning_rate": 4.752104770813845e-06, + "loss": 1.6332, + "step": 358 + }, + { + "epoch": 0.4532828282828283, + "grad_norm": 3.016479969024658, + "learning_rate": 4.74976613657624e-06, + "loss": 1.6692, + "step": 359 + }, + { + "epoch": 0.45454545454545453, + "grad_norm": 3.2790887355804443, + "learning_rate": 4.747427502338635e-06, + "loss": 1.704, + "step": 360 + }, + { + "epoch": 0.45454545454545453, + "eval_loss": 1.848663330078125, + "eval_runtime": 37.4817, + "eval_samples_per_second": 26.68, + "eval_steps_per_second": 1.121, + "step": 360 + }, + { + "epoch": 0.45580808080808083, + "grad_norm": 3.1137545108795166, + "learning_rate": 4.745088868101029e-06, + "loss": 1.6797, + "step": 361 + }, + { + "epoch": 0.45707070707070707, + "grad_norm": 2.7598798274993896, + "learning_rate": 4.742750233863424e-06, + "loss": 1.6477, + "step": 362 + }, + { + "epoch": 0.4583333333333333, + "grad_norm": 3.4851579666137695, + "learning_rate": 4.740411599625819e-06, + "loss": 1.6481, + "step": 363 + }, + { + "epoch": 0.4595959595959596, + "grad_norm": 2.9285919666290283, + "learning_rate": 4.7380729653882135e-06, + "loss": 1.7094, + "step": 364 + }, + { + "epoch": 0.46085858585858586, + "grad_norm": 2.901458501815796, + "learning_rate": 4.735734331150608e-06, + "loss": 1.6906, + "step": 365 + }, + { + "epoch": 0.4621212121212121, + "grad_norm": 2.8737902641296387, + "learning_rate": 4.733395696913003e-06, + "loss": 1.6819, + "step": 366 + }, + { + "epoch": 0.4633838383838384, + "grad_norm": 3.2480952739715576, + "learning_rate": 4.7310570626753984e-06, + "loss": 1.6827, + "step": 367 + }, + { + "epoch": 0.46464646464646464, + "grad_norm": 2.8714661598205566, + "learning_rate": 4.728718428437793e-06, + "loss": 1.7349, + "step": 368 + }, + { + "epoch": 0.4659090909090909, + "grad_norm": 2.8410098552703857, + "learning_rate": 4.726379794200187e-06, + "loss": 1.7214, + "step": 369 + }, + { + "epoch": 0.4671717171717172, + "grad_norm": 2.744807720184326, + "learning_rate": 4.7240411599625826e-06, + "loss": 1.6742, + "step": 370 + }, + { + "epoch": 0.4671717171717172, + "eval_loss": 1.842191457748413, + "eval_runtime": 37.454, + "eval_samples_per_second": 26.699, + "eval_steps_per_second": 1.121, + "step": 370 + }, + { + "epoch": 0.4684343434343434, + "grad_norm": 2.8135552406311035, + "learning_rate": 4.721702525724977e-06, + "loss": 1.6957, + "step": 371 + }, + { + "epoch": 0.4696969696969697, + "grad_norm": 2.856764078140259, + "learning_rate": 4.719363891487372e-06, + "loss": 1.6802, + "step": 372 + }, + { + "epoch": 0.47095959595959597, + "grad_norm": 3.0707688331604004, + "learning_rate": 4.717025257249767e-06, + "loss": 1.7383, + "step": 373 + }, + { + "epoch": 0.4722222222222222, + "grad_norm": 2.8832876682281494, + "learning_rate": 4.714686623012161e-06, + "loss": 1.6309, + "step": 374 + }, + { + "epoch": 0.4734848484848485, + "grad_norm": 2.6840622425079346, + "learning_rate": 4.712347988774556e-06, + "loss": 1.6624, + "step": 375 + }, + { + "epoch": 0.47474747474747475, + "grad_norm": 3.2920548915863037, + "learning_rate": 4.710009354536951e-06, + "loss": 1.6887, + "step": 376 + }, + { + "epoch": 0.476010101010101, + "grad_norm": 3.0317189693450928, + "learning_rate": 4.707670720299345e-06, + "loss": 1.6514, + "step": 377 + }, + { + "epoch": 0.4772727272727273, + "grad_norm": 2.7284793853759766, + "learning_rate": 4.7053320860617404e-06, + "loss": 1.6595, + "step": 378 + }, + { + "epoch": 0.47853535353535354, + "grad_norm": 2.86665415763855, + "learning_rate": 4.702993451824135e-06, + "loss": 1.7219, + "step": 379 + }, + { + "epoch": 0.4797979797979798, + "grad_norm": 3.041191577911377, + "learning_rate": 4.700654817586529e-06, + "loss": 1.6959, + "step": 380 + }, + { + "epoch": 0.4797979797979798, + "eval_loss": 1.8414148092269897, + "eval_runtime": 37.2537, + "eval_samples_per_second": 26.843, + "eval_steps_per_second": 1.127, + "step": 380 + }, + { + "epoch": 0.4810606060606061, + "grad_norm": 3.1729421615600586, + "learning_rate": 4.6983161833489246e-06, + "loss": 1.6538, + "step": 381 + }, + { + "epoch": 0.4823232323232323, + "grad_norm": 2.8562991619110107, + "learning_rate": 4.69597754911132e-06, + "loss": 1.6667, + "step": 382 + }, + { + "epoch": 0.48358585858585856, + "grad_norm": 3.0348711013793945, + "learning_rate": 4.693638914873714e-06, + "loss": 1.6888, + "step": 383 + }, + { + "epoch": 0.48484848484848486, + "grad_norm": 2.9156477451324463, + "learning_rate": 4.6913002806361095e-06, + "loss": 1.628, + "step": 384 + }, + { + "epoch": 0.4861111111111111, + "grad_norm": 2.832689046859741, + "learning_rate": 4.688961646398504e-06, + "loss": 1.664, + "step": 385 + }, + { + "epoch": 0.48737373737373735, + "grad_norm": 2.9858946800231934, + "learning_rate": 4.686623012160898e-06, + "loss": 1.6696, + "step": 386 + }, + { + "epoch": 0.48863636363636365, + "grad_norm": 2.732468843460083, + "learning_rate": 4.684284377923294e-06, + "loss": 1.6444, + "step": 387 + }, + { + "epoch": 0.4898989898989899, + "grad_norm": 2.7801132202148438, + "learning_rate": 4.681945743685688e-06, + "loss": 1.7084, + "step": 388 + }, + { + "epoch": 0.4911616161616162, + "grad_norm": 2.646303415298462, + "learning_rate": 4.6796071094480824e-06, + "loss": 1.6963, + "step": 389 + }, + { + "epoch": 0.49242424242424243, + "grad_norm": 2.8750250339508057, + "learning_rate": 4.677268475210478e-06, + "loss": 1.6814, + "step": 390 + }, + { + "epoch": 0.49242424242424243, + "eval_loss": 1.8283894062042236, + "eval_runtime": 37.2058, + "eval_samples_per_second": 26.877, + "eval_steps_per_second": 1.129, + "step": 390 + }, + { + "epoch": 0.4936868686868687, + "grad_norm": 2.5373270511627197, + "learning_rate": 4.674929840972872e-06, + "loss": 1.5956, + "step": 391 + }, + { + "epoch": 0.494949494949495, + "grad_norm": 2.837170124053955, + "learning_rate": 4.6725912067352666e-06, + "loss": 1.6237, + "step": 392 + }, + { + "epoch": 0.4962121212121212, + "grad_norm": 3.199463367462158, + "learning_rate": 4.670252572497662e-06, + "loss": 1.7087, + "step": 393 + }, + { + "epoch": 0.49747474747474746, + "grad_norm": 2.9850947856903076, + "learning_rate": 4.667913938260056e-06, + "loss": 1.5913, + "step": 394 + }, + { + "epoch": 0.49873737373737376, + "grad_norm": 2.954148292541504, + "learning_rate": 4.665575304022451e-06, + "loss": 1.6272, + "step": 395 + }, + { + "epoch": 0.5, + "grad_norm": 2.8625757694244385, + "learning_rate": 4.663236669784846e-06, + "loss": 1.6623, + "step": 396 + }, + { + "epoch": 0.5012626262626263, + "grad_norm": 3.196354627609253, + "learning_rate": 4.66089803554724e-06, + "loss": 1.7249, + "step": 397 + }, + { + "epoch": 0.5025252525252525, + "grad_norm": 2.7257323265075684, + "learning_rate": 4.658559401309636e-06, + "loss": 1.665, + "step": 398 + }, + { + "epoch": 0.5037878787878788, + "grad_norm": 3.0976016521453857, + "learning_rate": 4.656220767072031e-06, + "loss": 1.6763, + "step": 399 + }, + { + "epoch": 0.5050505050505051, + "grad_norm": 3.0398128032684326, + "learning_rate": 4.653882132834425e-06, + "loss": 1.6375, + "step": 400 + }, + { + "epoch": 0.5050505050505051, + "eval_loss": 1.8324880599975586, + "eval_runtime": 37.1754, + "eval_samples_per_second": 26.9, + "eval_steps_per_second": 1.13, + "step": 400 + }, + { + "epoch": 0.5063131313131313, + "grad_norm": 3.314958095550537, + "learning_rate": 4.65154349859682e-06, + "loss": 1.6699, + "step": 401 + }, + { + "epoch": 0.5075757575757576, + "grad_norm": 2.8478870391845703, + "learning_rate": 4.649204864359215e-06, + "loss": 1.6495, + "step": 402 + }, + { + "epoch": 0.5088383838383839, + "grad_norm": 2.8476510047912598, + "learning_rate": 4.646866230121609e-06, + "loss": 1.633, + "step": 403 + }, + { + "epoch": 0.51010101010101, + "grad_norm": 2.916667938232422, + "learning_rate": 4.644527595884004e-06, + "loss": 1.6092, + "step": 404 + }, + { + "epoch": 0.5113636363636364, + "grad_norm": 2.977226495742798, + "learning_rate": 4.642188961646399e-06, + "loss": 1.6588, + "step": 405 + }, + { + "epoch": 0.5126262626262627, + "grad_norm": 3.0540225505828857, + "learning_rate": 4.6398503274087935e-06, + "loss": 1.6529, + "step": 406 + }, + { + "epoch": 0.5138888888888888, + "grad_norm": 2.902035713195801, + "learning_rate": 4.637511693171188e-06, + "loss": 1.5983, + "step": 407 + }, + { + "epoch": 0.5151515151515151, + "grad_norm": 2.9009487628936768, + "learning_rate": 4.635173058933583e-06, + "loss": 1.6389, + "step": 408 + }, + { + "epoch": 0.5164141414141414, + "grad_norm": 2.904982805252075, + "learning_rate": 4.632834424695978e-06, + "loss": 1.6971, + "step": 409 + }, + { + "epoch": 0.5176767676767676, + "grad_norm": 2.8340721130371094, + "learning_rate": 4.630495790458373e-06, + "loss": 1.6336, + "step": 410 + }, + { + "epoch": 0.5176767676767676, + "eval_loss": 1.8218064308166504, + "eval_runtime": 36.9963, + "eval_samples_per_second": 27.03, + "eval_steps_per_second": 1.135, + "step": 410 + }, + { + "epoch": 0.5189393939393939, + "grad_norm": 2.9507815837860107, + "learning_rate": 4.628157156220767e-06, + "loss": 1.6795, + "step": 411 + }, + { + "epoch": 0.5202020202020202, + "grad_norm": 2.868793249130249, + "learning_rate": 4.625818521983162e-06, + "loss": 1.6494, + "step": 412 + }, + { + "epoch": 0.5214646464646465, + "grad_norm": 2.786395788192749, + "learning_rate": 4.623479887745557e-06, + "loss": 1.7156, + "step": 413 + }, + { + "epoch": 0.5227272727272727, + "grad_norm": 2.7456934452056885, + "learning_rate": 4.621141253507952e-06, + "loss": 1.6528, + "step": 414 + }, + { + "epoch": 0.523989898989899, + "grad_norm": 3.1211695671081543, + "learning_rate": 4.618802619270347e-06, + "loss": 1.6891, + "step": 415 + }, + { + "epoch": 0.5252525252525253, + "grad_norm": 3.2319254875183105, + "learning_rate": 4.616463985032741e-06, + "loss": 1.6452, + "step": 416 + }, + { + "epoch": 0.5265151515151515, + "grad_norm": 2.6718053817749023, + "learning_rate": 4.614125350795136e-06, + "loss": 1.6442, + "step": 417 + }, + { + "epoch": 0.5277777777777778, + "grad_norm": 2.985020637512207, + "learning_rate": 4.611786716557531e-06, + "loss": 1.6935, + "step": 418 + }, + { + "epoch": 0.5290404040404041, + "grad_norm": 2.798720598220825, + "learning_rate": 4.609448082319925e-06, + "loss": 1.6303, + "step": 419 + }, + { + "epoch": 0.5303030303030303, + "grad_norm": 2.860403299331665, + "learning_rate": 4.6071094480823205e-06, + "loss": 1.6445, + "step": 420 + }, + { + "epoch": 0.5303030303030303, + "eval_loss": 1.8173401355743408, + "eval_runtime": 37.1334, + "eval_samples_per_second": 26.93, + "eval_steps_per_second": 1.131, + "step": 420 + }, + { + "epoch": 0.5315656565656566, + "grad_norm": 2.883209705352783, + "learning_rate": 4.604770813844715e-06, + "loss": 1.6182, + "step": 421 + }, + { + "epoch": 0.5328282828282829, + "grad_norm": 3.0330183506011963, + "learning_rate": 4.60243217960711e-06, + "loss": 1.6208, + "step": 422 + }, + { + "epoch": 0.5340909090909091, + "grad_norm": 3.224184274673462, + "learning_rate": 4.6000935453695046e-06, + "loss": 1.6687, + "step": 423 + }, + { + "epoch": 0.5353535353535354, + "grad_norm": 3.0345349311828613, + "learning_rate": 4.597754911131899e-06, + "loss": 1.6654, + "step": 424 + }, + { + "epoch": 0.5366161616161617, + "grad_norm": 3.1425750255584717, + "learning_rate": 4.595416276894294e-06, + "loss": 1.6569, + "step": 425 + }, + { + "epoch": 0.5378787878787878, + "grad_norm": 2.6387979984283447, + "learning_rate": 4.593077642656689e-06, + "loss": 1.6868, + "step": 426 + }, + { + "epoch": 0.5391414141414141, + "grad_norm": 2.959446430206299, + "learning_rate": 4.590739008419083e-06, + "loss": 1.621, + "step": 427 + }, + { + "epoch": 0.5404040404040404, + "grad_norm": 3.009673833847046, + "learning_rate": 4.588400374181478e-06, + "loss": 1.6414, + "step": 428 + }, + { + "epoch": 0.5416666666666666, + "grad_norm": 2.827531576156616, + "learning_rate": 4.586061739943874e-06, + "loss": 1.6561, + "step": 429 + }, + { + "epoch": 0.5429292929292929, + "grad_norm": 2.8770735263824463, + "learning_rate": 4.583723105706268e-06, + "loss": 1.6685, + "step": 430 + }, + { + "epoch": 0.5429292929292929, + "eval_loss": 1.8234716653823853, + "eval_runtime": 37.1647, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 430 + }, + { + "epoch": 0.5441919191919192, + "grad_norm": 3.1215782165527344, + "learning_rate": 4.5813844714686625e-06, + "loss": 1.6548, + "step": 431 + }, + { + "epoch": 0.5454545454545454, + "grad_norm": 2.6587517261505127, + "learning_rate": 4.579045837231058e-06, + "loss": 1.649, + "step": 432 + }, + { + "epoch": 0.5467171717171717, + "grad_norm": 2.7059030532836914, + "learning_rate": 4.576707202993452e-06, + "loss": 1.6293, + "step": 433 + }, + { + "epoch": 0.547979797979798, + "grad_norm": 2.7390925884246826, + "learning_rate": 4.574368568755847e-06, + "loss": 1.6957, + "step": 434 + }, + { + "epoch": 0.5492424242424242, + "grad_norm": 2.9457757472991943, + "learning_rate": 4.572029934518242e-06, + "loss": 1.6714, + "step": 435 + }, + { + "epoch": 0.5505050505050505, + "grad_norm": 2.6871347427368164, + "learning_rate": 4.569691300280636e-06, + "loss": 1.6648, + "step": 436 + }, + { + "epoch": 0.5517676767676768, + "grad_norm": 3.1708056926727295, + "learning_rate": 4.5673526660430315e-06, + "loss": 1.6415, + "step": 437 + }, + { + "epoch": 0.553030303030303, + "grad_norm": 2.8767378330230713, + "learning_rate": 4.565014031805426e-06, + "loss": 1.6475, + "step": 438 + }, + { + "epoch": 0.5542929292929293, + "grad_norm": 3.083566188812256, + "learning_rate": 4.56267539756782e-06, + "loss": 1.632, + "step": 439 + }, + { + "epoch": 0.5555555555555556, + "grad_norm": 2.584460735321045, + "learning_rate": 4.560336763330216e-06, + "loss": 1.6469, + "step": 440 + }, + { + "epoch": 0.5555555555555556, + "eval_loss": 1.8140095472335815, + "eval_runtime": 37.1986, + "eval_samples_per_second": 26.883, + "eval_steps_per_second": 1.129, + "step": 440 + }, + { + "epoch": 0.5568181818181818, + "grad_norm": 2.5885088443756104, + "learning_rate": 4.55799812909261e-06, + "loss": 1.6058, + "step": 441 + }, + { + "epoch": 0.5580808080808081, + "grad_norm": 2.8467836380004883, + "learning_rate": 4.5556594948550045e-06, + "loss": 1.6959, + "step": 442 + }, + { + "epoch": 0.5593434343434344, + "grad_norm": 2.764326333999634, + "learning_rate": 4.5533208606174e-06, + "loss": 1.6565, + "step": 443 + }, + { + "epoch": 0.5606060606060606, + "grad_norm": 2.88089919090271, + "learning_rate": 4.550982226379795e-06, + "loss": 1.6651, + "step": 444 + }, + { + "epoch": 0.5618686868686869, + "grad_norm": 2.6239469051361084, + "learning_rate": 4.548643592142189e-06, + "loss": 1.6255, + "step": 445 + }, + { + "epoch": 0.5631313131313131, + "grad_norm": 2.7748732566833496, + "learning_rate": 4.546304957904585e-06, + "loss": 1.6587, + "step": 446 + }, + { + "epoch": 0.5643939393939394, + "grad_norm": 2.7075202465057373, + "learning_rate": 4.543966323666979e-06, + "loss": 1.6448, + "step": 447 + }, + { + "epoch": 0.5656565656565656, + "grad_norm": 2.9291634559631348, + "learning_rate": 4.5416276894293735e-06, + "loss": 1.6564, + "step": 448 + }, + { + "epoch": 0.5669191919191919, + "grad_norm": 3.0039470195770264, + "learning_rate": 4.539289055191769e-06, + "loss": 1.623, + "step": 449 + }, + { + "epoch": 0.5681818181818182, + "grad_norm": 3.085277557373047, + "learning_rate": 4.536950420954163e-06, + "loss": 1.6176, + "step": 450 + }, + { + "epoch": 0.5681818181818182, + "eval_loss": 1.8022958040237427, + "eval_runtime": 37.2204, + "eval_samples_per_second": 26.867, + "eval_steps_per_second": 1.128, + "step": 450 + }, + { + "epoch": 0.5694444444444444, + "grad_norm": 2.93575119972229, + "learning_rate": 4.534611786716558e-06, + "loss": 1.615, + "step": 451 + }, + { + "epoch": 0.5707070707070707, + "grad_norm": 2.847141742706299, + "learning_rate": 4.532273152478953e-06, + "loss": 1.6474, + "step": 452 + }, + { + "epoch": 0.571969696969697, + "grad_norm": 3.0299603939056396, + "learning_rate": 4.529934518241347e-06, + "loss": 1.6424, + "step": 453 + }, + { + "epoch": 0.5732323232323232, + "grad_norm": 3.0307669639587402, + "learning_rate": 4.527595884003742e-06, + "loss": 1.6473, + "step": 454 + }, + { + "epoch": 0.5744949494949495, + "grad_norm": 2.864269971847534, + "learning_rate": 4.525257249766137e-06, + "loss": 1.6631, + "step": 455 + }, + { + "epoch": 0.5757575757575758, + "grad_norm": 2.606637716293335, + "learning_rate": 4.522918615528531e-06, + "loss": 1.6118, + "step": 456 + }, + { + "epoch": 0.577020202020202, + "grad_norm": 3.127891778945923, + "learning_rate": 4.520579981290926e-06, + "loss": 1.6286, + "step": 457 + }, + { + "epoch": 0.5782828282828283, + "grad_norm": 2.8809893131256104, + "learning_rate": 4.518241347053321e-06, + "loss": 1.6158, + "step": 458 + }, + { + "epoch": 0.5795454545454546, + "grad_norm": 3.0662436485290527, + "learning_rate": 4.515902712815716e-06, + "loss": 1.7336, + "step": 459 + }, + { + "epoch": 0.5808080808080808, + "grad_norm": 3.0795772075653076, + "learning_rate": 4.513564078578111e-06, + "loss": 1.6974, + "step": 460 + }, + { + "epoch": 0.5808080808080808, + "eval_loss": 1.7981404066085815, + "eval_runtime": 37.299, + "eval_samples_per_second": 26.81, + "eval_steps_per_second": 1.126, + "step": 460 + }, + { + "epoch": 0.5820707070707071, + "grad_norm": 3.091688871383667, + "learning_rate": 4.511225444340506e-06, + "loss": 1.6803, + "step": 461 + }, + { + "epoch": 0.5833333333333334, + "grad_norm": 2.5597071647644043, + "learning_rate": 4.5088868101029005e-06, + "loss": 1.6104, + "step": 462 + }, + { + "epoch": 0.5845959595959596, + "grad_norm": 2.849806070327759, + "learning_rate": 4.506548175865295e-06, + "loss": 1.6147, + "step": 463 + }, + { + "epoch": 0.5858585858585859, + "grad_norm": 2.9791204929351807, + "learning_rate": 4.50420954162769e-06, + "loss": 1.6376, + "step": 464 + }, + { + "epoch": 0.5871212121212122, + "grad_norm": 2.8582305908203125, + "learning_rate": 4.501870907390085e-06, + "loss": 1.6831, + "step": 465 + }, + { + "epoch": 0.5883838383838383, + "grad_norm": 2.8445184230804443, + "learning_rate": 4.499532273152479e-06, + "loss": 1.6463, + "step": 466 + }, + { + "epoch": 0.5896464646464646, + "grad_norm": 2.858168125152588, + "learning_rate": 4.497193638914874e-06, + "loss": 1.6662, + "step": 467 + }, + { + "epoch": 0.5909090909090909, + "grad_norm": 2.829702854156494, + "learning_rate": 4.494855004677269e-06, + "loss": 1.6058, + "step": 468 + }, + { + "epoch": 0.5921717171717171, + "grad_norm": 2.9292726516723633, + "learning_rate": 4.492516370439663e-06, + "loss": 1.6341, + "step": 469 + }, + { + "epoch": 0.5934343434343434, + "grad_norm": 2.9006292819976807, + "learning_rate": 4.490177736202058e-06, + "loss": 1.5726, + "step": 470 + }, + { + "epoch": 0.5934343434343434, + "eval_loss": 1.7953916788101196, + "eval_runtime": 37.5058, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 470 + }, + { + "epoch": 0.5946969696969697, + "grad_norm": 2.8116230964660645, + "learning_rate": 4.487839101964453e-06, + "loss": 1.6083, + "step": 471 + }, + { + "epoch": 0.5959595959595959, + "grad_norm": 2.540607452392578, + "learning_rate": 4.485500467726848e-06, + "loss": 1.6208, + "step": 472 + }, + { + "epoch": 0.5972222222222222, + "grad_norm": 2.7339096069335938, + "learning_rate": 4.4831618334892425e-06, + "loss": 1.6839, + "step": 473 + }, + { + "epoch": 0.5984848484848485, + "grad_norm": 2.9094762802124023, + "learning_rate": 4.480823199251638e-06, + "loss": 1.6962, + "step": 474 + }, + { + "epoch": 0.5997474747474747, + "grad_norm": 2.7733027935028076, + "learning_rate": 4.478484565014032e-06, + "loss": 1.6252, + "step": 475 + }, + { + "epoch": 0.601010101010101, + "grad_norm": 2.5709428787231445, + "learning_rate": 4.4761459307764274e-06, + "loss": 1.5896, + "step": 476 + }, + { + "epoch": 0.6022727272727273, + "grad_norm": 2.8358278274536133, + "learning_rate": 4.473807296538822e-06, + "loss": 1.6247, + "step": 477 + }, + { + "epoch": 0.6035353535353535, + "grad_norm": 2.7504897117614746, + "learning_rate": 4.471468662301216e-06, + "loss": 1.7138, + "step": 478 + }, + { + "epoch": 0.6047979797979798, + "grad_norm": 2.6965620517730713, + "learning_rate": 4.4691300280636115e-06, + "loss": 1.6054, + "step": 479 + }, + { + "epoch": 0.6060606060606061, + "grad_norm": 2.9401891231536865, + "learning_rate": 4.466791393826006e-06, + "loss": 1.6172, + "step": 480 + }, + { + "epoch": 0.6060606060606061, + "eval_loss": 1.7946635484695435, + "eval_runtime": 37.5366, + "eval_samples_per_second": 26.641, + "eval_steps_per_second": 1.119, + "step": 480 + }, + { + "epoch": 0.6073232323232324, + "grad_norm": 2.71535587310791, + "learning_rate": 4.4644527595884e-06, + "loss": 1.61, + "step": 481 + }, + { + "epoch": 0.6085858585858586, + "grad_norm": 2.7910654544830322, + "learning_rate": 4.462114125350796e-06, + "loss": 1.6193, + "step": 482 + }, + { + "epoch": 0.6098484848484849, + "grad_norm": 2.809241533279419, + "learning_rate": 4.45977549111319e-06, + "loss": 1.602, + "step": 483 + }, + { + "epoch": 0.6111111111111112, + "grad_norm": 2.7152137756347656, + "learning_rate": 4.457436856875585e-06, + "loss": 1.6516, + "step": 484 + }, + { + "epoch": 0.6123737373737373, + "grad_norm": 2.7653465270996094, + "learning_rate": 4.45509822263798e-06, + "loss": 1.5904, + "step": 485 + }, + { + "epoch": 0.6136363636363636, + "grad_norm": 2.729876756668091, + "learning_rate": 4.452759588400374e-06, + "loss": 1.6259, + "step": 486 + }, + { + "epoch": 0.61489898989899, + "grad_norm": 2.6290180683135986, + "learning_rate": 4.4504209541627694e-06, + "loss": 1.5796, + "step": 487 + }, + { + "epoch": 0.6161616161616161, + "grad_norm": 2.939884662628174, + "learning_rate": 4.448082319925164e-06, + "loss": 1.6774, + "step": 488 + }, + { + "epoch": 0.6174242424242424, + "grad_norm": 2.7554702758789062, + "learning_rate": 4.445743685687559e-06, + "loss": 1.6226, + "step": 489 + }, + { + "epoch": 0.6186868686868687, + "grad_norm": 2.890458106994629, + "learning_rate": 4.4434050514499535e-06, + "loss": 1.592, + "step": 490 + }, + { + "epoch": 0.6186868686868687, + "eval_loss": 1.7868437767028809, + "eval_runtime": 37.5267, + "eval_samples_per_second": 26.648, + "eval_steps_per_second": 1.119, + "step": 490 + }, + { + "epoch": 0.6199494949494949, + "grad_norm": 2.757178783416748, + "learning_rate": 4.441066417212349e-06, + "loss": 1.6297, + "step": 491 + }, + { + "epoch": 0.6212121212121212, + "grad_norm": 2.7179954051971436, + "learning_rate": 4.438727782974743e-06, + "loss": 1.5685, + "step": 492 + }, + { + "epoch": 0.6224747474747475, + "grad_norm": 2.824514389038086, + "learning_rate": 4.436389148737138e-06, + "loss": 1.6625, + "step": 493 + }, + { + "epoch": 0.6237373737373737, + "grad_norm": 3.0062615871429443, + "learning_rate": 4.434050514499533e-06, + "loss": 1.5889, + "step": 494 + }, + { + "epoch": 0.625, + "grad_norm": 2.962176561355591, + "learning_rate": 4.431711880261927e-06, + "loss": 1.6813, + "step": 495 + }, + { + "epoch": 0.6262626262626263, + "grad_norm": 2.7531204223632812, + "learning_rate": 4.429373246024323e-06, + "loss": 1.6152, + "step": 496 + }, + { + "epoch": 0.6275252525252525, + "grad_norm": 2.764021873474121, + "learning_rate": 4.427034611786717e-06, + "loss": 1.6164, + "step": 497 + }, + { + "epoch": 0.6287878787878788, + "grad_norm": 2.836826801300049, + "learning_rate": 4.4246959775491114e-06, + "loss": 1.6192, + "step": 498 + }, + { + "epoch": 0.6300505050505051, + "grad_norm": 2.871220827102661, + "learning_rate": 4.422357343311507e-06, + "loss": 1.6096, + "step": 499 + }, + { + "epoch": 0.6313131313131313, + "grad_norm": 2.8129167556762695, + "learning_rate": 4.420018709073901e-06, + "loss": 1.6674, + "step": 500 + }, + { + "epoch": 0.6313131313131313, + "eval_loss": 1.785156011581421, + "eval_runtime": 37.2971, + "eval_samples_per_second": 26.812, + "eval_steps_per_second": 1.126, + "step": 500 + }, + { + "epoch": 0.6325757575757576, + "grad_norm": 2.775108814239502, + "learning_rate": 4.4176800748362955e-06, + "loss": 1.5851, + "step": 501 + }, + { + "epoch": 0.6338383838383839, + "grad_norm": 2.7775518894195557, + "learning_rate": 4.415341440598691e-06, + "loss": 1.6839, + "step": 502 + }, + { + "epoch": 0.63510101010101, + "grad_norm": 2.7997918128967285, + "learning_rate": 4.413002806361085e-06, + "loss": 1.5849, + "step": 503 + }, + { + "epoch": 0.6363636363636364, + "grad_norm": 2.91103196144104, + "learning_rate": 4.4106641721234805e-06, + "loss": 1.6369, + "step": 504 + }, + { + "epoch": 0.6376262626262627, + "grad_norm": 2.6800882816314697, + "learning_rate": 4.408325537885875e-06, + "loss": 1.6049, + "step": 505 + }, + { + "epoch": 0.6388888888888888, + "grad_norm": 2.7451021671295166, + "learning_rate": 4.40598690364827e-06, + "loss": 1.658, + "step": 506 + }, + { + "epoch": 0.6401515151515151, + "grad_norm": 2.794813632965088, + "learning_rate": 4.403648269410665e-06, + "loss": 1.6015, + "step": 507 + }, + { + "epoch": 0.6414141414141414, + "grad_norm": 2.769373655319214, + "learning_rate": 4.401309635173059e-06, + "loss": 1.6217, + "step": 508 + }, + { + "epoch": 0.6426767676767676, + "grad_norm": 2.9081523418426514, + "learning_rate": 4.398971000935454e-06, + "loss": 1.6235, + "step": 509 + }, + { + "epoch": 0.6439393939393939, + "grad_norm": 2.707618236541748, + "learning_rate": 4.396632366697849e-06, + "loss": 1.6003, + "step": 510 + }, + { + "epoch": 0.6439393939393939, + "eval_loss": 1.7876425981521606, + "eval_runtime": 36.9832, + "eval_samples_per_second": 27.039, + "eval_steps_per_second": 1.136, + "step": 510 + }, + { + "epoch": 0.6452020202020202, + "grad_norm": 2.7923099994659424, + "learning_rate": 4.394293732460244e-06, + "loss": 1.5647, + "step": 511 + }, + { + "epoch": 0.6464646464646465, + "grad_norm": 2.8449175357818604, + "learning_rate": 4.391955098222638e-06, + "loss": 1.6456, + "step": 512 + }, + { + "epoch": 0.6477272727272727, + "grad_norm": 2.755856990814209, + "learning_rate": 4.389616463985033e-06, + "loss": 1.6962, + "step": 513 + }, + { + "epoch": 0.648989898989899, + "grad_norm": 2.5578861236572266, + "learning_rate": 4.387277829747428e-06, + "loss": 1.6097, + "step": 514 + }, + { + "epoch": 0.6502525252525253, + "grad_norm": 2.8804502487182617, + "learning_rate": 4.3849391955098225e-06, + "loss": 1.6212, + "step": 515 + }, + { + "epoch": 0.6515151515151515, + "grad_norm": 2.684215545654297, + "learning_rate": 4.382600561272217e-06, + "loss": 1.5529, + "step": 516 + }, + { + "epoch": 0.6527777777777778, + "grad_norm": 2.9128026962280273, + "learning_rate": 4.380261927034612e-06, + "loss": 1.6424, + "step": 517 + }, + { + "epoch": 0.6540404040404041, + "grad_norm": 2.733415126800537, + "learning_rate": 4.377923292797007e-06, + "loss": 1.6201, + "step": 518 + }, + { + "epoch": 0.6553030303030303, + "grad_norm": 2.7211062908172607, + "learning_rate": 4.375584658559401e-06, + "loss": 1.5833, + "step": 519 + }, + { + "epoch": 0.6565656565656566, + "grad_norm": 2.7308316230773926, + "learning_rate": 4.373246024321796e-06, + "loss": 1.6288, + "step": 520 + }, + { + "epoch": 0.6565656565656566, + "eval_loss": 1.7787725925445557, + "eval_runtime": 37.1089, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 520 + }, + { + "epoch": 0.6578282828282829, + "grad_norm": 2.760362148284912, + "learning_rate": 4.3709073900841915e-06, + "loss": 1.5814, + "step": 521 + }, + { + "epoch": 0.6590909090909091, + "grad_norm": 2.7625153064727783, + "learning_rate": 4.368568755846586e-06, + "loss": 1.6239, + "step": 522 + }, + { + "epoch": 0.6603535353535354, + "grad_norm": 2.715710401535034, + "learning_rate": 4.366230121608981e-06, + "loss": 1.6069, + "step": 523 + }, + { + "epoch": 0.6616161616161617, + "grad_norm": 2.7753846645355225, + "learning_rate": 4.363891487371376e-06, + "loss": 1.6178, + "step": 524 + }, + { + "epoch": 0.6628787878787878, + "grad_norm": 2.650716543197632, + "learning_rate": 4.36155285313377e-06, + "loss": 1.6534, + "step": 525 + }, + { + "epoch": 0.6641414141414141, + "grad_norm": 2.573969602584839, + "learning_rate": 4.359214218896165e-06, + "loss": 1.6721, + "step": 526 + }, + { + "epoch": 0.6654040404040404, + "grad_norm": 2.9017715454101562, + "learning_rate": 4.35687558465856e-06, + "loss": 1.6273, + "step": 527 + }, + { + "epoch": 0.6666666666666666, + "grad_norm": 2.691643714904785, + "learning_rate": 4.354536950420954e-06, + "loss": 1.6199, + "step": 528 + }, + { + "epoch": 0.6679292929292929, + "grad_norm": 2.818662166595459, + "learning_rate": 4.3521983161833494e-06, + "loss": 1.5758, + "step": 529 + }, + { + "epoch": 0.6691919191919192, + "grad_norm": 3.017738103866577, + "learning_rate": 4.349859681945744e-06, + "loss": 1.6437, + "step": 530 + }, + { + "epoch": 0.6691919191919192, + "eval_loss": 1.7835458517074585, + "eval_runtime": 37.1359, + "eval_samples_per_second": 26.928, + "eval_steps_per_second": 1.131, + "step": 530 + }, + { + "epoch": 0.6704545454545454, + "grad_norm": 2.8027284145355225, + "learning_rate": 4.347521047708138e-06, + "loss": 1.6688, + "step": 531 + }, + { + "epoch": 0.6717171717171717, + "grad_norm": 2.6642649173736572, + "learning_rate": 4.3451824134705335e-06, + "loss": 1.6327, + "step": 532 + }, + { + "epoch": 0.672979797979798, + "grad_norm": 2.7406396865844727, + "learning_rate": 4.342843779232928e-06, + "loss": 1.6297, + "step": 533 + }, + { + "epoch": 0.6742424242424242, + "grad_norm": 2.758538007736206, + "learning_rate": 4.340505144995323e-06, + "loss": 1.5276, + "step": 534 + }, + { + "epoch": 0.6755050505050505, + "grad_norm": 2.815817356109619, + "learning_rate": 4.3381665107577185e-06, + "loss": 1.6192, + "step": 535 + }, + { + "epoch": 0.6767676767676768, + "grad_norm": 2.8416755199432373, + "learning_rate": 4.335827876520113e-06, + "loss": 1.5966, + "step": 536 + }, + { + "epoch": 0.678030303030303, + "grad_norm": 3.014401435852051, + "learning_rate": 4.333489242282507e-06, + "loss": 1.6104, + "step": 537 + }, + { + "epoch": 0.6792929292929293, + "grad_norm": 2.731356143951416, + "learning_rate": 4.331150608044903e-06, + "loss": 1.6177, + "step": 538 + }, + { + "epoch": 0.6805555555555556, + "grad_norm": 2.633310317993164, + "learning_rate": 4.328811973807297e-06, + "loss": 1.5004, + "step": 539 + }, + { + "epoch": 0.6818181818181818, + "grad_norm": 2.977476119995117, + "learning_rate": 4.3264733395696914e-06, + "loss": 1.6171, + "step": 540 + }, + { + "epoch": 0.6818181818181818, + "eval_loss": 1.7790141105651855, + "eval_runtime": 37.12, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 540 + }, + { + "epoch": 0.6830808080808081, + "grad_norm": 2.769422769546509, + "learning_rate": 4.324134705332087e-06, + "loss": 1.5591, + "step": 541 + }, + { + "epoch": 0.6843434343434344, + "grad_norm": 3.088426351547241, + "learning_rate": 4.321796071094481e-06, + "loss": 1.5603, + "step": 542 + }, + { + "epoch": 0.6856060606060606, + "grad_norm": 2.6784212589263916, + "learning_rate": 4.3194574368568755e-06, + "loss": 1.6061, + "step": 543 + }, + { + "epoch": 0.6868686868686869, + "grad_norm": 3.209821939468384, + "learning_rate": 4.317118802619271e-06, + "loss": 1.585, + "step": 544 + }, + { + "epoch": 0.6881313131313131, + "grad_norm": 2.940791368484497, + "learning_rate": 4.314780168381665e-06, + "loss": 1.6092, + "step": 545 + }, + { + "epoch": 0.6893939393939394, + "grad_norm": 2.679884672164917, + "learning_rate": 4.3124415341440605e-06, + "loss": 1.6207, + "step": 546 + }, + { + "epoch": 0.6906565656565656, + "grad_norm": 2.853809118270874, + "learning_rate": 4.310102899906455e-06, + "loss": 1.5674, + "step": 547 + }, + { + "epoch": 0.6919191919191919, + "grad_norm": 2.693838119506836, + "learning_rate": 4.307764265668849e-06, + "loss": 1.5886, + "step": 548 + }, + { + "epoch": 0.6931818181818182, + "grad_norm": 2.871680736541748, + "learning_rate": 4.305425631431245e-06, + "loss": 1.6309, + "step": 549 + }, + { + "epoch": 0.6944444444444444, + "grad_norm": 2.9908292293548584, + "learning_rate": 4.30308699719364e-06, + "loss": 1.5577, + "step": 550 + }, + { + "epoch": 0.6944444444444444, + "eval_loss": 1.7790372371673584, + "eval_runtime": 37.114, + "eval_samples_per_second": 26.944, + "eval_steps_per_second": 1.132, + "step": 550 + }, + { + "epoch": 0.6957070707070707, + "grad_norm": 2.883481740951538, + "learning_rate": 4.300748362956034e-06, + "loss": 1.5749, + "step": 551 + }, + { + "epoch": 0.696969696969697, + "grad_norm": 2.7523458003997803, + "learning_rate": 4.298409728718429e-06, + "loss": 1.624, + "step": 552 + }, + { + "epoch": 0.6982323232323232, + "grad_norm": 2.69509220123291, + "learning_rate": 4.296071094480824e-06, + "loss": 1.5856, + "step": 553 + }, + { + "epoch": 0.6994949494949495, + "grad_norm": 2.6340222358703613, + "learning_rate": 4.293732460243218e-06, + "loss": 1.5857, + "step": 554 + }, + { + "epoch": 0.7007575757575758, + "grad_norm": 2.4998607635498047, + "learning_rate": 4.291393826005613e-06, + "loss": 1.6038, + "step": 555 + }, + { + "epoch": 0.702020202020202, + "grad_norm": 2.7913858890533447, + "learning_rate": 4.289055191768008e-06, + "loss": 1.6411, + "step": 556 + }, + { + "epoch": 0.7032828282828283, + "grad_norm": 2.7149479389190674, + "learning_rate": 4.2867165575304025e-06, + "loss": 1.592, + "step": 557 + }, + { + "epoch": 0.7045454545454546, + "grad_norm": 2.7187132835388184, + "learning_rate": 4.284377923292797e-06, + "loss": 1.5413, + "step": 558 + }, + { + "epoch": 0.7058080808080808, + "grad_norm": 2.8913674354553223, + "learning_rate": 4.282039289055192e-06, + "loss": 1.6105, + "step": 559 + }, + { + "epoch": 0.7070707070707071, + "grad_norm": 2.8470897674560547, + "learning_rate": 4.279700654817587e-06, + "loss": 1.6481, + "step": 560 + }, + { + "epoch": 0.7070707070707071, + "eval_loss": 1.7711386680603027, + "eval_runtime": 37.1243, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 560 + }, + { + "epoch": 0.7083333333333334, + "grad_norm": 2.7525687217712402, + "learning_rate": 4.277362020579982e-06, + "loss": 1.5749, + "step": 561 + }, + { + "epoch": 0.7095959595959596, + "grad_norm": 2.6935207843780518, + "learning_rate": 4.275023386342376e-06, + "loss": 1.637, + "step": 562 + }, + { + "epoch": 0.7108585858585859, + "grad_norm": 2.6935207843780518, + "learning_rate": 4.275023386342376e-06, + "loss": 1.5903, + "step": 563 + }, + { + "epoch": 0.7121212121212122, + "grad_norm": 2.992992401123047, + "learning_rate": 4.272684752104771e-06, + "loss": 1.6654, + "step": 564 + }, + { + "epoch": 0.7133838383838383, + "grad_norm": 2.7314348220825195, + "learning_rate": 4.270346117867166e-06, + "loss": 1.6775, + "step": 565 + }, + { + "epoch": 0.7146464646464646, + "grad_norm": 2.6952013969421387, + "learning_rate": 4.26800748362956e-06, + "loss": 1.5936, + "step": 566 + }, + { + "epoch": 0.7159090909090909, + "grad_norm": 2.8244829177856445, + "learning_rate": 4.265668849391956e-06, + "loss": 1.6228, + "step": 567 + }, + { + "epoch": 0.7171717171717171, + "grad_norm": 2.9110143184661865, + "learning_rate": 4.26333021515435e-06, + "loss": 1.6016, + "step": 568 + }, + { + "epoch": 0.7184343434343434, + "grad_norm": 2.7953436374664307, + "learning_rate": 4.260991580916745e-06, + "loss": 1.611, + "step": 569 + }, + { + "epoch": 0.7196969696969697, + "grad_norm": 2.854447364807129, + "learning_rate": 4.25865294667914e-06, + "loss": 1.6365, + "step": 570 + }, + { + "epoch": 0.7196969696969697, + "eval_loss": 1.7711069583892822, + "eval_runtime": 37.1374, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 570 + }, + { + "epoch": 0.7209595959595959, + "grad_norm": 2.7885239124298096, + "learning_rate": 4.256314312441534e-06, + "loss": 1.6106, + "step": 571 + }, + { + "epoch": 0.7222222222222222, + "grad_norm": 2.8012993335723877, + "learning_rate": 4.2539756782039295e-06, + "loss": 1.6081, + "step": 572 + }, + { + "epoch": 0.7234848484848485, + "grad_norm": 2.9223906993865967, + "learning_rate": 4.251637043966324e-06, + "loss": 1.5836, + "step": 573 + }, + { + "epoch": 0.7247474747474747, + "grad_norm": 2.885830879211426, + "learning_rate": 4.249298409728719e-06, + "loss": 1.5616, + "step": 574 + }, + { + "epoch": 0.726010101010101, + "grad_norm": 2.687372922897339, + "learning_rate": 4.2469597754911136e-06, + "loss": 1.5873, + "step": 575 + }, + { + "epoch": 0.7272727272727273, + "grad_norm": 3.0156168937683105, + "learning_rate": 4.244621141253508e-06, + "loss": 1.5763, + "step": 576 + }, + { + "epoch": 0.7285353535353535, + "grad_norm": 3.061842918395996, + "learning_rate": 4.242282507015903e-06, + "loss": 1.6148, + "step": 577 + }, + { + "epoch": 0.7297979797979798, + "grad_norm": 2.7467715740203857, + "learning_rate": 4.239943872778298e-06, + "loss": 1.577, + "step": 578 + }, + { + "epoch": 0.7310606060606061, + "grad_norm": 2.669501543045044, + "learning_rate": 4.237605238540692e-06, + "loss": 1.5653, + "step": 579 + }, + { + "epoch": 0.7323232323232324, + "grad_norm": 2.874349355697632, + "learning_rate": 4.235266604303087e-06, + "loss": 1.6523, + "step": 580 + }, + { + "epoch": 0.7323232323232324, + "eval_loss": 1.7680890560150146, + "eval_runtime": 37.2931, + "eval_samples_per_second": 26.815, + "eval_steps_per_second": 1.126, + "step": 580 + }, + { + "epoch": 0.7335858585858586, + "grad_norm": 2.7468504905700684, + "learning_rate": 4.232927970065482e-06, + "loss": 1.5935, + "step": 581 + }, + { + "epoch": 0.7348484848484849, + "grad_norm": 2.936697244644165, + "learning_rate": 4.230589335827877e-06, + "loss": 1.6335, + "step": 582 + }, + { + "epoch": 0.7361111111111112, + "grad_norm": 2.6878974437713623, + "learning_rate": 4.2282507015902715e-06, + "loss": 1.5699, + "step": 583 + }, + { + "epoch": 0.7373737373737373, + "grad_norm": 3.3578853607177734, + "learning_rate": 4.225912067352667e-06, + "loss": 1.5979, + "step": 584 + }, + { + "epoch": 0.7386363636363636, + "grad_norm": 2.8441176414489746, + "learning_rate": 4.223573433115061e-06, + "loss": 1.5736, + "step": 585 + }, + { + "epoch": 0.73989898989899, + "grad_norm": 2.7186129093170166, + "learning_rate": 4.221234798877456e-06, + "loss": 1.636, + "step": 586 + }, + { + "epoch": 0.7411616161616161, + "grad_norm": 2.940002918243408, + "learning_rate": 4.218896164639851e-06, + "loss": 1.5721, + "step": 587 + }, + { + "epoch": 0.7424242424242424, + "grad_norm": 2.676284074783325, + "learning_rate": 4.216557530402245e-06, + "loss": 1.5972, + "step": 588 + }, + { + "epoch": 0.7436868686868687, + "grad_norm": 2.9176571369171143, + "learning_rate": 4.2142188961646405e-06, + "loss": 1.6045, + "step": 589 + }, + { + "epoch": 0.7449494949494949, + "grad_norm": 2.7716288566589355, + "learning_rate": 4.211880261927035e-06, + "loss": 1.5988, + "step": 590 + }, + { + "epoch": 0.7449494949494949, + "eval_loss": 1.775631070137024, + "eval_runtime": 37.4113, + "eval_samples_per_second": 26.73, + "eval_steps_per_second": 1.123, + "step": 590 + }, + { + "epoch": 0.7462121212121212, + "grad_norm": 2.7201945781707764, + "learning_rate": 4.209541627689429e-06, + "loss": 1.5767, + "step": 591 + }, + { + "epoch": 0.7474747474747475, + "grad_norm": 2.838362455368042, + "learning_rate": 4.207202993451825e-06, + "loss": 1.5793, + "step": 592 + }, + { + "epoch": 0.7487373737373737, + "grad_norm": 2.6511716842651367, + "learning_rate": 4.204864359214219e-06, + "loss": 1.5776, + "step": 593 + }, + { + "epoch": 0.75, + "grad_norm": 2.7062957286834717, + "learning_rate": 4.2025257249766135e-06, + "loss": 1.5593, + "step": 594 + }, + { + "epoch": 0.7512626262626263, + "grad_norm": 2.808570384979248, + "learning_rate": 4.200187090739009e-06, + "loss": 1.6249, + "step": 595 + }, + { + "epoch": 0.7525252525252525, + "grad_norm": 2.600278377532959, + "learning_rate": 4.197848456501403e-06, + "loss": 1.6171, + "step": 596 + }, + { + "epoch": 0.7537878787878788, + "grad_norm": 2.9669382572174072, + "learning_rate": 4.195509822263798e-06, + "loss": 1.608, + "step": 597 + }, + { + "epoch": 0.7550505050505051, + "grad_norm": 2.6576340198516846, + "learning_rate": 4.193171188026194e-06, + "loss": 1.5986, + "step": 598 + }, + { + "epoch": 0.7563131313131313, + "grad_norm": 2.787060022354126, + "learning_rate": 4.190832553788588e-06, + "loss": 1.548, + "step": 599 + }, + { + "epoch": 0.7575757575757576, + "grad_norm": 2.854971408843994, + "learning_rate": 4.1884939195509825e-06, + "loss": 1.6421, + "step": 600 + }, + { + "epoch": 0.7575757575757576, + "eval_loss": 1.7721046209335327, + "eval_runtime": 37.4471, + "eval_samples_per_second": 26.704, + "eval_steps_per_second": 1.122, + "step": 600 + }, + { + "epoch": 0.7588383838383839, + "grad_norm": 2.7417967319488525, + "learning_rate": 4.186155285313378e-06, + "loss": 1.6399, + "step": 601 + }, + { + "epoch": 0.76010101010101, + "grad_norm": 3.0036182403564453, + "learning_rate": 4.183816651075772e-06, + "loss": 1.5953, + "step": 602 + }, + { + "epoch": 0.7613636363636364, + "grad_norm": 2.502861261367798, + "learning_rate": 4.181478016838167e-06, + "loss": 1.5284, + "step": 603 + }, + { + "epoch": 0.7626262626262627, + "grad_norm": 2.7262094020843506, + "learning_rate": 4.179139382600562e-06, + "loss": 1.619, + "step": 604 + }, + { + "epoch": 0.7638888888888888, + "grad_norm": 2.8748724460601807, + "learning_rate": 4.176800748362956e-06, + "loss": 1.5625, + "step": 605 + }, + { + "epoch": 0.7651515151515151, + "grad_norm": 2.7272109985351562, + "learning_rate": 4.174462114125351e-06, + "loss": 1.5935, + "step": 606 + }, + { + "epoch": 0.7664141414141414, + "grad_norm": 2.7930479049682617, + "learning_rate": 4.172123479887746e-06, + "loss": 1.61, + "step": 607 + }, + { + "epoch": 0.7676767676767676, + "grad_norm": 2.952056407928467, + "learning_rate": 4.16978484565014e-06, + "loss": 1.6369, + "step": 608 + }, + { + "epoch": 0.7689393939393939, + "grad_norm": 2.979564905166626, + "learning_rate": 4.167446211412535e-06, + "loss": 1.5748, + "step": 609 + }, + { + "epoch": 0.7702020202020202, + "grad_norm": 2.7945773601531982, + "learning_rate": 4.16510757717493e-06, + "loss": 1.6096, + "step": 610 + }, + { + "epoch": 0.7702020202020202, + "eval_loss": 1.7568180561065674, + "eval_runtime": 36.969, + "eval_samples_per_second": 27.05, + "eval_steps_per_second": 1.136, + "step": 610 + }, + { + "epoch": 0.7714646464646465, + "grad_norm": 2.9217135906219482, + "learning_rate": 4.1627689429373245e-06, + "loss": 1.5902, + "step": 611 + }, + { + "epoch": 0.7727272727272727, + "grad_norm": 2.8184452056884766, + "learning_rate": 4.16043030869972e-06, + "loss": 1.5775, + "step": 612 + }, + { + "epoch": 0.773989898989899, + "grad_norm": 2.7121212482452393, + "learning_rate": 4.158091674462115e-06, + "loss": 1.5778, + "step": 613 + }, + { + "epoch": 0.7752525252525253, + "grad_norm": 2.7600955963134766, + "learning_rate": 4.1557530402245095e-06, + "loss": 1.6145, + "step": 614 + }, + { + "epoch": 0.7765151515151515, + "grad_norm": 2.6388845443725586, + "learning_rate": 4.153414405986904e-06, + "loss": 1.6475, + "step": 615 + }, + { + "epoch": 0.7777777777777778, + "grad_norm": 2.684379816055298, + "learning_rate": 4.151075771749299e-06, + "loss": 1.6187, + "step": 616 + }, + { + "epoch": 0.7790404040404041, + "grad_norm": 2.6907055377960205, + "learning_rate": 4.1487371375116936e-06, + "loss": 1.6108, + "step": 617 + }, + { + "epoch": 0.7803030303030303, + "grad_norm": 2.8458995819091797, + "learning_rate": 4.146398503274088e-06, + "loss": 1.5596, + "step": 618 + }, + { + "epoch": 0.7815656565656566, + "grad_norm": 2.68843674659729, + "learning_rate": 4.144059869036483e-06, + "loss": 1.5912, + "step": 619 + }, + { + "epoch": 0.7828282828282829, + "grad_norm": 2.7620413303375244, + "learning_rate": 4.141721234798878e-06, + "loss": 1.627, + "step": 620 + }, + { + "epoch": 0.7828282828282829, + "eval_loss": 1.757076621055603, + "eval_runtime": 37.0489, + "eval_samples_per_second": 26.991, + "eval_steps_per_second": 1.134, + "step": 620 + }, + { + "epoch": 0.7840909090909091, + "grad_norm": 2.7084949016571045, + "learning_rate": 4.139382600561272e-06, + "loss": 1.5249, + "step": 621 + }, + { + "epoch": 0.7853535353535354, + "grad_norm": 3.0410406589508057, + "learning_rate": 4.137043966323667e-06, + "loss": 1.6458, + "step": 622 + }, + { + "epoch": 0.7866161616161617, + "grad_norm": 2.8445630073547363, + "learning_rate": 4.134705332086062e-06, + "loss": 1.6404, + "step": 623 + }, + { + "epoch": 0.7878787878787878, + "grad_norm": 3.0764353275299072, + "learning_rate": 4.132366697848457e-06, + "loss": 1.5041, + "step": 624 + }, + { + "epoch": 0.7891414141414141, + "grad_norm": 2.626544237136841, + "learning_rate": 4.1300280636108515e-06, + "loss": 1.6145, + "step": 625 + }, + { + "epoch": 0.7904040404040404, + "grad_norm": 2.9735796451568604, + "learning_rate": 4.127689429373246e-06, + "loss": 1.5672, + "step": 626 + }, + { + "epoch": 0.7916666666666666, + "grad_norm": 2.7798562049865723, + "learning_rate": 4.125350795135641e-06, + "loss": 1.6272, + "step": 627 + }, + { + "epoch": 0.7929292929292929, + "grad_norm": 2.7438931465148926, + "learning_rate": 4.123012160898036e-06, + "loss": 1.5318, + "step": 628 + }, + { + "epoch": 0.7941919191919192, + "grad_norm": 3.103142499923706, + "learning_rate": 4.120673526660431e-06, + "loss": 1.6, + "step": 629 + }, + { + "epoch": 0.7954545454545454, + "grad_norm": 2.9056239128112793, + "learning_rate": 4.118334892422825e-06, + "loss": 1.6581, + "step": 630 + }, + { + "epoch": 0.7954545454545454, + "eval_loss": 1.7552870512008667, + "eval_runtime": 37.1223, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 630 + }, + { + "epoch": 0.7967171717171717, + "grad_norm": 2.7320070266723633, + "learning_rate": 4.1159962581852205e-06, + "loss": 1.5808, + "step": 631 + }, + { + "epoch": 0.797979797979798, + "grad_norm": 2.9518637657165527, + "learning_rate": 4.113657623947615e-06, + "loss": 1.6034, + "step": 632 + }, + { + "epoch": 0.7992424242424242, + "grad_norm": 2.7135164737701416, + "learning_rate": 4.111318989710009e-06, + "loss": 1.5684, + "step": 633 + }, + { + "epoch": 0.8005050505050505, + "grad_norm": 2.948230266571045, + "learning_rate": 4.108980355472405e-06, + "loss": 1.6378, + "step": 634 + }, + { + "epoch": 0.8017676767676768, + "grad_norm": 2.816206216812134, + "learning_rate": 4.106641721234799e-06, + "loss": 1.6118, + "step": 635 + }, + { + "epoch": 0.803030303030303, + "grad_norm": 2.8032639026641846, + "learning_rate": 4.104303086997194e-06, + "loss": 1.5627, + "step": 636 + }, + { + "epoch": 0.8042929292929293, + "grad_norm": 3.032348871231079, + "learning_rate": 4.101964452759589e-06, + "loss": 1.6192, + "step": 637 + }, + { + "epoch": 0.8055555555555556, + "grad_norm": 2.9512710571289062, + "learning_rate": 4.099625818521983e-06, + "loss": 1.5655, + "step": 638 + }, + { + "epoch": 0.8068181818181818, + "grad_norm": 2.738973379135132, + "learning_rate": 4.097287184284378e-06, + "loss": 1.5686, + "step": 639 + }, + { + "epoch": 0.8080808080808081, + "grad_norm": 2.6510915756225586, + "learning_rate": 4.094948550046773e-06, + "loss": 1.5889, + "step": 640 + }, + { + "epoch": 0.8080808080808081, + "eval_loss": 1.762879490852356, + "eval_runtime": 37.1457, + "eval_samples_per_second": 26.921, + "eval_steps_per_second": 1.131, + "step": 640 + }, + { + "epoch": 0.8093434343434344, + "grad_norm": 2.6933884620666504, + "learning_rate": 4.092609915809167e-06, + "loss": 1.5436, + "step": 641 + }, + { + "epoch": 0.8106060606060606, + "grad_norm": 2.703784227371216, + "learning_rate": 4.0902712815715625e-06, + "loss": 1.5869, + "step": 642 + }, + { + "epoch": 0.8118686868686869, + "grad_norm": 2.7742321491241455, + "learning_rate": 4.087932647333958e-06, + "loss": 1.5779, + "step": 643 + }, + { + "epoch": 0.8131313131313131, + "grad_norm": 2.836832284927368, + "learning_rate": 4.085594013096352e-06, + "loss": 1.5511, + "step": 644 + }, + { + "epoch": 0.8143939393939394, + "grad_norm": 2.9212028980255127, + "learning_rate": 4.083255378858747e-06, + "loss": 1.5604, + "step": 645 + }, + { + "epoch": 0.8156565656565656, + "grad_norm": 2.8608527183532715, + "learning_rate": 4.080916744621142e-06, + "loss": 1.5975, + "step": 646 + }, + { + "epoch": 0.8169191919191919, + "grad_norm": 2.8246703147888184, + "learning_rate": 4.078578110383536e-06, + "loss": 1.5962, + "step": 647 + }, + { + "epoch": 0.8181818181818182, + "grad_norm": 3.0648651123046875, + "learning_rate": 4.076239476145932e-06, + "loss": 1.6267, + "step": 648 + }, + { + "epoch": 0.8194444444444444, + "grad_norm": 2.849240303039551, + "learning_rate": 4.073900841908326e-06, + "loss": 1.6271, + "step": 649 + }, + { + "epoch": 0.8207070707070707, + "grad_norm": 2.8817555904388428, + "learning_rate": 4.07156220767072e-06, + "loss": 1.5916, + "step": 650 + }, + { + "epoch": 0.8207070707070707, + "eval_loss": 1.7502237558364868, + "eval_runtime": 37.1497, + "eval_samples_per_second": 26.918, + "eval_steps_per_second": 1.131, + "step": 650 + }, + { + "epoch": 0.821969696969697, + "grad_norm": 2.5747952461242676, + "learning_rate": 4.069223573433116e-06, + "loss": 1.5661, + "step": 651 + }, + { + "epoch": 0.8232323232323232, + "grad_norm": 2.7721974849700928, + "learning_rate": 4.06688493919551e-06, + "loss": 1.5903, + "step": 652 + }, + { + "epoch": 0.8244949494949495, + "grad_norm": 2.9606690406799316, + "learning_rate": 4.0645463049579045e-06, + "loss": 1.5431, + "step": 653 + }, + { + "epoch": 0.8257575757575758, + "grad_norm": 2.941169261932373, + "learning_rate": 4.0622076707203e-06, + "loss": 1.5711, + "step": 654 + }, + { + "epoch": 0.827020202020202, + "grad_norm": 2.8408572673797607, + "learning_rate": 4.059869036482694e-06, + "loss": 1.5975, + "step": 655 + }, + { + "epoch": 0.8282828282828283, + "grad_norm": 2.619875431060791, + "learning_rate": 4.057530402245089e-06, + "loss": 1.5194, + "step": 656 + }, + { + "epoch": 0.8295454545454546, + "grad_norm": 2.910982847213745, + "learning_rate": 4.055191768007484e-06, + "loss": 1.6318, + "step": 657 + }, + { + "epoch": 0.8308080808080808, + "grad_norm": 2.747422456741333, + "learning_rate": 4.052853133769879e-06, + "loss": 1.6205, + "step": 658 + }, + { + "epoch": 0.8320707070707071, + "grad_norm": 2.6889796257019043, + "learning_rate": 4.050514499532274e-06, + "loss": 1.5537, + "step": 659 + }, + { + "epoch": 0.8333333333333334, + "grad_norm": 2.9849202632904053, + "learning_rate": 4.048175865294669e-06, + "loss": 1.5438, + "step": 660 + }, + { + "epoch": 0.8333333333333334, + "eval_loss": 1.7484922409057617, + "eval_runtime": 37.161, + "eval_samples_per_second": 26.91, + "eval_steps_per_second": 1.13, + "step": 660 + }, + { + "epoch": 0.8345959595959596, + "grad_norm": 2.816930055618286, + "learning_rate": 4.045837231057063e-06, + "loss": 1.555, + "step": 661 + }, + { + "epoch": 0.8358585858585859, + "grad_norm": 2.990126371383667, + "learning_rate": 4.043498596819458e-06, + "loss": 1.6263, + "step": 662 + }, + { + "epoch": 0.8371212121212122, + "grad_norm": 2.898263692855835, + "learning_rate": 4.041159962581853e-06, + "loss": 1.5522, + "step": 663 + }, + { + "epoch": 0.8383838383838383, + "grad_norm": 2.7824151515960693, + "learning_rate": 4.038821328344247e-06, + "loss": 1.5366, + "step": 664 + }, + { + "epoch": 0.8396464646464646, + "grad_norm": 2.9943954944610596, + "learning_rate": 4.036482694106642e-06, + "loss": 1.6214, + "step": 665 + }, + { + "epoch": 0.8409090909090909, + "grad_norm": 2.628308057785034, + "learning_rate": 4.034144059869037e-06, + "loss": 1.5797, + "step": 666 + }, + { + "epoch": 0.8421717171717171, + "grad_norm": 2.7565176486968994, + "learning_rate": 4.0318054256314315e-06, + "loss": 1.5567, + "step": 667 + }, + { + "epoch": 0.8434343434343434, + "grad_norm": 2.688354969024658, + "learning_rate": 4.029466791393826e-06, + "loss": 1.5516, + "step": 668 + }, + { + "epoch": 0.8446969696969697, + "grad_norm": 2.899620771408081, + "learning_rate": 4.027128157156221e-06, + "loss": 1.6014, + "step": 669 + }, + { + "epoch": 0.8459595959595959, + "grad_norm": 2.7894718647003174, + "learning_rate": 4.024789522918616e-06, + "loss": 1.5688, + "step": 670 + }, + { + "epoch": 0.8459595959595959, + "eval_loss": 1.7486555576324463, + "eval_runtime": 37.1824, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 670 + }, + { + "epoch": 0.8472222222222222, + "grad_norm": 2.7441444396972656, + "learning_rate": 4.02245088868101e-06, + "loss": 1.5949, + "step": 671 + }, + { + "epoch": 0.8484848484848485, + "grad_norm": 2.7088451385498047, + "learning_rate": 4.020112254443405e-06, + "loss": 1.5878, + "step": 672 + }, + { + "epoch": 0.8497474747474747, + "grad_norm": 2.756920576095581, + "learning_rate": 4.0177736202058005e-06, + "loss": 1.6356, + "step": 673 + }, + { + "epoch": 0.851010101010101, + "grad_norm": 2.5605697631835938, + "learning_rate": 4.015434985968195e-06, + "loss": 1.5586, + "step": 674 + }, + { + "epoch": 0.8522727272727273, + "grad_norm": 2.946573257446289, + "learning_rate": 4.01309635173059e-06, + "loss": 1.5537, + "step": 675 + }, + { + "epoch": 0.8535353535353535, + "grad_norm": 2.65500545501709, + "learning_rate": 4.010757717492985e-06, + "loss": 1.5875, + "step": 676 + }, + { + "epoch": 0.8547979797979798, + "grad_norm": 2.6632118225097656, + "learning_rate": 4.008419083255379e-06, + "loss": 1.5329, + "step": 677 + }, + { + "epoch": 0.8560606060606061, + "grad_norm": 2.8909032344818115, + "learning_rate": 4.006080449017774e-06, + "loss": 1.6396, + "step": 678 + }, + { + "epoch": 0.8573232323232324, + "grad_norm": 2.8780951499938965, + "learning_rate": 4.003741814780169e-06, + "loss": 1.6189, + "step": 679 + }, + { + "epoch": 0.8585858585858586, + "grad_norm": 2.7757010459899902, + "learning_rate": 4.001403180542563e-06, + "loss": 1.5968, + "step": 680 + }, + { + "epoch": 0.8585858585858586, + "eval_loss": 1.7423925399780273, + "eval_runtime": 37.1537, + "eval_samples_per_second": 26.915, + "eval_steps_per_second": 1.13, + "step": 680 + }, + { + "epoch": 0.8598484848484849, + "grad_norm": 2.6610794067382812, + "learning_rate": 3.9990645463049584e-06, + "loss": 1.5258, + "step": 681 + }, + { + "epoch": 0.8611111111111112, + "grad_norm": 2.9292140007019043, + "learning_rate": 3.996725912067353e-06, + "loss": 1.5669, + "step": 682 + }, + { + "epoch": 0.8623737373737373, + "grad_norm": 2.7885360717773438, + "learning_rate": 3.994387277829747e-06, + "loss": 1.5737, + "step": 683 + }, + { + "epoch": 0.8636363636363636, + "grad_norm": 2.7511544227600098, + "learning_rate": 3.9920486435921425e-06, + "loss": 1.5715, + "step": 684 + }, + { + "epoch": 0.86489898989899, + "grad_norm": 2.8561315536499023, + "learning_rate": 3.989710009354537e-06, + "loss": 1.5965, + "step": 685 + }, + { + "epoch": 0.8661616161616161, + "grad_norm": 2.9127237796783447, + "learning_rate": 3.987371375116932e-06, + "loss": 1.5407, + "step": 686 + }, + { + "epoch": 0.8674242424242424, + "grad_norm": 2.735577344894409, + "learning_rate": 3.985032740879327e-06, + "loss": 1.5449, + "step": 687 + }, + { + "epoch": 0.8686868686868687, + "grad_norm": 2.6626689434051514, + "learning_rate": 3.982694106641721e-06, + "loss": 1.5173, + "step": 688 + }, + { + "epoch": 0.8699494949494949, + "grad_norm": 2.901592969894409, + "learning_rate": 3.980355472404116e-06, + "loss": 1.5757, + "step": 689 + }, + { + "epoch": 0.8712121212121212, + "grad_norm": 2.8801231384277344, + "learning_rate": 3.978016838166512e-06, + "loss": 1.5633, + "step": 690 + }, + { + "epoch": 0.8712121212121212, + "eval_loss": 1.7453432083129883, + "eval_runtime": 37.4368, + "eval_samples_per_second": 26.712, + "eval_steps_per_second": 1.122, + "step": 690 + }, + { + "epoch": 0.8724747474747475, + "grad_norm": 2.7463626861572266, + "learning_rate": 3.975678203928906e-06, + "loss": 1.5786, + "step": 691 + }, + { + "epoch": 0.8737373737373737, + "grad_norm": 2.659820318222046, + "learning_rate": 3.9733395696913004e-06, + "loss": 1.5923, + "step": 692 + }, + { + "epoch": 0.875, + "grad_norm": 2.759030342102051, + "learning_rate": 3.971000935453696e-06, + "loss": 1.6217, + "step": 693 + }, + { + "epoch": 0.8762626262626263, + "grad_norm": 2.689967632293701, + "learning_rate": 3.96866230121609e-06, + "loss": 1.5533, + "step": 694 + }, + { + "epoch": 0.8775252525252525, + "grad_norm": 3.1230740547180176, + "learning_rate": 3.9663236669784845e-06, + "loss": 1.6538, + "step": 695 + }, + { + "epoch": 0.8787878787878788, + "grad_norm": 2.657022476196289, + "learning_rate": 3.96398503274088e-06, + "loss": 1.5432, + "step": 696 + }, + { + "epoch": 0.8800505050505051, + "grad_norm": 2.919379711151123, + "learning_rate": 3.961646398503274e-06, + "loss": 1.5867, + "step": 697 + }, + { + "epoch": 0.8813131313131313, + "grad_norm": 2.607135057449341, + "learning_rate": 3.9593077642656695e-06, + "loss": 1.5079, + "step": 698 + }, + { + "epoch": 0.8825757575757576, + "grad_norm": 2.9978976249694824, + "learning_rate": 3.956969130028064e-06, + "loss": 1.5905, + "step": 699 + }, + { + "epoch": 0.8838383838383839, + "grad_norm": 2.8117148876190186, + "learning_rate": 3.954630495790458e-06, + "loss": 1.5588, + "step": 700 + }, + { + "epoch": 0.8838383838383839, + "eval_loss": 1.7376645803451538, + "eval_runtime": 37.5045, + "eval_samples_per_second": 26.663, + "eval_steps_per_second": 1.12, + "step": 700 + }, + { + "epoch": 0.88510101010101, + "grad_norm": 2.7135095596313477, + "learning_rate": 3.952291861552854e-06, + "loss": 1.6013, + "step": 701 + }, + { + "epoch": 0.8863636363636364, + "grad_norm": 2.812061071395874, + "learning_rate": 3.949953227315248e-06, + "loss": 1.5518, + "step": 702 + }, + { + "epoch": 0.8876262626262627, + "grad_norm": 3.0532422065734863, + "learning_rate": 3.9476145930776424e-06, + "loss": 1.5579, + "step": 703 + }, + { + "epoch": 0.8888888888888888, + "grad_norm": 2.8306567668914795, + "learning_rate": 3.945275958840038e-06, + "loss": 1.5478, + "step": 704 + }, + { + "epoch": 0.8901515151515151, + "grad_norm": 2.737921953201294, + "learning_rate": 3.942937324602433e-06, + "loss": 1.4958, + "step": 705 + }, + { + "epoch": 0.8914141414141414, + "grad_norm": 2.902211904525757, + "learning_rate": 3.940598690364827e-06, + "loss": 1.5715, + "step": 706 + }, + { + "epoch": 0.8926767676767676, + "grad_norm": 2.725818395614624, + "learning_rate": 3.938260056127222e-06, + "loss": 1.5502, + "step": 707 + }, + { + "epoch": 0.8939393939393939, + "grad_norm": 2.9530506134033203, + "learning_rate": 3.935921421889617e-06, + "loss": 1.564, + "step": 708 + }, + { + "epoch": 0.8952020202020202, + "grad_norm": 2.798231363296509, + "learning_rate": 3.9335827876520115e-06, + "loss": 1.5445, + "step": 709 + }, + { + "epoch": 0.8964646464646465, + "grad_norm": 3.0233471393585205, + "learning_rate": 3.931244153414407e-06, + "loss": 1.6193, + "step": 710 + }, + { + "epoch": 0.8964646464646465, + "eval_loss": 1.742077350616455, + "eval_runtime": 37.0255, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 710 + }, + { + "epoch": 0.8977272727272727, + "grad_norm": 2.84309458732605, + "learning_rate": 3.928905519176801e-06, + "loss": 1.5379, + "step": 711 + }, + { + "epoch": 0.898989898989899, + "grad_norm": 2.7412185668945312, + "learning_rate": 3.926566884939196e-06, + "loss": 1.5837, + "step": 712 + }, + { + "epoch": 0.9002525252525253, + "grad_norm": 3.1094822883605957, + "learning_rate": 3.924228250701591e-06, + "loss": 1.5621, + "step": 713 + }, + { + "epoch": 0.9015151515151515, + "grad_norm": 2.776059627532959, + "learning_rate": 3.921889616463985e-06, + "loss": 1.5714, + "step": 714 + }, + { + "epoch": 0.9027777777777778, + "grad_norm": 2.7650489807128906, + "learning_rate": 3.91955098222638e-06, + "loss": 1.5418, + "step": 715 + }, + { + "epoch": 0.9040404040404041, + "grad_norm": 2.8586974143981934, + "learning_rate": 3.917212347988775e-06, + "loss": 1.5891, + "step": 716 + }, + { + "epoch": 0.9053030303030303, + "grad_norm": 2.74332594871521, + "learning_rate": 3.914873713751169e-06, + "loss": 1.5492, + "step": 717 + }, + { + "epoch": 0.9065656565656566, + "grad_norm": 2.931682586669922, + "learning_rate": 3.912535079513564e-06, + "loss": 1.54, + "step": 718 + }, + { + "epoch": 0.9078282828282829, + "grad_norm": 2.678812026977539, + "learning_rate": 3.910196445275959e-06, + "loss": 1.5677, + "step": 719 + }, + { + "epoch": 0.9090909090909091, + "grad_norm": 2.7282893657684326, + "learning_rate": 3.907857811038354e-06, + "loss": 1.5413, + "step": 720 + }, + { + "epoch": 0.9090909090909091, + "eval_loss": 1.7409753799438477, + "eval_runtime": 37.1883, + "eval_samples_per_second": 26.89, + "eval_steps_per_second": 1.129, + "step": 720 + }, + { + "epoch": 0.9103535353535354, + "grad_norm": 3.073997735977173, + "learning_rate": 3.905519176800749e-06, + "loss": 1.5734, + "step": 721 + }, + { + "epoch": 0.9116161616161617, + "grad_norm": 2.6249935626983643, + "learning_rate": 3.903180542563144e-06, + "loss": 1.5063, + "step": 722 + }, + { + "epoch": 0.9128787878787878, + "grad_norm": 2.7758615016937256, + "learning_rate": 3.9008419083255384e-06, + "loss": 1.5112, + "step": 723 + }, + { + "epoch": 0.9141414141414141, + "grad_norm": 2.8954596519470215, + "learning_rate": 3.898503274087933e-06, + "loss": 1.5508, + "step": 724 + }, + { + "epoch": 0.9154040404040404, + "grad_norm": 2.79073166847229, + "learning_rate": 3.896164639850328e-06, + "loss": 1.5773, + "step": 725 + }, + { + "epoch": 0.9166666666666666, + "grad_norm": 2.687027931213379, + "learning_rate": 3.8938260056127226e-06, + "loss": 1.5243, + "step": 726 + }, + { + "epoch": 0.9179292929292929, + "grad_norm": 2.5704851150512695, + "learning_rate": 3.891487371375117e-06, + "loss": 1.5075, + "step": 727 + }, + { + "epoch": 0.9191919191919192, + "grad_norm": 2.8684048652648926, + "learning_rate": 3.889148737137512e-06, + "loss": 1.6471, + "step": 728 + }, + { + "epoch": 0.9204545454545454, + "grad_norm": 2.8995590209960938, + "learning_rate": 3.886810102899907e-06, + "loss": 1.5429, + "step": 729 + }, + { + "epoch": 0.9217171717171717, + "grad_norm": 2.7575254440307617, + "learning_rate": 3.884471468662301e-06, + "loss": 1.5241, + "step": 730 + }, + { + "epoch": 0.9217171717171717, + "eval_loss": 1.7294232845306396, + "eval_runtime": 37.1941, + "eval_samples_per_second": 26.886, + "eval_steps_per_second": 1.129, + "step": 730 + }, + { + "epoch": 0.922979797979798, + "grad_norm": 2.7057321071624756, + "learning_rate": 3.882132834424696e-06, + "loss": 1.5498, + "step": 731 + }, + { + "epoch": 0.9242424242424242, + "grad_norm": 2.7679436206817627, + "learning_rate": 3.879794200187091e-06, + "loss": 1.563, + "step": 732 + }, + { + "epoch": 0.9255050505050505, + "grad_norm": 2.7258687019348145, + "learning_rate": 3.877455565949485e-06, + "loss": 1.5016, + "step": 733 + }, + { + "epoch": 0.9267676767676768, + "grad_norm": 2.8474221229553223, + "learning_rate": 3.8751169317118804e-06, + "loss": 1.5318, + "step": 734 + }, + { + "epoch": 0.928030303030303, + "grad_norm": 2.7718396186828613, + "learning_rate": 3.872778297474276e-06, + "loss": 1.5919, + "step": 735 + }, + { + "epoch": 0.9292929292929293, + "grad_norm": 2.7785022258758545, + "learning_rate": 3.87043966323667e-06, + "loss": 1.5399, + "step": 736 + }, + { + "epoch": 0.9305555555555556, + "grad_norm": 2.742476463317871, + "learning_rate": 3.868101028999065e-06, + "loss": 1.5428, + "step": 737 + }, + { + "epoch": 0.9318181818181818, + "grad_norm": 2.7050726413726807, + "learning_rate": 3.86576239476146e-06, + "loss": 1.5604, + "step": 738 + }, + { + "epoch": 0.9330808080808081, + "grad_norm": 2.6755242347717285, + "learning_rate": 3.863423760523854e-06, + "loss": 1.5409, + "step": 739 + }, + { + "epoch": 0.9343434343434344, + "grad_norm": 2.6947829723358154, + "learning_rate": 3.8610851262862495e-06, + "loss": 1.5537, + "step": 740 + }, + { + "epoch": 0.9343434343434344, + "eval_loss": 1.7396037578582764, + "eval_runtime": 37.1912, + "eval_samples_per_second": 26.888, + "eval_steps_per_second": 1.129, + "step": 740 + }, + { + "epoch": 0.9356060606060606, + "grad_norm": 2.729816436767578, + "learning_rate": 3.858746492048644e-06, + "loss": 1.5581, + "step": 741 + }, + { + "epoch": 0.9368686868686869, + "grad_norm": 2.6951887607574463, + "learning_rate": 3.856407857811038e-06, + "loss": 1.5929, + "step": 742 + }, + { + "epoch": 0.9381313131313131, + "grad_norm": 2.6601064205169678, + "learning_rate": 3.854069223573434e-06, + "loss": 1.4881, + "step": 743 + }, + { + "epoch": 0.9393939393939394, + "grad_norm": 2.623898983001709, + "learning_rate": 3.851730589335828e-06, + "loss": 1.5499, + "step": 744 + }, + { + "epoch": 0.9406565656565656, + "grad_norm": 2.703871250152588, + "learning_rate": 3.8493919550982224e-06, + "loss": 1.5585, + "step": 745 + }, + { + "epoch": 0.9419191919191919, + "grad_norm": 2.674893379211426, + "learning_rate": 3.847053320860618e-06, + "loss": 1.5586, + "step": 746 + }, + { + "epoch": 0.9431818181818182, + "grad_norm": 2.8353004455566406, + "learning_rate": 3.844714686623012e-06, + "loss": 1.5937, + "step": 747 + }, + { + "epoch": 0.9444444444444444, + "grad_norm": 2.68009614944458, + "learning_rate": 3.842376052385407e-06, + "loss": 1.5192, + "step": 748 + }, + { + "epoch": 0.9457070707070707, + "grad_norm": 2.6847407817840576, + "learning_rate": 3.840037418147802e-06, + "loss": 1.512, + "step": 749 + }, + { + "epoch": 0.946969696969697, + "grad_norm": 2.624290943145752, + "learning_rate": 3.837698783910197e-06, + "loss": 1.5348, + "step": 750 + }, + { + "epoch": 0.946969696969697, + "eval_loss": 1.738332748413086, + "eval_runtime": 37.1782, + "eval_samples_per_second": 26.898, + "eval_steps_per_second": 1.13, + "step": 750 + }, + { + "epoch": 0.9482323232323232, + "grad_norm": 2.7467777729034424, + "learning_rate": 3.8353601496725915e-06, + "loss": 1.5603, + "step": 751 + }, + { + "epoch": 0.9494949494949495, + "grad_norm": 2.851987600326538, + "learning_rate": 3.833021515434987e-06, + "loss": 1.5974, + "step": 752 + }, + { + "epoch": 0.9507575757575758, + "grad_norm": 2.6642227172851562, + "learning_rate": 3.830682881197381e-06, + "loss": 1.5138, + "step": 753 + }, + { + "epoch": 0.952020202020202, + "grad_norm": 2.7351491451263428, + "learning_rate": 3.828344246959776e-06, + "loss": 1.5656, + "step": 754 + }, + { + "epoch": 0.9532828282828283, + "grad_norm": 2.593748092651367, + "learning_rate": 3.826005612722171e-06, + "loss": 1.5395, + "step": 755 + }, + { + "epoch": 0.9545454545454546, + "grad_norm": 2.639180898666382, + "learning_rate": 3.823666978484565e-06, + "loss": 1.5349, + "step": 756 + }, + { + "epoch": 0.9558080808080808, + "grad_norm": 2.805148124694824, + "learning_rate": 3.82132834424696e-06, + "loss": 1.5805, + "step": 757 + }, + { + "epoch": 0.9570707070707071, + "grad_norm": 2.5619425773620605, + "learning_rate": 3.818989710009355e-06, + "loss": 1.5389, + "step": 758 + }, + { + "epoch": 0.9583333333333334, + "grad_norm": 2.7224011421203613, + "learning_rate": 3.816651075771749e-06, + "loss": 1.5316, + "step": 759 + }, + { + "epoch": 0.9595959595959596, + "grad_norm": 2.4667656421661377, + "learning_rate": 3.8143124415341442e-06, + "loss": 1.4942, + "step": 760 + }, + { + "epoch": 0.9595959595959596, + "eval_loss": 1.7205606698989868, + "eval_runtime": 37.1654, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 760 + }, + { + "epoch": 0.9608585858585859, + "grad_norm": 2.456643581390381, + "learning_rate": 3.811973807296539e-06, + "loss": 1.538, + "step": 761 + }, + { + "epoch": 0.9621212121212122, + "grad_norm": 2.696720838546753, + "learning_rate": 3.8096351730589335e-06, + "loss": 1.5462, + "step": 762 + }, + { + "epoch": 0.9633838383838383, + "grad_norm": 2.7314131259918213, + "learning_rate": 3.8072965388213284e-06, + "loss": 1.5931, + "step": 763 + }, + { + "epoch": 0.9646464646464646, + "grad_norm": 2.616532325744629, + "learning_rate": 3.804957904583723e-06, + "loss": 1.5469, + "step": 764 + }, + { + "epoch": 0.9659090909090909, + "grad_norm": 2.7471885681152344, + "learning_rate": 3.8026192703461185e-06, + "loss": 1.5431, + "step": 765 + }, + { + "epoch": 0.9671717171717171, + "grad_norm": 2.7478253841400146, + "learning_rate": 3.8002806361085133e-06, + "loss": 1.5452, + "step": 766 + }, + { + "epoch": 0.9684343434343434, + "grad_norm": 2.543790340423584, + "learning_rate": 3.7979420018709077e-06, + "loss": 1.6091, + "step": 767 + }, + { + "epoch": 0.9696969696969697, + "grad_norm": 2.6169509887695312, + "learning_rate": 3.7956033676333026e-06, + "loss": 1.5557, + "step": 768 + }, + { + "epoch": 0.9709595959595959, + "grad_norm": 2.4768588542938232, + "learning_rate": 3.7932647333956974e-06, + "loss": 1.5247, + "step": 769 + }, + { + "epoch": 0.9722222222222222, + "grad_norm": 2.654648780822754, + "learning_rate": 3.7909260991580923e-06, + "loss": 1.5281, + "step": 770 + }, + { + "epoch": 0.9722222222222222, + "eval_loss": 1.7289395332336426, + "eval_runtime": 37.1374, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 770 + }, + { + "epoch": 0.9734848484848485, + "grad_norm": 2.893073081970215, + "learning_rate": 3.7885874649204867e-06, + "loss": 1.5749, + "step": 771 + }, + { + "epoch": 0.9747474747474747, + "grad_norm": 2.696341037750244, + "learning_rate": 3.7862488306828815e-06, + "loss": 1.5025, + "step": 772 + }, + { + "epoch": 0.976010101010101, + "grad_norm": 2.678222417831421, + "learning_rate": 3.7839101964452764e-06, + "loss": 1.538, + "step": 773 + }, + { + "epoch": 0.9772727272727273, + "grad_norm": 2.7772746086120605, + "learning_rate": 3.7815715622076708e-06, + "loss": 1.5648, + "step": 774 + }, + { + "epoch": 0.9785353535353535, + "grad_norm": 3.076620101928711, + "learning_rate": 3.7792329279700656e-06, + "loss": 1.6114, + "step": 775 + }, + { + "epoch": 0.9797979797979798, + "grad_norm": 2.700056791305542, + "learning_rate": 3.7768942937324605e-06, + "loss": 1.5881, + "step": 776 + }, + { + "epoch": 0.9810606060606061, + "grad_norm": 2.601419687271118, + "learning_rate": 3.7745556594948553e-06, + "loss": 1.5319, + "step": 777 + }, + { + "epoch": 0.9823232323232324, + "grad_norm": 2.6664950847625732, + "learning_rate": 3.7722170252572497e-06, + "loss": 1.5497, + "step": 778 + }, + { + "epoch": 0.9835858585858586, + "grad_norm": 2.7374589443206787, + "learning_rate": 3.7698783910196446e-06, + "loss": 1.5488, + "step": 779 + }, + { + "epoch": 0.9848484848484849, + "grad_norm": 2.797938823699951, + "learning_rate": 3.76753975678204e-06, + "loss": 1.5376, + "step": 780 + }, + { + "epoch": 0.9848484848484849, + "eval_loss": 1.7334709167480469, + "eval_runtime": 37.1372, + "eval_samples_per_second": 26.927, + "eval_steps_per_second": 1.131, + "step": 780 + }, + { + "epoch": 0.9861111111111112, + "grad_norm": 2.881591796875, + "learning_rate": 3.7652011225444347e-06, + "loss": 1.5405, + "step": 781 + }, + { + "epoch": 0.9873737373737373, + "grad_norm": 2.6965644359588623, + "learning_rate": 3.7628624883068295e-06, + "loss": 1.5066, + "step": 782 + }, + { + "epoch": 0.9886363636363636, + "grad_norm": 2.7286899089813232, + "learning_rate": 3.760523854069224e-06, + "loss": 1.5277, + "step": 783 + }, + { + "epoch": 0.98989898989899, + "grad_norm": 2.795644998550415, + "learning_rate": 3.7581852198316188e-06, + "loss": 1.5335, + "step": 784 + }, + { + "epoch": 0.9911616161616161, + "grad_norm": 2.8160440921783447, + "learning_rate": 3.7558465855940136e-06, + "loss": 1.585, + "step": 785 + }, + { + "epoch": 0.9924242424242424, + "grad_norm": 2.7984936237335205, + "learning_rate": 3.753507951356408e-06, + "loss": 1.5569, + "step": 786 + }, + { + "epoch": 0.9936868686868687, + "grad_norm": 2.7596278190612793, + "learning_rate": 3.751169317118803e-06, + "loss": 1.4838, + "step": 787 + }, + { + "epoch": 0.9949494949494949, + "grad_norm": 2.666383743286133, + "learning_rate": 3.7488306828811977e-06, + "loss": 1.5656, + "step": 788 + }, + { + "epoch": 0.9962121212121212, + "grad_norm": 2.8139331340789795, + "learning_rate": 3.7464920486435926e-06, + "loss": 1.5753, + "step": 789 + }, + { + "epoch": 0.9974747474747475, + "grad_norm": 2.7714269161224365, + "learning_rate": 3.744153414405987e-06, + "loss": 1.5187, + "step": 790 + }, + { + "epoch": 0.9974747474747475, + "eval_loss": 1.7257990837097168, + "eval_runtime": 37.1065, + "eval_samples_per_second": 26.949, + "eval_steps_per_second": 1.132, + "step": 790 + }, + { + "epoch": 0.9987373737373737, + "grad_norm": 2.697606325149536, + "learning_rate": 3.741814780168382e-06, + "loss": 1.56, + "step": 791 + }, + { + "epoch": 1.0, + "grad_norm": 2.7685229778289795, + "learning_rate": 3.7394761459307767e-06, + "loss": 1.5705, + "step": 792 + }, + { + "epoch": 1.0012626262626263, + "grad_norm": 2.778200626373291, + "learning_rate": 3.737137511693171e-06, + "loss": 1.5288, + "step": 793 + }, + { + "epoch": 1.0025252525252526, + "grad_norm": 2.7211337089538574, + "learning_rate": 3.734798877455566e-06, + "loss": 1.5348, + "step": 794 + }, + { + "epoch": 1.003787878787879, + "grad_norm": 2.690739154815674, + "learning_rate": 3.7324602432179608e-06, + "loss": 1.4734, + "step": 795 + }, + { + "epoch": 1.005050505050505, + "grad_norm": 2.6809420585632324, + "learning_rate": 3.730121608980356e-06, + "loss": 1.4852, + "step": 796 + }, + { + "epoch": 1.0063131313131313, + "grad_norm": 2.8550117015838623, + "learning_rate": 3.727782974742751e-06, + "loss": 1.4562, + "step": 797 + }, + { + "epoch": 1.0075757575757576, + "grad_norm": 2.713935613632202, + "learning_rate": 3.7254443405051453e-06, + "loss": 1.5005, + "step": 798 + }, + { + "epoch": 1.0088383838383839, + "grad_norm": 2.668363332748413, + "learning_rate": 3.72310570626754e-06, + "loss": 1.485, + "step": 799 + }, + { + "epoch": 1.0101010101010102, + "grad_norm": 2.840769052505493, + "learning_rate": 3.720767072029935e-06, + "loss": 1.4531, + "step": 800 + }, + { + "epoch": 1.0101010101010102, + "eval_loss": 1.725937008857727, + "eval_runtime": 37.2471, + "eval_samples_per_second": 26.848, + "eval_steps_per_second": 1.128, + "step": 800 + }, + { + "epoch": 1.0113636363636365, + "grad_norm": 2.7345798015594482, + "learning_rate": 3.71842843779233e-06, + "loss": 1.5091, + "step": 801 + }, + { + "epoch": 1.0126262626262625, + "grad_norm": 2.9052624702453613, + "learning_rate": 3.7160898035547243e-06, + "loss": 1.5039, + "step": 802 + }, + { + "epoch": 1.0138888888888888, + "grad_norm": 2.893073558807373, + "learning_rate": 3.713751169317119e-06, + "loss": 1.4613, + "step": 803 + }, + { + "epoch": 1.0151515151515151, + "grad_norm": 2.8505427837371826, + "learning_rate": 3.711412535079514e-06, + "loss": 1.4932, + "step": 804 + }, + { + "epoch": 1.0164141414141414, + "grad_norm": 2.9065170288085938, + "learning_rate": 3.7090739008419084e-06, + "loss": 1.5082, + "step": 805 + }, + { + "epoch": 1.0176767676767677, + "grad_norm": 2.888068675994873, + "learning_rate": 3.706735266604303e-06, + "loss": 1.5675, + "step": 806 + }, + { + "epoch": 1.018939393939394, + "grad_norm": 2.7325260639190674, + "learning_rate": 3.704396632366698e-06, + "loss": 1.4649, + "step": 807 + }, + { + "epoch": 1.02020202020202, + "grad_norm": 2.8861136436462402, + "learning_rate": 3.702057998129093e-06, + "loss": 1.4517, + "step": 808 + }, + { + "epoch": 1.0214646464646464, + "grad_norm": 2.8837692737579346, + "learning_rate": 3.6997193638914873e-06, + "loss": 1.5198, + "step": 809 + }, + { + "epoch": 1.0227272727272727, + "grad_norm": 2.790330410003662, + "learning_rate": 3.697380729653882e-06, + "loss": 1.5066, + "step": 810 + }, + { + "epoch": 1.0227272727272727, + "eval_loss": 1.7269084453582764, + "eval_runtime": 36.89, + "eval_samples_per_second": 27.108, + "eval_steps_per_second": 1.139, + "step": 810 + }, + { + "epoch": 1.023989898989899, + "grad_norm": 3.03643536567688, + "learning_rate": 3.6950420954162774e-06, + "loss": 1.4928, + "step": 811 + }, + { + "epoch": 1.0252525252525253, + "grad_norm": 2.841055154800415, + "learning_rate": 3.6927034611786723e-06, + "loss": 1.5273, + "step": 812 + }, + { + "epoch": 1.0265151515151516, + "grad_norm": 2.84134578704834, + "learning_rate": 3.690364826941067e-06, + "loss": 1.5443, + "step": 813 + }, + { + "epoch": 1.0277777777777777, + "grad_norm": 2.837758779525757, + "learning_rate": 3.6880261927034615e-06, + "loss": 1.5586, + "step": 814 + }, + { + "epoch": 1.029040404040404, + "grad_norm": 2.731412410736084, + "learning_rate": 3.6856875584658564e-06, + "loss": 1.4465, + "step": 815 + }, + { + "epoch": 1.0303030303030303, + "grad_norm": 2.8979599475860596, + "learning_rate": 3.683348924228251e-06, + "loss": 1.5663, + "step": 816 + }, + { + "epoch": 1.0315656565656566, + "grad_norm": 2.7740373611450195, + "learning_rate": 3.6810102899906456e-06, + "loss": 1.4578, + "step": 817 + }, + { + "epoch": 1.0328282828282829, + "grad_norm": 2.8984005451202393, + "learning_rate": 3.6786716557530405e-06, + "loss": 1.493, + "step": 818 + }, + { + "epoch": 1.0340909090909092, + "grad_norm": 2.880680561065674, + "learning_rate": 3.6763330215154353e-06, + "loss": 1.4858, + "step": 819 + }, + { + "epoch": 1.0353535353535352, + "grad_norm": 3.0220961570739746, + "learning_rate": 3.67399438727783e-06, + "loss": 1.5009, + "step": 820 + }, + { + "epoch": 1.0353535353535352, + "eval_loss": 1.7245674133300781, + "eval_runtime": 37.0235, + "eval_samples_per_second": 27.01, + "eval_steps_per_second": 1.134, + "step": 820 + }, + { + "epoch": 1.0366161616161615, + "grad_norm": 2.7905962467193604, + "learning_rate": 3.6716557530402246e-06, + "loss": 1.4728, + "step": 821 + }, + { + "epoch": 1.0378787878787878, + "grad_norm": 2.856935501098633, + "learning_rate": 3.6693171188026194e-06, + "loss": 1.521, + "step": 822 + }, + { + "epoch": 1.0391414141414141, + "grad_norm": 2.956584930419922, + "learning_rate": 3.6669784845650143e-06, + "loss": 1.4725, + "step": 823 + }, + { + "epoch": 1.0404040404040404, + "grad_norm": 3.130960702896118, + "learning_rate": 3.6646398503274087e-06, + "loss": 1.6127, + "step": 824 + }, + { + "epoch": 1.0416666666666667, + "grad_norm": 2.8855528831481934, + "learning_rate": 3.6623012160898035e-06, + "loss": 1.5078, + "step": 825 + }, + { + "epoch": 1.0429292929292928, + "grad_norm": 2.8030972480773926, + "learning_rate": 3.659962581852199e-06, + "loss": 1.4635, + "step": 826 + }, + { + "epoch": 1.0441919191919191, + "grad_norm": 3.061593770980835, + "learning_rate": 3.6576239476145936e-06, + "loss": 1.4944, + "step": 827 + }, + { + "epoch": 1.0454545454545454, + "grad_norm": 2.9983572959899902, + "learning_rate": 3.6552853133769885e-06, + "loss": 1.505, + "step": 828 + }, + { + "epoch": 1.0467171717171717, + "grad_norm": 2.7681095600128174, + "learning_rate": 3.652946679139383e-06, + "loss": 1.4774, + "step": 829 + }, + { + "epoch": 1.047979797979798, + "grad_norm": 3.0530946254730225, + "learning_rate": 3.6506080449017777e-06, + "loss": 1.4676, + "step": 830 + }, + { + "epoch": 1.047979797979798, + "eval_loss": 1.7242811918258667, + "eval_runtime": 37.0752, + "eval_samples_per_second": 26.972, + "eval_steps_per_second": 1.133, + "step": 830 + }, + { + "epoch": 1.0492424242424243, + "grad_norm": 2.8383126258850098, + "learning_rate": 3.6482694106641726e-06, + "loss": 1.486, + "step": 831 + }, + { + "epoch": 1.0505050505050506, + "grad_norm": 3.095928907394409, + "learning_rate": 3.6459307764265674e-06, + "loss": 1.5331, + "step": 832 + }, + { + "epoch": 1.0517676767676767, + "grad_norm": 2.9225456714630127, + "learning_rate": 3.643592142188962e-06, + "loss": 1.556, + "step": 833 + }, + { + "epoch": 1.053030303030303, + "grad_norm": 2.9037210941314697, + "learning_rate": 3.6412535079513567e-06, + "loss": 1.4834, + "step": 834 + }, + { + "epoch": 1.0542929292929293, + "grad_norm": 2.915022373199463, + "learning_rate": 3.6389148737137515e-06, + "loss": 1.5346, + "step": 835 + }, + { + "epoch": 1.0555555555555556, + "grad_norm": 3.250411033630371, + "learning_rate": 3.636576239476146e-06, + "loss": 1.4885, + "step": 836 + }, + { + "epoch": 1.0568181818181819, + "grad_norm": 2.759061336517334, + "learning_rate": 3.634237605238541e-06, + "loss": 1.456, + "step": 837 + }, + { + "epoch": 1.0580808080808082, + "grad_norm": 2.955981731414795, + "learning_rate": 3.6318989710009356e-06, + "loss": 1.4889, + "step": 838 + }, + { + "epoch": 1.0593434343434343, + "grad_norm": 3.243338108062744, + "learning_rate": 3.6295603367633305e-06, + "loss": 1.5207, + "step": 839 + }, + { + "epoch": 1.0606060606060606, + "grad_norm": 3.075840950012207, + "learning_rate": 3.627221702525725e-06, + "loss": 1.4906, + "step": 840 + }, + { + "epoch": 1.0606060606060606, + "eval_loss": 1.7290290594100952, + "eval_runtime": 37.097, + "eval_samples_per_second": 26.956, + "eval_steps_per_second": 1.132, + "step": 840 + }, + { + "epoch": 1.0618686868686869, + "grad_norm": 3.149125099182129, + "learning_rate": 3.62488306828812e-06, + "loss": 1.4755, + "step": 841 + }, + { + "epoch": 1.0631313131313131, + "grad_norm": 3.099318504333496, + "learning_rate": 3.622544434050515e-06, + "loss": 1.534, + "step": 842 + }, + { + "epoch": 1.0643939393939394, + "grad_norm": 2.8942148685455322, + "learning_rate": 3.62020579981291e-06, + "loss": 1.4844, + "step": 843 + }, + { + "epoch": 1.0656565656565657, + "grad_norm": 3.142664670944214, + "learning_rate": 3.6178671655753047e-06, + "loss": 1.5179, + "step": 844 + }, + { + "epoch": 1.0669191919191918, + "grad_norm": 3.106552839279175, + "learning_rate": 3.615528531337699e-06, + "loss": 1.4667, + "step": 845 + }, + { + "epoch": 1.0681818181818181, + "grad_norm": 2.8507936000823975, + "learning_rate": 3.613189897100094e-06, + "loss": 1.4463, + "step": 846 + }, + { + "epoch": 1.0694444444444444, + "grad_norm": 3.2198901176452637, + "learning_rate": 3.610851262862489e-06, + "loss": 1.5146, + "step": 847 + }, + { + "epoch": 1.0707070707070707, + "grad_norm": 3.146388292312622, + "learning_rate": 3.6085126286248832e-06, + "loss": 1.4927, + "step": 848 + }, + { + "epoch": 1.071969696969697, + "grad_norm": 2.8304760456085205, + "learning_rate": 3.606173994387278e-06, + "loss": 1.4971, + "step": 849 + }, + { + "epoch": 1.0732323232323233, + "grad_norm": 3.1714529991149902, + "learning_rate": 3.603835360149673e-06, + "loss": 1.5418, + "step": 850 + }, + { + "epoch": 1.0732323232323233, + "eval_loss": 1.725188970565796, + "eval_runtime": 37.117, + "eval_samples_per_second": 26.942, + "eval_steps_per_second": 1.132, + "step": 850 + }, + { + "epoch": 1.0744949494949494, + "grad_norm": 3.06162428855896, + "learning_rate": 3.6014967259120677e-06, + "loss": 1.4959, + "step": 851 + }, + { + "epoch": 1.0757575757575757, + "grad_norm": 2.812206268310547, + "learning_rate": 3.599158091674462e-06, + "loss": 1.5249, + "step": 852 + }, + { + "epoch": 1.077020202020202, + "grad_norm": 3.236210584640503, + "learning_rate": 3.596819457436857e-06, + "loss": 1.5166, + "step": 853 + }, + { + "epoch": 1.0782828282828283, + "grad_norm": 2.8397765159606934, + "learning_rate": 3.594480823199252e-06, + "loss": 1.5043, + "step": 854 + }, + { + "epoch": 1.0795454545454546, + "grad_norm": 2.9215919971466064, + "learning_rate": 3.5921421889616463e-06, + "loss": 1.5235, + "step": 855 + }, + { + "epoch": 1.0808080808080809, + "grad_norm": 2.8219058513641357, + "learning_rate": 3.589803554724041e-06, + "loss": 1.5091, + "step": 856 + }, + { + "epoch": 1.0820707070707072, + "grad_norm": 2.9344675540924072, + "learning_rate": 3.5874649204864364e-06, + "loss": 1.5355, + "step": 857 + }, + { + "epoch": 1.0833333333333333, + "grad_norm": 2.949094533920288, + "learning_rate": 3.5851262862488312e-06, + "loss": 1.5273, + "step": 858 + }, + { + "epoch": 1.0845959595959596, + "grad_norm": 2.9801411628723145, + "learning_rate": 3.582787652011226e-06, + "loss": 1.5078, + "step": 859 + }, + { + "epoch": 1.0858585858585859, + "grad_norm": 2.802182912826538, + "learning_rate": 3.5804490177736205e-06, + "loss": 1.5015, + "step": 860 + }, + { + "epoch": 1.0858585858585859, + "eval_loss": 1.7209502458572388, + "eval_runtime": 37.1238, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 860 + }, + { + "epoch": 1.0871212121212122, + "grad_norm": 2.857457160949707, + "learning_rate": 3.5781103835360153e-06, + "loss": 1.5057, + "step": 861 + }, + { + "epoch": 1.0883838383838385, + "grad_norm": 2.794478178024292, + "learning_rate": 3.57577174929841e-06, + "loss": 1.4594, + "step": 862 + }, + { + "epoch": 1.0896464646464645, + "grad_norm": 3.0028598308563232, + "learning_rate": 3.573433115060805e-06, + "loss": 1.4666, + "step": 863 + }, + { + "epoch": 1.0909090909090908, + "grad_norm": 2.7701406478881836, + "learning_rate": 3.5710944808231994e-06, + "loss": 1.4596, + "step": 864 + }, + { + "epoch": 1.0921717171717171, + "grad_norm": 2.9705960750579834, + "learning_rate": 3.5687558465855943e-06, + "loss": 1.5092, + "step": 865 + }, + { + "epoch": 1.0934343434343434, + "grad_norm": 2.9820821285247803, + "learning_rate": 3.566417212347989e-06, + "loss": 1.5246, + "step": 866 + }, + { + "epoch": 1.0946969696969697, + "grad_norm": 2.988356113433838, + "learning_rate": 3.5640785781103835e-06, + "loss": 1.5035, + "step": 867 + }, + { + "epoch": 1.095959595959596, + "grad_norm": 2.901599884033203, + "learning_rate": 3.5617399438727784e-06, + "loss": 1.5097, + "step": 868 + }, + { + "epoch": 1.0972222222222223, + "grad_norm": 2.9128799438476562, + "learning_rate": 3.5594013096351732e-06, + "loss": 1.5424, + "step": 869 + }, + { + "epoch": 1.0984848484848484, + "grad_norm": 2.719289779663086, + "learning_rate": 3.557062675397568e-06, + "loss": 1.4313, + "step": 870 + }, + { + "epoch": 1.0984848484848484, + "eval_loss": 1.7159698009490967, + "eval_runtime": 37.122, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 870 + }, + { + "epoch": 1.0997474747474747, + "grad_norm": 3.0080976486206055, + "learning_rate": 3.5547240411599625e-06, + "loss": 1.4483, + "step": 871 + }, + { + "epoch": 1.101010101010101, + "grad_norm": 2.856523036956787, + "learning_rate": 3.5523854069223578e-06, + "loss": 1.5463, + "step": 872 + }, + { + "epoch": 1.1022727272727273, + "grad_norm": 2.8585214614868164, + "learning_rate": 3.5500467726847526e-06, + "loss": 1.5391, + "step": 873 + }, + { + "epoch": 1.1035353535353536, + "grad_norm": 2.8440098762512207, + "learning_rate": 3.5477081384471474e-06, + "loss": 1.499, + "step": 874 + }, + { + "epoch": 1.10479797979798, + "grad_norm": 2.8114542961120605, + "learning_rate": 3.545369504209542e-06, + "loss": 1.4691, + "step": 875 + }, + { + "epoch": 1.106060606060606, + "grad_norm": 2.980273962020874, + "learning_rate": 3.5430308699719367e-06, + "loss": 1.4541, + "step": 876 + }, + { + "epoch": 1.1073232323232323, + "grad_norm": 3.0053417682647705, + "learning_rate": 3.5406922357343315e-06, + "loss": 1.4895, + "step": 877 + }, + { + "epoch": 1.1085858585858586, + "grad_norm": 2.924046754837036, + "learning_rate": 3.5383536014967264e-06, + "loss": 1.5186, + "step": 878 + }, + { + "epoch": 1.1098484848484849, + "grad_norm": 2.969461679458618, + "learning_rate": 3.536014967259121e-06, + "loss": 1.5213, + "step": 879 + }, + { + "epoch": 1.1111111111111112, + "grad_norm": 2.8936445713043213, + "learning_rate": 3.5336763330215156e-06, + "loss": 1.5585, + "step": 880 + }, + { + "epoch": 1.1111111111111112, + "eval_loss": 1.7150688171386719, + "eval_runtime": 37.1105, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 880 + }, + { + "epoch": 1.1123737373737375, + "grad_norm": 2.992034435272217, + "learning_rate": 3.5313376987839105e-06, + "loss": 1.5264, + "step": 881 + }, + { + "epoch": 1.1136363636363635, + "grad_norm": 3.0453898906707764, + "learning_rate": 3.5289990645463053e-06, + "loss": 1.5482, + "step": 882 + }, + { + "epoch": 1.1148989898989898, + "grad_norm": 2.7480897903442383, + "learning_rate": 3.5266604303086998e-06, + "loss": 1.4851, + "step": 883 + }, + { + "epoch": 1.1161616161616161, + "grad_norm": 2.9408013820648193, + "learning_rate": 3.5243217960710946e-06, + "loss": 1.536, + "step": 884 + }, + { + "epoch": 1.1174242424242424, + "grad_norm": 2.854438543319702, + "learning_rate": 3.5219831618334894e-06, + "loss": 1.5814, + "step": 885 + }, + { + "epoch": 1.1186868686868687, + "grad_norm": 2.9060397148132324, + "learning_rate": 3.519644527595884e-06, + "loss": 1.5019, + "step": 886 + }, + { + "epoch": 1.119949494949495, + "grad_norm": 2.8741681575775146, + "learning_rate": 3.517305893358279e-06, + "loss": 1.4713, + "step": 887 + }, + { + "epoch": 1.121212121212121, + "grad_norm": 2.972510576248169, + "learning_rate": 3.514967259120674e-06, + "loss": 1.5104, + "step": 888 + }, + { + "epoch": 1.1224747474747474, + "grad_norm": 2.8113558292388916, + "learning_rate": 3.512628624883069e-06, + "loss": 1.4973, + "step": 889 + }, + { + "epoch": 1.1237373737373737, + "grad_norm": 2.9687416553497314, + "learning_rate": 3.5102899906454637e-06, + "loss": 1.4774, + "step": 890 + }, + { + "epoch": 1.1237373737373737, + "eval_loss": 1.7169743776321411, + "eval_runtime": 37.0911, + "eval_samples_per_second": 26.961, + "eval_steps_per_second": 1.132, + "step": 890 + }, + { + "epoch": 1.125, + "grad_norm": 2.911409854888916, + "learning_rate": 3.507951356407858e-06, + "loss": 1.4536, + "step": 891 + }, + { + "epoch": 1.1262626262626263, + "grad_norm": 2.8209526538848877, + "learning_rate": 3.505612722170253e-06, + "loss": 1.5357, + "step": 892 + }, + { + "epoch": 1.1275252525252526, + "grad_norm": 2.811971426010132, + "learning_rate": 3.5032740879326478e-06, + "loss": 1.4889, + "step": 893 + }, + { + "epoch": 1.128787878787879, + "grad_norm": 2.969878673553467, + "learning_rate": 3.5009354536950426e-06, + "loss": 1.4758, + "step": 894 + }, + { + "epoch": 1.130050505050505, + "grad_norm": 2.858635425567627, + "learning_rate": 3.498596819457437e-06, + "loss": 1.5355, + "step": 895 + }, + { + "epoch": 1.1313131313131313, + "grad_norm": 3.018164873123169, + "learning_rate": 3.496258185219832e-06, + "loss": 1.5421, + "step": 896 + }, + { + "epoch": 1.1325757575757576, + "grad_norm": 2.846277952194214, + "learning_rate": 3.4939195509822267e-06, + "loss": 1.4936, + "step": 897 + }, + { + "epoch": 1.1338383838383839, + "grad_norm": 2.965245246887207, + "learning_rate": 3.491580916744621e-06, + "loss": 1.4903, + "step": 898 + }, + { + "epoch": 1.1351010101010102, + "grad_norm": 2.8500895500183105, + "learning_rate": 3.489242282507016e-06, + "loss": 1.5152, + "step": 899 + }, + { + "epoch": 1.1363636363636362, + "grad_norm": 2.8242156505584717, + "learning_rate": 3.486903648269411e-06, + "loss": 1.5531, + "step": 900 + }, + { + "epoch": 1.1363636363636362, + "eval_loss": 1.7180876731872559, + "eval_runtime": 37.4145, + "eval_samples_per_second": 26.728, + "eval_steps_per_second": 1.123, + "step": 900 + } + ], + "logging_steps": 1, + "max_steps": 2376, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 2.805578468425728e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_hop_words4_10M_seed0/runs/checkpoint-900/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/2d73a6863086ff9d491c28e49df9fb697cd92c2b.lock b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/2d73a6863086ff9d491c28e49df9fb697cd92c2b.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/4719a04514ec2f060240711b7c33ab21187cac730ecaba3040b7a0fd95a9cefb.lock b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/4719a04514ec2f060240711b7c33ab21187cac730ecaba3040b7a0fd95a9cefb.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/47d4a5aa69cdef91a53b77f5c5583647a578ca0e.lock b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/47d4a5aa69cdef91a53b77f5c5583647a578ca0e.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/584d8d3e3f82f7964955174dfe5e3b1cf117a9d859f022cfdf7fcb884856e002.lock b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/584d8d3e3f82f7964955174dfe5e3b1cf117a9d859f022cfdf7fcb884856e002.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58.lock b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58.lock new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/adapter_config.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/adapter_config.json new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/2d73a6863086ff9d491c28e49df9fb697cd92c2b b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/2d73a6863086ff9d491c28e49df9fb697cd92c2b new file mode 100644 index 0000000000000000000000000000000000000000..2d73a6863086ff9d491c28e49df9fb697cd92c2b --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/2d73a6863086ff9d491c28e49df9fb697cd92c2b @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "eos_token_id": 128001, + "transformers_version": "4.45.0.dev0", + "do_sample": true, + "temperature": 0.6, + "top_p": 0.9 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/47d4a5aa69cdef91a53b77f5c5583647a578ca0e b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/47d4a5aa69cdef91a53b77f5c5583647a578ca0e new file mode 100644 index 0000000000000000000000000000000000000000..47d4a5aa69cdef91a53b77f5c5583647a578ca0e --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/47d4a5aa69cdef91a53b77f5c5583647a578ca0e @@ -0,0 +1,35 @@ +{ + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "bfloat16", + "transformers_version": "4.45.0.dev0", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 new file mode 100644 index 0000000000000000000000000000000000000000..d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 @@ -0,0 +1,261 @@ +{ + "metadata": { + "total_size": 6425499648 + }, + "weight_map": { + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/refs/main b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/refs/main new file mode 100644 index 0000000000000000000000000000000000000000..940c46e96db631910d2cc5186ab76cd83e32b124 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/refs/main @@ -0,0 +1 @@ +13afe5124825b4f3751f836b40dafda64c1ed062 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/config.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/config.json new file mode 100644 index 0000000000000000000000000000000000000000..47d4a5aa69cdef91a53b77f5c5583647a578ca0e --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/config.json @@ -0,0 +1,35 @@ +{ + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "bfloat16", + "transformers_version": "4.45.0.dev0", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..2d73a6863086ff9d491c28e49df9fb697cd92c2b --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "eos_token_id": 128001, + "transformers_version": "4.45.0.dev0", + "do_sample": true, + "temperature": 0.6, + "top_p": 0.9 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors.index.json @@ -0,0 +1,261 @@ +{ + "metadata": { + "total_size": 6425499648 + }, + "weight_map": { + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/config.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/latest b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/latest new file mode 100644 index 0000000000000000000000000000000000000000..adbdf3f70e1e0db88a47bff3edcb0592ed0a2e61 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/latest @@ -0,0 +1 @@ +global_step2200 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..0b782d4602b361f7577abb52b710a8c0e088588d --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/tokenizer_config.json @@ -0,0 +1,2078 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128257": { + "content": "🅁", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3989327e79e9d40b04f4713eab2770710e88b0a3 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/trainer_state.json @@ -0,0 +1,17193 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.2598870056497176, + "eval_steps": 10, + "global_step": 2200, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001027221366204417, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7525, + "step": 1 + }, + { + "epoch": 0.002054442732408834, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7605, + "step": 2 + }, + { + "epoch": 0.0030816640986132513, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8008, + "step": 3 + }, + { + "epoch": 0.004108885464817668, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.809, + "step": 4 + }, + { + "epoch": 0.005136106831022085, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7841, + "step": 5 + }, + { + "epoch": 0.0061633281972265025, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7924, + "step": 6 + }, + { + "epoch": 0.007190549563430919, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7732, + "step": 7 + }, + { + "epoch": 0.008217770929635337, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7953, + "step": 8 + }, + { + "epoch": 0.009244992295839754, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7747, + "step": 9 + }, + { + "epoch": 0.01027221366204417, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7909, + "step": 10 + }, + { + "epoch": 0.01027221366204417, + "eval_loss": 2.811410665512085, + "eval_runtime": 37.9307, + "eval_samples_per_second": 26.364, + "eval_steps_per_second": 1.107, + "step": 10 + }, + { + "epoch": 0.011299435028248588, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7592, + "step": 11 + }, + { + "epoch": 0.012326656394453005, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7394, + "step": 12 + }, + { + "epoch": 0.01335387776065742, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8111, + "step": 13 + }, + { + "epoch": 0.014381099126861838, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7786, + "step": 14 + }, + { + "epoch": 0.015408320493066256, + "grad_norm": 5.814355373382568, + "learning_rate": 1.7123287671232876e-08, + "loss": 2.7711, + "step": 15 + }, + { + "epoch": 0.016435541859270673, + "grad_norm": 6.2984514236450195, + "learning_rate": 3.424657534246575e-08, + "loss": 2.7517, + "step": 16 + }, + { + "epoch": 0.01746276322547509, + "grad_norm": 6.235473155975342, + "learning_rate": 5.136986301369863e-08, + "loss": 2.7191, + "step": 17 + }, + { + "epoch": 0.01848998459167951, + "grad_norm": 6.235473155975342, + "learning_rate": 5.136986301369863e-08, + "loss": 2.806, + "step": 18 + }, + { + "epoch": 0.019517205957883924, + "grad_norm": 6.621068000793457, + "learning_rate": 6.84931506849315e-08, + "loss": 2.7192, + "step": 19 + }, + { + "epoch": 0.02054442732408834, + "grad_norm": 6.781496524810791, + "learning_rate": 8.561643835616439e-08, + "loss": 2.7862, + "step": 20 + }, + { + "epoch": 0.02054442732408834, + "eval_loss": 2.8109023571014404, + "eval_runtime": 37.527, + "eval_samples_per_second": 26.647, + "eval_steps_per_second": 1.119, + "step": 20 + }, + { + "epoch": 0.02157164869029276, + "grad_norm": 6.336451530456543, + "learning_rate": 1.0273972602739726e-07, + "loss": 2.7997, + "step": 21 + }, + { + "epoch": 0.022598870056497175, + "grad_norm": 6.184426307678223, + "learning_rate": 1.1986301369863014e-07, + "loss": 2.7938, + "step": 22 + }, + { + "epoch": 0.02362609142270159, + "grad_norm": 6.443874359130859, + "learning_rate": 1.36986301369863e-07, + "loss": 2.7485, + "step": 23 + }, + { + "epoch": 0.02465331278890601, + "grad_norm": 6.17667293548584, + "learning_rate": 1.541095890410959e-07, + "loss": 2.7779, + "step": 24 + }, + { + "epoch": 0.025680534155110426, + "grad_norm": 5.850281715393066, + "learning_rate": 1.7123287671232878e-07, + "loss": 2.7598, + "step": 25 + }, + { + "epoch": 0.02670775552131484, + "grad_norm": 5.788112640380859, + "learning_rate": 1.8835616438356165e-07, + "loss": 2.7837, + "step": 26 + }, + { + "epoch": 0.02773497688751926, + "grad_norm": 5.17655086517334, + "learning_rate": 2.0547945205479452e-07, + "loss": 2.7513, + "step": 27 + }, + { + "epoch": 0.028762198253723677, + "grad_norm": 4.698268890380859, + "learning_rate": 2.226027397260274e-07, + "loss": 2.7714, + "step": 28 + }, + { + "epoch": 0.029789419619928096, + "grad_norm": 4.562674522399902, + "learning_rate": 2.397260273972603e-07, + "loss": 2.7759, + "step": 29 + }, + { + "epoch": 0.030816640986132512, + "grad_norm": 4.138142108917236, + "learning_rate": 2.568493150684932e-07, + "loss": 2.7656, + "step": 30 + }, + { + "epoch": 0.030816640986132512, + "eval_loss": 2.784187078475952, + "eval_runtime": 36.9691, + "eval_samples_per_second": 27.05, + "eval_steps_per_second": 1.136, + "step": 30 + }, + { + "epoch": 0.03184386235233693, + "grad_norm": 4.39779806137085, + "learning_rate": 2.73972602739726e-07, + "loss": 2.7377, + "step": 31 + }, + { + "epoch": 0.03287108371854135, + "grad_norm": 3.9160122871398926, + "learning_rate": 2.910958904109589e-07, + "loss": 2.7753, + "step": 32 + }, + { + "epoch": 0.03389830508474576, + "grad_norm": 4.828811168670654, + "learning_rate": 3.082191780821918e-07, + "loss": 2.7311, + "step": 33 + }, + { + "epoch": 0.03492552645095018, + "grad_norm": 4.6300435066223145, + "learning_rate": 3.2534246575342466e-07, + "loss": 2.7287, + "step": 34 + }, + { + "epoch": 0.035952747817154594, + "grad_norm": 4.47259521484375, + "learning_rate": 3.4246575342465755e-07, + "loss": 2.7117, + "step": 35 + }, + { + "epoch": 0.03697996918335902, + "grad_norm": 4.557999610900879, + "learning_rate": 3.595890410958904e-07, + "loss": 2.6291, + "step": 36 + }, + { + "epoch": 0.03800719054956343, + "grad_norm": 4.3974761962890625, + "learning_rate": 3.767123287671233e-07, + "loss": 2.7421, + "step": 37 + }, + { + "epoch": 0.03903441191576785, + "grad_norm": 3.6151702404022217, + "learning_rate": 3.938356164383562e-07, + "loss": 2.6744, + "step": 38 + }, + { + "epoch": 0.040061633281972264, + "grad_norm": 3.165523052215576, + "learning_rate": 4.1095890410958903e-07, + "loss": 2.6694, + "step": 39 + }, + { + "epoch": 0.04108885464817668, + "grad_norm": 3.2827563285827637, + "learning_rate": 4.2808219178082193e-07, + "loss": 2.7177, + "step": 40 + }, + { + "epoch": 0.04108885464817668, + "eval_loss": 2.705791711807251, + "eval_runtime": 36.995, + "eval_samples_per_second": 27.031, + "eval_steps_per_second": 1.135, + "step": 40 + }, + { + "epoch": 0.042116076014381096, + "grad_norm": 3.241729736328125, + "learning_rate": 4.452054794520548e-07, + "loss": 2.6703, + "step": 41 + }, + { + "epoch": 0.04314329738058552, + "grad_norm": 3.401049852371216, + "learning_rate": 4.6232876712328767e-07, + "loss": 2.6036, + "step": 42 + }, + { + "epoch": 0.044170518746789934, + "grad_norm": 3.2498552799224854, + "learning_rate": 4.794520547945206e-07, + "loss": 2.6666, + "step": 43 + }, + { + "epoch": 0.04519774011299435, + "grad_norm": 3.1376283168792725, + "learning_rate": 4.965753424657534e-07, + "loss": 2.6331, + "step": 44 + }, + { + "epoch": 0.046224961479198766, + "grad_norm": 2.885286569595337, + "learning_rate": 5.136986301369864e-07, + "loss": 2.6609, + "step": 45 + }, + { + "epoch": 0.04725218284540318, + "grad_norm": 2.636420249938965, + "learning_rate": 5.308219178082192e-07, + "loss": 2.6462, + "step": 46 + }, + { + "epoch": 0.048279404211607604, + "grad_norm": 2.5609490871429443, + "learning_rate": 5.47945205479452e-07, + "loss": 2.6821, + "step": 47 + }, + { + "epoch": 0.04930662557781202, + "grad_norm": 2.7683653831481934, + "learning_rate": 5.65068493150685e-07, + "loss": 2.61, + "step": 48 + }, + { + "epoch": 0.050333846944016436, + "grad_norm": 3.3999135494232178, + "learning_rate": 5.821917808219178e-07, + "loss": 2.6166, + "step": 49 + }, + { + "epoch": 0.05136106831022085, + "grad_norm": 2.922145128250122, + "learning_rate": 5.993150684931507e-07, + "loss": 2.5657, + "step": 50 + }, + { + "epoch": 0.05136106831022085, + "eval_loss": 2.6353814601898193, + "eval_runtime": 37.0591, + "eval_samples_per_second": 26.984, + "eval_steps_per_second": 1.133, + "step": 50 + }, + { + "epoch": 0.05238828967642527, + "grad_norm": 2.5821900367736816, + "learning_rate": 6.164383561643836e-07, + "loss": 2.5512, + "step": 51 + }, + { + "epoch": 0.05341551104262968, + "grad_norm": 2.2299063205718994, + "learning_rate": 6.335616438356165e-07, + "loss": 2.6082, + "step": 52 + }, + { + "epoch": 0.054442732408834106, + "grad_norm": 2.0938310623168945, + "learning_rate": 6.506849315068493e-07, + "loss": 2.5301, + "step": 53 + }, + { + "epoch": 0.05546995377503852, + "grad_norm": 2.186629295349121, + "learning_rate": 6.678082191780823e-07, + "loss": 2.5755, + "step": 54 + }, + { + "epoch": 0.05649717514124294, + "grad_norm": 2.1747493743896484, + "learning_rate": 6.849315068493151e-07, + "loss": 2.5451, + "step": 55 + }, + { + "epoch": 0.05752439650744735, + "grad_norm": 2.5470354557037354, + "learning_rate": 7.020547945205481e-07, + "loss": 2.5159, + "step": 56 + }, + { + "epoch": 0.05855161787365177, + "grad_norm": 2.5056023597717285, + "learning_rate": 7.191780821917808e-07, + "loss": 2.5431, + "step": 57 + }, + { + "epoch": 0.05957883923985619, + "grad_norm": 2.6059749126434326, + "learning_rate": 7.363013698630137e-07, + "loss": 2.4684, + "step": 58 + }, + { + "epoch": 0.06060606060606061, + "grad_norm": 1.87150239944458, + "learning_rate": 7.534246575342466e-07, + "loss": 2.5269, + "step": 59 + }, + { + "epoch": 0.061633281972265024, + "grad_norm": 1.8777480125427246, + "learning_rate": 7.705479452054795e-07, + "loss": 2.5253, + "step": 60 + }, + { + "epoch": 0.061633281972265024, + "eval_loss": 2.5705313682556152, + "eval_runtime": 37.0929, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 60 + }, + { + "epoch": 0.06266050333846944, + "grad_norm": 1.8500003814697266, + "learning_rate": 7.876712328767124e-07, + "loss": 2.4987, + "step": 61 + }, + { + "epoch": 0.06368772470467386, + "grad_norm": 1.9473347663879395, + "learning_rate": 8.047945205479453e-07, + "loss": 2.4866, + "step": 62 + }, + { + "epoch": 0.06471494607087827, + "grad_norm": 2.2350873947143555, + "learning_rate": 8.219178082191781e-07, + "loss": 2.5343, + "step": 63 + }, + { + "epoch": 0.0657421674370827, + "grad_norm": 2.1791350841522217, + "learning_rate": 8.39041095890411e-07, + "loss": 2.5185, + "step": 64 + }, + { + "epoch": 0.0667693888032871, + "grad_norm": 1.8400903940200806, + "learning_rate": 8.561643835616439e-07, + "loss": 2.3897, + "step": 65 + }, + { + "epoch": 0.06779661016949153, + "grad_norm": 1.7193644046783447, + "learning_rate": 8.732876712328768e-07, + "loss": 2.485, + "step": 66 + }, + { + "epoch": 0.06882383153569595, + "grad_norm": 1.8478909730911255, + "learning_rate": 8.904109589041097e-07, + "loss": 2.4404, + "step": 67 + }, + { + "epoch": 0.06985105290190036, + "grad_norm": 1.8803719282150269, + "learning_rate": 9.075342465753426e-07, + "loss": 2.4933, + "step": 68 + }, + { + "epoch": 0.07087827426810478, + "grad_norm": 1.6866114139556885, + "learning_rate": 9.246575342465753e-07, + "loss": 2.4441, + "step": 69 + }, + { + "epoch": 0.07190549563430919, + "grad_norm": 1.807783842086792, + "learning_rate": 9.417808219178083e-07, + "loss": 2.5477, + "step": 70 + }, + { + "epoch": 0.07190549563430919, + "eval_loss": 2.5140035152435303, + "eval_runtime": 37.1384, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 70 + }, + { + "epoch": 0.07293271700051361, + "grad_norm": 1.7754000425338745, + "learning_rate": 9.589041095890411e-07, + "loss": 2.44, + "step": 71 + }, + { + "epoch": 0.07395993836671803, + "grad_norm": 1.7819652557373047, + "learning_rate": 9.76027397260274e-07, + "loss": 2.402, + "step": 72 + }, + { + "epoch": 0.07498715973292244, + "grad_norm": 1.6666278839111328, + "learning_rate": 9.931506849315068e-07, + "loss": 2.4511, + "step": 73 + }, + { + "epoch": 0.07601438109912687, + "grad_norm": 2.0625369548797607, + "learning_rate": 1.0102739726027399e-06, + "loss": 2.424, + "step": 74 + }, + { + "epoch": 0.07704160246533127, + "grad_norm": 1.9408000707626343, + "learning_rate": 1.0273972602739727e-06, + "loss": 2.4286, + "step": 75 + }, + { + "epoch": 0.0780688238315357, + "grad_norm": 1.7670842409133911, + "learning_rate": 1.0445205479452056e-06, + "loss": 2.407, + "step": 76 + }, + { + "epoch": 0.07909604519774012, + "grad_norm": 3.045717716217041, + "learning_rate": 1.0616438356164384e-06, + "loss": 2.425, + "step": 77 + }, + { + "epoch": 0.08012326656394453, + "grad_norm": 1.9278619289398193, + "learning_rate": 1.0787671232876712e-06, + "loss": 2.4306, + "step": 78 + }, + { + "epoch": 0.08115048793014895, + "grad_norm": 2.405146837234497, + "learning_rate": 1.095890410958904e-06, + "loss": 2.3499, + "step": 79 + }, + { + "epoch": 0.08217770929635336, + "grad_norm": 2.1083860397338867, + "learning_rate": 1.1130136986301371e-06, + "loss": 2.4115, + "step": 80 + }, + { + "epoch": 0.08217770929635336, + "eval_loss": 2.4501075744628906, + "eval_runtime": 37.1247, + "eval_samples_per_second": 26.936, + "eval_steps_per_second": 1.131, + "step": 80 + }, + { + "epoch": 0.08320493066255778, + "grad_norm": 2.181162118911743, + "learning_rate": 1.13013698630137e-06, + "loss": 2.3517, + "step": 81 + }, + { + "epoch": 0.08423215202876219, + "grad_norm": 2.1833999156951904, + "learning_rate": 1.1472602739726028e-06, + "loss": 2.38, + "step": 82 + }, + { + "epoch": 0.08525937339496661, + "grad_norm": 2.670384645462036, + "learning_rate": 1.1643835616438357e-06, + "loss": 2.3833, + "step": 83 + }, + { + "epoch": 0.08628659476117104, + "grad_norm": 2.438676118850708, + "learning_rate": 1.1815068493150685e-06, + "loss": 2.3399, + "step": 84 + }, + { + "epoch": 0.08731381612737545, + "grad_norm": 2.7584919929504395, + "learning_rate": 1.1986301369863014e-06, + "loss": 2.31, + "step": 85 + }, + { + "epoch": 0.08834103749357987, + "grad_norm": 2.4129064083099365, + "learning_rate": 1.2157534246575344e-06, + "loss": 2.3758, + "step": 86 + }, + { + "epoch": 0.08936825885978428, + "grad_norm": 2.366154193878174, + "learning_rate": 1.2328767123287673e-06, + "loss": 2.3154, + "step": 87 + }, + { + "epoch": 0.0903954802259887, + "grad_norm": 2.4916300773620605, + "learning_rate": 1.25e-06, + "loss": 2.3022, + "step": 88 + }, + { + "epoch": 0.09142270159219312, + "grad_norm": 2.4606640338897705, + "learning_rate": 1.267123287671233e-06, + "loss": 2.3103, + "step": 89 + }, + { + "epoch": 0.09244992295839753, + "grad_norm": 2.4134910106658936, + "learning_rate": 1.284246575342466e-06, + "loss": 2.2672, + "step": 90 + }, + { + "epoch": 0.09244992295839753, + "eval_loss": 2.3742051124572754, + "eval_runtime": 37.1258, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 90 + }, + { + "epoch": 0.09347714432460195, + "grad_norm": 2.302776336669922, + "learning_rate": 1.3013698630136986e-06, + "loss": 2.2798, + "step": 91 + }, + { + "epoch": 0.09450436569080636, + "grad_norm": 3.397819757461548, + "learning_rate": 1.3184931506849317e-06, + "loss": 2.2947, + "step": 92 + }, + { + "epoch": 0.09553158705701079, + "grad_norm": 2.6030948162078857, + "learning_rate": 1.3356164383561645e-06, + "loss": 2.2734, + "step": 93 + }, + { + "epoch": 0.09655880842321521, + "grad_norm": 2.9380528926849365, + "learning_rate": 1.3527397260273976e-06, + "loss": 2.2635, + "step": 94 + }, + { + "epoch": 0.09758602978941962, + "grad_norm": 2.9196345806121826, + "learning_rate": 1.3698630136986302e-06, + "loss": 2.3161, + "step": 95 + }, + { + "epoch": 0.09861325115562404, + "grad_norm": 2.2380659580230713, + "learning_rate": 1.386986301369863e-06, + "loss": 2.21, + "step": 96 + }, + { + "epoch": 0.09964047252182845, + "grad_norm": 2.730910539627075, + "learning_rate": 1.4041095890410961e-06, + "loss": 2.2562, + "step": 97 + }, + { + "epoch": 0.10066769388803287, + "grad_norm": 2.994748830795288, + "learning_rate": 1.421232876712329e-06, + "loss": 2.1597, + "step": 98 + }, + { + "epoch": 0.1016949152542373, + "grad_norm": 2.2228171825408936, + "learning_rate": 1.4383561643835616e-06, + "loss": 2.1835, + "step": 99 + }, + { + "epoch": 0.1027221366204417, + "grad_norm": 2.715527296066284, + "learning_rate": 1.4554794520547946e-06, + "loss": 2.2074, + "step": 100 + }, + { + "epoch": 0.1027221366204417, + "eval_loss": 2.3061234951019287, + "eval_runtime": 37.1393, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 100 + }, + { + "epoch": 0.10374935798664613, + "grad_norm": 2.7054452896118164, + "learning_rate": 1.4726027397260275e-06, + "loss": 2.2089, + "step": 101 + }, + { + "epoch": 0.10477657935285054, + "grad_norm": 2.795694589614868, + "learning_rate": 1.4897260273972605e-06, + "loss": 2.1802, + "step": 102 + }, + { + "epoch": 0.10580380071905496, + "grad_norm": 2.634822368621826, + "learning_rate": 1.5068493150684932e-06, + "loss": 2.1428, + "step": 103 + }, + { + "epoch": 0.10683102208525937, + "grad_norm": 2.2418437004089355, + "learning_rate": 1.5239726027397262e-06, + "loss": 2.1706, + "step": 104 + }, + { + "epoch": 0.10785824345146379, + "grad_norm": 2.471231698989868, + "learning_rate": 1.541095890410959e-06, + "loss": 2.2248, + "step": 105 + }, + { + "epoch": 0.10888546481766821, + "grad_norm": 2.4154577255249023, + "learning_rate": 1.5582191780821921e-06, + "loss": 2.1873, + "step": 106 + }, + { + "epoch": 0.10991268618387262, + "grad_norm": 2.2823996543884277, + "learning_rate": 1.5753424657534248e-06, + "loss": 2.2237, + "step": 107 + }, + { + "epoch": 0.11093990755007704, + "grad_norm": 2.8695461750030518, + "learning_rate": 1.5924657534246576e-06, + "loss": 2.1966, + "step": 108 + }, + { + "epoch": 0.11196712891628145, + "grad_norm": 3.319650650024414, + "learning_rate": 1.6095890410958907e-06, + "loss": 2.066, + "step": 109 + }, + { + "epoch": 0.11299435028248588, + "grad_norm": 2.815196990966797, + "learning_rate": 1.6267123287671235e-06, + "loss": 2.1355, + "step": 110 + }, + { + "epoch": 0.11299435028248588, + "eval_loss": 2.250187635421753, + "eval_runtime": 37.1645, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 110 + }, + { + "epoch": 0.1140215716486903, + "grad_norm": 2.9195759296417236, + "learning_rate": 1.6438356164383561e-06, + "loss": 2.0713, + "step": 111 + }, + { + "epoch": 0.1150487930148947, + "grad_norm": 2.5600695610046387, + "learning_rate": 1.6609589041095892e-06, + "loss": 2.1573, + "step": 112 + }, + { + "epoch": 0.11607601438109913, + "grad_norm": 2.7566795349121094, + "learning_rate": 1.678082191780822e-06, + "loss": 2.0909, + "step": 113 + }, + { + "epoch": 0.11710323574730354, + "grad_norm": 2.5593106746673584, + "learning_rate": 1.695205479452055e-06, + "loss": 2.144, + "step": 114 + }, + { + "epoch": 0.11813045711350796, + "grad_norm": 2.594931125640869, + "learning_rate": 1.7123287671232877e-06, + "loss": 2.0606, + "step": 115 + }, + { + "epoch": 0.11915767847971238, + "grad_norm": 2.769916534423828, + "learning_rate": 1.7294520547945206e-06, + "loss": 2.1737, + "step": 116 + }, + { + "epoch": 0.12018489984591679, + "grad_norm": 2.523179054260254, + "learning_rate": 1.7465753424657536e-06, + "loss": 2.1247, + "step": 117 + }, + { + "epoch": 0.12121212121212122, + "grad_norm": 2.8153676986694336, + "learning_rate": 1.7636986301369865e-06, + "loss": 2.1257, + "step": 118 + }, + { + "epoch": 0.12223934257832562, + "grad_norm": 2.5389721393585205, + "learning_rate": 1.7808219178082193e-06, + "loss": 2.1049, + "step": 119 + }, + { + "epoch": 0.12326656394453005, + "grad_norm": 2.9753856658935547, + "learning_rate": 1.7979452054794521e-06, + "loss": 2.1871, + "step": 120 + }, + { + "epoch": 0.12326656394453005, + "eval_loss": 2.2089297771453857, + "eval_runtime": 37.897, + "eval_samples_per_second": 26.387, + "eval_steps_per_second": 1.108, + "step": 120 + }, + { + "epoch": 0.12429378531073447, + "grad_norm": 2.737180471420288, + "learning_rate": 1.8150684931506852e-06, + "loss": 2.1166, + "step": 121 + }, + { + "epoch": 0.12532100667693888, + "grad_norm": 2.783801794052124, + "learning_rate": 1.832191780821918e-06, + "loss": 2.0932, + "step": 122 + }, + { + "epoch": 0.1263482280431433, + "grad_norm": 2.9615402221679688, + "learning_rate": 1.8493150684931507e-06, + "loss": 2.0327, + "step": 123 + }, + { + "epoch": 0.12737544940934772, + "grad_norm": 3.053622007369995, + "learning_rate": 1.8664383561643837e-06, + "loss": 2.0665, + "step": 124 + }, + { + "epoch": 0.12840267077555212, + "grad_norm": 3.265355348587036, + "learning_rate": 1.8835616438356166e-06, + "loss": 2.0243, + "step": 125 + }, + { + "epoch": 0.12942989214175654, + "grad_norm": 2.4237143993377686, + "learning_rate": 1.9006849315068496e-06, + "loss": 2.0902, + "step": 126 + }, + { + "epoch": 0.13045711350796096, + "grad_norm": 2.926950693130493, + "learning_rate": 1.9178082191780823e-06, + "loss": 2.0057, + "step": 127 + }, + { + "epoch": 0.1314843348741654, + "grad_norm": 2.5317165851593018, + "learning_rate": 1.9349315068493153e-06, + "loss": 2.0719, + "step": 128 + }, + { + "epoch": 0.1325115562403698, + "grad_norm": 2.4999701976776123, + "learning_rate": 1.952054794520548e-06, + "loss": 2.0969, + "step": 129 + }, + { + "epoch": 0.1335387776065742, + "grad_norm": 2.7063920497894287, + "learning_rate": 1.969178082191781e-06, + "loss": 2.0902, + "step": 130 + }, + { + "epoch": 0.1335387776065742, + "eval_loss": 2.175224542617798, + "eval_runtime": 37.4519, + "eval_samples_per_second": 26.701, + "eval_steps_per_second": 1.121, + "step": 130 + }, + { + "epoch": 0.13456599897277863, + "grad_norm": 2.650860071182251, + "learning_rate": 1.9863013698630136e-06, + "loss": 2.0191, + "step": 131 + }, + { + "epoch": 0.13559322033898305, + "grad_norm": 2.79182505607605, + "learning_rate": 2.0034246575342467e-06, + "loss": 2.0856, + "step": 132 + }, + { + "epoch": 0.13662044170518747, + "grad_norm": 2.453162431716919, + "learning_rate": 2.0205479452054797e-06, + "loss": 2.0679, + "step": 133 + }, + { + "epoch": 0.1376476630713919, + "grad_norm": 2.9630799293518066, + "learning_rate": 2.037671232876713e-06, + "loss": 2.0038, + "step": 134 + }, + { + "epoch": 0.1386748844375963, + "grad_norm": 3.0807266235351562, + "learning_rate": 2.0547945205479454e-06, + "loss": 2.0367, + "step": 135 + }, + { + "epoch": 0.1397021058038007, + "grad_norm": 2.9712910652160645, + "learning_rate": 2.071917808219178e-06, + "loss": 2.0571, + "step": 136 + }, + { + "epoch": 0.14072932717000514, + "grad_norm": 3.0804853439331055, + "learning_rate": 2.089041095890411e-06, + "loss": 1.9725, + "step": 137 + }, + { + "epoch": 0.14175654853620956, + "grad_norm": 2.659608840942383, + "learning_rate": 2.106164383561644e-06, + "loss": 2.0572, + "step": 138 + }, + { + "epoch": 0.14278376990241398, + "grad_norm": 2.8093109130859375, + "learning_rate": 2.123287671232877e-06, + "loss": 2.0178, + "step": 139 + }, + { + "epoch": 0.14381099126861838, + "grad_norm": 2.8163881301879883, + "learning_rate": 2.14041095890411e-06, + "loss": 2.0085, + "step": 140 + }, + { + "epoch": 0.14381099126861838, + "eval_loss": 2.147092342376709, + "eval_runtime": 37.2453, + "eval_samples_per_second": 26.849, + "eval_steps_per_second": 1.128, + "step": 140 + }, + { + "epoch": 0.1448382126348228, + "grad_norm": 3.795241594314575, + "learning_rate": 2.1575342465753425e-06, + "loss": 2.0149, + "step": 141 + }, + { + "epoch": 0.14586543400102722, + "grad_norm": 3.3287174701690674, + "learning_rate": 2.1746575342465755e-06, + "loss": 1.9658, + "step": 142 + }, + { + "epoch": 0.14689265536723164, + "grad_norm": 2.803213596343994, + "learning_rate": 2.191780821917808e-06, + "loss": 2.0983, + "step": 143 + }, + { + "epoch": 0.14791987673343607, + "grad_norm": 3.2640199661254883, + "learning_rate": 2.2089041095890412e-06, + "loss": 1.9865, + "step": 144 + }, + { + "epoch": 0.14894709809964046, + "grad_norm": 2.95979380607605, + "learning_rate": 2.2260273972602743e-06, + "loss": 1.9669, + "step": 145 + }, + { + "epoch": 0.14997431946584489, + "grad_norm": 3.737175226211548, + "learning_rate": 2.243150684931507e-06, + "loss": 1.9691, + "step": 146 + }, + { + "epoch": 0.1510015408320493, + "grad_norm": 2.7248430252075195, + "learning_rate": 2.26027397260274e-06, + "loss": 1.9764, + "step": 147 + }, + { + "epoch": 0.15202876219825373, + "grad_norm": 2.929856061935425, + "learning_rate": 2.2773972602739726e-06, + "loss": 1.9967, + "step": 148 + }, + { + "epoch": 0.15305598356445815, + "grad_norm": 2.8998641967773438, + "learning_rate": 2.2945205479452057e-06, + "loss": 2.0842, + "step": 149 + }, + { + "epoch": 0.15408320493066255, + "grad_norm": 2.9672694206237793, + "learning_rate": 2.3116438356164387e-06, + "loss": 1.9915, + "step": 150 + }, + { + "epoch": 0.15408320493066255, + "eval_loss": 2.122715473175049, + "eval_runtime": 37.1938, + "eval_samples_per_second": 26.886, + "eval_steps_per_second": 1.129, + "step": 150 + }, + { + "epoch": 0.15511042629686697, + "grad_norm": 2.8529887199401855, + "learning_rate": 2.3287671232876713e-06, + "loss": 1.9022, + "step": 151 + }, + { + "epoch": 0.1561376476630714, + "grad_norm": 2.914121389389038, + "learning_rate": 2.3458904109589044e-06, + "loss": 1.956, + "step": 152 + }, + { + "epoch": 0.15716486902927582, + "grad_norm": 3.6286494731903076, + "learning_rate": 2.363013698630137e-06, + "loss": 2.0118, + "step": 153 + }, + { + "epoch": 0.15819209039548024, + "grad_norm": 2.789562702178955, + "learning_rate": 2.38013698630137e-06, + "loss": 1.9935, + "step": 154 + }, + { + "epoch": 0.15921931176168463, + "grad_norm": 3.367119550704956, + "learning_rate": 2.3972602739726027e-06, + "loss": 1.9519, + "step": 155 + }, + { + "epoch": 0.16024653312788906, + "grad_norm": 2.9345803260803223, + "learning_rate": 2.4143835616438358e-06, + "loss": 1.9017, + "step": 156 + }, + { + "epoch": 0.16127375449409348, + "grad_norm": 2.8674590587615967, + "learning_rate": 2.431506849315069e-06, + "loss": 1.9731, + "step": 157 + }, + { + "epoch": 0.1623009758602979, + "grad_norm": 3.144766092300415, + "learning_rate": 2.4486301369863015e-06, + "loss": 1.9811, + "step": 158 + }, + { + "epoch": 0.1633281972265023, + "grad_norm": 3.845395088195801, + "learning_rate": 2.4657534246575345e-06, + "loss": 1.9523, + "step": 159 + }, + { + "epoch": 0.16435541859270672, + "grad_norm": 2.776700973510742, + "learning_rate": 2.482876712328767e-06, + "loss": 1.9687, + "step": 160 + }, + { + "epoch": 0.16435541859270672, + "eval_loss": 2.0929579734802246, + "eval_runtime": 37.1917, + "eval_samples_per_second": 26.888, + "eval_steps_per_second": 1.129, + "step": 160 + }, + { + "epoch": 0.16538263995891114, + "grad_norm": 3.4260904788970947, + "learning_rate": 2.5e-06, + "loss": 1.9665, + "step": 161 + }, + { + "epoch": 0.16640986132511557, + "grad_norm": 3.1217620372772217, + "learning_rate": 2.5171232876712333e-06, + "loss": 1.9361, + "step": 162 + }, + { + "epoch": 0.16743708269132, + "grad_norm": 3.342068910598755, + "learning_rate": 2.534246575342466e-06, + "loss": 1.9084, + "step": 163 + }, + { + "epoch": 0.16846430405752438, + "grad_norm": 3.6485838890075684, + "learning_rate": 2.551369863013699e-06, + "loss": 1.9542, + "step": 164 + }, + { + "epoch": 0.1694915254237288, + "grad_norm": 3.1247780323028564, + "learning_rate": 2.568493150684932e-06, + "loss": 1.9717, + "step": 165 + }, + { + "epoch": 0.17051874678993323, + "grad_norm": 2.9856860637664795, + "learning_rate": 2.585616438356164e-06, + "loss": 1.9574, + "step": 166 + }, + { + "epoch": 0.17154596815613765, + "grad_norm": 3.337902069091797, + "learning_rate": 2.6027397260273973e-06, + "loss": 1.8884, + "step": 167 + }, + { + "epoch": 0.17257318952234207, + "grad_norm": 2.844534158706665, + "learning_rate": 2.6198630136986303e-06, + "loss": 1.9545, + "step": 168 + }, + { + "epoch": 0.17360041088854647, + "grad_norm": 3.75907564163208, + "learning_rate": 2.6369863013698634e-06, + "loss": 1.9907, + "step": 169 + }, + { + "epoch": 0.1746276322547509, + "grad_norm": 2.9321961402893066, + "learning_rate": 2.654109589041096e-06, + "loss": 1.9663, + "step": 170 + }, + { + "epoch": 0.1746276322547509, + "eval_loss": 2.0753843784332275, + "eval_runtime": 37.2389, + "eval_samples_per_second": 26.854, + "eval_steps_per_second": 1.128, + "step": 170 + }, + { + "epoch": 0.17565485362095531, + "grad_norm": 2.952523946762085, + "learning_rate": 2.671232876712329e-06, + "loss": 1.9065, + "step": 171 + }, + { + "epoch": 0.17668207498715974, + "grad_norm": 2.952463150024414, + "learning_rate": 2.688356164383562e-06, + "loss": 1.8994, + "step": 172 + }, + { + "epoch": 0.17770929635336416, + "grad_norm": 3.1352880001068115, + "learning_rate": 2.705479452054795e-06, + "loss": 1.8902, + "step": 173 + }, + { + "epoch": 0.17873651771956856, + "grad_norm": 3.5727322101593018, + "learning_rate": 2.7226027397260274e-06, + "loss": 1.8454, + "step": 174 + }, + { + "epoch": 0.17976373908577298, + "grad_norm": 3.0562493801116943, + "learning_rate": 2.7397260273972604e-06, + "loss": 1.984, + "step": 175 + }, + { + "epoch": 0.1807909604519774, + "grad_norm": 3.367924690246582, + "learning_rate": 2.7568493150684935e-06, + "loss": 1.9398, + "step": 176 + }, + { + "epoch": 0.18181818181818182, + "grad_norm": 3.1982340812683105, + "learning_rate": 2.773972602739726e-06, + "loss": 1.8566, + "step": 177 + }, + { + "epoch": 0.18284540318438625, + "grad_norm": 3.8962080478668213, + "learning_rate": 2.791095890410959e-06, + "loss": 1.9459, + "step": 178 + }, + { + "epoch": 0.18387262455059064, + "grad_norm": 3.3503036499023438, + "learning_rate": 2.8082191780821922e-06, + "loss": 1.8348, + "step": 179 + }, + { + "epoch": 0.18489984591679506, + "grad_norm": 3.276139497756958, + "learning_rate": 2.8253424657534253e-06, + "loss": 1.9849, + "step": 180 + }, + { + "epoch": 0.18489984591679506, + "eval_loss": 2.0605413913726807, + "eval_runtime": 37.2276, + "eval_samples_per_second": 26.862, + "eval_steps_per_second": 1.128, + "step": 180 + }, + { + "epoch": 0.1859270672829995, + "grad_norm": 3.3787975311279297, + "learning_rate": 2.842465753424658e-06, + "loss": 1.9517, + "step": 181 + }, + { + "epoch": 0.1869542886492039, + "grad_norm": 3.054290294647217, + "learning_rate": 2.8595890410958905e-06, + "loss": 1.9572, + "step": 182 + }, + { + "epoch": 0.18798151001540833, + "grad_norm": 3.0737054347991943, + "learning_rate": 2.876712328767123e-06, + "loss": 1.9674, + "step": 183 + }, + { + "epoch": 0.18900873138161273, + "grad_norm": 3.2833855152130127, + "learning_rate": 2.8938356164383562e-06, + "loss": 1.8185, + "step": 184 + }, + { + "epoch": 0.19003595274781715, + "grad_norm": 2.70452880859375, + "learning_rate": 2.9109589041095893e-06, + "loss": 1.8552, + "step": 185 + }, + { + "epoch": 0.19106317411402157, + "grad_norm": 3.2723655700683594, + "learning_rate": 2.9280821917808223e-06, + "loss": 1.8433, + "step": 186 + }, + { + "epoch": 0.192090395480226, + "grad_norm": 2.844536066055298, + "learning_rate": 2.945205479452055e-06, + "loss": 1.8488, + "step": 187 + }, + { + "epoch": 0.19311761684643042, + "grad_norm": 2.944545269012451, + "learning_rate": 2.962328767123288e-06, + "loss": 1.9131, + "step": 188 + }, + { + "epoch": 0.1941448382126348, + "grad_norm": 2.812204122543335, + "learning_rate": 2.979452054794521e-06, + "loss": 1.8977, + "step": 189 + }, + { + "epoch": 0.19517205957883924, + "grad_norm": 2.8999950885772705, + "learning_rate": 2.9965753424657533e-06, + "loss": 1.8477, + "step": 190 + }, + { + "epoch": 0.19517205957883924, + "eval_loss": 2.051161527633667, + "eval_runtime": 37.2363, + "eval_samples_per_second": 26.856, + "eval_steps_per_second": 1.128, + "step": 190 + }, + { + "epoch": 0.19619928094504366, + "grad_norm": 2.7814741134643555, + "learning_rate": 3.0136986301369864e-06, + "loss": 1.8989, + "step": 191 + }, + { + "epoch": 0.19722650231124808, + "grad_norm": 2.93346905708313, + "learning_rate": 3.0308219178082194e-06, + "loss": 1.8748, + "step": 192 + }, + { + "epoch": 0.1982537236774525, + "grad_norm": 3.00400447845459, + "learning_rate": 3.0479452054794525e-06, + "loss": 1.838, + "step": 193 + }, + { + "epoch": 0.1992809450436569, + "grad_norm": 2.949695110321045, + "learning_rate": 3.065068493150685e-06, + "loss": 1.8817, + "step": 194 + }, + { + "epoch": 0.20030816640986132, + "grad_norm": 2.8858301639556885, + "learning_rate": 3.082191780821918e-06, + "loss": 1.8333, + "step": 195 + }, + { + "epoch": 0.20133538777606574, + "grad_norm": 3.123706102371216, + "learning_rate": 3.099315068493151e-06, + "loss": 1.8533, + "step": 196 + }, + { + "epoch": 0.20236260914227017, + "grad_norm": 3.0130505561828613, + "learning_rate": 3.1164383561643843e-06, + "loss": 1.842, + "step": 197 + }, + { + "epoch": 0.2033898305084746, + "grad_norm": 2.8823280334472656, + "learning_rate": 3.1335616438356165e-06, + "loss": 1.8025, + "step": 198 + }, + { + "epoch": 0.20441705187467898, + "grad_norm": 2.920884132385254, + "learning_rate": 3.1506849315068495e-06, + "loss": 1.9488, + "step": 199 + }, + { + "epoch": 0.2054442732408834, + "grad_norm": 2.9152207374572754, + "learning_rate": 3.167808219178082e-06, + "loss": 1.8163, + "step": 200 + }, + { + "epoch": 0.2054442732408834, + "eval_loss": 2.033583641052246, + "eval_runtime": 37.127, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 200 + }, + { + "epoch": 0.20647149460708783, + "grad_norm": 2.904230833053589, + "learning_rate": 3.184931506849315e-06, + "loss": 1.8207, + "step": 201 + }, + { + "epoch": 0.20749871597329225, + "grad_norm": 3.128448486328125, + "learning_rate": 3.2020547945205483e-06, + "loss": 1.8051, + "step": 202 + }, + { + "epoch": 0.20852593733949668, + "grad_norm": 3.0049452781677246, + "learning_rate": 3.2191780821917813e-06, + "loss": 1.7823, + "step": 203 + }, + { + "epoch": 0.20955315870570107, + "grad_norm": 2.7612497806549072, + "learning_rate": 3.236301369863014e-06, + "loss": 1.7703, + "step": 204 + }, + { + "epoch": 0.2105803800719055, + "grad_norm": 2.9178879261016846, + "learning_rate": 3.253424657534247e-06, + "loss": 1.7427, + "step": 205 + }, + { + "epoch": 0.21160760143810992, + "grad_norm": 2.836967945098877, + "learning_rate": 3.2705479452054796e-06, + "loss": 1.8347, + "step": 206 + }, + { + "epoch": 0.21263482280431434, + "grad_norm": 2.8437204360961914, + "learning_rate": 3.2876712328767123e-06, + "loss": 1.8147, + "step": 207 + }, + { + "epoch": 0.21366204417051873, + "grad_norm": 3.0728225708007812, + "learning_rate": 3.3047945205479453e-06, + "loss": 1.7823, + "step": 208 + }, + { + "epoch": 0.21468926553672316, + "grad_norm": 3.357631206512451, + "learning_rate": 3.3219178082191784e-06, + "loss": 1.8013, + "step": 209 + }, + { + "epoch": 0.21571648690292758, + "grad_norm": 2.7996842861175537, + "learning_rate": 3.3390410958904114e-06, + "loss": 1.8279, + "step": 210 + }, + { + "epoch": 0.21571648690292758, + "eval_loss": 2.024500846862793, + "eval_runtime": 37.0122, + "eval_samples_per_second": 27.018, + "eval_steps_per_second": 1.135, + "step": 210 + }, + { + "epoch": 0.216743708269132, + "grad_norm": 3.3635380268096924, + "learning_rate": 3.356164383561644e-06, + "loss": 1.8509, + "step": 211 + }, + { + "epoch": 0.21777092963533642, + "grad_norm": 2.703932523727417, + "learning_rate": 3.373287671232877e-06, + "loss": 1.8984, + "step": 212 + }, + { + "epoch": 0.21879815100154082, + "grad_norm": 2.9341695308685303, + "learning_rate": 3.39041095890411e-06, + "loss": 1.8381, + "step": 213 + }, + { + "epoch": 0.21982537236774524, + "grad_norm": 3.072448253631592, + "learning_rate": 3.4075342465753424e-06, + "loss": 1.8779, + "step": 214 + }, + { + "epoch": 0.22085259373394966, + "grad_norm": 2.850555419921875, + "learning_rate": 3.4246575342465754e-06, + "loss": 1.7656, + "step": 215 + }, + { + "epoch": 0.2218798151001541, + "grad_norm": 3.0218873023986816, + "learning_rate": 3.4417808219178085e-06, + "loss": 1.829, + "step": 216 + }, + { + "epoch": 0.2229070364663585, + "grad_norm": 2.8475091457366943, + "learning_rate": 3.458904109589041e-06, + "loss": 1.8121, + "step": 217 + }, + { + "epoch": 0.2239342578325629, + "grad_norm": 2.8760111331939697, + "learning_rate": 3.476027397260274e-06, + "loss": 1.8281, + "step": 218 + }, + { + "epoch": 0.22496147919876733, + "grad_norm": 2.976670265197754, + "learning_rate": 3.4931506849315072e-06, + "loss": 1.8274, + "step": 219 + }, + { + "epoch": 0.22598870056497175, + "grad_norm": 2.912505865097046, + "learning_rate": 3.5102739726027403e-06, + "loss": 1.8316, + "step": 220 + }, + { + "epoch": 0.22598870056497175, + "eval_loss": 2.0203635692596436, + "eval_runtime": 37.0433, + "eval_samples_per_second": 26.995, + "eval_steps_per_second": 1.134, + "step": 220 + }, + { + "epoch": 0.22701592193117617, + "grad_norm": 3.1133334636688232, + "learning_rate": 3.527397260273973e-06, + "loss": 1.8277, + "step": 221 + }, + { + "epoch": 0.2280431432973806, + "grad_norm": 3.289062261581421, + "learning_rate": 3.5445205479452056e-06, + "loss": 1.7893, + "step": 222 + }, + { + "epoch": 0.229070364663585, + "grad_norm": 2.814558267593384, + "learning_rate": 3.5616438356164386e-06, + "loss": 1.8095, + "step": 223 + }, + { + "epoch": 0.2300975860297894, + "grad_norm": 3.48117733001709, + "learning_rate": 3.5787671232876712e-06, + "loss": 1.8701, + "step": 224 + }, + { + "epoch": 0.23112480739599384, + "grad_norm": 2.922144651412964, + "learning_rate": 3.5958904109589043e-06, + "loss": 1.8479, + "step": 225 + }, + { + "epoch": 0.23215202876219826, + "grad_norm": 3.552428960800171, + "learning_rate": 3.6130136986301373e-06, + "loss": 1.7822, + "step": 226 + }, + { + "epoch": 0.23317925012840268, + "grad_norm": 3.058274269104004, + "learning_rate": 3.6301369863013704e-06, + "loss": 1.8228, + "step": 227 + }, + { + "epoch": 0.23420647149460708, + "grad_norm": 3.365823268890381, + "learning_rate": 3.647260273972603e-06, + "loss": 1.8415, + "step": 228 + }, + { + "epoch": 0.2352336928608115, + "grad_norm": 2.87333607673645, + "learning_rate": 3.664383561643836e-06, + "loss": 1.8474, + "step": 229 + }, + { + "epoch": 0.23626091422701592, + "grad_norm": 3.185807228088379, + "learning_rate": 3.6815068493150687e-06, + "loss": 1.7854, + "step": 230 + }, + { + "epoch": 0.23626091422701592, + "eval_loss": 2.0132367610931396, + "eval_runtime": 37.3246, + "eval_samples_per_second": 26.792, + "eval_steps_per_second": 1.125, + "step": 230 + }, + { + "epoch": 0.23728813559322035, + "grad_norm": 3.02864933013916, + "learning_rate": 3.6986301369863014e-06, + "loss": 1.8081, + "step": 231 + }, + { + "epoch": 0.23831535695942477, + "grad_norm": 2.892451286315918, + "learning_rate": 3.7157534246575344e-06, + "loss": 1.7619, + "step": 232 + }, + { + "epoch": 0.23934257832562916, + "grad_norm": 3.24947452545166, + "learning_rate": 3.7328767123287675e-06, + "loss": 1.8729, + "step": 233 + }, + { + "epoch": 0.24036979969183359, + "grad_norm": 3.13212513923645, + "learning_rate": 3.7500000000000005e-06, + "loss": 1.7849, + "step": 234 + }, + { + "epoch": 0.241397021058038, + "grad_norm": 3.004814863204956, + "learning_rate": 3.767123287671233e-06, + "loss": 1.8217, + "step": 235 + }, + { + "epoch": 0.24242424242424243, + "grad_norm": 2.845555067062378, + "learning_rate": 3.784246575342466e-06, + "loss": 1.8438, + "step": 236 + }, + { + "epoch": 0.24345146379044685, + "grad_norm": 3.1796751022338867, + "learning_rate": 3.8013698630136993e-06, + "loss": 1.7807, + "step": 237 + }, + { + "epoch": 0.24447868515665125, + "grad_norm": 2.7663002014160156, + "learning_rate": 3.818493150684932e-06, + "loss": 1.7981, + "step": 238 + }, + { + "epoch": 0.24550590652285567, + "grad_norm": 3.0547749996185303, + "learning_rate": 3.8356164383561645e-06, + "loss": 1.7543, + "step": 239 + }, + { + "epoch": 0.2465331278890601, + "grad_norm": 2.6994550228118896, + "learning_rate": 3.852739726027397e-06, + "loss": 1.8612, + "step": 240 + }, + { + "epoch": 0.2465331278890601, + "eval_loss": 2.0076475143432617, + "eval_runtime": 37.3223, + "eval_samples_per_second": 26.794, + "eval_steps_per_second": 1.125, + "step": 240 + }, + { + "epoch": 0.24756034925526452, + "grad_norm": 2.769717216491699, + "learning_rate": 3.869863013698631e-06, + "loss": 1.8352, + "step": 241 + }, + { + "epoch": 0.24858757062146894, + "grad_norm": 2.7526769638061523, + "learning_rate": 3.886986301369863e-06, + "loss": 1.7775, + "step": 242 + }, + { + "epoch": 0.24961479198767333, + "grad_norm": 3.0213067531585693, + "learning_rate": 3.904109589041096e-06, + "loss": 1.7901, + "step": 243 + }, + { + "epoch": 0.25064201335387776, + "grad_norm": 2.8562498092651367, + "learning_rate": 3.921232876712329e-06, + "loss": 1.7483, + "step": 244 + }, + { + "epoch": 0.2516692347200822, + "grad_norm": 2.812774896621704, + "learning_rate": 3.938356164383562e-06, + "loss": 1.7643, + "step": 245 + }, + { + "epoch": 0.2526964560862866, + "grad_norm": 3.2509777545928955, + "learning_rate": 3.9554794520547955e-06, + "loss": 1.729, + "step": 246 + }, + { + "epoch": 0.253723677452491, + "grad_norm": 3.6394171714782715, + "learning_rate": 3.972602739726027e-06, + "loss": 1.8253, + "step": 247 + }, + { + "epoch": 0.25475089881869545, + "grad_norm": 3.1234781742095947, + "learning_rate": 3.989726027397261e-06, + "loss": 1.7457, + "step": 248 + }, + { + "epoch": 0.25577812018489987, + "grad_norm": 2.8587522506713867, + "learning_rate": 4.006849315068493e-06, + "loss": 1.804, + "step": 249 + }, + { + "epoch": 0.25680534155110424, + "grad_norm": 2.9718286991119385, + "learning_rate": 4.023972602739726e-06, + "loss": 1.8051, + "step": 250 + }, + { + "epoch": 0.25680534155110424, + "eval_loss": 2.0024874210357666, + "eval_runtime": 37.2884, + "eval_samples_per_second": 26.818, + "eval_steps_per_second": 1.126, + "step": 250 + }, + { + "epoch": 0.25783256291730866, + "grad_norm": 3.007124423980713, + "learning_rate": 4.0410958904109595e-06, + "loss": 1.7946, + "step": 251 + }, + { + "epoch": 0.2588597842835131, + "grad_norm": 3.1715426445007324, + "learning_rate": 4.058219178082192e-06, + "loss": 1.7811, + "step": 252 + }, + { + "epoch": 0.2598870056497175, + "grad_norm": 2.814269542694092, + "learning_rate": 4.075342465753426e-06, + "loss": 1.8024, + "step": 253 + }, + { + "epoch": 0.26091422701592193, + "grad_norm": 2.841212034225464, + "learning_rate": 4.092465753424658e-06, + "loss": 1.7713, + "step": 254 + }, + { + "epoch": 0.26194144838212635, + "grad_norm": 3.0878522396087646, + "learning_rate": 4.109589041095891e-06, + "loss": 1.7579, + "step": 255 + }, + { + "epoch": 0.2629686697483308, + "grad_norm": 2.946709394454956, + "learning_rate": 4.1267123287671235e-06, + "loss": 1.823, + "step": 256 + }, + { + "epoch": 0.2639958911145352, + "grad_norm": 2.8136606216430664, + "learning_rate": 4.143835616438356e-06, + "loss": 1.7731, + "step": 257 + }, + { + "epoch": 0.2650231124807396, + "grad_norm": 2.745762825012207, + "learning_rate": 4.16095890410959e-06, + "loss": 1.8445, + "step": 258 + }, + { + "epoch": 0.26605033384694404, + "grad_norm": 3.0986413955688477, + "learning_rate": 4.178082191780822e-06, + "loss": 1.8231, + "step": 259 + }, + { + "epoch": 0.2670775552131484, + "grad_norm": 2.6990394592285156, + "learning_rate": 4.195205479452055e-06, + "loss": 1.7635, + "step": 260 + }, + { + "epoch": 0.2670775552131484, + "eval_loss": 1.9839800596237183, + "eval_runtime": 37.1612, + "eval_samples_per_second": 26.91, + "eval_steps_per_second": 1.13, + "step": 260 + }, + { + "epoch": 0.26810477657935283, + "grad_norm": 3.281421184539795, + "learning_rate": 4.212328767123288e-06, + "loss": 1.7587, + "step": 261 + }, + { + "epoch": 0.26913199794555726, + "grad_norm": 3.1233623027801514, + "learning_rate": 4.229452054794521e-06, + "loss": 1.7673, + "step": 262 + }, + { + "epoch": 0.2701592193117617, + "grad_norm": 3.3131625652313232, + "learning_rate": 4.246575342465754e-06, + "loss": 1.7755, + "step": 263 + }, + { + "epoch": 0.2711864406779661, + "grad_norm": 3.3002350330352783, + "learning_rate": 4.263698630136986e-06, + "loss": 1.7575, + "step": 264 + }, + { + "epoch": 0.2722136620441705, + "grad_norm": 2.965397357940674, + "learning_rate": 4.28082191780822e-06, + "loss": 1.7162, + "step": 265 + }, + { + "epoch": 0.27324088341037495, + "grad_norm": 2.9904608726501465, + "learning_rate": 4.297945205479452e-06, + "loss": 1.8794, + "step": 266 + }, + { + "epoch": 0.27426810477657937, + "grad_norm": 2.707996129989624, + "learning_rate": 4.315068493150685e-06, + "loss": 1.7646, + "step": 267 + }, + { + "epoch": 0.2752953261427838, + "grad_norm": 2.8878190517425537, + "learning_rate": 4.3321917808219185e-06, + "loss": 1.7559, + "step": 268 + }, + { + "epoch": 0.2763225475089882, + "grad_norm": 3.1676602363586426, + "learning_rate": 4.349315068493151e-06, + "loss": 1.7806, + "step": 269 + }, + { + "epoch": 0.2773497688751926, + "grad_norm": 2.9179935455322266, + "learning_rate": 4.3664383561643846e-06, + "loss": 1.7493, + "step": 270 + }, + { + "epoch": 0.2773497688751926, + "eval_loss": 1.9736579656600952, + "eval_runtime": 37.0983, + "eval_samples_per_second": 26.955, + "eval_steps_per_second": 1.132, + "step": 270 + }, + { + "epoch": 0.278376990241397, + "grad_norm": 3.0081794261932373, + "learning_rate": 4.383561643835616e-06, + "loss": 1.7405, + "step": 271 + }, + { + "epoch": 0.2794042116076014, + "grad_norm": 3.240462064743042, + "learning_rate": 4.40068493150685e-06, + "loss": 1.8326, + "step": 272 + }, + { + "epoch": 0.28043143297380585, + "grad_norm": 3.324059247970581, + "learning_rate": 4.4178082191780825e-06, + "loss": 1.7819, + "step": 273 + }, + { + "epoch": 0.2814586543400103, + "grad_norm": 2.8438684940338135, + "learning_rate": 4.434931506849315e-06, + "loss": 1.7435, + "step": 274 + }, + { + "epoch": 0.2824858757062147, + "grad_norm": 2.9525997638702393, + "learning_rate": 4.4520547945205486e-06, + "loss": 1.7123, + "step": 275 + }, + { + "epoch": 0.2835130970724191, + "grad_norm": 2.972714900970459, + "learning_rate": 4.469178082191781e-06, + "loss": 1.7693, + "step": 276 + }, + { + "epoch": 0.28454031843862354, + "grad_norm": 3.4248580932617188, + "learning_rate": 4.486301369863014e-06, + "loss": 1.7502, + "step": 277 + }, + { + "epoch": 0.28556753980482796, + "grad_norm": 2.9928481578826904, + "learning_rate": 4.503424657534247e-06, + "loss": 1.8064, + "step": 278 + }, + { + "epoch": 0.28659476117103233, + "grad_norm": 2.944093942642212, + "learning_rate": 4.52054794520548e-06, + "loss": 1.7075, + "step": 279 + }, + { + "epoch": 0.28762198253723675, + "grad_norm": 3.2028305530548096, + "learning_rate": 4.537671232876713e-06, + "loss": 1.7366, + "step": 280 + }, + { + "epoch": 0.28762198253723675, + "eval_loss": 1.9768972396850586, + "eval_runtime": 37.1078, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 280 + }, + { + "epoch": 0.2886492039034412, + "grad_norm": 2.8642148971557617, + "learning_rate": 4.554794520547945e-06, + "loss": 1.7909, + "step": 281 + }, + { + "epoch": 0.2896764252696456, + "grad_norm": 3.2716004848480225, + "learning_rate": 4.571917808219179e-06, + "loss": 1.7906, + "step": 282 + }, + { + "epoch": 0.29070364663585, + "grad_norm": 2.9601588249206543, + "learning_rate": 4.589041095890411e-06, + "loss": 1.7731, + "step": 283 + }, + { + "epoch": 0.29173086800205444, + "grad_norm": 3.1850085258483887, + "learning_rate": 4.606164383561644e-06, + "loss": 1.7335, + "step": 284 + }, + { + "epoch": 0.29275808936825887, + "grad_norm": 3.128908634185791, + "learning_rate": 4.6232876712328774e-06, + "loss": 1.748, + "step": 285 + }, + { + "epoch": 0.2937853107344633, + "grad_norm": 3.063014507293701, + "learning_rate": 4.64041095890411e-06, + "loss": 1.7632, + "step": 286 + }, + { + "epoch": 0.2948125321006677, + "grad_norm": 3.467817783355713, + "learning_rate": 4.657534246575343e-06, + "loss": 1.7295, + "step": 287 + }, + { + "epoch": 0.29583975346687214, + "grad_norm": 2.7431066036224365, + "learning_rate": 4.674657534246575e-06, + "loss": 1.7728, + "step": 288 + }, + { + "epoch": 0.2968669748330765, + "grad_norm": 3.271421432495117, + "learning_rate": 4.691780821917809e-06, + "loss": 1.7403, + "step": 289 + }, + { + "epoch": 0.2978941961992809, + "grad_norm": 2.908351421356201, + "learning_rate": 4.7089041095890414e-06, + "loss": 1.7983, + "step": 290 + }, + { + "epoch": 0.2978941961992809, + "eval_loss": 1.955810308456421, + "eval_runtime": 37.0914, + "eval_samples_per_second": 26.96, + "eval_steps_per_second": 1.132, + "step": 290 + }, + { + "epoch": 0.29892141756548535, + "grad_norm": 3.046114683151245, + "learning_rate": 4.726027397260274e-06, + "loss": 1.7612, + "step": 291 + }, + { + "epoch": 0.29994863893168977, + "grad_norm": 3.275486707687378, + "learning_rate": 4.7431506849315075e-06, + "loss": 1.7548, + "step": 292 + }, + { + "epoch": 0.3009758602978942, + "grad_norm": 2.819552183151245, + "learning_rate": 4.76027397260274e-06, + "loss": 1.7645, + "step": 293 + }, + { + "epoch": 0.3020030816640986, + "grad_norm": 3.234227418899536, + "learning_rate": 4.777397260273973e-06, + "loss": 1.715, + "step": 294 + }, + { + "epoch": 0.30303030303030304, + "grad_norm": 2.988713502883911, + "learning_rate": 4.7945205479452054e-06, + "loss": 1.7343, + "step": 295 + }, + { + "epoch": 0.30405752439650746, + "grad_norm": 3.195387601852417, + "learning_rate": 4.811643835616439e-06, + "loss": 1.7421, + "step": 296 + }, + { + "epoch": 0.3050847457627119, + "grad_norm": 3.617765426635742, + "learning_rate": 4.8287671232876716e-06, + "loss": 1.6661, + "step": 297 + }, + { + "epoch": 0.3061119671289163, + "grad_norm": 2.741631269454956, + "learning_rate": 4.845890410958904e-06, + "loss": 1.7188, + "step": 298 + }, + { + "epoch": 0.3071391884951207, + "grad_norm": 3.3230361938476562, + "learning_rate": 4.863013698630138e-06, + "loss": 1.721, + "step": 299 + }, + { + "epoch": 0.3081664098613251, + "grad_norm": 2.696352005004883, + "learning_rate": 4.88013698630137e-06, + "loss": 1.7137, + "step": 300 + }, + { + "epoch": 0.3081664098613251, + "eval_loss": 1.9677459001541138, + "eval_runtime": 37.1101, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 300 + }, + { + "epoch": 0.3091936312275295, + "grad_norm": 3.162966251373291, + "learning_rate": 4.897260273972603e-06, + "loss": 1.6956, + "step": 301 + }, + { + "epoch": 0.31022085259373394, + "grad_norm": 2.6055824756622314, + "learning_rate": 4.914383561643836e-06, + "loss": 1.8174, + "step": 302 + }, + { + "epoch": 0.31124807395993837, + "grad_norm": 2.816882610321045, + "learning_rate": 4.931506849315069e-06, + "loss": 1.7444, + "step": 303 + }, + { + "epoch": 0.3122752953261428, + "grad_norm": 2.9913320541381836, + "learning_rate": 4.948630136986302e-06, + "loss": 1.7744, + "step": 304 + }, + { + "epoch": 0.3133025166923472, + "grad_norm": 2.992614507675171, + "learning_rate": 4.965753424657534e-06, + "loss": 1.8155, + "step": 305 + }, + { + "epoch": 0.31432973805855163, + "grad_norm": 3.2482688426971436, + "learning_rate": 4.982876712328768e-06, + "loss": 1.6903, + "step": 306 + }, + { + "epoch": 0.31535695942475606, + "grad_norm": 2.7701141834259033, + "learning_rate": 5e-06, + "loss": 1.7228, + "step": 307 + }, + { + "epoch": 0.3163841807909605, + "grad_norm": 3.179245948791504, + "learning_rate": 4.998096688237534e-06, + "loss": 1.6644, + "step": 308 + }, + { + "epoch": 0.31741140215716485, + "grad_norm": 2.845163106918335, + "learning_rate": 4.996193376475067e-06, + "loss": 1.772, + "step": 309 + }, + { + "epoch": 0.31843862352336927, + "grad_norm": 3.1509721279144287, + "learning_rate": 4.9942900647126005e-06, + "loss": 1.6662, + "step": 310 + }, + { + "epoch": 0.31843862352336927, + "eval_loss": 1.9684854745864868, + "eval_runtime": 37.0507, + "eval_samples_per_second": 26.99, + "eval_steps_per_second": 1.134, + "step": 310 + }, + { + "epoch": 0.3194658448895737, + "grad_norm": 2.8594536781311035, + "learning_rate": 4.992386752950134e-06, + "loss": 1.6587, + "step": 311 + }, + { + "epoch": 0.3204930662557781, + "grad_norm": 3.358464002609253, + "learning_rate": 4.990483441187667e-06, + "loss": 1.7021, + "step": 312 + }, + { + "epoch": 0.32152028762198254, + "grad_norm": 2.6954872608184814, + "learning_rate": 4.9885801294252e-06, + "loss": 1.7301, + "step": 313 + }, + { + "epoch": 0.32254750898818696, + "grad_norm": 2.9256083965301514, + "learning_rate": 4.986676817662733e-06, + "loss": 1.7316, + "step": 314 + }, + { + "epoch": 0.3235747303543914, + "grad_norm": 2.840461254119873, + "learning_rate": 4.9847735059002665e-06, + "loss": 1.6848, + "step": 315 + }, + { + "epoch": 0.3246019517205958, + "grad_norm": 2.9138107299804688, + "learning_rate": 4.9828701941378e-06, + "loss": 1.765, + "step": 316 + }, + { + "epoch": 0.3256291730868002, + "grad_norm": 2.858368396759033, + "learning_rate": 4.980966882375333e-06, + "loss": 1.7078, + "step": 317 + }, + { + "epoch": 0.3266563944530046, + "grad_norm": 2.905238628387451, + "learning_rate": 4.979063570612867e-06, + "loss": 1.6765, + "step": 318 + }, + { + "epoch": 0.327683615819209, + "grad_norm": 3.117845296859741, + "learning_rate": 4.9771602588504e-06, + "loss": 1.6419, + "step": 319 + }, + { + "epoch": 0.32871083718541344, + "grad_norm": 2.5665884017944336, + "learning_rate": 4.975256947087933e-06, + "loss": 1.7424, + "step": 320 + }, + { + "epoch": 0.32871083718541344, + "eval_loss": 1.9528932571411133, + "eval_runtime": 37.2179, + "eval_samples_per_second": 26.869, + "eval_steps_per_second": 1.128, + "step": 320 + }, + { + "epoch": 0.32973805855161786, + "grad_norm": 2.499392032623291, + "learning_rate": 4.973353635325467e-06, + "loss": 1.6544, + "step": 321 + }, + { + "epoch": 0.3307652799178223, + "grad_norm": 2.8056774139404297, + "learning_rate": 4.971450323563e-06, + "loss": 1.7085, + "step": 322 + }, + { + "epoch": 0.3317925012840267, + "grad_norm": 2.850602388381958, + "learning_rate": 4.969547011800533e-06, + "loss": 1.6489, + "step": 323 + }, + { + "epoch": 0.33281972265023113, + "grad_norm": 2.783855438232422, + "learning_rate": 4.967643700038067e-06, + "loss": 1.7153, + "step": 324 + }, + { + "epoch": 0.33384694401643555, + "grad_norm": 2.7178030014038086, + "learning_rate": 4.9657403882756e-06, + "loss": 1.7104, + "step": 325 + }, + { + "epoch": 0.33487416538264, + "grad_norm": 2.7497682571411133, + "learning_rate": 4.9638370765131335e-06, + "loss": 1.6778, + "step": 326 + }, + { + "epoch": 0.3359013867488444, + "grad_norm": 2.688120126724243, + "learning_rate": 4.961933764750667e-06, + "loss": 1.6673, + "step": 327 + }, + { + "epoch": 0.33692860811504877, + "grad_norm": 2.661578416824341, + "learning_rate": 4.9600304529882e-06, + "loss": 1.7353, + "step": 328 + }, + { + "epoch": 0.3379558294812532, + "grad_norm": 2.759965419769287, + "learning_rate": 4.9581271412257335e-06, + "loss": 1.7506, + "step": 329 + }, + { + "epoch": 0.3389830508474576, + "grad_norm": 2.972900390625, + "learning_rate": 4.956223829463266e-06, + "loss": 1.7114, + "step": 330 + }, + { + "epoch": 0.3389830508474576, + "eval_loss": 1.9369672536849976, + "eval_runtime": 37.2876, + "eval_samples_per_second": 26.819, + "eval_steps_per_second": 1.126, + "step": 330 + }, + { + "epoch": 0.34001027221366203, + "grad_norm": 2.7194814682006836, + "learning_rate": 4.954320517700799e-06, + "loss": 1.621, + "step": 331 + }, + { + "epoch": 0.34103749357986646, + "grad_norm": 2.822052001953125, + "learning_rate": 4.952417205938333e-06, + "loss": 1.7215, + "step": 332 + }, + { + "epoch": 0.3420647149460709, + "grad_norm": 2.778348207473755, + "learning_rate": 4.950513894175866e-06, + "loss": 1.7049, + "step": 333 + }, + { + "epoch": 0.3430919363122753, + "grad_norm": 2.8513247966766357, + "learning_rate": 4.9486105824133995e-06, + "loss": 1.684, + "step": 334 + }, + { + "epoch": 0.3441191576784797, + "grad_norm": 2.5596683025360107, + "learning_rate": 4.946707270650933e-06, + "loss": 1.6992, + "step": 335 + }, + { + "epoch": 0.34514637904468415, + "grad_norm": 2.850728750228882, + "learning_rate": 4.944803958888466e-06, + "loss": 1.7095, + "step": 336 + }, + { + "epoch": 0.34617360041088857, + "grad_norm": 2.8497488498687744, + "learning_rate": 4.942900647126e-06, + "loss": 1.7361, + "step": 337 + }, + { + "epoch": 0.34720082177709294, + "grad_norm": 2.9133079051971436, + "learning_rate": 4.940997335363533e-06, + "loss": 1.6783, + "step": 338 + }, + { + "epoch": 0.34822804314329736, + "grad_norm": 2.7611589431762695, + "learning_rate": 4.939094023601066e-06, + "loss": 1.6407, + "step": 339 + }, + { + "epoch": 0.3492552645095018, + "grad_norm": 2.7190754413604736, + "learning_rate": 4.9371907118386e-06, + "loss": 1.6886, + "step": 340 + }, + { + "epoch": 0.3492552645095018, + "eval_loss": 1.9296809434890747, + "eval_runtime": 37.4339, + "eval_samples_per_second": 26.714, + "eval_steps_per_second": 1.122, + "step": 340 + }, + { + "epoch": 0.3502824858757062, + "grad_norm": 2.863070487976074, + "learning_rate": 4.935287400076133e-06, + "loss": 1.6912, + "step": 341 + }, + { + "epoch": 0.35130970724191063, + "grad_norm": 2.5978336334228516, + "learning_rate": 4.933384088313666e-06, + "loss": 1.7107, + "step": 342 + }, + { + "epoch": 0.35233692860811505, + "grad_norm": 2.8323776721954346, + "learning_rate": 4.9314807765512e-06, + "loss": 1.6723, + "step": 343 + }, + { + "epoch": 0.3533641499743195, + "grad_norm": 2.7128350734710693, + "learning_rate": 4.929577464788733e-06, + "loss": 1.5993, + "step": 344 + }, + { + "epoch": 0.3543913713405239, + "grad_norm": 2.618936061859131, + "learning_rate": 4.9276741530262665e-06, + "loss": 1.6769, + "step": 345 + }, + { + "epoch": 0.3554185927067283, + "grad_norm": 2.9575355052948, + "learning_rate": 4.9257708412638e-06, + "loss": 1.6743, + "step": 346 + }, + { + "epoch": 0.35644581407293274, + "grad_norm": 2.756464719772339, + "learning_rate": 4.923867529501332e-06, + "loss": 1.6858, + "step": 347 + }, + { + "epoch": 0.3574730354391371, + "grad_norm": 2.6581053733825684, + "learning_rate": 4.921964217738866e-06, + "loss": 1.6515, + "step": 348 + }, + { + "epoch": 0.35850025680534153, + "grad_norm": 2.890632390975952, + "learning_rate": 4.920060905976399e-06, + "loss": 1.7378, + "step": 349 + }, + { + "epoch": 0.35952747817154596, + "grad_norm": 2.7895255088806152, + "learning_rate": 4.9181575942139325e-06, + "loss": 1.7131, + "step": 350 + }, + { + "epoch": 0.35952747817154596, + "eval_loss": 1.9226924180984497, + "eval_runtime": 37.5186, + "eval_samples_per_second": 26.653, + "eval_steps_per_second": 1.119, + "step": 350 + }, + { + "epoch": 0.3605546995377504, + "grad_norm": 2.7698557376861572, + "learning_rate": 4.916254282451466e-06, + "loss": 1.71, + "step": 351 + }, + { + "epoch": 0.3615819209039548, + "grad_norm": 2.956592321395874, + "learning_rate": 4.914350970688999e-06, + "loss": 1.7487, + "step": 352 + }, + { + "epoch": 0.3626091422701592, + "grad_norm": 2.6062943935394287, + "learning_rate": 4.9124476589265325e-06, + "loss": 1.6486, + "step": 353 + }, + { + "epoch": 0.36363636363636365, + "grad_norm": 2.588418483734131, + "learning_rate": 4.910544347164066e-06, + "loss": 1.6468, + "step": 354 + }, + { + "epoch": 0.36466358500256807, + "grad_norm": 2.4665889739990234, + "learning_rate": 4.908641035401599e-06, + "loss": 1.7264, + "step": 355 + }, + { + "epoch": 0.3656908063687725, + "grad_norm": 2.701957941055298, + "learning_rate": 4.906737723639133e-06, + "loss": 1.7914, + "step": 356 + }, + { + "epoch": 0.3667180277349769, + "grad_norm": 2.68169903755188, + "learning_rate": 4.904834411876665e-06, + "loss": 1.6609, + "step": 357 + }, + { + "epoch": 0.3677452491011813, + "grad_norm": 2.633120536804199, + "learning_rate": 4.902931100114199e-06, + "loss": 1.6998, + "step": 358 + }, + { + "epoch": 0.3687724704673857, + "grad_norm": 2.549065113067627, + "learning_rate": 4.901027788351733e-06, + "loss": 1.6227, + "step": 359 + }, + { + "epoch": 0.3697996918335901, + "grad_norm": 2.6537365913391113, + "learning_rate": 4.899124476589266e-06, + "loss": 1.6678, + "step": 360 + }, + { + "epoch": 0.3697996918335901, + "eval_loss": 1.9240167140960693, + "eval_runtime": 37.3212, + "eval_samples_per_second": 26.794, + "eval_steps_per_second": 1.125, + "step": 360 + }, + { + "epoch": 0.37082691319979455, + "grad_norm": 2.8710854053497314, + "learning_rate": 4.8972211648267994e-06, + "loss": 1.6377, + "step": 361 + }, + { + "epoch": 0.371854134565999, + "grad_norm": 2.8168065547943115, + "learning_rate": 4.895317853064333e-06, + "loss": 1.6888, + "step": 362 + }, + { + "epoch": 0.3728813559322034, + "grad_norm": 2.7674975395202637, + "learning_rate": 4.893414541301866e-06, + "loss": 1.6532, + "step": 363 + }, + { + "epoch": 0.3739085772984078, + "grad_norm": 2.5604195594787598, + "learning_rate": 4.891511229539399e-06, + "loss": 1.6513, + "step": 364 + }, + { + "epoch": 0.37493579866461224, + "grad_norm": 2.5669636726379395, + "learning_rate": 4.889607917776932e-06, + "loss": 1.6832, + "step": 365 + }, + { + "epoch": 0.37596302003081666, + "grad_norm": 2.650104522705078, + "learning_rate": 4.887704606014465e-06, + "loss": 1.6836, + "step": 366 + }, + { + "epoch": 0.37699024139702103, + "grad_norm": 2.702894926071167, + "learning_rate": 4.885801294251999e-06, + "loss": 1.6916, + "step": 367 + }, + { + "epoch": 0.37801746276322545, + "grad_norm": 2.991715431213379, + "learning_rate": 4.883897982489532e-06, + "loss": 1.6146, + "step": 368 + }, + { + "epoch": 0.3790446841294299, + "grad_norm": 2.725093126296997, + "learning_rate": 4.8819946707270655e-06, + "loss": 1.7124, + "step": 369 + }, + { + "epoch": 0.3800719054956343, + "grad_norm": 2.8419744968414307, + "learning_rate": 4.880091358964599e-06, + "loss": 1.7002, + "step": 370 + }, + { + "epoch": 0.3800719054956343, + "eval_loss": 1.9024126529693604, + "eval_runtime": 37.231, + "eval_samples_per_second": 26.859, + "eval_steps_per_second": 1.128, + "step": 370 + }, + { + "epoch": 0.3810991268618387, + "grad_norm": 2.3915929794311523, + "learning_rate": 4.878188047202132e-06, + "loss": 1.6096, + "step": 371 + }, + { + "epoch": 0.38212634822804314, + "grad_norm": 2.577662944793701, + "learning_rate": 4.876284735439666e-06, + "loss": 1.772, + "step": 372 + }, + { + "epoch": 0.38315356959424757, + "grad_norm": 2.6686151027679443, + "learning_rate": 4.874381423677199e-06, + "loss": 1.7375, + "step": 373 + }, + { + "epoch": 0.384180790960452, + "grad_norm": 2.4534966945648193, + "learning_rate": 4.8724781119147315e-06, + "loss": 1.6603, + "step": 374 + }, + { + "epoch": 0.3852080123266564, + "grad_norm": 2.7308576107025146, + "learning_rate": 4.870574800152265e-06, + "loss": 1.6999, + "step": 375 + }, + { + "epoch": 0.38623523369286084, + "grad_norm": 2.551203727722168, + "learning_rate": 4.868671488389798e-06, + "loss": 1.6193, + "step": 376 + }, + { + "epoch": 0.3872624550590652, + "grad_norm": 2.78924822807312, + "learning_rate": 4.8667681766273316e-06, + "loss": 1.6359, + "step": 377 + }, + { + "epoch": 0.3882896764252696, + "grad_norm": 2.7972333431243896, + "learning_rate": 4.864864864864866e-06, + "loss": 1.6057, + "step": 378 + }, + { + "epoch": 0.38931689779147405, + "grad_norm": 2.721069097518921, + "learning_rate": 4.862961553102399e-06, + "loss": 1.6661, + "step": 379 + }, + { + "epoch": 0.39034411915767847, + "grad_norm": 2.817729949951172, + "learning_rate": 4.8610582413399325e-06, + "loss": 1.7053, + "step": 380 + }, + { + "epoch": 0.39034411915767847, + "eval_loss": 1.9177663326263428, + "eval_runtime": 37.2158, + "eval_samples_per_second": 26.87, + "eval_steps_per_second": 1.129, + "step": 380 + }, + { + "epoch": 0.3913713405238829, + "grad_norm": 2.800873279571533, + "learning_rate": 4.859154929577465e-06, + "loss": 1.6584, + "step": 381 + }, + { + "epoch": 0.3923985618900873, + "grad_norm": 2.6885769367218018, + "learning_rate": 4.857251617814998e-06, + "loss": 1.661, + "step": 382 + }, + { + "epoch": 0.39342578325629174, + "grad_norm": 2.56797456741333, + "learning_rate": 4.855348306052532e-06, + "loss": 1.7058, + "step": 383 + }, + { + "epoch": 0.39445300462249616, + "grad_norm": 2.5235650539398193, + "learning_rate": 4.853444994290065e-06, + "loss": 1.6977, + "step": 384 + }, + { + "epoch": 0.3954802259887006, + "grad_norm": 2.713711738586426, + "learning_rate": 4.8515416825275985e-06, + "loss": 1.6449, + "step": 385 + }, + { + "epoch": 0.396507447354905, + "grad_norm": 2.7639107704162598, + "learning_rate": 4.849638370765132e-06, + "loss": 1.7169, + "step": 386 + }, + { + "epoch": 0.3975346687211094, + "grad_norm": 2.4775755405426025, + "learning_rate": 4.847735059002665e-06, + "loss": 1.6423, + "step": 387 + }, + { + "epoch": 0.3985618900873138, + "grad_norm": 2.574907064437866, + "learning_rate": 4.8458317472401985e-06, + "loss": 1.5911, + "step": 388 + }, + { + "epoch": 0.3995891114535182, + "grad_norm": 2.667907953262329, + "learning_rate": 4.843928435477732e-06, + "loss": 1.6883, + "step": 389 + }, + { + "epoch": 0.40061633281972264, + "grad_norm": 2.592353343963623, + "learning_rate": 4.842025123715265e-06, + "loss": 1.639, + "step": 390 + }, + { + "epoch": 0.40061633281972264, + "eval_loss": 1.9100708961486816, + "eval_runtime": 37.2379, + "eval_samples_per_second": 26.854, + "eval_steps_per_second": 1.128, + "step": 390 + }, + { + "epoch": 0.40164355418592707, + "grad_norm": 2.8238866329193115, + "learning_rate": 4.840121811952798e-06, + "loss": 1.6835, + "step": 391 + }, + { + "epoch": 0.4026707755521315, + "grad_norm": 2.769272804260254, + "learning_rate": 4.838218500190331e-06, + "loss": 1.6651, + "step": 392 + }, + { + "epoch": 0.4036979969183359, + "grad_norm": 2.749093532562256, + "learning_rate": 4.8363151884278645e-06, + "loss": 1.6815, + "step": 393 + }, + { + "epoch": 0.40472521828454033, + "grad_norm": 2.7952616214752197, + "learning_rate": 4.834411876665398e-06, + "loss": 1.6622, + "step": 394 + }, + { + "epoch": 0.40575243965074476, + "grad_norm": 2.5004849433898926, + "learning_rate": 4.832508564902931e-06, + "loss": 1.6768, + "step": 395 + }, + { + "epoch": 0.4067796610169492, + "grad_norm": 2.938015937805176, + "learning_rate": 4.830605253140465e-06, + "loss": 1.6171, + "step": 396 + }, + { + "epoch": 0.40780688238315355, + "grad_norm": 2.737666130065918, + "learning_rate": 4.828701941377998e-06, + "loss": 1.6572, + "step": 397 + }, + { + "epoch": 0.40883410374935797, + "grad_norm": 2.6246249675750732, + "learning_rate": 4.826798629615531e-06, + "loss": 1.609, + "step": 398 + }, + { + "epoch": 0.4098613251155624, + "grad_norm": 2.6501305103302, + "learning_rate": 4.824895317853065e-06, + "loss": 1.6235, + "step": 399 + }, + { + "epoch": 0.4108885464817668, + "grad_norm": 2.89320707321167, + "learning_rate": 4.822992006090598e-06, + "loss": 1.6494, + "step": 400 + }, + { + "epoch": 0.4108885464817668, + "eval_loss": 1.8916113376617432, + "eval_runtime": 37.1849, + "eval_samples_per_second": 26.893, + "eval_steps_per_second": 1.129, + "step": 400 + }, + { + "epoch": 0.41191576784797124, + "grad_norm": 2.4857001304626465, + "learning_rate": 4.821088694328131e-06, + "loss": 1.594, + "step": 401 + }, + { + "epoch": 0.41294298921417566, + "grad_norm": 2.903836250305176, + "learning_rate": 4.819185382565665e-06, + "loss": 1.6319, + "step": 402 + }, + { + "epoch": 0.4139702105803801, + "grad_norm": 2.69596266746521, + "learning_rate": 4.817282070803198e-06, + "loss": 1.6843, + "step": 403 + }, + { + "epoch": 0.4149974319465845, + "grad_norm": 2.5105295181274414, + "learning_rate": 4.8153787590407315e-06, + "loss": 1.6725, + "step": 404 + }, + { + "epoch": 0.41602465331278893, + "grad_norm": 2.7997353076934814, + "learning_rate": 4.813475447278265e-06, + "loss": 1.6017, + "step": 405 + }, + { + "epoch": 0.41705187467899335, + "grad_norm": 2.6215803623199463, + "learning_rate": 4.811572135515798e-06, + "loss": 1.6824, + "step": 406 + }, + { + "epoch": 0.4180790960451977, + "grad_norm": 2.76534104347229, + "learning_rate": 4.809668823753332e-06, + "loss": 1.6111, + "step": 407 + }, + { + "epoch": 0.41910631741140214, + "grad_norm": 2.6977133750915527, + "learning_rate": 4.807765511990864e-06, + "loss": 1.6165, + "step": 408 + }, + { + "epoch": 0.42013353877760656, + "grad_norm": 2.6771748065948486, + "learning_rate": 4.8058622002283975e-06, + "loss": 1.648, + "step": 409 + }, + { + "epoch": 0.421160760143811, + "grad_norm": 2.746731996536255, + "learning_rate": 4.803958888465931e-06, + "loss": 1.6946, + "step": 410 + }, + { + "epoch": 0.421160760143811, + "eval_loss": 1.878269076347351, + "eval_runtime": 37.0935, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 410 + }, + { + "epoch": 0.4221879815100154, + "grad_norm": 2.688074827194214, + "learning_rate": 4.802055576703464e-06, + "loss": 1.6545, + "step": 411 + }, + { + "epoch": 0.42321520287621983, + "grad_norm": 2.7336084842681885, + "learning_rate": 4.8001522649409976e-06, + "loss": 1.7305, + "step": 412 + }, + { + "epoch": 0.42424242424242425, + "grad_norm": 2.6902427673339844, + "learning_rate": 4.798248953178531e-06, + "loss": 1.612, + "step": 413 + }, + { + "epoch": 0.4252696456086287, + "grad_norm": 2.5339174270629883, + "learning_rate": 4.796345641416064e-06, + "loss": 1.5947, + "step": 414 + }, + { + "epoch": 0.4262968669748331, + "grad_norm": 2.6804542541503906, + "learning_rate": 4.794442329653598e-06, + "loss": 1.634, + "step": 415 + }, + { + "epoch": 0.42732408834103747, + "grad_norm": 2.5764853954315186, + "learning_rate": 4.792539017891131e-06, + "loss": 1.6947, + "step": 416 + }, + { + "epoch": 0.4283513097072419, + "grad_norm": 2.4067094326019287, + "learning_rate": 4.790635706128664e-06, + "loss": 1.616, + "step": 417 + }, + { + "epoch": 0.4293785310734463, + "grad_norm": 2.576206684112549, + "learning_rate": 4.788732394366197e-06, + "loss": 1.6529, + "step": 418 + }, + { + "epoch": 0.43040575243965074, + "grad_norm": 2.57826828956604, + "learning_rate": 4.786829082603731e-06, + "loss": 1.5771, + "step": 419 + }, + { + "epoch": 0.43143297380585516, + "grad_norm": 2.5002858638763428, + "learning_rate": 4.7849257708412645e-06, + "loss": 1.5814, + "step": 420 + }, + { + "epoch": 0.43143297380585516, + "eval_loss": 1.8940727710723877, + "eval_runtime": 37.0932, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 420 + }, + { + "epoch": 0.4324601951720596, + "grad_norm": 2.618067502975464, + "learning_rate": 4.783022459078798e-06, + "loss": 1.6645, + "step": 421 + }, + { + "epoch": 0.433487416538264, + "grad_norm": 2.5745837688446045, + "learning_rate": 4.781119147316331e-06, + "loss": 1.602, + "step": 422 + }, + { + "epoch": 0.4345146379044684, + "grad_norm": 2.8009719848632812, + "learning_rate": 4.7792158355538645e-06, + "loss": 1.6269, + "step": 423 + }, + { + "epoch": 0.43554185927067285, + "grad_norm": 2.8486008644104004, + "learning_rate": 4.777312523791398e-06, + "loss": 1.6402, + "step": 424 + }, + { + "epoch": 0.43656908063687727, + "grad_norm": 2.5588459968566895, + "learning_rate": 4.77540921202893e-06, + "loss": 1.6122, + "step": 425 + }, + { + "epoch": 0.43759630200308164, + "grad_norm": 3.1455588340759277, + "learning_rate": 4.773505900266464e-06, + "loss": 1.6542, + "step": 426 + }, + { + "epoch": 0.43862352336928606, + "grad_norm": 2.6221370697021484, + "learning_rate": 4.771602588503997e-06, + "loss": 1.6339, + "step": 427 + }, + { + "epoch": 0.4396507447354905, + "grad_norm": 3.006816864013672, + "learning_rate": 4.7696992767415305e-06, + "loss": 1.6381, + "step": 428 + }, + { + "epoch": 0.4406779661016949, + "grad_norm": 2.7317590713500977, + "learning_rate": 4.767795964979064e-06, + "loss": 1.6689, + "step": 429 + }, + { + "epoch": 0.44170518746789933, + "grad_norm": 2.9259109497070312, + "learning_rate": 4.765892653216597e-06, + "loss": 1.6793, + "step": 430 + }, + { + "epoch": 0.44170518746789933, + "eval_loss": 1.8887399435043335, + "eval_runtime": 37.1334, + "eval_samples_per_second": 26.93, + "eval_steps_per_second": 1.131, + "step": 430 + }, + { + "epoch": 0.44273240883410375, + "grad_norm": 2.8546087741851807, + "learning_rate": 4.763989341454131e-06, + "loss": 1.6706, + "step": 431 + }, + { + "epoch": 0.4437596302003082, + "grad_norm": 2.6085422039031982, + "learning_rate": 4.762086029691664e-06, + "loss": 1.5855, + "step": 432 + }, + { + "epoch": 0.4447868515665126, + "grad_norm": 2.622277021408081, + "learning_rate": 4.760182717929197e-06, + "loss": 1.6068, + "step": 433 + }, + { + "epoch": 0.445814072932717, + "grad_norm": 2.5327494144439697, + "learning_rate": 4.758279406166731e-06, + "loss": 1.6303, + "step": 434 + }, + { + "epoch": 0.44684129429892144, + "grad_norm": 2.7626633644104004, + "learning_rate": 4.756376094404263e-06, + "loss": 1.5567, + "step": 435 + }, + { + "epoch": 0.4478685156651258, + "grad_norm": 2.613476514816284, + "learning_rate": 4.7544727826417966e-06, + "loss": 1.6345, + "step": 436 + }, + { + "epoch": 0.44889573703133023, + "grad_norm": 2.58027720451355, + "learning_rate": 4.75256947087933e-06, + "loss": 1.675, + "step": 437 + }, + { + "epoch": 0.44992295839753466, + "grad_norm": 2.6709048748016357, + "learning_rate": 4.750666159116863e-06, + "loss": 1.6335, + "step": 438 + }, + { + "epoch": 0.4509501797637391, + "grad_norm": 2.7493093013763428, + "learning_rate": 4.7487628473543975e-06, + "loss": 1.6061, + "step": 439 + }, + { + "epoch": 0.4519774011299435, + "grad_norm": 2.725672960281372, + "learning_rate": 4.746859535591931e-06, + "loss": 1.5515, + "step": 440 + }, + { + "epoch": 0.4519774011299435, + "eval_loss": 1.8779852390289307, + "eval_runtime": 37.3135, + "eval_samples_per_second": 26.8, + "eval_steps_per_second": 1.126, + "step": 440 + }, + { + "epoch": 0.4530046224961479, + "grad_norm": 2.784477472305298, + "learning_rate": 4.744956223829464e-06, + "loss": 1.6146, + "step": 441 + }, + { + "epoch": 0.45403184386235235, + "grad_norm": 2.758119821548462, + "learning_rate": 4.743052912066997e-06, + "loss": 1.5945, + "step": 442 + }, + { + "epoch": 0.45505906522855677, + "grad_norm": 2.9006264209747314, + "learning_rate": 4.74114960030453e-06, + "loss": 1.6461, + "step": 443 + }, + { + "epoch": 0.4560862865947612, + "grad_norm": 2.8473031520843506, + "learning_rate": 4.7392462885420635e-06, + "loss": 1.6565, + "step": 444 + }, + { + "epoch": 0.4571135079609656, + "grad_norm": 2.715806722640991, + "learning_rate": 4.737342976779597e-06, + "loss": 1.5414, + "step": 445 + }, + { + "epoch": 0.45814072932717, + "grad_norm": 2.6588337421417236, + "learning_rate": 4.73543966501713e-06, + "loss": 1.6284, + "step": 446 + }, + { + "epoch": 0.4591679506933744, + "grad_norm": 3.1152243614196777, + "learning_rate": 4.7335363532546636e-06, + "loss": 1.6237, + "step": 447 + }, + { + "epoch": 0.4601951720595788, + "grad_norm": 2.545490264892578, + "learning_rate": 4.731633041492197e-06, + "loss": 1.6149, + "step": 448 + }, + { + "epoch": 0.46122239342578325, + "grad_norm": 3.0324277877807617, + "learning_rate": 4.72972972972973e-06, + "loss": 1.6277, + "step": 449 + }, + { + "epoch": 0.4622496147919877, + "grad_norm": 2.885970115661621, + "learning_rate": 4.727826417967264e-06, + "loss": 1.6625, + "step": 450 + }, + { + "epoch": 0.4622496147919877, + "eval_loss": 1.8651624917984009, + "eval_runtime": 37.3486, + "eval_samples_per_second": 26.775, + "eval_steps_per_second": 1.125, + "step": 450 + }, + { + "epoch": 0.4632768361581921, + "grad_norm": 2.7258739471435547, + "learning_rate": 4.725923106204797e-06, + "loss": 1.6263, + "step": 451 + }, + { + "epoch": 0.4643040575243965, + "grad_norm": 2.631122589111328, + "learning_rate": 4.72401979444233e-06, + "loss": 1.6081, + "step": 452 + }, + { + "epoch": 0.46533127889060094, + "grad_norm": 2.5882441997528076, + "learning_rate": 4.722116482679863e-06, + "loss": 1.6237, + "step": 453 + }, + { + "epoch": 0.46635850025680536, + "grad_norm": 2.66310715675354, + "learning_rate": 4.720213170917396e-06, + "loss": 1.6306, + "step": 454 + }, + { + "epoch": 0.46738572162300973, + "grad_norm": 2.6906657218933105, + "learning_rate": 4.71830985915493e-06, + "loss": 1.6492, + "step": 455 + }, + { + "epoch": 0.46841294298921415, + "grad_norm": 2.6957409381866455, + "learning_rate": 4.716406547392463e-06, + "loss": 1.5993, + "step": 456 + }, + { + "epoch": 0.4694401643554186, + "grad_norm": 2.904311418533325, + "learning_rate": 4.714503235629996e-06, + "loss": 1.6633, + "step": 457 + }, + { + "epoch": 0.470467385721623, + "grad_norm": 2.6853256225585938, + "learning_rate": 4.71259992386753e-06, + "loss": 1.5824, + "step": 458 + }, + { + "epoch": 0.4714946070878274, + "grad_norm": 2.4767656326293945, + "learning_rate": 4.710696612105063e-06, + "loss": 1.6145, + "step": 459 + }, + { + "epoch": 0.47252182845403184, + "grad_norm": 2.5569417476654053, + "learning_rate": 4.708793300342596e-06, + "loss": 1.6116, + "step": 460 + }, + { + "epoch": 0.47252182845403184, + "eval_loss": 1.8785003423690796, + "eval_runtime": 37.401, + "eval_samples_per_second": 26.737, + "eval_steps_per_second": 1.123, + "step": 460 + }, + { + "epoch": 0.47354904982023627, + "grad_norm": 2.6568074226379395, + "learning_rate": 4.70688998858013e-06, + "loss": 1.6097, + "step": 461 + }, + { + "epoch": 0.4745762711864407, + "grad_norm": 2.6002280712127686, + "learning_rate": 4.704986676817663e-06, + "loss": 1.6025, + "step": 462 + }, + { + "epoch": 0.4756034925526451, + "grad_norm": 2.6055355072021484, + "learning_rate": 4.7030833650551965e-06, + "loss": 1.6322, + "step": 463 + }, + { + "epoch": 0.47663071391884954, + "grad_norm": 2.6288487911224365, + "learning_rate": 4.70118005329273e-06, + "loss": 1.594, + "step": 464 + }, + { + "epoch": 0.4776579352850539, + "grad_norm": 2.6981899738311768, + "learning_rate": 4.699276741530263e-06, + "loss": 1.6707, + "step": 465 + }, + { + "epoch": 0.4786851566512583, + "grad_norm": 2.6832222938537598, + "learning_rate": 4.697373429767797e-06, + "loss": 1.6435, + "step": 466 + }, + { + "epoch": 0.47971237801746275, + "grad_norm": 2.563159227371216, + "learning_rate": 4.69547011800533e-06, + "loss": 1.6249, + "step": 467 + }, + { + "epoch": 0.48073959938366717, + "grad_norm": 2.545778751373291, + "learning_rate": 4.693566806242863e-06, + "loss": 1.6309, + "step": 468 + }, + { + "epoch": 0.4817668207498716, + "grad_norm": 2.555227518081665, + "learning_rate": 4.691663494480397e-06, + "loss": 1.6128, + "step": 469 + }, + { + "epoch": 0.482794042116076, + "grad_norm": 2.612576723098755, + "learning_rate": 4.689760182717929e-06, + "loss": 1.6484, + "step": 470 + }, + { + "epoch": 0.482794042116076, + "eval_loss": 1.8684974908828735, + "eval_runtime": 37.2607, + "eval_samples_per_second": 26.838, + "eval_steps_per_second": 1.127, + "step": 470 + }, + { + "epoch": 0.48382126348228044, + "grad_norm": 2.4981396198272705, + "learning_rate": 4.6878568709554626e-06, + "loss": 1.5393, + "step": 471 + }, + { + "epoch": 0.48484848484848486, + "grad_norm": 2.793700695037842, + "learning_rate": 4.685953559192996e-06, + "loss": 1.608, + "step": 472 + }, + { + "epoch": 0.4858757062146893, + "grad_norm": 2.7605607509613037, + "learning_rate": 4.684050247430529e-06, + "loss": 1.6236, + "step": 473 + }, + { + "epoch": 0.4869029275808937, + "grad_norm": 2.413719892501831, + "learning_rate": 4.682146935668063e-06, + "loss": 1.5555, + "step": 474 + }, + { + "epoch": 0.4879301489470981, + "grad_norm": 2.9032249450683594, + "learning_rate": 4.680243623905596e-06, + "loss": 1.6042, + "step": 475 + }, + { + "epoch": 0.4889573703133025, + "grad_norm": 2.5445175170898438, + "learning_rate": 4.678340312143129e-06, + "loss": 1.6265, + "step": 476 + }, + { + "epoch": 0.4899845916795069, + "grad_norm": 2.6955716609954834, + "learning_rate": 4.676437000380663e-06, + "loss": 1.5879, + "step": 477 + }, + { + "epoch": 0.49101181304571134, + "grad_norm": 2.8399500846862793, + "learning_rate": 4.674533688618196e-06, + "loss": 1.7009, + "step": 478 + }, + { + "epoch": 0.49203903441191577, + "grad_norm": 2.635035991668701, + "learning_rate": 4.6726303768557295e-06, + "loss": 1.6381, + "step": 479 + }, + { + "epoch": 0.4930662557781202, + "grad_norm": 2.7609052658081055, + "learning_rate": 4.670727065093263e-06, + "loss": 1.5492, + "step": 480 + }, + { + "epoch": 0.4930662557781202, + "eval_loss": 1.8574503660202026, + "eval_runtime": 37.2173, + "eval_samples_per_second": 26.869, + "eval_steps_per_second": 1.129, + "step": 480 + }, + { + "epoch": 0.4940934771443246, + "grad_norm": 2.593766927719116, + "learning_rate": 4.668823753330796e-06, + "loss": 1.5693, + "step": 481 + }, + { + "epoch": 0.49512069851052903, + "grad_norm": 2.8000242710113525, + "learning_rate": 4.6669204415683295e-06, + "loss": 1.5998, + "step": 482 + }, + { + "epoch": 0.49614791987673346, + "grad_norm": 2.522688627243042, + "learning_rate": 4.665017129805863e-06, + "loss": 1.6653, + "step": 483 + }, + { + "epoch": 0.4971751412429379, + "grad_norm": 2.854534864425659, + "learning_rate": 4.663113818043396e-06, + "loss": 1.6393, + "step": 484 + }, + { + "epoch": 0.49820236260914225, + "grad_norm": 2.5333967208862305, + "learning_rate": 4.66121050628093e-06, + "loss": 1.606, + "step": 485 + }, + { + "epoch": 0.49922958397534667, + "grad_norm": 2.565260171890259, + "learning_rate": 4.659307194518463e-06, + "loss": 1.6471, + "step": 486 + }, + { + "epoch": 0.5002568053415511, + "grad_norm": 2.570220947265625, + "learning_rate": 4.6574038827559955e-06, + "loss": 1.6023, + "step": 487 + }, + { + "epoch": 0.5012840267077555, + "grad_norm": 2.460740089416504, + "learning_rate": 4.655500570993529e-06, + "loss": 1.6208, + "step": 488 + }, + { + "epoch": 0.50231124807396, + "grad_norm": 2.6339426040649414, + "learning_rate": 4.653597259231062e-06, + "loss": 1.6533, + "step": 489 + }, + { + "epoch": 0.5033384694401644, + "grad_norm": 2.681044578552246, + "learning_rate": 4.651693947468596e-06, + "loss": 1.5437, + "step": 490 + }, + { + "epoch": 0.5033384694401644, + "eval_loss": 1.8701337575912476, + "eval_runtime": 37.1714, + "eval_samples_per_second": 26.902, + "eval_steps_per_second": 1.13, + "step": 490 + }, + { + "epoch": 0.5043656908063687, + "grad_norm": 2.6900649070739746, + "learning_rate": 4.649790635706129e-06, + "loss": 1.6816, + "step": 491 + }, + { + "epoch": 0.5053929121725732, + "grad_norm": 2.6106340885162354, + "learning_rate": 4.647887323943662e-06, + "loss": 1.5784, + "step": 492 + }, + { + "epoch": 0.5064201335387776, + "grad_norm": 3.0704469680786133, + "learning_rate": 4.645984012181196e-06, + "loss": 1.6065, + "step": 493 + }, + { + "epoch": 0.507447354904982, + "grad_norm": 2.648425817489624, + "learning_rate": 4.644080700418729e-06, + "loss": 1.5134, + "step": 494 + }, + { + "epoch": 0.5084745762711864, + "grad_norm": 2.509066343307495, + "learning_rate": 4.642177388656262e-06, + "loss": 1.6324, + "step": 495 + }, + { + "epoch": 0.5095017976373909, + "grad_norm": 2.8628857135772705, + "learning_rate": 4.640274076893796e-06, + "loss": 1.5957, + "step": 496 + }, + { + "epoch": 0.5105290190035953, + "grad_norm": 2.83579683303833, + "learning_rate": 4.638370765131328e-06, + "loss": 1.6556, + "step": 497 + }, + { + "epoch": 0.5115562403697997, + "grad_norm": 2.6618151664733887, + "learning_rate": 4.636467453368862e-06, + "loss": 1.5738, + "step": 498 + }, + { + "epoch": 0.5125834617360041, + "grad_norm": 2.936943769454956, + "learning_rate": 4.634564141606396e-06, + "loss": 1.5597, + "step": 499 + }, + { + "epoch": 0.5136106831022085, + "grad_norm": 2.8520913124084473, + "learning_rate": 4.632660829843929e-06, + "loss": 1.6316, + "step": 500 + }, + { + "epoch": 0.5136106831022085, + "eval_loss": 1.8548446893692017, + "eval_runtime": 37.1878, + "eval_samples_per_second": 26.891, + "eval_steps_per_second": 1.129, + "step": 500 + }, + { + "epoch": 0.514637904468413, + "grad_norm": 2.446373462677002, + "learning_rate": 4.630757518081463e-06, + "loss": 1.5533, + "step": 501 + }, + { + "epoch": 0.5156651258346173, + "grad_norm": 2.8339035511016846, + "learning_rate": 4.628854206318996e-06, + "loss": 1.6904, + "step": 502 + }, + { + "epoch": 0.5166923472008218, + "grad_norm": 2.664249897003174, + "learning_rate": 4.626950894556529e-06, + "loss": 1.5926, + "step": 503 + }, + { + "epoch": 0.5177195685670262, + "grad_norm": 2.681185483932495, + "learning_rate": 4.625047582794062e-06, + "loss": 1.6348, + "step": 504 + }, + { + "epoch": 0.5187467899332306, + "grad_norm": 2.468263864517212, + "learning_rate": 4.623144271031595e-06, + "loss": 1.5477, + "step": 505 + }, + { + "epoch": 0.519774011299435, + "grad_norm": 2.539066791534424, + "learning_rate": 4.6212409592691286e-06, + "loss": 1.5615, + "step": 506 + }, + { + "epoch": 0.5208012326656395, + "grad_norm": 2.822235345840454, + "learning_rate": 4.619337647506662e-06, + "loss": 1.6398, + "step": 507 + }, + { + "epoch": 0.5218284540318439, + "grad_norm": 2.6063523292541504, + "learning_rate": 4.617434335744195e-06, + "loss": 1.5992, + "step": 508 + }, + { + "epoch": 0.5228556753980482, + "grad_norm": 2.624810218811035, + "learning_rate": 4.615531023981729e-06, + "loss": 1.5708, + "step": 509 + }, + { + "epoch": 0.5238828967642527, + "grad_norm": 2.6324198246002197, + "learning_rate": 4.613627712219262e-06, + "loss": 1.5865, + "step": 510 + }, + { + "epoch": 0.5238828967642527, + "eval_loss": 1.8442509174346924, + "eval_runtime": 36.9899, + "eval_samples_per_second": 27.034, + "eval_steps_per_second": 1.135, + "step": 510 + }, + { + "epoch": 0.5249101181304571, + "grad_norm": 2.5496973991394043, + "learning_rate": 4.611724400456795e-06, + "loss": 1.6048, + "step": 511 + }, + { + "epoch": 0.5259373394966615, + "grad_norm": 2.4150168895721436, + "learning_rate": 4.609821088694329e-06, + "loss": 1.6009, + "step": 512 + }, + { + "epoch": 0.5269645608628659, + "grad_norm": 2.8799378871917725, + "learning_rate": 4.607917776931862e-06, + "loss": 1.5761, + "step": 513 + }, + { + "epoch": 0.5279917822290704, + "grad_norm": 2.749016761779785, + "learning_rate": 4.606014465169395e-06, + "loss": 1.6202, + "step": 514 + }, + { + "epoch": 0.5290190035952748, + "grad_norm": 2.6993141174316406, + "learning_rate": 4.604111153406928e-06, + "loss": 1.632, + "step": 515 + }, + { + "epoch": 0.5300462249614792, + "grad_norm": 2.9027445316314697, + "learning_rate": 4.602207841644461e-06, + "loss": 1.5352, + "step": 516 + }, + { + "epoch": 0.5310734463276836, + "grad_norm": 2.874267816543579, + "learning_rate": 4.600304529881995e-06, + "loss": 1.642, + "step": 517 + }, + { + "epoch": 0.5321006676938881, + "grad_norm": 2.6057357788085938, + "learning_rate": 4.598401218119528e-06, + "loss": 1.6179, + "step": 518 + }, + { + "epoch": 0.5331278890600925, + "grad_norm": 2.658812999725342, + "learning_rate": 4.5964979063570614e-06, + "loss": 1.6285, + "step": 519 + }, + { + "epoch": 0.5341551104262968, + "grad_norm": 2.667215585708618, + "learning_rate": 4.594594594594596e-06, + "loss": 1.5694, + "step": 520 + }, + { + "epoch": 0.5341551104262968, + "eval_loss": 1.8490455150604248, + "eval_runtime": 37.0914, + "eval_samples_per_second": 26.96, + "eval_steps_per_second": 1.132, + "step": 520 + }, + { + "epoch": 0.5351823317925013, + "grad_norm": 2.5979607105255127, + "learning_rate": 4.592691282832128e-06, + "loss": 1.6091, + "step": 521 + }, + { + "epoch": 0.5362095531587057, + "grad_norm": 2.552816152572632, + "learning_rate": 4.5907879710696615e-06, + "loss": 1.6145, + "step": 522 + }, + { + "epoch": 0.5372367745249101, + "grad_norm": 2.575045347213745, + "learning_rate": 4.588884659307195e-06, + "loss": 1.5444, + "step": 523 + }, + { + "epoch": 0.5382639958911145, + "grad_norm": 2.650202751159668, + "learning_rate": 4.586981347544728e-06, + "loss": 1.6181, + "step": 524 + }, + { + "epoch": 0.539291217257319, + "grad_norm": 2.7869739532470703, + "learning_rate": 4.585078035782262e-06, + "loss": 1.563, + "step": 525 + }, + { + "epoch": 0.5403184386235234, + "grad_norm": 2.6828651428222656, + "learning_rate": 4.583174724019795e-06, + "loss": 1.6004, + "step": 526 + }, + { + "epoch": 0.5413456599897278, + "grad_norm": 2.552387237548828, + "learning_rate": 4.581271412257328e-06, + "loss": 1.5351, + "step": 527 + }, + { + "epoch": 0.5423728813559322, + "grad_norm": 2.785661458969116, + "learning_rate": 4.579368100494862e-06, + "loss": 1.5264, + "step": 528 + }, + { + "epoch": 0.5434001027221366, + "grad_norm": 2.832939863204956, + "learning_rate": 4.577464788732395e-06, + "loss": 1.6261, + "step": 529 + }, + { + "epoch": 0.544427324088341, + "grad_norm": 2.641766309738159, + "learning_rate": 4.575561476969928e-06, + "loss": 1.6019, + "step": 530 + }, + { + "epoch": 0.544427324088341, + "eval_loss": 1.8477164506912231, + "eval_runtime": 37.1444, + "eval_samples_per_second": 26.922, + "eval_steps_per_second": 1.131, + "step": 530 + }, + { + "epoch": 0.5454545454545454, + "grad_norm": 3.0502524375915527, + "learning_rate": 4.573658165207461e-06, + "loss": 1.6436, + "step": 531 + }, + { + "epoch": 0.5464817668207499, + "grad_norm": 2.64772891998291, + "learning_rate": 4.571754853444994e-06, + "loss": 1.6085, + "step": 532 + }, + { + "epoch": 0.5475089881869543, + "grad_norm": 2.549692392349243, + "learning_rate": 4.569851541682528e-06, + "loss": 1.5977, + "step": 533 + }, + { + "epoch": 0.5485362095531587, + "grad_norm": 2.719322681427002, + "learning_rate": 4.567948229920061e-06, + "loss": 1.5659, + "step": 534 + }, + { + "epoch": 0.5495634309193631, + "grad_norm": 2.5005698204040527, + "learning_rate": 4.566044918157594e-06, + "loss": 1.5301, + "step": 535 + }, + { + "epoch": 0.5505906522855676, + "grad_norm": 2.7379472255706787, + "learning_rate": 4.564141606395128e-06, + "loss": 1.5721, + "step": 536 + }, + { + "epoch": 0.551617873651772, + "grad_norm": 2.6673905849456787, + "learning_rate": 4.562238294632661e-06, + "loss": 1.6105, + "step": 537 + }, + { + "epoch": 0.5526450950179764, + "grad_norm": 2.6745669841766357, + "learning_rate": 4.5603349828701945e-06, + "loss": 1.577, + "step": 538 + }, + { + "epoch": 0.5536723163841808, + "grad_norm": 2.593475341796875, + "learning_rate": 4.558431671107728e-06, + "loss": 1.6084, + "step": 539 + }, + { + "epoch": 0.5546995377503852, + "grad_norm": 2.6577041149139404, + "learning_rate": 4.556528359345261e-06, + "loss": 1.6417, + "step": 540 + }, + { + "epoch": 0.5546995377503852, + "eval_loss": 1.839761734008789, + "eval_runtime": 37.1765, + "eval_samples_per_second": 26.899, + "eval_steps_per_second": 1.13, + "step": 540 + }, + { + "epoch": 0.5557267591165896, + "grad_norm": 2.6834933757781982, + "learning_rate": 4.5546250475827946e-06, + "loss": 1.5803, + "step": 541 + }, + { + "epoch": 0.556753980482794, + "grad_norm": 2.464665174484253, + "learning_rate": 4.552721735820328e-06, + "loss": 1.5324, + "step": 542 + }, + { + "epoch": 0.5577812018489985, + "grad_norm": 2.4488964080810547, + "learning_rate": 4.550818424057861e-06, + "loss": 1.5767, + "step": 543 + }, + { + "epoch": 0.5588084232152029, + "grad_norm": 2.763124465942383, + "learning_rate": 4.548915112295395e-06, + "loss": 1.5903, + "step": 544 + }, + { + "epoch": 0.5598356445814073, + "grad_norm": 2.7473502159118652, + "learning_rate": 4.547011800532928e-06, + "loss": 1.5966, + "step": 545 + }, + { + "epoch": 0.5608628659476117, + "grad_norm": 3.050267219543457, + "learning_rate": 4.545108488770461e-06, + "loss": 1.5492, + "step": 546 + }, + { + "epoch": 0.5618900873138162, + "grad_norm": 2.910439968109131, + "learning_rate": 4.543205177007995e-06, + "loss": 1.6357, + "step": 547 + }, + { + "epoch": 0.5629173086800205, + "grad_norm": 2.5621109008789062, + "learning_rate": 4.541301865245527e-06, + "loss": 1.5693, + "step": 548 + }, + { + "epoch": 0.5639445300462249, + "grad_norm": 2.7492010593414307, + "learning_rate": 4.539398553483061e-06, + "loss": 1.615, + "step": 549 + }, + { + "epoch": 0.5649717514124294, + "grad_norm": 2.6651856899261475, + "learning_rate": 4.537495241720594e-06, + "loss": 1.5804, + "step": 550 + }, + { + "epoch": 0.5649717514124294, + "eval_loss": 1.8344007730484009, + "eval_runtime": 37.1218, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 550 + }, + { + "epoch": 0.5659989727786338, + "grad_norm": 2.5933356285095215, + "learning_rate": 4.535591929958127e-06, + "loss": 1.6108, + "step": 551 + }, + { + "epoch": 0.5670261941448382, + "grad_norm": 2.660522222518921, + "learning_rate": 4.533688618195661e-06, + "loss": 1.5977, + "step": 552 + }, + { + "epoch": 0.5680534155110426, + "grad_norm": 2.5636332035064697, + "learning_rate": 4.531785306433194e-06, + "loss": 1.5802, + "step": 553 + }, + { + "epoch": 0.5690806368772471, + "grad_norm": 2.759533166885376, + "learning_rate": 4.529881994670727e-06, + "loss": 1.5656, + "step": 554 + }, + { + "epoch": 0.5701078582434514, + "grad_norm": 2.6530680656433105, + "learning_rate": 4.527978682908261e-06, + "loss": 1.5693, + "step": 555 + }, + { + "epoch": 0.5711350796096559, + "grad_norm": 2.688427448272705, + "learning_rate": 4.526075371145794e-06, + "loss": 1.5769, + "step": 556 + }, + { + "epoch": 0.5721623009758603, + "grad_norm": 2.5163164138793945, + "learning_rate": 4.5241720593833275e-06, + "loss": 1.593, + "step": 557 + }, + { + "epoch": 0.5731895223420647, + "grad_norm": 2.612612724304199, + "learning_rate": 4.52226874762086e-06, + "loss": 1.5854, + "step": 558 + }, + { + "epoch": 0.5742167437082691, + "grad_norm": 2.647853374481201, + "learning_rate": 4.520365435858393e-06, + "loss": 1.6085, + "step": 559 + }, + { + "epoch": 0.5752439650744735, + "grad_norm": 2.7160937786102295, + "learning_rate": 4.518462124095928e-06, + "loss": 1.6353, + "step": 560 + }, + { + "epoch": 0.5752439650744735, + "eval_loss": 1.8585809469223022, + "eval_runtime": 37.328, + "eval_samples_per_second": 26.79, + "eval_steps_per_second": 1.125, + "step": 560 + }, + { + "epoch": 0.576271186440678, + "grad_norm": 2.9025261402130127, + "learning_rate": 4.516558812333461e-06, + "loss": 1.5903, + "step": 561 + }, + { + "epoch": 0.5772984078068824, + "grad_norm": 2.5665547847747803, + "learning_rate": 4.514655500570994e-06, + "loss": 1.5775, + "step": 562 + }, + { + "epoch": 0.5783256291730868, + "grad_norm": 2.591381311416626, + "learning_rate": 4.512752188808528e-06, + "loss": 1.6387, + "step": 563 + }, + { + "epoch": 0.5793528505392912, + "grad_norm": 2.7540411949157715, + "learning_rate": 4.510848877046061e-06, + "loss": 1.6432, + "step": 564 + }, + { + "epoch": 0.5803800719054957, + "grad_norm": 2.6911778450012207, + "learning_rate": 4.5089455652835936e-06, + "loss": 1.6078, + "step": 565 + }, + { + "epoch": 0.5814072932717, + "grad_norm": 2.609619140625, + "learning_rate": 4.507042253521127e-06, + "loss": 1.5709, + "step": 566 + }, + { + "epoch": 0.5824345146379045, + "grad_norm": 2.4930672645568848, + "learning_rate": 4.50513894175866e-06, + "loss": 1.5305, + "step": 567 + }, + { + "epoch": 0.5834617360041089, + "grad_norm": 2.490992307662964, + "learning_rate": 4.503235629996194e-06, + "loss": 1.61, + "step": 568 + }, + { + "epoch": 0.5844889573703133, + "grad_norm": 2.7229626178741455, + "learning_rate": 4.501332318233727e-06, + "loss": 1.5996, + "step": 569 + }, + { + "epoch": 0.5855161787365177, + "grad_norm": 2.5289995670318604, + "learning_rate": 4.49942900647126e-06, + "loss": 1.5624, + "step": 570 + }, + { + "epoch": 0.5855161787365177, + "eval_loss": 1.867126226425171, + "eval_runtime": 37.4117, + "eval_samples_per_second": 26.73, + "eval_steps_per_second": 1.123, + "step": 570 + }, + { + "epoch": 0.5865434001027221, + "grad_norm": 2.6244945526123047, + "learning_rate": 4.497525694708794e-06, + "loss": 1.6154, + "step": 571 + }, + { + "epoch": 0.5875706214689266, + "grad_norm": 2.4343698024749756, + "learning_rate": 4.495622382946327e-06, + "loss": 1.5967, + "step": 572 + }, + { + "epoch": 0.588597842835131, + "grad_norm": 2.5899314880371094, + "learning_rate": 4.4937190711838605e-06, + "loss": 1.5601, + "step": 573 + }, + { + "epoch": 0.5896250642013354, + "grad_norm": 2.829028367996216, + "learning_rate": 4.491815759421394e-06, + "loss": 1.6315, + "step": 574 + }, + { + "epoch": 0.5906522855675398, + "grad_norm": 2.748849868774414, + "learning_rate": 4.489912447658926e-06, + "loss": 1.5591, + "step": 575 + }, + { + "epoch": 0.5916795069337443, + "grad_norm": 2.7016541957855225, + "learning_rate": 4.48800913589646e-06, + "loss": 1.5575, + "step": 576 + }, + { + "epoch": 0.5927067282999486, + "grad_norm": 2.6260931491851807, + "learning_rate": 4.486105824133993e-06, + "loss": 1.5823, + "step": 577 + }, + { + "epoch": 0.593733949666153, + "grad_norm": 2.797224998474121, + "learning_rate": 4.4842025123715264e-06, + "loss": 1.5902, + "step": 578 + }, + { + "epoch": 0.5947611710323575, + "grad_norm": 2.671342611312866, + "learning_rate": 4.48229920060906e-06, + "loss": 1.551, + "step": 579 + }, + { + "epoch": 0.5957883923985619, + "grad_norm": 2.626988172531128, + "learning_rate": 4.480395888846594e-06, + "loss": 1.5312, + "step": 580 + }, + { + "epoch": 0.5957883923985619, + "eval_loss": 1.8685871362686157, + "eval_runtime": 37.388, + "eval_samples_per_second": 26.747, + "eval_steps_per_second": 1.123, + "step": 580 + }, + { + "epoch": 0.5968156137647663, + "grad_norm": 2.8422539234161377, + "learning_rate": 4.478492577084127e-06, + "loss": 1.6393, + "step": 581 + }, + { + "epoch": 0.5978428351309707, + "grad_norm": 2.6877822875976562, + "learning_rate": 4.47658926532166e-06, + "loss": 1.5376, + "step": 582 + }, + { + "epoch": 0.5988700564971752, + "grad_norm": 2.736452341079712, + "learning_rate": 4.474685953559193e-06, + "loss": 1.5798, + "step": 583 + }, + { + "epoch": 0.5998972778633795, + "grad_norm": 2.732513904571533, + "learning_rate": 4.472782641796727e-06, + "loss": 1.6897, + "step": 584 + }, + { + "epoch": 0.600924499229584, + "grad_norm": 2.8732001781463623, + "learning_rate": 4.47087933003426e-06, + "loss": 1.533, + "step": 585 + }, + { + "epoch": 0.6019517205957884, + "grad_norm": 2.6099846363067627, + "learning_rate": 4.468976018271793e-06, + "loss": 1.5481, + "step": 586 + }, + { + "epoch": 0.6029789419619929, + "grad_norm": 2.754856586456299, + "learning_rate": 4.467072706509327e-06, + "loss": 1.5497, + "step": 587 + }, + { + "epoch": 0.6040061633281972, + "grad_norm": 2.6755611896514893, + "learning_rate": 4.46516939474686e-06, + "loss": 1.5882, + "step": 588 + }, + { + "epoch": 0.6050333846944016, + "grad_norm": 2.632248640060425, + "learning_rate": 4.463266082984393e-06, + "loss": 1.5481, + "step": 589 + }, + { + "epoch": 0.6060606060606061, + "grad_norm": 2.6425793170928955, + "learning_rate": 4.461362771221927e-06, + "loss": 1.6607, + "step": 590 + }, + { + "epoch": 0.6060606060606061, + "eval_loss": 1.863004446029663, + "eval_runtime": 37.2064, + "eval_samples_per_second": 26.877, + "eval_steps_per_second": 1.129, + "step": 590 + }, + { + "epoch": 0.6070878274268104, + "grad_norm": 2.6340153217315674, + "learning_rate": 4.45945945945946e-06, + "loss": 1.5474, + "step": 591 + }, + { + "epoch": 0.6081150487930149, + "grad_norm": 2.7601606845855713, + "learning_rate": 4.457556147696993e-06, + "loss": 1.5569, + "step": 592 + }, + { + "epoch": 0.6091422701592193, + "grad_norm": 2.8151822090148926, + "learning_rate": 4.455652835934526e-06, + "loss": 1.6441, + "step": 593 + }, + { + "epoch": 0.6101694915254238, + "grad_norm": 2.757577419281006, + "learning_rate": 4.453749524172059e-06, + "loss": 1.6339, + "step": 594 + }, + { + "epoch": 0.6111967128916281, + "grad_norm": 2.6828997135162354, + "learning_rate": 4.451846212409593e-06, + "loss": 1.5885, + "step": 595 + }, + { + "epoch": 0.6122239342578326, + "grad_norm": 2.6645700931549072, + "learning_rate": 4.449942900647126e-06, + "loss": 1.5866, + "step": 596 + }, + { + "epoch": 0.613251155624037, + "grad_norm": 2.572789430618286, + "learning_rate": 4.4480395888846595e-06, + "loss": 1.5255, + "step": 597 + }, + { + "epoch": 0.6142783769902413, + "grad_norm": 2.6281981468200684, + "learning_rate": 4.446136277122193e-06, + "loss": 1.5905, + "step": 598 + }, + { + "epoch": 0.6153055983564458, + "grad_norm": 2.6164042949676514, + "learning_rate": 4.444232965359726e-06, + "loss": 1.5274, + "step": 599 + }, + { + "epoch": 0.6163328197226502, + "grad_norm": 2.606343984603882, + "learning_rate": 4.4423296535972596e-06, + "loss": 1.5846, + "step": 600 + }, + { + "epoch": 0.6163328197226502, + "eval_loss": 1.8634170293807983, + "eval_runtime": 37.1568, + "eval_samples_per_second": 26.913, + "eval_steps_per_second": 1.13, + "step": 600 + }, + { + "epoch": 0.6173600410888547, + "grad_norm": 2.8743226528167725, + "learning_rate": 4.440426341834793e-06, + "loss": 1.538, + "step": 601 + }, + { + "epoch": 0.618387262455059, + "grad_norm": 2.5392212867736816, + "learning_rate": 4.438523030072326e-06, + "loss": 1.585, + "step": 602 + }, + { + "epoch": 0.6194144838212635, + "grad_norm": 2.715684413909912, + "learning_rate": 4.43661971830986e-06, + "loss": 1.6028, + "step": 603 + }, + { + "epoch": 0.6204417051874679, + "grad_norm": 2.514934778213501, + "learning_rate": 4.434716406547393e-06, + "loss": 1.5525, + "step": 604 + }, + { + "epoch": 0.6214689265536724, + "grad_norm": 2.685807943344116, + "learning_rate": 4.432813094784926e-06, + "loss": 1.5456, + "step": 605 + }, + { + "epoch": 0.6224961479198767, + "grad_norm": 2.848557949066162, + "learning_rate": 4.43090978302246e-06, + "loss": 1.6017, + "step": 606 + }, + { + "epoch": 0.6235233692860811, + "grad_norm": 2.6820950508117676, + "learning_rate": 4.429006471259993e-06, + "loss": 1.573, + "step": 607 + }, + { + "epoch": 0.6245505906522856, + "grad_norm": 2.493544340133667, + "learning_rate": 4.4271031594975265e-06, + "loss": 1.5576, + "step": 608 + }, + { + "epoch": 0.6255778120184899, + "grad_norm": 2.6240968704223633, + "learning_rate": 4.425199847735059e-06, + "loss": 1.5312, + "step": 609 + }, + { + "epoch": 0.6266050333846944, + "grad_norm": 2.516113042831421, + "learning_rate": 4.423296535972592e-06, + "loss": 1.5284, + "step": 610 + }, + { + "epoch": 0.6266050333846944, + "eval_loss": 1.8370482921600342, + "eval_runtime": 36.9401, + "eval_samples_per_second": 27.071, + "eval_steps_per_second": 1.137, + "step": 610 + }, + { + "epoch": 0.6276322547508988, + "grad_norm": 2.8341944217681885, + "learning_rate": 4.421393224210126e-06, + "loss": 1.5609, + "step": 611 + }, + { + "epoch": 0.6286594761171033, + "grad_norm": 2.5433242321014404, + "learning_rate": 4.419489912447659e-06, + "loss": 1.5253, + "step": 612 + }, + { + "epoch": 0.6296866974833076, + "grad_norm": 2.6761138439178467, + "learning_rate": 4.4175866006851924e-06, + "loss": 1.6171, + "step": 613 + }, + { + "epoch": 0.6307139188495121, + "grad_norm": 2.570707321166992, + "learning_rate": 4.415683288922726e-06, + "loss": 1.5766, + "step": 614 + }, + { + "epoch": 0.6317411402157165, + "grad_norm": 2.5421109199523926, + "learning_rate": 4.413779977160259e-06, + "loss": 1.5616, + "step": 615 + }, + { + "epoch": 0.632768361581921, + "grad_norm": 2.6767091751098633, + "learning_rate": 4.4118766653977925e-06, + "loss": 1.5531, + "step": 616 + }, + { + "epoch": 0.6337955829481253, + "grad_norm": 2.7730000019073486, + "learning_rate": 4.409973353635326e-06, + "loss": 1.6224, + "step": 617 + }, + { + "epoch": 0.6348228043143297, + "grad_norm": 2.7248494625091553, + "learning_rate": 4.408070041872859e-06, + "loss": 1.5752, + "step": 618 + }, + { + "epoch": 0.6358500256805342, + "grad_norm": 2.5071568489074707, + "learning_rate": 4.406166730110393e-06, + "loss": 1.5178, + "step": 619 + }, + { + "epoch": 0.6368772470467385, + "grad_norm": 2.5583243370056152, + "learning_rate": 4.404263418347925e-06, + "loss": 1.5298, + "step": 620 + }, + { + "epoch": 0.6368772470467385, + "eval_loss": 1.833156943321228, + "eval_runtime": 37.1108, + "eval_samples_per_second": 26.946, + "eval_steps_per_second": 1.132, + "step": 620 + }, + { + "epoch": 0.637904468412943, + "grad_norm": 2.492332696914673, + "learning_rate": 4.402360106585459e-06, + "loss": 1.5443, + "step": 621 + }, + { + "epoch": 0.6389316897791474, + "grad_norm": 2.5734798908233643, + "learning_rate": 4.400456794822993e-06, + "loss": 1.5362, + "step": 622 + }, + { + "epoch": 0.6399589111453519, + "grad_norm": 2.7597742080688477, + "learning_rate": 4.398553483060526e-06, + "loss": 1.56, + "step": 623 + }, + { + "epoch": 0.6409861325115562, + "grad_norm": 2.6911630630493164, + "learning_rate": 4.396650171298059e-06, + "loss": 1.5138, + "step": 624 + }, + { + "epoch": 0.6420133538777607, + "grad_norm": 2.596557378768921, + "learning_rate": 4.394746859535593e-06, + "loss": 1.5381, + "step": 625 + }, + { + "epoch": 0.6430405752439651, + "grad_norm": 2.5787081718444824, + "learning_rate": 4.392843547773125e-06, + "loss": 1.5284, + "step": 626 + }, + { + "epoch": 0.6440677966101694, + "grad_norm": 2.5975289344787598, + "learning_rate": 4.390940236010659e-06, + "loss": 1.5359, + "step": 627 + }, + { + "epoch": 0.6450950179763739, + "grad_norm": 2.5175442695617676, + "learning_rate": 4.389036924248192e-06, + "loss": 1.5608, + "step": 628 + }, + { + "epoch": 0.6461222393425783, + "grad_norm": 2.535139322280884, + "learning_rate": 4.387133612485725e-06, + "loss": 1.5856, + "step": 629 + }, + { + "epoch": 0.6471494607087828, + "grad_norm": 2.5596396923065186, + "learning_rate": 4.385230300723259e-06, + "loss": 1.5467, + "step": 630 + }, + { + "epoch": 0.6471494607087828, + "eval_loss": 1.835809588432312, + "eval_runtime": 37.1073, + "eval_samples_per_second": 26.949, + "eval_steps_per_second": 1.132, + "step": 630 + }, + { + "epoch": 0.6481766820749871, + "grad_norm": 2.691441297531128, + "learning_rate": 4.383326988960792e-06, + "loss": 1.5516, + "step": 631 + }, + { + "epoch": 0.6492039034411916, + "grad_norm": 2.534902811050415, + "learning_rate": 4.3814236771983255e-06, + "loss": 1.5215, + "step": 632 + }, + { + "epoch": 0.650231124807396, + "grad_norm": 2.626211643218994, + "learning_rate": 4.379520365435859e-06, + "loss": 1.5811, + "step": 633 + }, + { + "epoch": 0.6512583461736005, + "grad_norm": 2.6646480560302734, + "learning_rate": 4.377617053673392e-06, + "loss": 1.5674, + "step": 634 + }, + { + "epoch": 0.6522855675398048, + "grad_norm": 2.5728375911712646, + "learning_rate": 4.3757137419109256e-06, + "loss": 1.5655, + "step": 635 + }, + { + "epoch": 0.6533127889060092, + "grad_norm": 2.72817325592041, + "learning_rate": 4.373810430148459e-06, + "loss": 1.6283, + "step": 636 + }, + { + "epoch": 0.6543400102722137, + "grad_norm": 2.5395050048828125, + "learning_rate": 4.3719071183859914e-06, + "loss": 1.5191, + "step": 637 + }, + { + "epoch": 0.655367231638418, + "grad_norm": 2.964163064956665, + "learning_rate": 4.370003806623525e-06, + "loss": 1.5839, + "step": 638 + }, + { + "epoch": 0.6563944530046225, + "grad_norm": 2.517214298248291, + "learning_rate": 4.368100494861058e-06, + "loss": 1.5278, + "step": 639 + }, + { + "epoch": 0.6574216743708269, + "grad_norm": 2.765145778656006, + "learning_rate": 4.3661971830985915e-06, + "loss": 1.5718, + "step": 640 + }, + { + "epoch": 0.6574216743708269, + "eval_loss": 1.8324981927871704, + "eval_runtime": 37.0883, + "eval_samples_per_second": 26.963, + "eval_steps_per_second": 1.132, + "step": 640 + }, + { + "epoch": 0.6584488957370314, + "grad_norm": 2.463439464569092, + "learning_rate": 4.364293871336126e-06, + "loss": 1.5378, + "step": 641 + }, + { + "epoch": 0.6594761171032357, + "grad_norm": 2.7756192684173584, + "learning_rate": 4.362390559573659e-06, + "loss": 1.5601, + "step": 642 + }, + { + "epoch": 0.6605033384694402, + "grad_norm": 2.4134342670440674, + "learning_rate": 4.3604872478111925e-06, + "loss": 1.5252, + "step": 643 + }, + { + "epoch": 0.6615305598356446, + "grad_norm": 2.4549670219421387, + "learning_rate": 4.358583936048725e-06, + "loss": 1.4602, + "step": 644 + }, + { + "epoch": 0.662557781201849, + "grad_norm": 2.6815311908721924, + "learning_rate": 4.356680624286258e-06, + "loss": 1.5865, + "step": 645 + }, + { + "epoch": 0.6635850025680534, + "grad_norm": 2.5216286182403564, + "learning_rate": 4.354777312523792e-06, + "loss": 1.5432, + "step": 646 + }, + { + "epoch": 0.6646122239342578, + "grad_norm": 2.5547335147857666, + "learning_rate": 4.352874000761325e-06, + "loss": 1.59, + "step": 647 + }, + { + "epoch": 0.6656394453004623, + "grad_norm": 2.587646484375, + "learning_rate": 4.3509706889988584e-06, + "loss": 1.5351, + "step": 648 + }, + { + "epoch": 0.6666666666666666, + "grad_norm": 2.530768871307373, + "learning_rate": 4.349067377236392e-06, + "loss": 1.603, + "step": 649 + }, + { + "epoch": 0.6676938880328711, + "grad_norm": 2.727213144302368, + "learning_rate": 4.347164065473925e-06, + "loss": 1.6032, + "step": 650 + }, + { + "epoch": 0.6676938880328711, + "eval_loss": 1.8441001176834106, + "eval_runtime": 37.0797, + "eval_samples_per_second": 26.969, + "eval_steps_per_second": 1.133, + "step": 650 + }, + { + "epoch": 0.6687211093990755, + "grad_norm": 2.750441312789917, + "learning_rate": 4.3452607537114585e-06, + "loss": 1.6172, + "step": 651 + }, + { + "epoch": 0.66974833076528, + "grad_norm": 2.571692705154419, + "learning_rate": 4.343357441948992e-06, + "loss": 1.528, + "step": 652 + }, + { + "epoch": 0.6707755521314843, + "grad_norm": 2.5415565967559814, + "learning_rate": 4.341454130186525e-06, + "loss": 1.5629, + "step": 653 + }, + { + "epoch": 0.6718027734976888, + "grad_norm": 2.4970591068267822, + "learning_rate": 4.339550818424058e-06, + "loss": 1.5299, + "step": 654 + }, + { + "epoch": 0.6728299948638932, + "grad_norm": 2.5283010005950928, + "learning_rate": 4.337647506661591e-06, + "loss": 1.5149, + "step": 655 + }, + { + "epoch": 0.6738572162300975, + "grad_norm": 2.5205166339874268, + "learning_rate": 4.3357441948991245e-06, + "loss": 1.5486, + "step": 656 + }, + { + "epoch": 0.674884437596302, + "grad_norm": 2.4621481895446777, + "learning_rate": 4.333840883136658e-06, + "loss": 1.5769, + "step": 657 + }, + { + "epoch": 0.6759116589625064, + "grad_norm": 2.5954174995422363, + "learning_rate": 4.331937571374191e-06, + "loss": 1.5242, + "step": 658 + }, + { + "epoch": 0.6769388803287109, + "grad_norm": 2.553523063659668, + "learning_rate": 4.3300342596117246e-06, + "loss": 1.5339, + "step": 659 + }, + { + "epoch": 0.6779661016949152, + "grad_norm": 2.5289323329925537, + "learning_rate": 4.328130947849258e-06, + "loss": 1.5371, + "step": 660 + }, + { + "epoch": 0.6779661016949152, + "eval_loss": 1.8562791347503662, + "eval_runtime": 37.1299, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 660 + }, + { + "epoch": 0.6789933230611197, + "grad_norm": 2.537010908126831, + "learning_rate": 4.326227636086791e-06, + "loss": 1.5005, + "step": 661 + }, + { + "epoch": 0.6800205444273241, + "grad_norm": 2.6716437339782715, + "learning_rate": 4.324324324324325e-06, + "loss": 1.5501, + "step": 662 + }, + { + "epoch": 0.6810477657935285, + "grad_norm": 2.4969863891601562, + "learning_rate": 4.322421012561858e-06, + "loss": 1.5474, + "step": 663 + }, + { + "epoch": 0.6820749871597329, + "grad_norm": 2.6803722381591797, + "learning_rate": 4.320517700799391e-06, + "loss": 1.5397, + "step": 664 + }, + { + "epoch": 0.6831022085259374, + "grad_norm": 2.6370725631713867, + "learning_rate": 4.318614389036925e-06, + "loss": 1.6274, + "step": 665 + }, + { + "epoch": 0.6841294298921418, + "grad_norm": 2.6329426765441895, + "learning_rate": 4.316711077274458e-06, + "loss": 1.6015, + "step": 666 + }, + { + "epoch": 0.6851566512583461, + "grad_norm": 2.6113548278808594, + "learning_rate": 4.3148077655119915e-06, + "loss": 1.5015, + "step": 667 + }, + { + "epoch": 0.6861838726245506, + "grad_norm": 2.718801498413086, + "learning_rate": 4.312904453749525e-06, + "loss": 1.597, + "step": 668 + }, + { + "epoch": 0.687211093990755, + "grad_norm": 2.638871908187866, + "learning_rate": 4.311001141987058e-06, + "loss": 1.5528, + "step": 669 + }, + { + "epoch": 0.6882383153569595, + "grad_norm": 2.74849534034729, + "learning_rate": 4.3090978302245916e-06, + "loss": 1.5627, + "step": 670 + }, + { + "epoch": 0.6882383153569595, + "eval_loss": 1.8493062257766724, + "eval_runtime": 37.2964, + "eval_samples_per_second": 26.812, + "eval_steps_per_second": 1.126, + "step": 670 + }, + { + "epoch": 0.6892655367231638, + "grad_norm": 2.6839728355407715, + "learning_rate": 4.307194518462124e-06, + "loss": 1.552, + "step": 671 + }, + { + "epoch": 0.6902927580893683, + "grad_norm": 2.844496488571167, + "learning_rate": 4.3052912066996574e-06, + "loss": 1.6218, + "step": 672 + }, + { + "epoch": 0.6913199794555727, + "grad_norm": 2.4792141914367676, + "learning_rate": 4.303387894937191e-06, + "loss": 1.592, + "step": 673 + }, + { + "epoch": 0.6923472008217771, + "grad_norm": 2.4852263927459717, + "learning_rate": 4.301484583174724e-06, + "loss": 1.5165, + "step": 674 + }, + { + "epoch": 0.6933744221879815, + "grad_norm": 2.5438406467437744, + "learning_rate": 4.2995812714122575e-06, + "loss": 1.5109, + "step": 675 + }, + { + "epoch": 0.6944016435541859, + "grad_norm": 2.5431931018829346, + "learning_rate": 4.297677959649791e-06, + "loss": 1.5345, + "step": 676 + }, + { + "epoch": 0.6954288649203904, + "grad_norm": 2.650897979736328, + "learning_rate": 4.295774647887324e-06, + "loss": 1.5607, + "step": 677 + }, + { + "epoch": 0.6964560862865947, + "grad_norm": 2.721167802810669, + "learning_rate": 4.293871336124858e-06, + "loss": 1.5764, + "step": 678 + }, + { + "epoch": 0.6974833076527992, + "grad_norm": 2.5402960777282715, + "learning_rate": 4.291968024362391e-06, + "loss": 1.5505, + "step": 679 + }, + { + "epoch": 0.6985105290190036, + "grad_norm": 2.6878864765167236, + "learning_rate": 4.290064712599924e-06, + "loss": 1.5821, + "step": 680 + }, + { + "epoch": 0.6985105290190036, + "eval_loss": 1.8400572538375854, + "eval_runtime": 37.3885, + "eval_samples_per_second": 26.746, + "eval_steps_per_second": 1.123, + "step": 680 + }, + { + "epoch": 0.699537750385208, + "grad_norm": 2.6703765392303467, + "learning_rate": 4.288161400837458e-06, + "loss": 1.5794, + "step": 681 + }, + { + "epoch": 0.7005649717514124, + "grad_norm": 2.626725196838379, + "learning_rate": 4.286258089074991e-06, + "loss": 1.4986, + "step": 682 + }, + { + "epoch": 0.7015921931176169, + "grad_norm": 2.6062862873077393, + "learning_rate": 4.284354777312524e-06, + "loss": 1.5606, + "step": 683 + }, + { + "epoch": 0.7026194144838213, + "grad_norm": 2.6359503269195557, + "learning_rate": 4.282451465550058e-06, + "loss": 1.5174, + "step": 684 + }, + { + "epoch": 0.7036466358500256, + "grad_norm": 2.8453221321105957, + "learning_rate": 4.280548153787591e-06, + "loss": 1.543, + "step": 685 + }, + { + "epoch": 0.7046738572162301, + "grad_norm": 2.5195140838623047, + "learning_rate": 4.2786448420251245e-06, + "loss": 1.5462, + "step": 686 + }, + { + "epoch": 0.7057010785824345, + "grad_norm": 2.757230281829834, + "learning_rate": 4.276741530262658e-06, + "loss": 1.5033, + "step": 687 + }, + { + "epoch": 0.706728299948639, + "grad_norm": 2.708148241043091, + "learning_rate": 4.27483821850019e-06, + "loss": 1.5286, + "step": 688 + }, + { + "epoch": 0.7077555213148433, + "grad_norm": 2.6836555004119873, + "learning_rate": 4.272934906737724e-06, + "loss": 1.5973, + "step": 689 + }, + { + "epoch": 0.7087827426810478, + "grad_norm": 2.6970176696777344, + "learning_rate": 4.271031594975257e-06, + "loss": 1.5844, + "step": 690 + }, + { + "epoch": 0.7087827426810478, + "eval_loss": 1.8414137363433838, + "eval_runtime": 37.3847, + "eval_samples_per_second": 26.749, + "eval_steps_per_second": 1.123, + "step": 690 + }, + { + "epoch": 0.7098099640472522, + "grad_norm": 2.69212007522583, + "learning_rate": 4.2691282832127905e-06, + "loss": 1.5817, + "step": 691 + }, + { + "epoch": 0.7108371854134566, + "grad_norm": 2.6203925609588623, + "learning_rate": 4.267224971450324e-06, + "loss": 1.5411, + "step": 692 + }, + { + "epoch": 0.711864406779661, + "grad_norm": 2.418916940689087, + "learning_rate": 4.265321659687857e-06, + "loss": 1.483, + "step": 693 + }, + { + "epoch": 0.7128916281458655, + "grad_norm": 2.784196615219116, + "learning_rate": 4.2634183479253906e-06, + "loss": 1.5286, + "step": 694 + }, + { + "epoch": 0.7139188495120699, + "grad_norm": 2.675043821334839, + "learning_rate": 4.261515036162924e-06, + "loss": 1.5067, + "step": 695 + }, + { + "epoch": 0.7149460708782742, + "grad_norm": 2.7075769901275635, + "learning_rate": 4.259611724400457e-06, + "loss": 1.5316, + "step": 696 + }, + { + "epoch": 0.7159732922444787, + "grad_norm": 2.911898612976074, + "learning_rate": 4.257708412637991e-06, + "loss": 1.5203, + "step": 697 + }, + { + "epoch": 0.7170005136106831, + "grad_norm": 2.748126983642578, + "learning_rate": 4.255805100875523e-06, + "loss": 1.465, + "step": 698 + }, + { + "epoch": 0.7180277349768875, + "grad_norm": 2.6017746925354004, + "learning_rate": 4.2539017891130565e-06, + "loss": 1.4956, + "step": 699 + }, + { + "epoch": 0.7190549563430919, + "grad_norm": 3.0877811908721924, + "learning_rate": 4.25199847735059e-06, + "loss": 1.587, + "step": 700 + }, + { + "epoch": 0.7190549563430919, + "eval_loss": 1.8209773302078247, + "eval_runtime": 37.1796, + "eval_samples_per_second": 26.896, + "eval_steps_per_second": 1.13, + "step": 700 + }, + { + "epoch": 0.7200821777092964, + "grad_norm": 2.6982834339141846, + "learning_rate": 4.250095165588123e-06, + "loss": 1.4984, + "step": 701 + }, + { + "epoch": 0.7211093990755008, + "grad_norm": 2.7425835132598877, + "learning_rate": 4.2481918538256575e-06, + "loss": 1.5596, + "step": 702 + }, + { + "epoch": 0.7221366204417052, + "grad_norm": 2.838890314102173, + "learning_rate": 4.246288542063191e-06, + "loss": 1.5207, + "step": 703 + }, + { + "epoch": 0.7231638418079096, + "grad_norm": 2.694871664047241, + "learning_rate": 4.244385230300724e-06, + "loss": 1.5592, + "step": 704 + }, + { + "epoch": 0.724191063174114, + "grad_norm": 2.630206823348999, + "learning_rate": 4.242481918538257e-06, + "loss": 1.5005, + "step": 705 + }, + { + "epoch": 0.7252182845403184, + "grad_norm": 2.9697742462158203, + "learning_rate": 4.24057860677579e-06, + "loss": 1.5329, + "step": 706 + }, + { + "epoch": 0.7262455059065228, + "grad_norm": 2.8384499549865723, + "learning_rate": 4.2386752950133234e-06, + "loss": 1.5731, + "step": 707 + }, + { + "epoch": 0.7272727272727273, + "grad_norm": 2.530134916305542, + "learning_rate": 4.236771983250857e-06, + "loss": 1.5187, + "step": 708 + }, + { + "epoch": 0.7282999486389317, + "grad_norm": 2.601384162902832, + "learning_rate": 4.23486867148839e-06, + "loss": 1.5069, + "step": 709 + }, + { + "epoch": 0.7293271700051361, + "grad_norm": 2.7746055126190186, + "learning_rate": 4.2329653597259235e-06, + "loss": 1.5377, + "step": 710 + }, + { + "epoch": 0.7293271700051361, + "eval_loss": 1.8144561052322388, + "eval_runtime": 36.9983, + "eval_samples_per_second": 27.028, + "eval_steps_per_second": 1.135, + "step": 710 + }, + { + "epoch": 0.7303543913713405, + "grad_norm": 2.860414505004883, + "learning_rate": 4.231062047963457e-06, + "loss": 1.5604, + "step": 711 + }, + { + "epoch": 0.731381612737545, + "grad_norm": 2.7479686737060547, + "learning_rate": 4.22915873620099e-06, + "loss": 1.578, + "step": 712 + }, + { + "epoch": 0.7324088341037494, + "grad_norm": 3.093319892883301, + "learning_rate": 4.227255424438524e-06, + "loss": 1.5609, + "step": 713 + }, + { + "epoch": 0.7334360554699538, + "grad_norm": 2.611778497695923, + "learning_rate": 4.225352112676057e-06, + "loss": 1.5775, + "step": 714 + }, + { + "epoch": 0.7344632768361582, + "grad_norm": 2.5740373134613037, + "learning_rate": 4.2234488009135895e-06, + "loss": 1.5059, + "step": 715 + }, + { + "epoch": 0.7354904982023626, + "grad_norm": 2.5564701557159424, + "learning_rate": 4.221545489151123e-06, + "loss": 1.4687, + "step": 716 + }, + { + "epoch": 0.736517719568567, + "grad_norm": 2.566359758377075, + "learning_rate": 4.219642177388656e-06, + "loss": 1.5084, + "step": 717 + }, + { + "epoch": 0.7375449409347714, + "grad_norm": 2.632016897201538, + "learning_rate": 4.2177388656261896e-06, + "loss": 1.5449, + "step": 718 + }, + { + "epoch": 0.7385721623009759, + "grad_norm": 2.525616407394409, + "learning_rate": 4.215835553863723e-06, + "loss": 1.4882, + "step": 719 + }, + { + "epoch": 0.7395993836671803, + "grad_norm": 2.674612522125244, + "learning_rate": 4.213932242101256e-06, + "loss": 1.4576, + "step": 720 + }, + { + "epoch": 0.7395993836671803, + "eval_loss": 1.8278547525405884, + "eval_runtime": 37.0938, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 720 + }, + { + "epoch": 0.7406266050333847, + "grad_norm": 2.8636388778686523, + "learning_rate": 4.21202893033879e-06, + "loss": 1.5149, + "step": 721 + }, + { + "epoch": 0.7416538263995891, + "grad_norm": 2.7943782806396484, + "learning_rate": 4.210125618576323e-06, + "loss": 1.5409, + "step": 722 + }, + { + "epoch": 0.7426810477657936, + "grad_norm": 2.5911595821380615, + "learning_rate": 4.208222306813856e-06, + "loss": 1.5324, + "step": 723 + }, + { + "epoch": 0.743708269131998, + "grad_norm": 2.7497329711914062, + "learning_rate": 4.20631899505139e-06, + "loss": 1.5842, + "step": 724 + }, + { + "epoch": 0.7447354904982023, + "grad_norm": 2.649540424346924, + "learning_rate": 4.204415683288923e-06, + "loss": 1.438, + "step": 725 + }, + { + "epoch": 0.7457627118644068, + "grad_norm": 2.696852684020996, + "learning_rate": 4.2025123715264565e-06, + "loss": 1.5538, + "step": 726 + }, + { + "epoch": 0.7467899332306112, + "grad_norm": 2.553107738494873, + "learning_rate": 4.20060905976399e-06, + "loss": 1.5115, + "step": 727 + }, + { + "epoch": 0.7478171545968156, + "grad_norm": 2.559718370437622, + "learning_rate": 4.198705748001523e-06, + "loss": 1.4613, + "step": 728 + }, + { + "epoch": 0.74884437596302, + "grad_norm": 2.615692377090454, + "learning_rate": 4.1968024362390566e-06, + "loss": 1.5655, + "step": 729 + }, + { + "epoch": 0.7498715973292245, + "grad_norm": 2.432598114013672, + "learning_rate": 4.19489912447659e-06, + "loss": 1.5192, + "step": 730 + }, + { + "epoch": 0.7498715973292245, + "eval_loss": 1.8231704235076904, + "eval_runtime": 37.1648, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 730 + }, + { + "epoch": 0.7508988186954288, + "grad_norm": 2.63838529586792, + "learning_rate": 4.192995812714123e-06, + "loss": 1.5888, + "step": 731 + }, + { + "epoch": 0.7519260400616333, + "grad_norm": 2.711975574493408, + "learning_rate": 4.191092500951656e-06, + "loss": 1.5309, + "step": 732 + }, + { + "epoch": 0.7529532614278377, + "grad_norm": 2.459561347961426, + "learning_rate": 4.189189189189189e-06, + "loss": 1.5141, + "step": 733 + }, + { + "epoch": 0.7539804827940421, + "grad_norm": 2.7000269889831543, + "learning_rate": 4.1872858774267225e-06, + "loss": 1.5148, + "step": 734 + }, + { + "epoch": 0.7550077041602465, + "grad_norm": 2.8983139991760254, + "learning_rate": 4.185382565664256e-06, + "loss": 1.5766, + "step": 735 + }, + { + "epoch": 0.7560349255264509, + "grad_norm": 2.600132942199707, + "learning_rate": 4.183479253901789e-06, + "loss": 1.4987, + "step": 736 + }, + { + "epoch": 0.7570621468926554, + "grad_norm": 2.579481363296509, + "learning_rate": 4.181575942139323e-06, + "loss": 1.4761, + "step": 737 + }, + { + "epoch": 0.7580893682588598, + "grad_norm": 2.708799123764038, + "learning_rate": 4.179672630376856e-06, + "loss": 1.5278, + "step": 738 + }, + { + "epoch": 0.7591165896250642, + "grad_norm": 2.4836950302124023, + "learning_rate": 4.177769318614389e-06, + "loss": 1.5056, + "step": 739 + }, + { + "epoch": 0.7601438109912686, + "grad_norm": 2.4556121826171875, + "learning_rate": 4.175866006851923e-06, + "loss": 1.5637, + "step": 740 + }, + { + "epoch": 0.7601438109912686, + "eval_loss": 1.8386361598968506, + "eval_runtime": 37.1304, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 740 + }, + { + "epoch": 0.7611710323574731, + "grad_norm": 2.504653215408325, + "learning_rate": 4.173962695089456e-06, + "loss": 1.5519, + "step": 741 + }, + { + "epoch": 0.7621982537236774, + "grad_norm": 2.50435733795166, + "learning_rate": 4.1720593833269894e-06, + "loss": 1.5379, + "step": 742 + }, + { + "epoch": 0.7632254750898819, + "grad_norm": 2.5818099975585938, + "learning_rate": 4.170156071564523e-06, + "loss": 1.5462, + "step": 743 + }, + { + "epoch": 0.7642526964560863, + "grad_norm": 2.6530778408050537, + "learning_rate": 4.168252759802056e-06, + "loss": 1.6112, + "step": 744 + }, + { + "epoch": 0.7652799178222907, + "grad_norm": 2.463993549346924, + "learning_rate": 4.1663494480395895e-06, + "loss": 1.5174, + "step": 745 + }, + { + "epoch": 0.7663071391884951, + "grad_norm": 2.6928186416625977, + "learning_rate": 4.164446136277123e-06, + "loss": 1.5639, + "step": 746 + }, + { + "epoch": 0.7673343605546995, + "grad_norm": 2.5539324283599854, + "learning_rate": 4.162542824514656e-06, + "loss": 1.5557, + "step": 747 + }, + { + "epoch": 0.768361581920904, + "grad_norm": 2.6068854331970215, + "learning_rate": 4.16063951275219e-06, + "loss": 1.5309, + "step": 748 + }, + { + "epoch": 0.7693888032871083, + "grad_norm": 2.5830016136169434, + "learning_rate": 4.158736200989722e-06, + "loss": 1.4876, + "step": 749 + }, + { + "epoch": 0.7704160246533128, + "grad_norm": 2.6156973838806152, + "learning_rate": 4.1568328892272555e-06, + "loss": 1.5284, + "step": 750 + }, + { + "epoch": 0.7704160246533128, + "eval_loss": 1.8401216268539429, + "eval_runtime": 37.1657, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 750 + }, + { + "epoch": 0.7714432460195172, + "grad_norm": 2.6343140602111816, + "learning_rate": 4.154929577464789e-06, + "loss": 1.5348, + "step": 751 + }, + { + "epoch": 0.7724704673857217, + "grad_norm": 2.483319044113159, + "learning_rate": 4.153026265702322e-06, + "loss": 1.5344, + "step": 752 + }, + { + "epoch": 0.773497688751926, + "grad_norm": 2.755953550338745, + "learning_rate": 4.1511229539398556e-06, + "loss": 1.4715, + "step": 753 + }, + { + "epoch": 0.7745249101181304, + "grad_norm": 2.5264573097229004, + "learning_rate": 4.149219642177389e-06, + "loss": 1.5236, + "step": 754 + }, + { + "epoch": 0.7755521314843349, + "grad_norm": 2.793328046798706, + "learning_rate": 4.147316330414922e-06, + "loss": 1.5316, + "step": 755 + }, + { + "epoch": 0.7765793528505393, + "grad_norm": 2.7607345581054688, + "learning_rate": 4.145413018652456e-06, + "loss": 1.4518, + "step": 756 + }, + { + "epoch": 0.7776065742167437, + "grad_norm": 2.5960099697113037, + "learning_rate": 4.143509706889989e-06, + "loss": 1.5499, + "step": 757 + }, + { + "epoch": 0.7786337955829481, + "grad_norm": 2.752039909362793, + "learning_rate": 4.141606395127522e-06, + "loss": 1.5433, + "step": 758 + }, + { + "epoch": 0.7796610169491526, + "grad_norm": 2.724228620529175, + "learning_rate": 4.139703083365056e-06, + "loss": 1.5242, + "step": 759 + }, + { + "epoch": 0.7806882383153569, + "grad_norm": 2.668635368347168, + "learning_rate": 4.137799771602588e-06, + "loss": 1.563, + "step": 760 + }, + { + "epoch": 0.7806882383153569, + "eval_loss": 1.8386328220367432, + "eval_runtime": 37.1498, + "eval_samples_per_second": 26.918, + "eval_steps_per_second": 1.131, + "step": 760 + }, + { + "epoch": 0.7817154596815614, + "grad_norm": 2.5619916915893555, + "learning_rate": 4.135896459840122e-06, + "loss": 1.5398, + "step": 761 + }, + { + "epoch": 0.7827426810477658, + "grad_norm": 2.642075538635254, + "learning_rate": 4.133993148077656e-06, + "loss": 1.4938, + "step": 762 + }, + { + "epoch": 0.7837699024139703, + "grad_norm": 2.700350522994995, + "learning_rate": 4.132089836315189e-06, + "loss": 1.5196, + "step": 763 + }, + { + "epoch": 0.7847971237801746, + "grad_norm": 2.9511470794677734, + "learning_rate": 4.1301865245527226e-06, + "loss": 1.531, + "step": 764 + }, + { + "epoch": 0.785824345146379, + "grad_norm": 2.681643486022949, + "learning_rate": 4.128283212790256e-06, + "loss": 1.4747, + "step": 765 + }, + { + "epoch": 0.7868515665125835, + "grad_norm": 2.5386908054351807, + "learning_rate": 4.1263799010277884e-06, + "loss": 1.4424, + "step": 766 + }, + { + "epoch": 0.7878787878787878, + "grad_norm": 2.8831303119659424, + "learning_rate": 4.124476589265322e-06, + "loss": 1.5677, + "step": 767 + }, + { + "epoch": 0.7889060092449923, + "grad_norm": 2.593442678451538, + "learning_rate": 4.122573277502855e-06, + "loss": 1.5043, + "step": 768 + }, + { + "epoch": 0.7899332306111967, + "grad_norm": 2.6778204441070557, + "learning_rate": 4.1206699657403885e-06, + "loss": 1.5053, + "step": 769 + }, + { + "epoch": 0.7909604519774012, + "grad_norm": 2.6590428352355957, + "learning_rate": 4.118766653977922e-06, + "loss": 1.5338, + "step": 770 + }, + { + "epoch": 0.7909604519774012, + "eval_loss": 1.8266831636428833, + "eval_runtime": 37.1428, + "eval_samples_per_second": 26.923, + "eval_steps_per_second": 1.131, + "step": 770 + }, + { + "epoch": 0.7919876733436055, + "grad_norm": 2.6895735263824463, + "learning_rate": 4.116863342215455e-06, + "loss": 1.5067, + "step": 771 + }, + { + "epoch": 0.79301489470981, + "grad_norm": 2.5284390449523926, + "learning_rate": 4.114960030452989e-06, + "loss": 1.4775, + "step": 772 + }, + { + "epoch": 0.7940421160760144, + "grad_norm": 2.610442876815796, + "learning_rate": 4.113056718690522e-06, + "loss": 1.519, + "step": 773 + }, + { + "epoch": 0.7950693374422187, + "grad_norm": 2.5558695793151855, + "learning_rate": 4.111153406928055e-06, + "loss": 1.4565, + "step": 774 + }, + { + "epoch": 0.7960965588084232, + "grad_norm": 2.6313698291778564, + "learning_rate": 4.109250095165589e-06, + "loss": 1.5626, + "step": 775 + }, + { + "epoch": 0.7971237801746276, + "grad_norm": 2.553133726119995, + "learning_rate": 4.107346783403122e-06, + "loss": 1.5591, + "step": 776 + }, + { + "epoch": 0.7981510015408321, + "grad_norm": 2.6599066257476807, + "learning_rate": 4.105443471640655e-06, + "loss": 1.5064, + "step": 777 + }, + { + "epoch": 0.7991782229070364, + "grad_norm": 2.5582847595214844, + "learning_rate": 4.103540159878188e-06, + "loss": 1.4824, + "step": 778 + }, + { + "epoch": 0.8002054442732409, + "grad_norm": 2.658006429672241, + "learning_rate": 4.101636848115721e-06, + "loss": 1.5187, + "step": 779 + }, + { + "epoch": 0.8012326656394453, + "grad_norm": 2.5531060695648193, + "learning_rate": 4.099733536353255e-06, + "loss": 1.4862, + "step": 780 + }, + { + "epoch": 0.8012326656394453, + "eval_loss": 1.8293416500091553, + "eval_runtime": 37.2532, + "eval_samples_per_second": 26.843, + "eval_steps_per_second": 1.127, + "step": 780 + }, + { + "epoch": 0.8022598870056498, + "grad_norm": 2.7279789447784424, + "learning_rate": 4.097830224590788e-06, + "loss": 1.5299, + "step": 781 + }, + { + "epoch": 0.8032871083718541, + "grad_norm": 2.7395598888397217, + "learning_rate": 4.095926912828322e-06, + "loss": 1.6016, + "step": 782 + }, + { + "epoch": 0.8043143297380585, + "grad_norm": 2.6175215244293213, + "learning_rate": 4.094023601065855e-06, + "loss": 1.5657, + "step": 783 + }, + { + "epoch": 0.805341551104263, + "grad_norm": 2.4423458576202393, + "learning_rate": 4.092120289303388e-06, + "loss": 1.5085, + "step": 784 + }, + { + "epoch": 0.8063687724704673, + "grad_norm": 2.569363594055176, + "learning_rate": 4.0902169775409215e-06, + "loss": 1.5272, + "step": 785 + }, + { + "epoch": 0.8073959938366718, + "grad_norm": 2.5281481742858887, + "learning_rate": 4.088313665778455e-06, + "loss": 1.473, + "step": 786 + }, + { + "epoch": 0.8084232152028762, + "grad_norm": 2.527724504470825, + "learning_rate": 4.086410354015988e-06, + "loss": 1.4978, + "step": 787 + }, + { + "epoch": 0.8094504365690807, + "grad_norm": 2.524348497390747, + "learning_rate": 4.0845070422535216e-06, + "loss": 1.4818, + "step": 788 + }, + { + "epoch": 0.810477657935285, + "grad_norm": 2.5176353454589844, + "learning_rate": 4.082603730491055e-06, + "loss": 1.4946, + "step": 789 + }, + { + "epoch": 0.8115048793014895, + "grad_norm": 2.63879132270813, + "learning_rate": 4.080700418728588e-06, + "loss": 1.5411, + "step": 790 + }, + { + "epoch": 0.8115048793014895, + "eval_loss": 1.8180036544799805, + "eval_runtime": 37.4282, + "eval_samples_per_second": 26.718, + "eval_steps_per_second": 1.122, + "step": 790 + }, + { + "epoch": 0.8125321006676939, + "grad_norm": 2.5492401123046875, + "learning_rate": 4.078797106966122e-06, + "loss": 1.5185, + "step": 791 + }, + { + "epoch": 0.8135593220338984, + "grad_norm": 2.668663501739502, + "learning_rate": 4.076893795203655e-06, + "loss": 1.4187, + "step": 792 + }, + { + "epoch": 0.8145865434001027, + "grad_norm": 2.500401496887207, + "learning_rate": 4.074990483441188e-06, + "loss": 1.4902, + "step": 793 + }, + { + "epoch": 0.8156137647663071, + "grad_norm": 2.7383482456207275, + "learning_rate": 4.073087171678721e-06, + "loss": 1.5454, + "step": 794 + }, + { + "epoch": 0.8166409861325116, + "grad_norm": 2.7144806385040283, + "learning_rate": 4.071183859916254e-06, + "loss": 1.5336, + "step": 795 + }, + { + "epoch": 0.8176682074987159, + "grad_norm": 2.8088390827178955, + "learning_rate": 4.069280548153788e-06, + "loss": 1.5107, + "step": 796 + }, + { + "epoch": 0.8186954288649204, + "grad_norm": 2.786731004714966, + "learning_rate": 4.067377236391321e-06, + "loss": 1.4762, + "step": 797 + }, + { + "epoch": 0.8197226502311248, + "grad_norm": 2.7185654640197754, + "learning_rate": 4.065473924628854e-06, + "loss": 1.5129, + "step": 798 + }, + { + "epoch": 0.8207498715973293, + "grad_norm": 2.5182769298553467, + "learning_rate": 4.063570612866388e-06, + "loss": 1.5392, + "step": 799 + }, + { + "epoch": 0.8217770929635336, + "grad_norm": 2.594088554382324, + "learning_rate": 4.061667301103921e-06, + "loss": 1.5389, + "step": 800 + }, + { + "epoch": 0.8217770929635336, + "eval_loss": 1.8078373670578003, + "eval_runtime": 37.4191, + "eval_samples_per_second": 26.724, + "eval_steps_per_second": 1.122, + "step": 800 + }, + { + "epoch": 0.8228043143297381, + "grad_norm": 2.5099830627441406, + "learning_rate": 4.0597639893414544e-06, + "loss": 1.5033, + "step": 801 + }, + { + "epoch": 0.8238315356959425, + "grad_norm": 2.514415740966797, + "learning_rate": 4.057860677578988e-06, + "loss": 1.5152, + "step": 802 + }, + { + "epoch": 0.8248587570621468, + "grad_norm": 2.738563060760498, + "learning_rate": 4.055957365816521e-06, + "loss": 1.5302, + "step": 803 + }, + { + "epoch": 0.8258859784283513, + "grad_norm": 2.633472204208374, + "learning_rate": 4.0540540540540545e-06, + "loss": 1.4735, + "step": 804 + }, + { + "epoch": 0.8269131997945557, + "grad_norm": 2.6635546684265137, + "learning_rate": 4.052150742291588e-06, + "loss": 1.4675, + "step": 805 + }, + { + "epoch": 0.8279404211607602, + "grad_norm": 2.515171527862549, + "learning_rate": 4.050247430529121e-06, + "loss": 1.4682, + "step": 806 + }, + { + "epoch": 0.8289676425269645, + "grad_norm": 2.689464569091797, + "learning_rate": 4.048344118766655e-06, + "loss": 1.5499, + "step": 807 + }, + { + "epoch": 0.829994863893169, + "grad_norm": 2.639822244644165, + "learning_rate": 4.046440807004188e-06, + "loss": 1.5297, + "step": 808 + }, + { + "epoch": 0.8310220852593734, + "grad_norm": 2.6470797061920166, + "learning_rate": 4.044537495241721e-06, + "loss": 1.5215, + "step": 809 + }, + { + "epoch": 0.8320493066255779, + "grad_norm": 2.4769506454467773, + "learning_rate": 4.042634183479255e-06, + "loss": 1.4859, + "step": 810 + }, + { + "epoch": 0.8320493066255779, + "eval_loss": 1.8010139465332031, + "eval_runtime": 36.974, + "eval_samples_per_second": 27.046, + "eval_steps_per_second": 1.136, + "step": 810 + }, + { + "epoch": 0.8330765279917822, + "grad_norm": 2.72485613822937, + "learning_rate": 4.040730871716787e-06, + "loss": 1.5301, + "step": 811 + }, + { + "epoch": 0.8341037493579867, + "grad_norm": 2.522390365600586, + "learning_rate": 4.038827559954321e-06, + "loss": 1.5199, + "step": 812 + }, + { + "epoch": 0.8351309707241911, + "grad_norm": 2.6865367889404297, + "learning_rate": 4.036924248191854e-06, + "loss": 1.5605, + "step": 813 + }, + { + "epoch": 0.8361581920903954, + "grad_norm": 2.6478004455566406, + "learning_rate": 4.035020936429387e-06, + "loss": 1.5291, + "step": 814 + }, + { + "epoch": 0.8371854134565999, + "grad_norm": 2.632477045059204, + "learning_rate": 4.033117624666921e-06, + "loss": 1.5134, + "step": 815 + }, + { + "epoch": 0.8382126348228043, + "grad_norm": 2.683198928833008, + "learning_rate": 4.031214312904454e-06, + "loss": 1.4787, + "step": 816 + }, + { + "epoch": 0.8392398561890088, + "grad_norm": 2.536470651626587, + "learning_rate": 4.029311001141987e-06, + "loss": 1.4717, + "step": 817 + }, + { + "epoch": 0.8402670775552131, + "grad_norm": 2.5733802318573, + "learning_rate": 4.027407689379521e-06, + "loss": 1.4507, + "step": 818 + }, + { + "epoch": 0.8412942989214176, + "grad_norm": 2.779752016067505, + "learning_rate": 4.025504377617054e-06, + "loss": 1.503, + "step": 819 + }, + { + "epoch": 0.842321520287622, + "grad_norm": 2.6932976245880127, + "learning_rate": 4.0236010658545875e-06, + "loss": 1.4815, + "step": 820 + }, + { + "epoch": 0.842321520287622, + "eval_loss": 1.7960872650146484, + "eval_runtime": 37.0481, + "eval_samples_per_second": 26.992, + "eval_steps_per_second": 1.134, + "step": 820 + }, + { + "epoch": 0.8433487416538265, + "grad_norm": 2.7531607151031494, + "learning_rate": 4.02169775409212e-06, + "loss": 1.5286, + "step": 821 + }, + { + "epoch": 0.8443759630200308, + "grad_norm": 2.8817851543426514, + "learning_rate": 4.019794442329653e-06, + "loss": 1.5765, + "step": 822 + }, + { + "epoch": 0.8454031843862352, + "grad_norm": 2.544451951980591, + "learning_rate": 4.0178911305671876e-06, + "loss": 1.452, + "step": 823 + }, + { + "epoch": 0.8464304057524397, + "grad_norm": 2.8142952919006348, + "learning_rate": 4.015987818804721e-06, + "loss": 1.5374, + "step": 824 + }, + { + "epoch": 0.847457627118644, + "grad_norm": 2.5938656330108643, + "learning_rate": 4.014084507042254e-06, + "loss": 1.4756, + "step": 825 + }, + { + "epoch": 0.8484848484848485, + "grad_norm": 2.7007956504821777, + "learning_rate": 4.012181195279788e-06, + "loss": 1.445, + "step": 826 + }, + { + "epoch": 0.8495120698510529, + "grad_norm": 2.9885144233703613, + "learning_rate": 4.010277883517321e-06, + "loss": 1.5516, + "step": 827 + }, + { + "epoch": 0.8505392912172574, + "grad_norm": 2.814607620239258, + "learning_rate": 4.0083745717548535e-06, + "loss": 1.5375, + "step": 828 + }, + { + "epoch": 0.8515665125834617, + "grad_norm": 2.9187042713165283, + "learning_rate": 4.006471259992387e-06, + "loss": 1.496, + "step": 829 + }, + { + "epoch": 0.8525937339496662, + "grad_norm": 2.7408974170684814, + "learning_rate": 4.00456794822992e-06, + "loss": 1.5195, + "step": 830 + }, + { + "epoch": 0.8525937339496662, + "eval_loss": 1.795318365097046, + "eval_runtime": 37.1058, + "eval_samples_per_second": 26.95, + "eval_steps_per_second": 1.132, + "step": 830 + }, + { + "epoch": 0.8536209553158706, + "grad_norm": 2.6824872493743896, + "learning_rate": 4.002664636467454e-06, + "loss": 1.5568, + "step": 831 + }, + { + "epoch": 0.8546481766820749, + "grad_norm": 2.7598671913146973, + "learning_rate": 4.000761324704987e-06, + "loss": 1.5134, + "step": 832 + }, + { + "epoch": 0.8556753980482794, + "grad_norm": 2.728449583053589, + "learning_rate": 3.99885801294252e-06, + "loss": 1.4675, + "step": 833 + }, + { + "epoch": 0.8567026194144838, + "grad_norm": 2.900674343109131, + "learning_rate": 3.996954701180054e-06, + "loss": 1.5259, + "step": 834 + }, + { + "epoch": 0.8577298407806883, + "grad_norm": 2.818105697631836, + "learning_rate": 3.995051389417587e-06, + "loss": 1.4533, + "step": 835 + }, + { + "epoch": 0.8587570621468926, + "grad_norm": 2.791207790374756, + "learning_rate": 3.9931480776551204e-06, + "loss": 1.4692, + "step": 836 + }, + { + "epoch": 0.8597842835130971, + "grad_norm": 2.584268808364868, + "learning_rate": 3.991244765892654e-06, + "loss": 1.554, + "step": 837 + }, + { + "epoch": 0.8608115048793015, + "grad_norm": 2.5646660327911377, + "learning_rate": 3.989341454130186e-06, + "loss": 1.4782, + "step": 838 + }, + { + "epoch": 0.861838726245506, + "grad_norm": 2.4912099838256836, + "learning_rate": 3.98743814236772e-06, + "loss": 1.4494, + "step": 839 + }, + { + "epoch": 0.8628659476117103, + "grad_norm": 2.4632041454315186, + "learning_rate": 3.985534830605253e-06, + "loss": 1.4911, + "step": 840 + }, + { + "epoch": 0.8628659476117103, + "eval_loss": 1.7955427169799805, + "eval_runtime": 37.0188, + "eval_samples_per_second": 27.013, + "eval_steps_per_second": 1.135, + "step": 840 + }, + { + "epoch": 0.8638931689779148, + "grad_norm": 2.6388566493988037, + "learning_rate": 3.983631518842786e-06, + "loss": 1.5135, + "step": 841 + }, + { + "epoch": 0.8649203903441192, + "grad_norm": 2.7619292736053467, + "learning_rate": 3.98172820708032e-06, + "loss": 1.5279, + "step": 842 + }, + { + "epoch": 0.8659476117103235, + "grad_norm": 2.510636806488037, + "learning_rate": 3.979824895317854e-06, + "loss": 1.4815, + "step": 843 + }, + { + "epoch": 0.866974833076528, + "grad_norm": 2.6847214698791504, + "learning_rate": 3.977921583555387e-06, + "loss": 1.4556, + "step": 844 + }, + { + "epoch": 0.8680020544427324, + "grad_norm": 2.5366628170013428, + "learning_rate": 3.97601827179292e-06, + "loss": 1.5297, + "step": 845 + }, + { + "epoch": 0.8690292758089369, + "grad_norm": 2.6018271446228027, + "learning_rate": 3.974114960030453e-06, + "loss": 1.4642, + "step": 846 + }, + { + "epoch": 0.8700564971751412, + "grad_norm": 2.612107276916504, + "learning_rate": 3.9722116482679866e-06, + "loss": 1.5065, + "step": 847 + }, + { + "epoch": 0.8710837185413457, + "grad_norm": 2.6163978576660156, + "learning_rate": 3.97030833650552e-06, + "loss": 1.4913, + "step": 848 + }, + { + "epoch": 0.8721109399075501, + "grad_norm": 2.516697645187378, + "learning_rate": 3.968405024743053e-06, + "loss": 1.5006, + "step": 849 + }, + { + "epoch": 0.8731381612737545, + "grad_norm": 2.6012990474700928, + "learning_rate": 3.966501712980587e-06, + "loss": 1.4276, + "step": 850 + }, + { + "epoch": 0.8731381612737545, + "eval_loss": 1.807686448097229, + "eval_runtime": 37.0357, + "eval_samples_per_second": 27.001, + "eval_steps_per_second": 1.134, + "step": 850 + }, + { + "epoch": 0.8741653826399589, + "grad_norm": 2.6806578636169434, + "learning_rate": 3.96459840121812e-06, + "loss": 1.4849, + "step": 851 + }, + { + "epoch": 0.8751926040061633, + "grad_norm": 2.5139150619506836, + "learning_rate": 3.962695089455653e-06, + "loss": 1.4919, + "step": 852 + }, + { + "epoch": 0.8762198253723678, + "grad_norm": 2.779398202896118, + "learning_rate": 3.960791777693187e-06, + "loss": 1.5446, + "step": 853 + }, + { + "epoch": 0.8772470467385721, + "grad_norm": 2.7080724239349365, + "learning_rate": 3.95888846593072e-06, + "loss": 1.5215, + "step": 854 + }, + { + "epoch": 0.8782742681047766, + "grad_norm": 2.658066749572754, + "learning_rate": 3.956985154168253e-06, + "loss": 1.5315, + "step": 855 + }, + { + "epoch": 0.879301489470981, + "grad_norm": 2.6813952922821045, + "learning_rate": 3.955081842405786e-06, + "loss": 1.549, + "step": 856 + }, + { + "epoch": 0.8803287108371854, + "grad_norm": 2.744917869567871, + "learning_rate": 3.953178530643319e-06, + "loss": 1.5717, + "step": 857 + }, + { + "epoch": 0.8813559322033898, + "grad_norm": 2.869288921356201, + "learning_rate": 3.951275218880853e-06, + "loss": 1.5882, + "step": 858 + }, + { + "epoch": 0.8823831535695943, + "grad_norm": 2.633666515350342, + "learning_rate": 3.949371907118386e-06, + "loss": 1.4332, + "step": 859 + }, + { + "epoch": 0.8834103749357987, + "grad_norm": 2.7296335697174072, + "learning_rate": 3.9474685953559194e-06, + "loss": 1.5349, + "step": 860 + }, + { + "epoch": 0.8834103749357987, + "eval_loss": 1.80648934841156, + "eval_runtime": 37.0387, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 860 + }, + { + "epoch": 0.884437596302003, + "grad_norm": 3.0827677249908447, + "learning_rate": 3.945565283593453e-06, + "loss": 1.5108, + "step": 861 + }, + { + "epoch": 0.8854648176682075, + "grad_norm": 2.81630539894104, + "learning_rate": 3.943661971830986e-06, + "loss": 1.493, + "step": 862 + }, + { + "epoch": 0.8864920390344119, + "grad_norm": 2.804215908050537, + "learning_rate": 3.9417586600685195e-06, + "loss": 1.5432, + "step": 863 + }, + { + "epoch": 0.8875192604006163, + "grad_norm": 2.9585464000701904, + "learning_rate": 3.939855348306053e-06, + "loss": 1.5202, + "step": 864 + }, + { + "epoch": 0.8885464817668207, + "grad_norm": 2.717324733734131, + "learning_rate": 3.937952036543586e-06, + "loss": 1.5172, + "step": 865 + }, + { + "epoch": 0.8895737031330252, + "grad_norm": 2.62406325340271, + "learning_rate": 3.93604872478112e-06, + "loss": 1.5064, + "step": 866 + }, + { + "epoch": 0.8906009244992296, + "grad_norm": 3.2518985271453857, + "learning_rate": 3.934145413018653e-06, + "loss": 1.5159, + "step": 867 + }, + { + "epoch": 0.891628145865434, + "grad_norm": 2.6304733753204346, + "learning_rate": 3.932242101256186e-06, + "loss": 1.4723, + "step": 868 + }, + { + "epoch": 0.8926553672316384, + "grad_norm": 2.6788930892944336, + "learning_rate": 3.93033878949372e-06, + "loss": 1.4759, + "step": 869 + }, + { + "epoch": 0.8936825885978429, + "grad_norm": 2.8033249378204346, + "learning_rate": 3.928435477731253e-06, + "loss": 1.4435, + "step": 870 + }, + { + "epoch": 0.8936825885978429, + "eval_loss": 1.8105196952819824, + "eval_runtime": 37.062, + "eval_samples_per_second": 26.982, + "eval_steps_per_second": 1.133, + "step": 870 + }, + { + "epoch": 0.8947098099640473, + "grad_norm": 2.8569960594177246, + "learning_rate": 3.9265321659687864e-06, + "loss": 1.5424, + "step": 871 + }, + { + "epoch": 0.8957370313302516, + "grad_norm": 2.8037121295928955, + "learning_rate": 3.924628854206319e-06, + "loss": 1.5368, + "step": 872 + }, + { + "epoch": 0.8967642526964561, + "grad_norm": 2.7444164752960205, + "learning_rate": 3.922725542443852e-06, + "loss": 1.5102, + "step": 873 + }, + { + "epoch": 0.8977914740626605, + "grad_norm": 2.8689560890197754, + "learning_rate": 3.920822230681386e-06, + "loss": 1.4649, + "step": 874 + }, + { + "epoch": 0.8988186954288649, + "grad_norm": 2.677800178527832, + "learning_rate": 3.918918918918919e-06, + "loss": 1.4815, + "step": 875 + }, + { + "epoch": 0.8998459167950693, + "grad_norm": 2.8248631954193115, + "learning_rate": 3.917015607156452e-06, + "loss": 1.5193, + "step": 876 + }, + { + "epoch": 0.9008731381612738, + "grad_norm": 2.8272030353546143, + "learning_rate": 3.915112295393986e-06, + "loss": 1.4968, + "step": 877 + }, + { + "epoch": 0.9019003595274782, + "grad_norm": 3.0890679359436035, + "learning_rate": 3.913208983631519e-06, + "loss": 1.5705, + "step": 878 + }, + { + "epoch": 0.9029275808936826, + "grad_norm": 2.7617108821868896, + "learning_rate": 3.9113056718690525e-06, + "loss": 1.5598, + "step": 879 + }, + { + "epoch": 0.903954802259887, + "grad_norm": 2.7723312377929688, + "learning_rate": 3.909402360106586e-06, + "loss": 1.4781, + "step": 880 + }, + { + "epoch": 0.903954802259887, + "eval_loss": 1.8097552061080933, + "eval_runtime": 37.0572, + "eval_samples_per_second": 26.985, + "eval_steps_per_second": 1.133, + "step": 880 + }, + { + "epoch": 0.9049820236260914, + "grad_norm": 2.82560133934021, + "learning_rate": 3.907499048344119e-06, + "loss": 1.5224, + "step": 881 + }, + { + "epoch": 0.9060092449922958, + "grad_norm": 2.881899118423462, + "learning_rate": 3.905595736581652e-06, + "loss": 1.5169, + "step": 882 + }, + { + "epoch": 0.9070364663585002, + "grad_norm": 2.8723788261413574, + "learning_rate": 3.903692424819186e-06, + "loss": 1.5225, + "step": 883 + }, + { + "epoch": 0.9080636877247047, + "grad_norm": 2.5932493209838867, + "learning_rate": 3.901789113056719e-06, + "loss": 1.4796, + "step": 884 + }, + { + "epoch": 0.9090909090909091, + "grad_norm": 2.6493358612060547, + "learning_rate": 3.899885801294253e-06, + "loss": 1.5324, + "step": 885 + }, + { + "epoch": 0.9101181304571135, + "grad_norm": 2.630830764770508, + "learning_rate": 3.897982489531786e-06, + "loss": 1.5113, + "step": 886 + }, + { + "epoch": 0.9111453518233179, + "grad_norm": 2.8183815479278564, + "learning_rate": 3.896079177769319e-06, + "loss": 1.5677, + "step": 887 + }, + { + "epoch": 0.9121725731895224, + "grad_norm": 2.811241865158081, + "learning_rate": 3.894175866006853e-06, + "loss": 1.5357, + "step": 888 + }, + { + "epoch": 0.9131997945557268, + "grad_norm": 2.6256868839263916, + "learning_rate": 3.892272554244385e-06, + "loss": 1.5603, + "step": 889 + }, + { + "epoch": 0.9142270159219312, + "grad_norm": 2.677891254425049, + "learning_rate": 3.890369242481919e-06, + "loss": 1.4748, + "step": 890 + }, + { + "epoch": 0.9142270159219312, + "eval_loss": 1.7937747240066528, + "eval_runtime": 37.3281, + "eval_samples_per_second": 26.789, + "eval_steps_per_second": 1.125, + "step": 890 + }, + { + "epoch": 0.9152542372881356, + "grad_norm": 2.639749765396118, + "learning_rate": 3.888465930719452e-06, + "loss": 1.5477, + "step": 891 + }, + { + "epoch": 0.91628145865434, + "grad_norm": 2.622887134552002, + "learning_rate": 3.886562618956985e-06, + "loss": 1.481, + "step": 892 + }, + { + "epoch": 0.9173086800205444, + "grad_norm": 2.659916639328003, + "learning_rate": 3.884659307194519e-06, + "loss": 1.5044, + "step": 893 + }, + { + "epoch": 0.9183359013867488, + "grad_norm": 2.524515151977539, + "learning_rate": 3.882755995432052e-06, + "loss": 1.4378, + "step": 894 + }, + { + "epoch": 0.9193631227529533, + "grad_norm": 2.5722432136535645, + "learning_rate": 3.8808526836695854e-06, + "loss": 1.5173, + "step": 895 + }, + { + "epoch": 0.9203903441191577, + "grad_norm": 2.6036641597747803, + "learning_rate": 3.878949371907119e-06, + "loss": 1.4823, + "step": 896 + }, + { + "epoch": 0.9214175654853621, + "grad_norm": 2.671830415725708, + "learning_rate": 3.877046060144652e-06, + "loss": 1.4679, + "step": 897 + }, + { + "epoch": 0.9224447868515665, + "grad_norm": 2.649601936340332, + "learning_rate": 3.8751427483821855e-06, + "loss": 1.4651, + "step": 898 + }, + { + "epoch": 0.923472008217771, + "grad_norm": 2.6315200328826904, + "learning_rate": 3.873239436619718e-06, + "loss": 1.5029, + "step": 899 + }, + { + "epoch": 0.9244992295839753, + "grad_norm": 2.7867085933685303, + "learning_rate": 3.871336124857251e-06, + "loss": 1.4635, + "step": 900 + }, + { + "epoch": 0.9244992295839753, + "eval_loss": 1.8004201650619507, + "eval_runtime": 37.3928, + "eval_samples_per_second": 26.743, + "eval_steps_per_second": 1.123, + "step": 900 + }, + { + "epoch": 0.9255264509501797, + "grad_norm": 2.6903293132781982, + "learning_rate": 3.869432813094785e-06, + "loss": 1.5043, + "step": 901 + }, + { + "epoch": 0.9265536723163842, + "grad_norm": 2.6979103088378906, + "learning_rate": 3.867529501332318e-06, + "loss": 1.5168, + "step": 902 + }, + { + "epoch": 0.9275808936825886, + "grad_norm": 2.594578266143799, + "learning_rate": 3.8656261895698515e-06, + "loss": 1.5141, + "step": 903 + }, + { + "epoch": 0.928608115048793, + "grad_norm": 2.774486780166626, + "learning_rate": 3.863722877807386e-06, + "loss": 1.4982, + "step": 904 + }, + { + "epoch": 0.9296353364149974, + "grad_norm": 2.6801764965057373, + "learning_rate": 3.861819566044919e-06, + "loss": 1.5294, + "step": 905 + }, + { + "epoch": 0.9306625577812019, + "grad_norm": 2.681169033050537, + "learning_rate": 3.859916254282452e-06, + "loss": 1.5687, + "step": 906 + }, + { + "epoch": 0.9316897791474062, + "grad_norm": 2.6592564582824707, + "learning_rate": 3.858012942519985e-06, + "loss": 1.4967, + "step": 907 + }, + { + "epoch": 0.9327170005136107, + "grad_norm": 2.636106491088867, + "learning_rate": 3.856109630757518e-06, + "loss": 1.5108, + "step": 908 + }, + { + "epoch": 0.9337442218798151, + "grad_norm": 2.6918840408325195, + "learning_rate": 3.854206318995052e-06, + "loss": 1.4821, + "step": 909 + }, + { + "epoch": 0.9347714432460195, + "grad_norm": 2.7018930912017822, + "learning_rate": 3.852303007232585e-06, + "loss": 1.5336, + "step": 910 + }, + { + "epoch": 0.9347714432460195, + "eval_loss": 1.789487600326538, + "eval_runtime": 36.9792, + "eval_samples_per_second": 27.042, + "eval_steps_per_second": 1.136, + "step": 910 + }, + { + "epoch": 0.9357986646122239, + "grad_norm": 2.5881776809692383, + "learning_rate": 3.850399695470118e-06, + "loss": 1.4632, + "step": 911 + }, + { + "epoch": 0.9368258859784283, + "grad_norm": 2.55487322807312, + "learning_rate": 3.848496383707652e-06, + "loss": 1.4825, + "step": 912 + }, + { + "epoch": 0.9378531073446328, + "grad_norm": 2.5630884170532227, + "learning_rate": 3.846593071945185e-06, + "loss": 1.463, + "step": 913 + }, + { + "epoch": 0.9388803287108372, + "grad_norm": 2.6740293502807617, + "learning_rate": 3.8446897601827185e-06, + "loss": 1.4499, + "step": 914 + }, + { + "epoch": 0.9399075500770416, + "grad_norm": 2.641127109527588, + "learning_rate": 3.842786448420252e-06, + "loss": 1.5618, + "step": 915 + }, + { + "epoch": 0.940934771443246, + "grad_norm": 2.669304370880127, + "learning_rate": 3.840883136657785e-06, + "loss": 1.5725, + "step": 916 + }, + { + "epoch": 0.9419619928094505, + "grad_norm": 2.7020649909973145, + "learning_rate": 3.838979824895318e-06, + "loss": 1.5587, + "step": 917 + }, + { + "epoch": 0.9429892141756548, + "grad_norm": 2.6244914531707764, + "learning_rate": 3.837076513132851e-06, + "loss": 1.4363, + "step": 918 + }, + { + "epoch": 0.9440164355418593, + "grad_norm": 2.7254905700683594, + "learning_rate": 3.8351732013703844e-06, + "loss": 1.5205, + "step": 919 + }, + { + "epoch": 0.9450436569080637, + "grad_norm": 2.4565985202789307, + "learning_rate": 3.833269889607918e-06, + "loss": 1.4785, + "step": 920 + }, + { + "epoch": 0.9450436569080637, + "eval_loss": 1.8004498481750488, + "eval_runtime": 37.0945, + "eval_samples_per_second": 26.958, + "eval_steps_per_second": 1.132, + "step": 920 + }, + { + "epoch": 0.9460708782742681, + "grad_norm": 2.8681352138519287, + "learning_rate": 3.831366577845451e-06, + "loss": 1.4598, + "step": 921 + }, + { + "epoch": 0.9470980996404725, + "grad_norm": 2.6068029403686523, + "learning_rate": 3.8294632660829845e-06, + "loss": 1.4565, + "step": 922 + }, + { + "epoch": 0.9481253210066769, + "grad_norm": 2.764873504638672, + "learning_rate": 3.827559954320518e-06, + "loss": 1.4882, + "step": 923 + }, + { + "epoch": 0.9491525423728814, + "grad_norm": 2.655008316040039, + "learning_rate": 3.825656642558051e-06, + "loss": 1.5004, + "step": 924 + }, + { + "epoch": 0.9501797637390857, + "grad_norm": 2.7953972816467285, + "learning_rate": 3.823753330795585e-06, + "loss": 1.5172, + "step": 925 + }, + { + "epoch": 0.9512069851052902, + "grad_norm": 2.698564052581787, + "learning_rate": 3.821850019033118e-06, + "loss": 1.5022, + "step": 926 + }, + { + "epoch": 0.9522342064714946, + "grad_norm": 2.4997222423553467, + "learning_rate": 3.819946707270651e-06, + "loss": 1.4373, + "step": 927 + }, + { + "epoch": 0.9532614278376991, + "grad_norm": 2.6832621097564697, + "learning_rate": 3.818043395508185e-06, + "loss": 1.49, + "step": 928 + }, + { + "epoch": 0.9542886492039034, + "grad_norm": 2.734297513961792, + "learning_rate": 3.816140083745718e-06, + "loss": 1.499, + "step": 929 + }, + { + "epoch": 0.9553158705701078, + "grad_norm": 2.6073720455169678, + "learning_rate": 3.8142367719832514e-06, + "loss": 1.5701, + "step": 930 + }, + { + "epoch": 0.9553158705701078, + "eval_loss": 1.8036372661590576, + "eval_runtime": 37.141, + "eval_samples_per_second": 26.924, + "eval_steps_per_second": 1.131, + "step": 930 + }, + { + "epoch": 0.9563430919363123, + "grad_norm": 2.6502838134765625, + "learning_rate": 3.8123334602207844e-06, + "loss": 1.4666, + "step": 931 + }, + { + "epoch": 0.9573703133025167, + "grad_norm": 2.6244993209838867, + "learning_rate": 3.8104301484583177e-06, + "loss": 1.4659, + "step": 932 + }, + { + "epoch": 0.9583975346687211, + "grad_norm": 2.5014829635620117, + "learning_rate": 3.808526836695851e-06, + "loss": 1.4562, + "step": 933 + }, + { + "epoch": 0.9594247560349255, + "grad_norm": 2.6359620094299316, + "learning_rate": 3.8066235249333845e-06, + "loss": 1.5223, + "step": 934 + }, + { + "epoch": 0.96045197740113, + "grad_norm": 2.648331642150879, + "learning_rate": 3.804720213170918e-06, + "loss": 1.4924, + "step": 935 + }, + { + "epoch": 0.9614791987673343, + "grad_norm": 2.731069564819336, + "learning_rate": 3.8028169014084508e-06, + "loss": 1.4959, + "step": 936 + }, + { + "epoch": 0.9625064201335388, + "grad_norm": 2.5789313316345215, + "learning_rate": 3.800913589645984e-06, + "loss": 1.5125, + "step": 937 + }, + { + "epoch": 0.9635336414997432, + "grad_norm": 2.6203935146331787, + "learning_rate": 3.7990102778835175e-06, + "loss": 1.5064, + "step": 938 + }, + { + "epoch": 0.9645608628659477, + "grad_norm": 2.7040021419525146, + "learning_rate": 3.797106966121051e-06, + "loss": 1.5092, + "step": 939 + }, + { + "epoch": 0.965588084232152, + "grad_norm": 2.575338125228882, + "learning_rate": 3.7952036543585842e-06, + "loss": 1.5653, + "step": 940 + }, + { + "epoch": 0.965588084232152, + "eval_loss": 1.7945882081985474, + "eval_runtime": 37.1973, + "eval_samples_per_second": 26.884, + "eval_steps_per_second": 1.129, + "step": 940 + }, + { + "epoch": 0.9666153055983564, + "grad_norm": 2.597348690032959, + "learning_rate": 3.793300342596117e-06, + "loss": 1.5312, + "step": 941 + }, + { + "epoch": 0.9676425269645609, + "grad_norm": 2.590437650680542, + "learning_rate": 3.7913970308336505e-06, + "loss": 1.5347, + "step": 942 + }, + { + "epoch": 0.9686697483307652, + "grad_norm": 2.6253085136413574, + "learning_rate": 3.789493719071184e-06, + "loss": 1.5265, + "step": 943 + }, + { + "epoch": 0.9696969696969697, + "grad_norm": 2.5652995109558105, + "learning_rate": 3.7875904073087177e-06, + "loss": 1.4738, + "step": 944 + }, + { + "epoch": 0.9707241910631741, + "grad_norm": 2.552037239074707, + "learning_rate": 3.785687095546251e-06, + "loss": 1.4311, + "step": 945 + }, + { + "epoch": 0.9717514124293786, + "grad_norm": 2.806792974472046, + "learning_rate": 3.7837837837837844e-06, + "loss": 1.4857, + "step": 946 + }, + { + "epoch": 0.9727786337955829, + "grad_norm": 2.696113348007202, + "learning_rate": 3.7818804720213178e-06, + "loss": 1.4604, + "step": 947 + }, + { + "epoch": 0.9738058551617874, + "grad_norm": 2.972555637359619, + "learning_rate": 3.7799771602588507e-06, + "loss": 1.4328, + "step": 948 + }, + { + "epoch": 0.9748330765279918, + "grad_norm": 2.7260053157806396, + "learning_rate": 3.778073848496384e-06, + "loss": 1.4315, + "step": 949 + }, + { + "epoch": 0.9758602978941961, + "grad_norm": 2.501840353012085, + "learning_rate": 3.7761705367339174e-06, + "loss": 1.4913, + "step": 950 + }, + { + "epoch": 0.9758602978941961, + "eval_loss": 1.7844858169555664, + "eval_runtime": 37.1439, + "eval_samples_per_second": 26.922, + "eval_steps_per_second": 1.131, + "step": 950 + }, + { + "epoch": 0.9768875192604006, + "grad_norm": 2.704096794128418, + "learning_rate": 3.7742672249714508e-06, + "loss": 1.5138, + "step": 951 + }, + { + "epoch": 0.977914740626605, + "grad_norm": 2.7086358070373535, + "learning_rate": 3.772363913208984e-06, + "loss": 1.5228, + "step": 952 + }, + { + "epoch": 0.9789419619928095, + "grad_norm": 2.6690590381622314, + "learning_rate": 3.770460601446517e-06, + "loss": 1.4828, + "step": 953 + }, + { + "epoch": 0.9799691833590138, + "grad_norm": 2.6134684085845947, + "learning_rate": 3.7685572896840504e-06, + "loss": 1.509, + "step": 954 + }, + { + "epoch": 0.9809964047252183, + "grad_norm": 2.7675423622131348, + "learning_rate": 3.766653977921584e-06, + "loss": 1.4544, + "step": 955 + }, + { + "epoch": 0.9820236260914227, + "grad_norm": 2.8518829345703125, + "learning_rate": 3.764750666159117e-06, + "loss": 1.4042, + "step": 956 + }, + { + "epoch": 0.9830508474576272, + "grad_norm": 2.6791179180145264, + "learning_rate": 3.7628473543966505e-06, + "loss": 1.4961, + "step": 957 + }, + { + "epoch": 0.9840780688238315, + "grad_norm": 3.014099597930908, + "learning_rate": 3.7609440426341835e-06, + "loss": 1.4582, + "step": 958 + }, + { + "epoch": 0.9851052901900359, + "grad_norm": 2.7100412845611572, + "learning_rate": 3.759040730871717e-06, + "loss": 1.5491, + "step": 959 + }, + { + "epoch": 0.9861325115562404, + "grad_norm": 2.565263271331787, + "learning_rate": 3.75713741910925e-06, + "loss": 1.4437, + "step": 960 + }, + { + "epoch": 0.9861325115562404, + "eval_loss": 1.7909018993377686, + "eval_runtime": 37.1828, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 960 + }, + { + "epoch": 0.9871597329224447, + "grad_norm": 2.8322439193725586, + "learning_rate": 3.7552341073467836e-06, + "loss": 1.4322, + "step": 961 + }, + { + "epoch": 0.9881869542886492, + "grad_norm": 2.738586902618408, + "learning_rate": 3.753330795584317e-06, + "loss": 1.4797, + "step": 962 + }, + { + "epoch": 0.9892141756548536, + "grad_norm": 2.711247205734253, + "learning_rate": 3.75142748382185e-06, + "loss": 1.5409, + "step": 963 + }, + { + "epoch": 0.9902413970210581, + "grad_norm": 2.5288846492767334, + "learning_rate": 3.749524172059384e-06, + "loss": 1.4408, + "step": 964 + }, + { + "epoch": 0.9912686183872624, + "grad_norm": 2.5907785892486572, + "learning_rate": 3.747620860296917e-06, + "loss": 1.4872, + "step": 965 + }, + { + "epoch": 0.9922958397534669, + "grad_norm": 2.4952504634857178, + "learning_rate": 3.7457175485344504e-06, + "loss": 1.5277, + "step": 966 + }, + { + "epoch": 0.9933230611196713, + "grad_norm": 2.633167266845703, + "learning_rate": 3.7438142367719837e-06, + "loss": 1.5258, + "step": 967 + }, + { + "epoch": 0.9943502824858758, + "grad_norm": 2.6295969486236572, + "learning_rate": 3.741910925009517e-06, + "loss": 1.5013, + "step": 968 + }, + { + "epoch": 0.9953775038520801, + "grad_norm": 2.4780771732330322, + "learning_rate": 3.7400076132470505e-06, + "loss": 1.4646, + "step": 969 + }, + { + "epoch": 0.9964047252182845, + "grad_norm": 2.585749626159668, + "learning_rate": 3.7381043014845834e-06, + "loss": 1.4814, + "step": 970 + }, + { + "epoch": 0.9964047252182845, + "eval_loss": 1.801538109779358, + "eval_runtime": 37.1396, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 970 + }, + { + "epoch": 0.997431946584489, + "grad_norm": 2.626293182373047, + "learning_rate": 3.7362009897221168e-06, + "loss": 1.5071, + "step": 971 + }, + { + "epoch": 0.9984591679506933, + "grad_norm": 2.7129111289978027, + "learning_rate": 3.73429767795965e-06, + "loss": 1.4911, + "step": 972 + }, + { + "epoch": 0.9994863893168978, + "grad_norm": 2.623265027999878, + "learning_rate": 3.7323943661971835e-06, + "loss": 1.506, + "step": 973 + }, + { + "epoch": 1.0005136106831023, + "grad_norm": 2.497767925262451, + "learning_rate": 3.730491054434717e-06, + "loss": 1.4119, + "step": 974 + }, + { + "epoch": 1.0015408320493067, + "grad_norm": 2.8463239669799805, + "learning_rate": 3.72858774267225e-06, + "loss": 1.3677, + "step": 975 + }, + { + "epoch": 1.002568053415511, + "grad_norm": 2.643601417541504, + "learning_rate": 3.726684430909783e-06, + "loss": 1.4573, + "step": 976 + }, + { + "epoch": 1.0035952747817154, + "grad_norm": 2.751330614089966, + "learning_rate": 3.7247811191473165e-06, + "loss": 1.4821, + "step": 977 + }, + { + "epoch": 1.00462249614792, + "grad_norm": 2.714946985244751, + "learning_rate": 3.72287780738485e-06, + "loss": 1.4898, + "step": 978 + }, + { + "epoch": 1.0056497175141244, + "grad_norm": 3.066772699356079, + "learning_rate": 3.7209744956223832e-06, + "loss": 1.464, + "step": 979 + }, + { + "epoch": 1.0066769388803287, + "grad_norm": 2.818673610687256, + "learning_rate": 3.719071183859916e-06, + "loss": 1.4743, + "step": 980 + }, + { + "epoch": 1.0066769388803287, + "eval_loss": 1.7895241975784302, + "eval_runtime": 37.1443, + "eval_samples_per_second": 26.922, + "eval_steps_per_second": 1.131, + "step": 980 + }, + { + "epoch": 1.007704160246533, + "grad_norm": 2.6721136569976807, + "learning_rate": 3.7171678720974495e-06, + "loss": 1.3996, + "step": 981 + }, + { + "epoch": 1.0087313816127375, + "grad_norm": 2.757756471633911, + "learning_rate": 3.715264560334983e-06, + "loss": 1.5009, + "step": 982 + }, + { + "epoch": 1.009758602978942, + "grad_norm": 2.8042495250701904, + "learning_rate": 3.7133612485725163e-06, + "loss": 1.4672, + "step": 983 + }, + { + "epoch": 1.0107858243451464, + "grad_norm": 2.662808418273926, + "learning_rate": 3.71145793681005e-06, + "loss": 1.4799, + "step": 984 + }, + { + "epoch": 1.0118130457113508, + "grad_norm": 2.6021053791046143, + "learning_rate": 3.7095546250475834e-06, + "loss": 1.4035, + "step": 985 + }, + { + "epoch": 1.0128402670775551, + "grad_norm": 2.7127013206481934, + "learning_rate": 3.7076513132851168e-06, + "loss": 1.5192, + "step": 986 + }, + { + "epoch": 1.0138674884437597, + "grad_norm": 2.6402854919433594, + "learning_rate": 3.7057480015226497e-06, + "loss": 1.4396, + "step": 987 + }, + { + "epoch": 1.014894709809964, + "grad_norm": 2.5758111476898193, + "learning_rate": 3.703844689760183e-06, + "loss": 1.4109, + "step": 988 + }, + { + "epoch": 1.0159219311761685, + "grad_norm": 2.7668240070343018, + "learning_rate": 3.7019413779977164e-06, + "loss": 1.4865, + "step": 989 + }, + { + "epoch": 1.0169491525423728, + "grad_norm": 2.840165615081787, + "learning_rate": 3.70003806623525e-06, + "loss": 1.4708, + "step": 990 + }, + { + "epoch": 1.0169491525423728, + "eval_loss": 1.7967512607574463, + "eval_runtime": 37.2322, + "eval_samples_per_second": 26.858, + "eval_steps_per_second": 1.128, + "step": 990 + }, + { + "epoch": 1.0179763739085772, + "grad_norm": 2.8218352794647217, + "learning_rate": 3.698134754472783e-06, + "loss": 1.4414, + "step": 991 + }, + { + "epoch": 1.0190035952747818, + "grad_norm": 2.742658853530884, + "learning_rate": 3.696231442710316e-06, + "loss": 1.5109, + "step": 992 + }, + { + "epoch": 1.0200308166409862, + "grad_norm": 2.6375293731689453, + "learning_rate": 3.6943281309478495e-06, + "loss": 1.4433, + "step": 993 + }, + { + "epoch": 1.0210580380071905, + "grad_norm": 2.914246082305908, + "learning_rate": 3.692424819185383e-06, + "loss": 1.4656, + "step": 994 + }, + { + "epoch": 1.022085259373395, + "grad_norm": 2.9181385040283203, + "learning_rate": 3.690521507422916e-06, + "loss": 1.4735, + "step": 995 + }, + { + "epoch": 1.0231124807395995, + "grad_norm": 2.755748748779297, + "learning_rate": 3.6886181956604496e-06, + "loss": 1.4196, + "step": 996 + }, + { + "epoch": 1.0241397021058039, + "grad_norm": 2.9245896339416504, + "learning_rate": 3.686714883897983e-06, + "loss": 1.4316, + "step": 997 + }, + { + "epoch": 1.0251669234720082, + "grad_norm": 2.839991569519043, + "learning_rate": 3.684811572135516e-06, + "loss": 1.4965, + "step": 998 + }, + { + "epoch": 1.0261941448382126, + "grad_norm": 2.904658079147339, + "learning_rate": 3.6829082603730492e-06, + "loss": 1.4962, + "step": 999 + }, + { + "epoch": 1.027221366204417, + "grad_norm": 2.811279535293579, + "learning_rate": 3.6810049486105826e-06, + "loss": 1.4514, + "step": 1000 + }, + { + "epoch": 1.027221366204417, + "eval_loss": 1.7917094230651855, + "eval_runtime": 37.315, + "eval_samples_per_second": 26.799, + "eval_steps_per_second": 1.126, + "step": 1000 + }, + { + "epoch": 1.0282485875706215, + "grad_norm": 2.730212926864624, + "learning_rate": 3.679101636848116e-06, + "loss": 1.4842, + "step": 1001 + }, + { + "epoch": 1.029275808936826, + "grad_norm": 2.8678042888641357, + "learning_rate": 3.6771983250856493e-06, + "loss": 1.4726, + "step": 1002 + }, + { + "epoch": 1.0303030303030303, + "grad_norm": 2.8228330612182617, + "learning_rate": 3.6752950133231822e-06, + "loss": 1.4501, + "step": 1003 + }, + { + "epoch": 1.0313302516692346, + "grad_norm": 2.8341903686523438, + "learning_rate": 3.6733917015607156e-06, + "loss": 1.4758, + "step": 1004 + }, + { + "epoch": 1.0323574730354392, + "grad_norm": 2.7462005615234375, + "learning_rate": 3.6714883897982494e-06, + "loss": 1.4783, + "step": 1005 + }, + { + "epoch": 1.0333846944016436, + "grad_norm": 2.920111894607544, + "learning_rate": 3.6695850780357828e-06, + "loss": 1.4393, + "step": 1006 + }, + { + "epoch": 1.034411915767848, + "grad_norm": 2.919542074203491, + "learning_rate": 3.667681766273316e-06, + "loss": 1.426, + "step": 1007 + }, + { + "epoch": 1.0354391371340523, + "grad_norm": 2.8268563747406006, + "learning_rate": 3.6657784545108495e-06, + "loss": 1.4391, + "step": 1008 + }, + { + "epoch": 1.0364663585002567, + "grad_norm": 2.711895227432251, + "learning_rate": 3.6638751427483824e-06, + "loss": 1.5107, + "step": 1009 + }, + { + "epoch": 1.0374935798664613, + "grad_norm": 2.7996091842651367, + "learning_rate": 3.6619718309859158e-06, + "loss": 1.4679, + "step": 1010 + }, + { + "epoch": 1.0374935798664613, + "eval_loss": 1.800447940826416, + "eval_runtime": 36.9218, + "eval_samples_per_second": 27.084, + "eval_steps_per_second": 1.138, + "step": 1010 + }, + { + "epoch": 1.0385208012326657, + "grad_norm": 2.8792128562927246, + "learning_rate": 3.660068519223449e-06, + "loss": 1.5156, + "step": 1011 + }, + { + "epoch": 1.03954802259887, + "grad_norm": 2.7948005199432373, + "learning_rate": 3.6581652074609825e-06, + "loss": 1.4709, + "step": 1012 + }, + { + "epoch": 1.0405752439650744, + "grad_norm": 2.7429840564727783, + "learning_rate": 3.656261895698516e-06, + "loss": 1.4414, + "step": 1013 + }, + { + "epoch": 1.041602465331279, + "grad_norm": 2.7512004375457764, + "learning_rate": 3.6543585839360492e-06, + "loss": 1.4326, + "step": 1014 + }, + { + "epoch": 1.0426296866974833, + "grad_norm": 2.7879600524902344, + "learning_rate": 3.652455272173582e-06, + "loss": 1.4292, + "step": 1015 + }, + { + "epoch": 1.0436569080636877, + "grad_norm": 2.568354368209839, + "learning_rate": 3.6505519604111155e-06, + "loss": 1.4383, + "step": 1016 + }, + { + "epoch": 1.044684129429892, + "grad_norm": 2.811404228210449, + "learning_rate": 3.648648648648649e-06, + "loss": 1.4573, + "step": 1017 + }, + { + "epoch": 1.0457113507960965, + "grad_norm": 2.909632921218872, + "learning_rate": 3.6467453368861823e-06, + "loss": 1.4934, + "step": 1018 + }, + { + "epoch": 1.046738572162301, + "grad_norm": 2.8014776706695557, + "learning_rate": 3.6448420251237156e-06, + "loss": 1.4433, + "step": 1019 + }, + { + "epoch": 1.0477657935285054, + "grad_norm": 2.8205742835998535, + "learning_rate": 3.6429387133612486e-06, + "loss": 1.4752, + "step": 1020 + }, + { + "epoch": 1.0477657935285054, + "eval_loss": 1.8070552349090576, + "eval_runtime": 37.1396, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 1020 + }, + { + "epoch": 1.0487930148947098, + "grad_norm": 2.768143653869629, + "learning_rate": 3.641035401598782e-06, + "loss": 1.4928, + "step": 1021 + }, + { + "epoch": 1.0498202362609141, + "grad_norm": 2.8027846813201904, + "learning_rate": 3.6391320898363153e-06, + "loss": 1.4416, + "step": 1022 + }, + { + "epoch": 1.0508474576271187, + "grad_norm": 2.832526683807373, + "learning_rate": 3.6372287780738487e-06, + "loss": 1.4796, + "step": 1023 + }, + { + "epoch": 1.051874678993323, + "grad_norm": 2.8225502967834473, + "learning_rate": 3.635325466311382e-06, + "loss": 1.4553, + "step": 1024 + }, + { + "epoch": 1.0529019003595275, + "grad_norm": 2.7793405055999756, + "learning_rate": 3.633422154548916e-06, + "loss": 1.4756, + "step": 1025 + }, + { + "epoch": 1.0539291217257318, + "grad_norm": 2.872873544692993, + "learning_rate": 3.6315188427864487e-06, + "loss": 1.496, + "step": 1026 + }, + { + "epoch": 1.0549563430919364, + "grad_norm": 2.7574472427368164, + "learning_rate": 3.629615531023982e-06, + "loss": 1.464, + "step": 1027 + }, + { + "epoch": 1.0559835644581408, + "grad_norm": 2.7946009635925293, + "learning_rate": 3.6277122192615155e-06, + "loss": 1.4149, + "step": 1028 + }, + { + "epoch": 1.0570107858243452, + "grad_norm": 2.764500617980957, + "learning_rate": 3.625808907499049e-06, + "loss": 1.4756, + "step": 1029 + }, + { + "epoch": 1.0580380071905495, + "grad_norm": 2.659693479537964, + "learning_rate": 3.623905595736582e-06, + "loss": 1.4385, + "step": 1030 + }, + { + "epoch": 1.0580380071905495, + "eval_loss": 1.8091230392456055, + "eval_runtime": 37.109, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 1030 + }, + { + "epoch": 1.059065228556754, + "grad_norm": 2.7500619888305664, + "learning_rate": 3.6220022839741156e-06, + "loss": 1.4334, + "step": 1031 + }, + { + "epoch": 1.0600924499229585, + "grad_norm": 2.7663776874542236, + "learning_rate": 3.6200989722116485e-06, + "loss": 1.4194, + "step": 1032 + }, + { + "epoch": 1.0611196712891628, + "grad_norm": 2.820913076400757, + "learning_rate": 3.618195660449182e-06, + "loss": 1.4453, + "step": 1033 + }, + { + "epoch": 1.0621468926553672, + "grad_norm": 2.8579916954040527, + "learning_rate": 3.6162923486867152e-06, + "loss": 1.4529, + "step": 1034 + }, + { + "epoch": 1.0631741140215716, + "grad_norm": 2.729501247406006, + "learning_rate": 3.6143890369242486e-06, + "loss": 1.3907, + "step": 1035 + }, + { + "epoch": 1.064201335387776, + "grad_norm": 2.7084193229675293, + "learning_rate": 3.612485725161782e-06, + "loss": 1.4252, + "step": 1036 + }, + { + "epoch": 1.0652285567539805, + "grad_norm": 2.7723402976989746, + "learning_rate": 3.610582413399315e-06, + "loss": 1.4339, + "step": 1037 + }, + { + "epoch": 1.066255778120185, + "grad_norm": 2.5925116539001465, + "learning_rate": 3.6086791016368482e-06, + "loss": 1.4252, + "step": 1038 + }, + { + "epoch": 1.0672829994863893, + "grad_norm": 2.7101070880889893, + "learning_rate": 3.6067757898743816e-06, + "loss": 1.4405, + "step": 1039 + }, + { + "epoch": 1.0683102208525936, + "grad_norm": 2.786576509475708, + "learning_rate": 3.604872478111915e-06, + "loss": 1.5049, + "step": 1040 + }, + { + "epoch": 1.0683102208525936, + "eval_loss": 1.8023844957351685, + "eval_runtime": 37.0823, + "eval_samples_per_second": 26.967, + "eval_steps_per_second": 1.133, + "step": 1040 + }, + { + "epoch": 1.0693374422187982, + "grad_norm": 2.8227438926696777, + "learning_rate": 3.6029691663494483e-06, + "loss": 1.5062, + "step": 1041 + }, + { + "epoch": 1.0703646635850026, + "grad_norm": 2.722092866897583, + "learning_rate": 3.6010658545869813e-06, + "loss": 1.4508, + "step": 1042 + }, + { + "epoch": 1.071391884951207, + "grad_norm": 2.719161033630371, + "learning_rate": 3.5991625428245146e-06, + "loss": 1.4791, + "step": 1043 + }, + { + "epoch": 1.0724191063174113, + "grad_norm": 2.8131134510040283, + "learning_rate": 3.597259231062048e-06, + "loss": 1.4811, + "step": 1044 + }, + { + "epoch": 1.073446327683616, + "grad_norm": 2.7374107837677, + "learning_rate": 3.5953559192995818e-06, + "loss": 1.4063, + "step": 1045 + }, + { + "epoch": 1.0744735490498203, + "grad_norm": 2.7651076316833496, + "learning_rate": 3.593452607537115e-06, + "loss": 1.4589, + "step": 1046 + }, + { + "epoch": 1.0755007704160247, + "grad_norm": 2.806370735168457, + "learning_rate": 3.5915492957746485e-06, + "loss": 1.5197, + "step": 1047 + }, + { + "epoch": 1.076527991782229, + "grad_norm": 2.8023886680603027, + "learning_rate": 3.589645984012182e-06, + "loss": 1.4427, + "step": 1048 + }, + { + "epoch": 1.0775552131484334, + "grad_norm": 2.7536637783050537, + "learning_rate": 3.587742672249715e-06, + "loss": 1.4753, + "step": 1049 + }, + { + "epoch": 1.078582434514638, + "grad_norm": 2.6758344173431396, + "learning_rate": 3.585839360487248e-06, + "loss": 1.4302, + "step": 1050 + }, + { + "epoch": 1.078582434514638, + "eval_loss": 1.7898778915405273, + "eval_runtime": 37.04, + "eval_samples_per_second": 26.998, + "eval_steps_per_second": 1.134, + "step": 1050 + }, + { + "epoch": 1.0796096558808423, + "grad_norm": 2.688101053237915, + "learning_rate": 3.5839360487247815e-06, + "loss": 1.4271, + "step": 1051 + }, + { + "epoch": 1.0806368772470467, + "grad_norm": 2.8667550086975098, + "learning_rate": 3.582032736962315e-06, + "loss": 1.4372, + "step": 1052 + }, + { + "epoch": 1.081664098613251, + "grad_norm": 2.9168481826782227, + "learning_rate": 3.5801294251998483e-06, + "loss": 1.5063, + "step": 1053 + }, + { + "epoch": 1.0826913199794557, + "grad_norm": 2.743194580078125, + "learning_rate": 3.578226113437381e-06, + "loss": 1.4041, + "step": 1054 + }, + { + "epoch": 1.08371854134566, + "grad_norm": 2.6846320629119873, + "learning_rate": 3.5763228016749146e-06, + "loss": 1.4427, + "step": 1055 + }, + { + "epoch": 1.0847457627118644, + "grad_norm": 3.1511945724487305, + "learning_rate": 3.574419489912448e-06, + "loss": 1.515, + "step": 1056 + }, + { + "epoch": 1.0857729840780688, + "grad_norm": 2.7326550483703613, + "learning_rate": 3.5725161781499813e-06, + "loss": 1.4529, + "step": 1057 + }, + { + "epoch": 1.0868002054442731, + "grad_norm": 2.7120213508605957, + "learning_rate": 3.5706128663875147e-06, + "loss": 1.4407, + "step": 1058 + }, + { + "epoch": 1.0878274268104777, + "grad_norm": 2.7596473693847656, + "learning_rate": 3.5687095546250476e-06, + "loss": 1.4606, + "step": 1059 + }, + { + "epoch": 1.088854648176682, + "grad_norm": 2.9972002506256104, + "learning_rate": 3.566806242862581e-06, + "loss": 1.4836, + "step": 1060 + }, + { + "epoch": 1.088854648176682, + "eval_loss": 1.7897655963897705, + "eval_runtime": 37.0403, + "eval_samples_per_second": 26.998, + "eval_steps_per_second": 1.134, + "step": 1060 + }, + { + "epoch": 1.0898818695428865, + "grad_norm": 2.8134756088256836, + "learning_rate": 3.5649029311001143e-06, + "loss": 1.4575, + "step": 1061 + }, + { + "epoch": 1.0909090909090908, + "grad_norm": 2.7311880588531494, + "learning_rate": 3.5629996193376477e-06, + "loss": 1.4326, + "step": 1062 + }, + { + "epoch": 1.0919363122752954, + "grad_norm": 2.7033157348632812, + "learning_rate": 3.561096307575181e-06, + "loss": 1.4457, + "step": 1063 + }, + { + "epoch": 1.0929635336414998, + "grad_norm": 2.8118138313293457, + "learning_rate": 3.559192995812714e-06, + "loss": 1.4869, + "step": 1064 + }, + { + "epoch": 1.0939907550077042, + "grad_norm": 2.8734021186828613, + "learning_rate": 3.557289684050248e-06, + "loss": 1.4964, + "step": 1065 + }, + { + "epoch": 1.0950179763739085, + "grad_norm": 2.8652546405792236, + "learning_rate": 3.555386372287781e-06, + "loss": 1.4856, + "step": 1066 + }, + { + "epoch": 1.0960451977401129, + "grad_norm": 2.6305630207061768, + "learning_rate": 3.5534830605253145e-06, + "loss": 1.49, + "step": 1067 + }, + { + "epoch": 1.0970724191063175, + "grad_norm": 2.9516141414642334, + "learning_rate": 3.551579748762848e-06, + "loss": 1.4617, + "step": 1068 + }, + { + "epoch": 1.0980996404725218, + "grad_norm": 2.862701416015625, + "learning_rate": 3.5496764370003812e-06, + "loss": 1.4481, + "step": 1069 + }, + { + "epoch": 1.0991268618387262, + "grad_norm": 2.829742193222046, + "learning_rate": 3.5477731252379146e-06, + "loss": 1.4309, + "step": 1070 + }, + { + "epoch": 1.0991268618387262, + "eval_loss": 1.7925201654434204, + "eval_runtime": 37.0708, + "eval_samples_per_second": 26.975, + "eval_steps_per_second": 1.133, + "step": 1070 + }, + { + "epoch": 1.1001540832049306, + "grad_norm": 2.786259174346924, + "learning_rate": 3.5458698134754475e-06, + "loss": 1.3945, + "step": 1071 + }, + { + "epoch": 1.1011813045711352, + "grad_norm": 2.766643762588501, + "learning_rate": 3.543966501712981e-06, + "loss": 1.419, + "step": 1072 + }, + { + "epoch": 1.1022085259373395, + "grad_norm": 2.830751419067383, + "learning_rate": 3.5420631899505142e-06, + "loss": 1.4222, + "step": 1073 + }, + { + "epoch": 1.103235747303544, + "grad_norm": 3.1409215927124023, + "learning_rate": 3.5401598781880476e-06, + "loss": 1.5262, + "step": 1074 + }, + { + "epoch": 1.1042629686697483, + "grad_norm": 2.678180694580078, + "learning_rate": 3.538256566425581e-06, + "loss": 1.3482, + "step": 1075 + }, + { + "epoch": 1.1052901900359529, + "grad_norm": 2.935002088546753, + "learning_rate": 3.536353254663114e-06, + "loss": 1.4864, + "step": 1076 + }, + { + "epoch": 1.1063174114021572, + "grad_norm": 2.91631817817688, + "learning_rate": 3.5344499429006473e-06, + "loss": 1.4334, + "step": 1077 + }, + { + "epoch": 1.1073446327683616, + "grad_norm": 2.6676135063171387, + "learning_rate": 3.5325466311381806e-06, + "loss": 1.4236, + "step": 1078 + }, + { + "epoch": 1.108371854134566, + "grad_norm": 2.779284715652466, + "learning_rate": 3.530643319375714e-06, + "loss": 1.463, + "step": 1079 + }, + { + "epoch": 1.1093990755007703, + "grad_norm": 2.9586293697357178, + "learning_rate": 3.5287400076132474e-06, + "loss": 1.3953, + "step": 1080 + }, + { + "epoch": 1.1093990755007703, + "eval_loss": 1.8054182529449463, + "eval_runtime": 37.0591, + "eval_samples_per_second": 26.984, + "eval_steps_per_second": 1.133, + "step": 1080 + }, + { + "epoch": 1.110426296866975, + "grad_norm": 2.990781545639038, + "learning_rate": 3.5268366958507803e-06, + "loss": 1.4249, + "step": 1081 + }, + { + "epoch": 1.1114535182331793, + "grad_norm": 2.8546760082244873, + "learning_rate": 3.5249333840883137e-06, + "loss": 1.4679, + "step": 1082 + }, + { + "epoch": 1.1124807395993837, + "grad_norm": 3.1174535751342773, + "learning_rate": 3.523030072325847e-06, + "loss": 1.4587, + "step": 1083 + }, + { + "epoch": 1.113507960965588, + "grad_norm": 3.0766773223876953, + "learning_rate": 3.5211267605633804e-06, + "loss": 1.422, + "step": 1084 + }, + { + "epoch": 1.1145351823317924, + "grad_norm": 2.842961311340332, + "learning_rate": 3.519223448800914e-06, + "loss": 1.4335, + "step": 1085 + }, + { + "epoch": 1.115562403697997, + "grad_norm": 2.682828426361084, + "learning_rate": 3.5173201370384475e-06, + "loss": 1.4401, + "step": 1086 + }, + { + "epoch": 1.1165896250642013, + "grad_norm": 2.9262278079986572, + "learning_rate": 3.515416825275981e-06, + "loss": 1.4399, + "step": 1087 + }, + { + "epoch": 1.1176168464304057, + "grad_norm": 3.0017175674438477, + "learning_rate": 3.513513513513514e-06, + "loss": 1.4466, + "step": 1088 + }, + { + "epoch": 1.11864406779661, + "grad_norm": 2.6254069805145264, + "learning_rate": 3.511610201751047e-06, + "loss": 1.4662, + "step": 1089 + }, + { + "epoch": 1.1196712891628147, + "grad_norm": 2.8206210136413574, + "learning_rate": 3.5097068899885806e-06, + "loss": 1.3802, + "step": 1090 + }, + { + "epoch": 1.1196712891628147, + "eval_loss": 1.7929506301879883, + "eval_runtime": 37.2033, + "eval_samples_per_second": 26.879, + "eval_steps_per_second": 1.129, + "step": 1090 + }, + { + "epoch": 1.120698510529019, + "grad_norm": 2.769279718399048, + "learning_rate": 3.507803578226114e-06, + "loss": 1.4483, + "step": 1091 + }, + { + "epoch": 1.1217257318952234, + "grad_norm": 2.872576951980591, + "learning_rate": 3.5059002664636473e-06, + "loss": 1.5144, + "step": 1092 + }, + { + "epoch": 1.1227529532614278, + "grad_norm": 2.690979480743408, + "learning_rate": 3.5039969547011802e-06, + "loss": 1.3938, + "step": 1093 + }, + { + "epoch": 1.1237801746276324, + "grad_norm": 2.7444441318511963, + "learning_rate": 3.5020936429387136e-06, + "loss": 1.4295, + "step": 1094 + }, + { + "epoch": 1.1248073959938367, + "grad_norm": 3.083296775817871, + "learning_rate": 3.500190331176247e-06, + "loss": 1.4406, + "step": 1095 + }, + { + "epoch": 1.125834617360041, + "grad_norm": 2.7987470626831055, + "learning_rate": 3.4982870194137803e-06, + "loss": 1.4937, + "step": 1096 + }, + { + "epoch": 1.1268618387262455, + "grad_norm": 2.9728713035583496, + "learning_rate": 3.4963837076513137e-06, + "loss": 1.4253, + "step": 1097 + }, + { + "epoch": 1.1278890600924498, + "grad_norm": 2.9592628479003906, + "learning_rate": 3.4944803958888466e-06, + "loss": 1.4339, + "step": 1098 + }, + { + "epoch": 1.1289162814586544, + "grad_norm": 2.6775944232940674, + "learning_rate": 3.49257708412638e-06, + "loss": 1.3673, + "step": 1099 + }, + { + "epoch": 1.1299435028248588, + "grad_norm": 2.9349329471588135, + "learning_rate": 3.4906737723639133e-06, + "loss": 1.4954, + "step": 1100 + }, + { + "epoch": 1.1299435028248588, + "eval_loss": 1.7869561910629272, + "eval_runtime": 37.3494, + "eval_samples_per_second": 26.774, + "eval_steps_per_second": 1.125, + "step": 1100 + }, + { + "epoch": 1.1309707241910631, + "grad_norm": 3.086017608642578, + "learning_rate": 3.4887704606014467e-06, + "loss": 1.5096, + "step": 1101 + }, + { + "epoch": 1.1319979455572675, + "grad_norm": 2.9253454208374023, + "learning_rate": 3.48686714883898e-06, + "loss": 1.4466, + "step": 1102 + }, + { + "epoch": 1.133025166923472, + "grad_norm": 2.7312660217285156, + "learning_rate": 3.484963837076513e-06, + "loss": 1.4934, + "step": 1103 + }, + { + "epoch": 1.1340523882896765, + "grad_norm": 2.9656598567962646, + "learning_rate": 3.4830605253140464e-06, + "loss": 1.4439, + "step": 1104 + }, + { + "epoch": 1.1350796096558808, + "grad_norm": 3.083383321762085, + "learning_rate": 3.4811572135515797e-06, + "loss": 1.3937, + "step": 1105 + }, + { + "epoch": 1.1361068310220852, + "grad_norm": 2.7466018199920654, + "learning_rate": 3.4792539017891135e-06, + "loss": 1.4244, + "step": 1106 + }, + { + "epoch": 1.1371340523882898, + "grad_norm": 2.755654811859131, + "learning_rate": 3.477350590026647e-06, + "loss": 1.3984, + "step": 1107 + }, + { + "epoch": 1.1381612737544942, + "grad_norm": 3.2099268436431885, + "learning_rate": 3.4754472782641802e-06, + "loss": 1.4775, + "step": 1108 + }, + { + "epoch": 1.1391884951206985, + "grad_norm": 2.843669891357422, + "learning_rate": 3.4735439665017136e-06, + "loss": 1.461, + "step": 1109 + }, + { + "epoch": 1.140215716486903, + "grad_norm": 2.816732883453369, + "learning_rate": 3.4716406547392465e-06, + "loss": 1.429, + "step": 1110 + }, + { + "epoch": 1.140215716486903, + "eval_loss": 1.7837378978729248, + "eval_runtime": 36.9198, + "eval_samples_per_second": 27.086, + "eval_steps_per_second": 1.138, + "step": 1110 + }, + { + "epoch": 1.1412429378531073, + "grad_norm": 2.885383367538452, + "learning_rate": 3.46973734297678e-06, + "loss": 1.4462, + "step": 1111 + }, + { + "epoch": 1.1422701592193119, + "grad_norm": 2.8761868476867676, + "learning_rate": 3.4678340312143133e-06, + "loss": 1.4468, + "step": 1112 + }, + { + "epoch": 1.1432973805855162, + "grad_norm": 2.9817779064178467, + "learning_rate": 3.4659307194518466e-06, + "loss": 1.4919, + "step": 1113 + }, + { + "epoch": 1.1443246019517206, + "grad_norm": 2.931018352508545, + "learning_rate": 3.46402740768938e-06, + "loss": 1.4836, + "step": 1114 + }, + { + "epoch": 1.145351823317925, + "grad_norm": 2.708481788635254, + "learning_rate": 3.462124095926913e-06, + "loss": 1.4092, + "step": 1115 + }, + { + "epoch": 1.1463790446841293, + "grad_norm": 2.8369429111480713, + "learning_rate": 3.4602207841644463e-06, + "loss": 1.4135, + "step": 1116 + }, + { + "epoch": 1.147406266050334, + "grad_norm": 2.8598222732543945, + "learning_rate": 3.4583174724019797e-06, + "loss": 1.4578, + "step": 1117 + }, + { + "epoch": 1.1484334874165383, + "grad_norm": 2.7808401584625244, + "learning_rate": 3.456414160639513e-06, + "loss": 1.5097, + "step": 1118 + }, + { + "epoch": 1.1494607087827426, + "grad_norm": 2.777153253555298, + "learning_rate": 3.4545108488770464e-06, + "loss": 1.4687, + "step": 1119 + }, + { + "epoch": 1.150487930148947, + "grad_norm": 2.750174045562744, + "learning_rate": 3.4526075371145793e-06, + "loss": 1.4513, + "step": 1120 + }, + { + "epoch": 1.150487930148947, + "eval_loss": 1.8012630939483643, + "eval_runtime": 37.0093, + "eval_samples_per_second": 27.02, + "eval_steps_per_second": 1.135, + "step": 1120 + }, + { + "epoch": 1.1515151515151516, + "grad_norm": 2.7932651042938232, + "learning_rate": 3.4507042253521127e-06, + "loss": 1.4574, + "step": 1121 + }, + { + "epoch": 1.152542372881356, + "grad_norm": 2.8366539478302, + "learning_rate": 3.448800913589646e-06, + "loss": 1.5257, + "step": 1122 + }, + { + "epoch": 1.1535695942475603, + "grad_norm": 2.7671217918395996, + "learning_rate": 3.4468976018271794e-06, + "loss": 1.4256, + "step": 1123 + }, + { + "epoch": 1.1545968156137647, + "grad_norm": 2.845287799835205, + "learning_rate": 3.4449942900647128e-06, + "loss": 1.5317, + "step": 1124 + }, + { + "epoch": 1.1556240369799693, + "grad_norm": 2.6672773361206055, + "learning_rate": 3.443090978302246e-06, + "loss": 1.4782, + "step": 1125 + }, + { + "epoch": 1.1566512583461737, + "grad_norm": 2.844306230545044, + "learning_rate": 3.44118766653978e-06, + "loss": 1.4771, + "step": 1126 + }, + { + "epoch": 1.157678479712378, + "grad_norm": 2.723320484161377, + "learning_rate": 3.439284354777313e-06, + "loss": 1.4568, + "step": 1127 + }, + { + "epoch": 1.1587057010785824, + "grad_norm": 2.725198984146118, + "learning_rate": 3.4373810430148462e-06, + "loss": 1.4682, + "step": 1128 + }, + { + "epoch": 1.1597329224447868, + "grad_norm": 2.8702471256256104, + "learning_rate": 3.4354777312523796e-06, + "loss": 1.4079, + "step": 1129 + }, + { + "epoch": 1.1607601438109914, + "grad_norm": 2.924783945083618, + "learning_rate": 3.433574419489913e-06, + "loss": 1.437, + "step": 1130 + }, + { + "epoch": 1.1607601438109914, + "eval_loss": 1.817325234413147, + "eval_runtime": 37.0048, + "eval_samples_per_second": 27.024, + "eval_steps_per_second": 1.135, + "step": 1130 + }, + { + "epoch": 1.1617873651771957, + "grad_norm": 2.8339531421661377, + "learning_rate": 3.4316711077274463e-06, + "loss": 1.469, + "step": 1131 + }, + { + "epoch": 1.1628145865434, + "grad_norm": 2.880077362060547, + "learning_rate": 3.4297677959649792e-06, + "loss": 1.4364, + "step": 1132 + }, + { + "epoch": 1.1638418079096045, + "grad_norm": 2.96932315826416, + "learning_rate": 3.4278644842025126e-06, + "loss": 1.4868, + "step": 1133 + }, + { + "epoch": 1.1648690292758088, + "grad_norm": 2.643693685531616, + "learning_rate": 3.425961172440046e-06, + "loss": 1.428, + "step": 1134 + }, + { + "epoch": 1.1658962506420134, + "grad_norm": 2.9450795650482178, + "learning_rate": 3.4240578606775793e-06, + "loss": 1.5155, + "step": 1135 + }, + { + "epoch": 1.1669234720082178, + "grad_norm": 2.7274043560028076, + "learning_rate": 3.4221545489151127e-06, + "loss": 1.4486, + "step": 1136 + }, + { + "epoch": 1.1679506933744221, + "grad_norm": 2.7501749992370605, + "learning_rate": 3.4202512371526456e-06, + "loss": 1.4254, + "step": 1137 + }, + { + "epoch": 1.1689779147406265, + "grad_norm": 3.0331339836120605, + "learning_rate": 3.418347925390179e-06, + "loss": 1.5338, + "step": 1138 + }, + { + "epoch": 1.170005136106831, + "grad_norm": 2.9705629348754883, + "learning_rate": 3.4164446136277124e-06, + "loss": 1.474, + "step": 1139 + }, + { + "epoch": 1.1710323574730355, + "grad_norm": 2.76196026802063, + "learning_rate": 3.4145413018652457e-06, + "loss": 1.4505, + "step": 1140 + }, + { + "epoch": 1.1710323574730355, + "eval_loss": 1.8043485879898071, + "eval_runtime": 37.0258, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 1140 + }, + { + "epoch": 1.1720595788392398, + "grad_norm": 2.6624460220336914, + "learning_rate": 3.412637990102779e-06, + "loss": 1.4739, + "step": 1141 + }, + { + "epoch": 1.1730868002054442, + "grad_norm": 2.7865853309631348, + "learning_rate": 3.4107346783403125e-06, + "loss": 1.4923, + "step": 1142 + }, + { + "epoch": 1.1741140215716488, + "grad_norm": 2.782196044921875, + "learning_rate": 3.4088313665778454e-06, + "loss": 1.4041, + "step": 1143 + }, + { + "epoch": 1.1751412429378532, + "grad_norm": 2.7796125411987305, + "learning_rate": 3.4069280548153788e-06, + "loss": 1.4446, + "step": 1144 + }, + { + "epoch": 1.1761684643040575, + "grad_norm": 2.8180930614471436, + "learning_rate": 3.405024743052912e-06, + "loss": 1.424, + "step": 1145 + }, + { + "epoch": 1.177195685670262, + "grad_norm": 2.8180930614471436, + "learning_rate": 3.405024743052912e-06, + "loss": 1.4457, + "step": 1146 + }, + { + "epoch": 1.1782229070364663, + "grad_norm": 2.922241449356079, + "learning_rate": 3.403121431290446e-06, + "loss": 1.4162, + "step": 1147 + }, + { + "epoch": 1.1792501284026708, + "grad_norm": 2.850069761276245, + "learning_rate": 3.4012181195279793e-06, + "loss": 1.432, + "step": 1148 + }, + { + "epoch": 1.1802773497688752, + "grad_norm": 2.863086462020874, + "learning_rate": 3.3993148077655126e-06, + "loss": 1.442, + "step": 1149 + }, + { + "epoch": 1.1813045711350796, + "grad_norm": 2.9545702934265137, + "learning_rate": 3.3974114960030456e-06, + "loss": 1.3931, + "step": 1150 + }, + { + "epoch": 1.1813045711350796, + "eval_loss": 1.812690258026123, + "eval_runtime": 37.0328, + "eval_samples_per_second": 27.003, + "eval_steps_per_second": 1.134, + "step": 1150 + }, + { + "epoch": 1.182331792501284, + "grad_norm": 2.9074196815490723, + "learning_rate": 3.395508184240579e-06, + "loss": 1.4199, + "step": 1151 + }, + { + "epoch": 1.1833590138674885, + "grad_norm": 3.1387181282043457, + "learning_rate": 3.3936048724781123e-06, + "loss": 1.509, + "step": 1152 + }, + { + "epoch": 1.184386235233693, + "grad_norm": 3.0677523612976074, + "learning_rate": 3.3917015607156457e-06, + "loss": 1.3952, + "step": 1153 + }, + { + "epoch": 1.1854134565998973, + "grad_norm": 2.8462986946105957, + "learning_rate": 3.389798248953179e-06, + "loss": 1.415, + "step": 1154 + }, + { + "epoch": 1.1864406779661016, + "grad_norm": 2.9588911533355713, + "learning_rate": 3.387894937190712e-06, + "loss": 1.464, + "step": 1155 + }, + { + "epoch": 1.1874678993323062, + "grad_norm": 2.7406837940216064, + "learning_rate": 3.3859916254282453e-06, + "loss": 1.4604, + "step": 1156 + }, + { + "epoch": 1.1884951206985106, + "grad_norm": 2.837615489959717, + "learning_rate": 3.3840883136657787e-06, + "loss": 1.5108, + "step": 1157 + }, + { + "epoch": 1.189522342064715, + "grad_norm": 2.6457152366638184, + "learning_rate": 3.382185001903312e-06, + "loss": 1.4247, + "step": 1158 + }, + { + "epoch": 1.1905495634309193, + "grad_norm": 2.887178659439087, + "learning_rate": 3.3802816901408454e-06, + "loss": 1.4609, + "step": 1159 + }, + { + "epoch": 1.1915767847971237, + "grad_norm": 2.6568970680236816, + "learning_rate": 3.3783783783783788e-06, + "loss": 1.4393, + "step": 1160 + }, + { + "epoch": 1.1915767847971237, + "eval_loss": 1.7956509590148926, + "eval_runtime": 37.0256, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 1160 + }, + { + "epoch": 1.1926040061633283, + "grad_norm": 2.731114625930786, + "learning_rate": 3.3764750666159117e-06, + "loss": 1.4399, + "step": 1161 + }, + { + "epoch": 1.1936312275295327, + "grad_norm": 2.8750929832458496, + "learning_rate": 3.374571754853445e-06, + "loss": 1.4318, + "step": 1162 + }, + { + "epoch": 1.194658448895737, + "grad_norm": 2.697359085083008, + "learning_rate": 3.3726684430909784e-06, + "loss": 1.3884, + "step": 1163 + }, + { + "epoch": 1.1956856702619414, + "grad_norm": 2.6850969791412354, + "learning_rate": 3.370765131328512e-06, + "loss": 1.4104, + "step": 1164 + }, + { + "epoch": 1.1967128916281458, + "grad_norm": 2.739121913909912, + "learning_rate": 3.368861819566045e-06, + "loss": 1.4227, + "step": 1165 + }, + { + "epoch": 1.1977401129943503, + "grad_norm": 2.7206380367279053, + "learning_rate": 3.366958507803578e-06, + "loss": 1.4587, + "step": 1166 + }, + { + "epoch": 1.1987673343605547, + "grad_norm": 2.651898145675659, + "learning_rate": 3.365055196041112e-06, + "loss": 1.4081, + "step": 1167 + }, + { + "epoch": 1.199794555726759, + "grad_norm": 2.7718563079833984, + "learning_rate": 3.3631518842786452e-06, + "loss": 1.4835, + "step": 1168 + }, + { + "epoch": 1.2008217770929634, + "grad_norm": 2.771209716796875, + "learning_rate": 3.3612485725161786e-06, + "loss": 1.4036, + "step": 1169 + }, + { + "epoch": 1.201848998459168, + "grad_norm": 2.6938540935516357, + "learning_rate": 3.359345260753712e-06, + "loss": 1.4987, + "step": 1170 + }, + { + "epoch": 1.201848998459168, + "eval_loss": 1.784244179725647, + "eval_runtime": 37.0121, + "eval_samples_per_second": 27.018, + "eval_steps_per_second": 1.135, + "step": 1170 + }, + { + "epoch": 1.2028762198253724, + "grad_norm": 2.884730339050293, + "learning_rate": 3.3574419489912453e-06, + "loss": 1.4778, + "step": 1171 + }, + { + "epoch": 1.2039034411915768, + "grad_norm": 3.020632266998291, + "learning_rate": 3.3555386372287783e-06, + "loss": 1.5133, + "step": 1172 + }, + { + "epoch": 1.2049306625577811, + "grad_norm": 2.8026037216186523, + "learning_rate": 3.3536353254663116e-06, + "loss": 1.405, + "step": 1173 + }, + { + "epoch": 1.2059578839239857, + "grad_norm": 2.763852596282959, + "learning_rate": 3.351732013703845e-06, + "loss": 1.4437, + "step": 1174 + }, + { + "epoch": 1.20698510529019, + "grad_norm": 2.897768259048462, + "learning_rate": 3.3498287019413784e-06, + "loss": 1.4548, + "step": 1175 + }, + { + "epoch": 1.2080123266563945, + "grad_norm": 2.8483922481536865, + "learning_rate": 3.3479253901789117e-06, + "loss": 1.4734, + "step": 1176 + }, + { + "epoch": 1.2090395480225988, + "grad_norm": 2.8507916927337646, + "learning_rate": 3.346022078416445e-06, + "loss": 1.378, + "step": 1177 + }, + { + "epoch": 1.2100667693888032, + "grad_norm": 2.74048113822937, + "learning_rate": 3.344118766653978e-06, + "loss": 1.4493, + "step": 1178 + }, + { + "epoch": 1.2110939907550078, + "grad_norm": 2.8064157962799072, + "learning_rate": 3.3422154548915114e-06, + "loss": 1.4221, + "step": 1179 + }, + { + "epoch": 1.2121212121212122, + "grad_norm": 2.721090793609619, + "learning_rate": 3.3403121431290448e-06, + "loss": 1.4603, + "step": 1180 + }, + { + "epoch": 1.2121212121212122, + "eval_loss": 1.7907730340957642, + "eval_runtime": 37.0168, + "eval_samples_per_second": 27.015, + "eval_steps_per_second": 1.135, + "step": 1180 + }, + { + "epoch": 1.2131484334874165, + "grad_norm": 2.706892967224121, + "learning_rate": 3.338408831366578e-06, + "loss": 1.3903, + "step": 1181 + }, + { + "epoch": 1.214175654853621, + "grad_norm": 3.0193893909454346, + "learning_rate": 3.3365055196041115e-06, + "loss": 1.4715, + "step": 1182 + }, + { + "epoch": 1.2152028762198253, + "grad_norm": 2.8713300228118896, + "learning_rate": 3.3346022078416444e-06, + "loss": 1.5017, + "step": 1183 + }, + { + "epoch": 1.2162300975860298, + "grad_norm": 2.7250728607177734, + "learning_rate": 3.3326988960791778e-06, + "loss": 1.4242, + "step": 1184 + }, + { + "epoch": 1.2172573189522342, + "grad_norm": 2.9367101192474365, + "learning_rate": 3.330795584316711e-06, + "loss": 1.3563, + "step": 1185 + }, + { + "epoch": 1.2182845403184386, + "grad_norm": 3.059776544570923, + "learning_rate": 3.3288922725542445e-06, + "loss": 1.4334, + "step": 1186 + }, + { + "epoch": 1.219311761684643, + "grad_norm": 2.996394157409668, + "learning_rate": 3.326988960791778e-06, + "loss": 1.4457, + "step": 1187 + }, + { + "epoch": 1.2203389830508475, + "grad_norm": 2.8958983421325684, + "learning_rate": 3.3250856490293117e-06, + "loss": 1.4904, + "step": 1188 + }, + { + "epoch": 1.221366204417052, + "grad_norm": 2.9497668743133545, + "learning_rate": 3.323182337266845e-06, + "loss": 1.4836, + "step": 1189 + }, + { + "epoch": 1.2223934257832563, + "grad_norm": 2.87447452545166, + "learning_rate": 3.321279025504378e-06, + "loss": 1.4664, + "step": 1190 + }, + { + "epoch": 1.2223934257832563, + "eval_loss": 1.7865233421325684, + "eval_runtime": 37.1144, + "eval_samples_per_second": 26.944, + "eval_steps_per_second": 1.132, + "step": 1190 + }, + { + "epoch": 1.2234206471494606, + "grad_norm": 2.7893576622009277, + "learning_rate": 3.3193757137419113e-06, + "loss": 1.4584, + "step": 1191 + }, + { + "epoch": 1.2244478685156652, + "grad_norm": 2.791938543319702, + "learning_rate": 3.3174724019794447e-06, + "loss": 1.4974, + "step": 1192 + }, + { + "epoch": 1.2254750898818696, + "grad_norm": 2.7816781997680664, + "learning_rate": 3.315569090216978e-06, + "loss": 1.4103, + "step": 1193 + }, + { + "epoch": 1.226502311248074, + "grad_norm": 2.938931703567505, + "learning_rate": 3.3136657784545114e-06, + "loss": 1.498, + "step": 1194 + }, + { + "epoch": 1.2275295326142783, + "grad_norm": 2.7500035762786865, + "learning_rate": 3.3117624666920443e-06, + "loss": 1.4044, + "step": 1195 + }, + { + "epoch": 1.2285567539804827, + "grad_norm": 2.758352756500244, + "learning_rate": 3.3098591549295777e-06, + "loss": 1.3968, + "step": 1196 + }, + { + "epoch": 1.2295839753466873, + "grad_norm": 3.144857406616211, + "learning_rate": 3.307955843167111e-06, + "loss": 1.459, + "step": 1197 + }, + { + "epoch": 1.2306111967128917, + "grad_norm": 2.777985095977783, + "learning_rate": 3.3060525314046444e-06, + "loss": 1.4419, + "step": 1198 + }, + { + "epoch": 1.231638418079096, + "grad_norm": 2.812570810317993, + "learning_rate": 3.304149219642178e-06, + "loss": 1.446, + "step": 1199 + }, + { + "epoch": 1.2326656394453004, + "grad_norm": 2.6672797203063965, + "learning_rate": 3.3022459078797107e-06, + "loss": 1.4615, + "step": 1200 + }, + { + "epoch": 1.2326656394453004, + "eval_loss": 1.7857756614685059, + "eval_runtime": 37.313, + "eval_samples_per_second": 26.8, + "eval_steps_per_second": 1.126, + "step": 1200 + }, + { + "epoch": 1.2336928608115048, + "grad_norm": 2.8981986045837402, + "learning_rate": 3.300342596117244e-06, + "loss": 1.4418, + "step": 1201 + }, + { + "epoch": 1.2347200821777093, + "grad_norm": 2.7738826274871826, + "learning_rate": 3.2984392843547775e-06, + "loss": 1.3926, + "step": 1202 + }, + { + "epoch": 1.2357473035439137, + "grad_norm": 2.735461950302124, + "learning_rate": 3.296535972592311e-06, + "loss": 1.3843, + "step": 1203 + }, + { + "epoch": 1.236774524910118, + "grad_norm": 2.965693950653076, + "learning_rate": 3.294632660829844e-06, + "loss": 1.4248, + "step": 1204 + }, + { + "epoch": 1.2378017462763227, + "grad_norm": 2.850008010864258, + "learning_rate": 3.292729349067377e-06, + "loss": 1.4277, + "step": 1205 + }, + { + "epoch": 1.238828967642527, + "grad_norm": 2.823054790496826, + "learning_rate": 3.2908260373049105e-06, + "loss": 1.4151, + "step": 1206 + }, + { + "epoch": 1.2398561890087314, + "grad_norm": 2.7468149662017822, + "learning_rate": 3.288922725542444e-06, + "loss": 1.4661, + "step": 1207 + }, + { + "epoch": 1.2408834103749358, + "grad_norm": 2.7379629611968994, + "learning_rate": 3.2870194137799776e-06, + "loss": 1.4219, + "step": 1208 + }, + { + "epoch": 1.2419106317411401, + "grad_norm": 2.7809221744537354, + "learning_rate": 3.285116102017511e-06, + "loss": 1.4002, + "step": 1209 + }, + { + "epoch": 1.2429378531073447, + "grad_norm": 2.9209654331207275, + "learning_rate": 3.2832127902550444e-06, + "loss": 1.4646, + "step": 1210 + }, + { + "epoch": 1.2429378531073447, + "eval_loss": 1.786431908607483, + "eval_runtime": 36.9218, + "eval_samples_per_second": 27.084, + "eval_steps_per_second": 1.138, + "step": 1210 + }, + { + "epoch": 1.243965074473549, + "grad_norm": 2.9019596576690674, + "learning_rate": 3.2813094784925777e-06, + "loss": 1.4526, + "step": 1211 + }, + { + "epoch": 1.2449922958397535, + "grad_norm": 2.7218263149261475, + "learning_rate": 3.2794061667301107e-06, + "loss": 1.3911, + "step": 1212 + }, + { + "epoch": 1.2460195172059578, + "grad_norm": 2.8646154403686523, + "learning_rate": 3.277502854967644e-06, + "loss": 1.4305, + "step": 1213 + }, + { + "epoch": 1.2470467385721622, + "grad_norm": 2.9065611362457275, + "learning_rate": 3.2755995432051774e-06, + "loss": 1.4239, + "step": 1214 + }, + { + "epoch": 1.2480739599383668, + "grad_norm": 2.7945196628570557, + "learning_rate": 3.2736962314427107e-06, + "loss": 1.4321, + "step": 1215 + }, + { + "epoch": 1.2491011813045712, + "grad_norm": 2.644244909286499, + "learning_rate": 3.271792919680244e-06, + "loss": 1.4411, + "step": 1216 + }, + { + "epoch": 1.2501284026707755, + "grad_norm": 2.9058117866516113, + "learning_rate": 3.269889607917777e-06, + "loss": 1.4942, + "step": 1217 + }, + { + "epoch": 1.25115562403698, + "grad_norm": 2.8242623805999756, + "learning_rate": 3.2679862961553104e-06, + "loss": 1.4033, + "step": 1218 + }, + { + "epoch": 1.2521828454031843, + "grad_norm": 2.859428644180298, + "learning_rate": 3.2660829843928438e-06, + "loss": 1.4298, + "step": 1219 + }, + { + "epoch": 1.2532100667693888, + "grad_norm": 2.7864537239074707, + "learning_rate": 3.264179672630377e-06, + "loss": 1.4827, + "step": 1220 + }, + { + "epoch": 1.2532100667693888, + "eval_loss": 1.7798999547958374, + "eval_runtime": 37.0394, + "eval_samples_per_second": 26.998, + "eval_steps_per_second": 1.134, + "step": 1220 + }, + { + "epoch": 1.2542372881355932, + "grad_norm": 2.7938263416290283, + "learning_rate": 3.2622763608679105e-06, + "loss": 1.4412, + "step": 1221 + }, + { + "epoch": 1.2552645095017976, + "grad_norm": 2.811281204223633, + "learning_rate": 3.2603730491054434e-06, + "loss": 1.4139, + "step": 1222 + }, + { + "epoch": 1.2562917308680022, + "grad_norm": 2.7984230518341064, + "learning_rate": 3.258469737342977e-06, + "loss": 1.4205, + "step": 1223 + }, + { + "epoch": 1.2573189522342065, + "grad_norm": 2.7457568645477295, + "learning_rate": 3.25656642558051e-06, + "loss": 1.396, + "step": 1224 + }, + { + "epoch": 1.258346173600411, + "grad_norm": 2.84529972076416, + "learning_rate": 3.2546631138180435e-06, + "loss": 1.4192, + "step": 1225 + }, + { + "epoch": 1.2593733949666153, + "grad_norm": 2.8183720111846924, + "learning_rate": 3.252759802055577e-06, + "loss": 1.4104, + "step": 1226 + }, + { + "epoch": 1.2604006163328196, + "grad_norm": 2.8872454166412354, + "learning_rate": 3.25085649029311e-06, + "loss": 1.4256, + "step": 1227 + }, + { + "epoch": 1.2614278376990242, + "grad_norm": 2.935094118118286, + "learning_rate": 3.248953178530644e-06, + "loss": 1.4558, + "step": 1228 + }, + { + "epoch": 1.2624550590652286, + "grad_norm": 2.69989013671875, + "learning_rate": 3.247049866768177e-06, + "loss": 1.4167, + "step": 1229 + }, + { + "epoch": 1.263482280431433, + "grad_norm": 2.781031847000122, + "learning_rate": 3.2451465550057103e-06, + "loss": 1.3745, + "step": 1230 + }, + { + "epoch": 1.263482280431433, + "eval_loss": 1.7843073606491089, + "eval_runtime": 37.0672, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 1230 + }, + { + "epoch": 1.2645095017976373, + "grad_norm": 2.974919080734253, + "learning_rate": 3.2432432432432437e-06, + "loss": 1.4221, + "step": 1231 + }, + { + "epoch": 1.2655367231638417, + "grad_norm": 3.095144271850586, + "learning_rate": 3.241339931480777e-06, + "loss": 1.4539, + "step": 1232 + }, + { + "epoch": 1.2665639445300463, + "grad_norm": 3.0127196311950684, + "learning_rate": 3.2394366197183104e-06, + "loss": 1.4559, + "step": 1233 + }, + { + "epoch": 1.2675911658962506, + "grad_norm": 2.7474801540374756, + "learning_rate": 3.2375333079558434e-06, + "loss": 1.4177, + "step": 1234 + }, + { + "epoch": 1.268618387262455, + "grad_norm": 2.743516445159912, + "learning_rate": 3.2356299961933767e-06, + "loss": 1.4471, + "step": 1235 + }, + { + "epoch": 1.2696456086286596, + "grad_norm": 2.7801740169525146, + "learning_rate": 3.23372668443091e-06, + "loss": 1.4588, + "step": 1236 + }, + { + "epoch": 1.270672829994864, + "grad_norm": 2.697547435760498, + "learning_rate": 3.2318233726684435e-06, + "loss": 1.4857, + "step": 1237 + }, + { + "epoch": 1.2717000513610683, + "grad_norm": 2.646890640258789, + "learning_rate": 3.229920060905977e-06, + "loss": 1.4068, + "step": 1238 + }, + { + "epoch": 1.2727272727272727, + "grad_norm": 2.694401979446411, + "learning_rate": 3.2280167491435098e-06, + "loss": 1.4148, + "step": 1239 + }, + { + "epoch": 1.273754494093477, + "grad_norm": 2.6139938831329346, + "learning_rate": 3.226113437381043e-06, + "loss": 1.4214, + "step": 1240 + }, + { + "epoch": 1.273754494093477, + "eval_loss": 1.7777915000915527, + "eval_runtime": 37.1013, + "eval_samples_per_second": 26.953, + "eval_steps_per_second": 1.132, + "step": 1240 + }, + { + "epoch": 1.2747817154596817, + "grad_norm": 2.7611944675445557, + "learning_rate": 3.2242101256185765e-06, + "loss": 1.4344, + "step": 1241 + }, + { + "epoch": 1.275808936825886, + "grad_norm": 2.7328102588653564, + "learning_rate": 3.22230681385611e-06, + "loss": 1.4891, + "step": 1242 + }, + { + "epoch": 1.2768361581920904, + "grad_norm": 2.822169065475464, + "learning_rate": 3.220403502093643e-06, + "loss": 1.4328, + "step": 1243 + }, + { + "epoch": 1.2778633795582948, + "grad_norm": 2.878005027770996, + "learning_rate": 3.218500190331176e-06, + "loss": 1.4779, + "step": 1244 + }, + { + "epoch": 1.2788906009244991, + "grad_norm": 2.725376844406128, + "learning_rate": 3.2165968785687095e-06, + "loss": 1.424, + "step": 1245 + }, + { + "epoch": 1.2799178222907037, + "grad_norm": 2.7598440647125244, + "learning_rate": 3.214693566806243e-06, + "loss": 1.4533, + "step": 1246 + }, + { + "epoch": 1.280945043656908, + "grad_norm": 2.8355820178985596, + "learning_rate": 3.2127902550437762e-06, + "loss": 1.4433, + "step": 1247 + }, + { + "epoch": 1.2819722650231125, + "grad_norm": 2.7520253658294678, + "learning_rate": 3.21088694328131e-06, + "loss": 1.4318, + "step": 1248 + }, + { + "epoch": 1.2829994863893168, + "grad_norm": 2.7606403827667236, + "learning_rate": 3.2089836315188434e-06, + "loss": 1.4242, + "step": 1249 + }, + { + "epoch": 1.2840267077555212, + "grad_norm": 2.784301280975342, + "learning_rate": 3.2070803197563767e-06, + "loss": 1.4253, + "step": 1250 + }, + { + "epoch": 1.2840267077555212, + "eval_loss": 1.7868103981018066, + "eval_runtime": 37.0911, + "eval_samples_per_second": 26.961, + "eval_steps_per_second": 1.132, + "step": 1250 + }, + { + "epoch": 1.2850539291217258, + "grad_norm": 2.8295185565948486, + "learning_rate": 3.2051770079939097e-06, + "loss": 1.4432, + "step": 1251 + }, + { + "epoch": 1.2860811504879301, + "grad_norm": 2.8488125801086426, + "learning_rate": 3.203273696231443e-06, + "loss": 1.4714, + "step": 1252 + }, + { + "epoch": 1.2871083718541345, + "grad_norm": 2.771249294281006, + "learning_rate": 3.2013703844689764e-06, + "loss": 1.4093, + "step": 1253 + }, + { + "epoch": 1.288135593220339, + "grad_norm": 2.782425880432129, + "learning_rate": 3.1994670727065098e-06, + "loss": 1.4597, + "step": 1254 + }, + { + "epoch": 1.2891628145865435, + "grad_norm": 2.88934588432312, + "learning_rate": 3.197563760944043e-06, + "loss": 1.4805, + "step": 1255 + }, + { + "epoch": 1.2901900359527478, + "grad_norm": 2.8694138526916504, + "learning_rate": 3.195660449181576e-06, + "loss": 1.4699, + "step": 1256 + }, + { + "epoch": 1.2912172573189522, + "grad_norm": 2.799614667892456, + "learning_rate": 3.1937571374191094e-06, + "loss": 1.4044, + "step": 1257 + }, + { + "epoch": 1.2922444786851566, + "grad_norm": 3.001746416091919, + "learning_rate": 3.191853825656643e-06, + "loss": 1.4305, + "step": 1258 + }, + { + "epoch": 1.2932717000513612, + "grad_norm": 2.8572680950164795, + "learning_rate": 3.189950513894176e-06, + "loss": 1.4252, + "step": 1259 + }, + { + "epoch": 1.2942989214175655, + "grad_norm": 2.942213535308838, + "learning_rate": 3.1880472021317095e-06, + "loss": 1.427, + "step": 1260 + }, + { + "epoch": 1.2942989214175655, + "eval_loss": 1.7929412126541138, + "eval_runtime": 37.087, + "eval_samples_per_second": 26.964, + "eval_steps_per_second": 1.132, + "step": 1260 + }, + { + "epoch": 1.29532614278377, + "grad_norm": 2.893172264099121, + "learning_rate": 3.1861438903692425e-06, + "loss": 1.379, + "step": 1261 + }, + { + "epoch": 1.2963533641499743, + "grad_norm": 2.9168691635131836, + "learning_rate": 3.184240578606776e-06, + "loss": 1.4124, + "step": 1262 + }, + { + "epoch": 1.2973805855161786, + "grad_norm": 2.7513487339019775, + "learning_rate": 3.182337266844309e-06, + "loss": 1.4343, + "step": 1263 + }, + { + "epoch": 1.2984078068823832, + "grad_norm": 2.8485500812530518, + "learning_rate": 3.1804339550818426e-06, + "loss": 1.467, + "step": 1264 + }, + { + "epoch": 1.2994350282485876, + "grad_norm": 2.6620397567749023, + "learning_rate": 3.178530643319376e-06, + "loss": 1.4191, + "step": 1265 + }, + { + "epoch": 1.300462249614792, + "grad_norm": 2.659393072128296, + "learning_rate": 3.176627331556909e-06, + "loss": 1.3934, + "step": 1266 + }, + { + "epoch": 1.3014894709809965, + "grad_norm": 2.8579442501068115, + "learning_rate": 3.1747240197944422e-06, + "loss": 1.4399, + "step": 1267 + }, + { + "epoch": 1.3025166923472007, + "grad_norm": 2.7417430877685547, + "learning_rate": 3.172820708031976e-06, + "loss": 1.3956, + "step": 1268 + }, + { + "epoch": 1.3035439137134053, + "grad_norm": 3.1082868576049805, + "learning_rate": 3.1709173962695094e-06, + "loss": 1.4768, + "step": 1269 + }, + { + "epoch": 1.3045711350796096, + "grad_norm": 3.099637985229492, + "learning_rate": 3.1690140845070427e-06, + "loss": 1.4679, + "step": 1270 + }, + { + "epoch": 1.3045711350796096, + "eval_loss": 1.7878395318984985, + "eval_runtime": 37.0795, + "eval_samples_per_second": 26.969, + "eval_steps_per_second": 1.133, + "step": 1270 + }, + { + "epoch": 1.305598356445814, + "grad_norm": 2.8106865882873535, + "learning_rate": 3.167110772744576e-06, + "loss": 1.4257, + "step": 1271 + }, + { + "epoch": 1.3066255778120186, + "grad_norm": 2.7419369220733643, + "learning_rate": 3.1652074609821095e-06, + "loss": 1.4197, + "step": 1272 + }, + { + "epoch": 1.307652799178223, + "grad_norm": 2.7077760696411133, + "learning_rate": 3.1633041492196424e-06, + "loss": 1.4188, + "step": 1273 + }, + { + "epoch": 1.3086800205444273, + "grad_norm": 2.6611413955688477, + "learning_rate": 3.1614008374571758e-06, + "loss": 1.4456, + "step": 1274 + }, + { + "epoch": 1.3097072419106317, + "grad_norm": 2.8461341857910156, + "learning_rate": 3.159497525694709e-06, + "loss": 1.4544, + "step": 1275 + }, + { + "epoch": 1.310734463276836, + "grad_norm": 2.7263684272766113, + "learning_rate": 3.1575942139322425e-06, + "loss": 1.4177, + "step": 1276 + }, + { + "epoch": 1.3117616846430407, + "grad_norm": 2.8571529388427734, + "learning_rate": 3.155690902169776e-06, + "loss": 1.4473, + "step": 1277 + }, + { + "epoch": 1.312788906009245, + "grad_norm": 2.914841651916504, + "learning_rate": 3.1537875904073088e-06, + "loss": 1.4117, + "step": 1278 + }, + { + "epoch": 1.3138161273754494, + "grad_norm": 2.7753429412841797, + "learning_rate": 3.151884278644842e-06, + "loss": 1.4359, + "step": 1279 + }, + { + "epoch": 1.3148433487416538, + "grad_norm": 2.624159097671509, + "learning_rate": 3.1499809668823755e-06, + "loss": 1.4072, + "step": 1280 + }, + { + "epoch": 1.3148433487416538, + "eval_loss": 1.7820065021514893, + "eval_runtime": 37.0679, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 1280 + }, + { + "epoch": 1.3158705701078581, + "grad_norm": 2.742490291595459, + "learning_rate": 3.148077655119909e-06, + "loss": 1.4681, + "step": 1281 + }, + { + "epoch": 1.3168977914740627, + "grad_norm": 2.6784660816192627, + "learning_rate": 3.1461743433574422e-06, + "loss": 1.406, + "step": 1282 + }, + { + "epoch": 1.317925012840267, + "grad_norm": 2.746730089187622, + "learning_rate": 3.144271031594975e-06, + "loss": 1.4538, + "step": 1283 + }, + { + "epoch": 1.3189522342064715, + "grad_norm": 2.746652603149414, + "learning_rate": 3.1423677198325085e-06, + "loss": 1.4574, + "step": 1284 + }, + { + "epoch": 1.319979455572676, + "grad_norm": 2.842672109603882, + "learning_rate": 3.140464408070042e-06, + "loss": 1.414, + "step": 1285 + }, + { + "epoch": 1.3210066769388804, + "grad_norm": 2.7332921028137207, + "learning_rate": 3.1385610963075753e-06, + "loss": 1.4115, + "step": 1286 + }, + { + "epoch": 1.3220338983050848, + "grad_norm": 2.6771159172058105, + "learning_rate": 3.1366577845451086e-06, + "loss": 1.4023, + "step": 1287 + }, + { + "epoch": 1.3230611196712891, + "grad_norm": 2.689605951309204, + "learning_rate": 3.134754472782642e-06, + "loss": 1.3897, + "step": 1288 + }, + { + "epoch": 1.3240883410374935, + "grad_norm": 2.7937700748443604, + "learning_rate": 3.1328511610201758e-06, + "loss": 1.3783, + "step": 1289 + }, + { + "epoch": 1.325115562403698, + "grad_norm": 2.643263816833496, + "learning_rate": 3.1309478492577087e-06, + "loss": 1.378, + "step": 1290 + }, + { + "epoch": 1.325115562403698, + "eval_loss": 1.7850252389907837, + "eval_runtime": 37.0816, + "eval_samples_per_second": 26.968, + "eval_steps_per_second": 1.133, + "step": 1290 + }, + { + "epoch": 1.3261427837699025, + "grad_norm": 2.7702224254608154, + "learning_rate": 3.129044537495242e-06, + "loss": 1.4236, + "step": 1291 + }, + { + "epoch": 1.3271700051361068, + "grad_norm": 2.783445358276367, + "learning_rate": 3.1271412257327754e-06, + "loss": 1.4454, + "step": 1292 + }, + { + "epoch": 1.3281972265023112, + "grad_norm": 2.556509494781494, + "learning_rate": 3.125237913970309e-06, + "loss": 1.3998, + "step": 1293 + }, + { + "epoch": 1.3292244478685156, + "grad_norm": 2.706526041030884, + "learning_rate": 3.123334602207842e-06, + "loss": 1.3781, + "step": 1294 + }, + { + "epoch": 1.3302516692347202, + "grad_norm": 2.9328322410583496, + "learning_rate": 3.121431290445375e-06, + "loss": 1.4747, + "step": 1295 + }, + { + "epoch": 1.3312788906009245, + "grad_norm": 3.0525310039520264, + "learning_rate": 3.1195279786829085e-06, + "loss": 1.5197, + "step": 1296 + }, + { + "epoch": 1.332306111967129, + "grad_norm": 2.715700149536133, + "learning_rate": 3.117624666920442e-06, + "loss": 1.4059, + "step": 1297 + }, + { + "epoch": 1.3333333333333333, + "grad_norm": 2.7344141006469727, + "learning_rate": 3.115721355157975e-06, + "loss": 1.4101, + "step": 1298 + }, + { + "epoch": 1.3343605546995376, + "grad_norm": 2.746990203857422, + "learning_rate": 3.1138180433955085e-06, + "loss": 1.3625, + "step": 1299 + }, + { + "epoch": 1.3353877760657422, + "grad_norm": 2.817866086959839, + "learning_rate": 3.1119147316330415e-06, + "loss": 1.4414, + "step": 1300 + }, + { + "epoch": 1.3353877760657422, + "eval_loss": 1.7872337102890015, + "eval_runtime": 37.2584, + "eval_samples_per_second": 26.84, + "eval_steps_per_second": 1.127, + "step": 1300 + }, + { + "epoch": 1.3364149974319466, + "grad_norm": 2.814664125442505, + "learning_rate": 3.110011419870575e-06, + "loss": 1.4768, + "step": 1301 + }, + { + "epoch": 1.337442218798151, + "grad_norm": 2.7719714641571045, + "learning_rate": 3.1081081081081082e-06, + "loss": 1.4128, + "step": 1302 + }, + { + "epoch": 1.3384694401643555, + "grad_norm": 2.8300364017486572, + "learning_rate": 3.1062047963456416e-06, + "loss": 1.4419, + "step": 1303 + }, + { + "epoch": 1.33949666153056, + "grad_norm": 2.943674087524414, + "learning_rate": 3.104301484583175e-06, + "loss": 1.4841, + "step": 1304 + }, + { + "epoch": 1.3405238828967643, + "grad_norm": 2.7583727836608887, + "learning_rate": 3.1023981728207083e-06, + "loss": 1.4674, + "step": 1305 + }, + { + "epoch": 1.3415511042629686, + "grad_norm": 2.716261386871338, + "learning_rate": 3.1004948610582412e-06, + "loss": 1.4292, + "step": 1306 + }, + { + "epoch": 1.342578325629173, + "grad_norm": 2.8233978748321533, + "learning_rate": 3.0985915492957746e-06, + "loss": 1.4208, + "step": 1307 + }, + { + "epoch": 1.3436055469953776, + "grad_norm": 2.829420804977417, + "learning_rate": 3.096688237533308e-06, + "loss": 1.3828, + "step": 1308 + }, + { + "epoch": 1.344632768361582, + "grad_norm": 2.791999101638794, + "learning_rate": 3.0947849257708418e-06, + "loss": 1.4751, + "step": 1309 + }, + { + "epoch": 1.3456599897277863, + "grad_norm": 2.691257953643799, + "learning_rate": 3.092881614008375e-06, + "loss": 1.423, + "step": 1310 + }, + { + "epoch": 1.3456599897277863, + "eval_loss": 1.7816330194473267, + "eval_runtime": 36.9065, + "eval_samples_per_second": 27.095, + "eval_steps_per_second": 1.138, + "step": 1310 + }, + { + "epoch": 1.3466872110939907, + "grad_norm": 2.8363473415374756, + "learning_rate": 3.0909783022459085e-06, + "loss": 1.4771, + "step": 1311 + }, + { + "epoch": 1.347714432460195, + "grad_norm": 2.720055103302002, + "learning_rate": 3.0890749904834414e-06, + "loss": 1.4126, + "step": 1312 + }, + { + "epoch": 1.3487416538263997, + "grad_norm": 2.8794236183166504, + "learning_rate": 3.0871716787209748e-06, + "loss": 1.4649, + "step": 1313 + }, + { + "epoch": 1.349768875192604, + "grad_norm": 2.760241746902466, + "learning_rate": 3.085268366958508e-06, + "loss": 1.4427, + "step": 1314 + }, + { + "epoch": 1.3507960965588084, + "grad_norm": 3.031611680984497, + "learning_rate": 3.0833650551960415e-06, + "loss": 1.4476, + "step": 1315 + }, + { + "epoch": 1.3518233179250128, + "grad_norm": 2.7548346519470215, + "learning_rate": 3.081461743433575e-06, + "loss": 1.4036, + "step": 1316 + }, + { + "epoch": 1.3528505392912171, + "grad_norm": 2.7737832069396973, + "learning_rate": 3.079558431671108e-06, + "loss": 1.4209, + "step": 1317 + }, + { + "epoch": 1.3538777606574217, + "grad_norm": 2.580578565597534, + "learning_rate": 3.077655119908641e-06, + "loss": 1.3899, + "step": 1318 + }, + { + "epoch": 1.354904982023626, + "grad_norm": 2.7918219566345215, + "learning_rate": 3.0757518081461745e-06, + "loss": 1.4205, + "step": 1319 + }, + { + "epoch": 1.3559322033898304, + "grad_norm": 2.807173490524292, + "learning_rate": 3.073848496383708e-06, + "loss": 1.4371, + "step": 1320 + }, + { + "epoch": 1.3559322033898304, + "eval_loss": 1.7834100723266602, + "eval_runtime": 37.0369, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 1320 + }, + { + "epoch": 1.356959424756035, + "grad_norm": 2.856569290161133, + "learning_rate": 3.0719451846212413e-06, + "loss": 1.4054, + "step": 1321 + }, + { + "epoch": 1.3579866461222394, + "grad_norm": 2.850403308868408, + "learning_rate": 3.0700418728587746e-06, + "loss": 1.4248, + "step": 1322 + }, + { + "epoch": 1.3590138674884438, + "grad_norm": 2.8618571758270264, + "learning_rate": 3.0681385610963076e-06, + "loss": 1.4066, + "step": 1323 + }, + { + "epoch": 1.3600410888546481, + "grad_norm": 3.0298640727996826, + "learning_rate": 3.066235249333841e-06, + "loss": 1.4479, + "step": 1324 + }, + { + "epoch": 1.3610683102208525, + "grad_norm": 2.9725687503814697, + "learning_rate": 3.0643319375713743e-06, + "loss": 1.4837, + "step": 1325 + }, + { + "epoch": 1.362095531587057, + "grad_norm": 2.8397116661071777, + "learning_rate": 3.0624286258089076e-06, + "loss": 1.4248, + "step": 1326 + }, + { + "epoch": 1.3631227529532615, + "grad_norm": 2.7257447242736816, + "learning_rate": 3.060525314046441e-06, + "loss": 1.4047, + "step": 1327 + }, + { + "epoch": 1.3641499743194658, + "grad_norm": 2.832534074783325, + "learning_rate": 3.058622002283974e-06, + "loss": 1.4491, + "step": 1328 + }, + { + "epoch": 1.3651771956856702, + "grad_norm": 2.6238863468170166, + "learning_rate": 3.0567186905215077e-06, + "loss": 1.457, + "step": 1329 + }, + { + "epoch": 1.3662044170518746, + "grad_norm": 2.7993323802948, + "learning_rate": 3.054815378759041e-06, + "loss": 1.458, + "step": 1330 + }, + { + "epoch": 1.3662044170518746, + "eval_loss": 1.7857885360717773, + "eval_runtime": 37.0668, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 1330 + }, + { + "epoch": 1.3672316384180792, + "grad_norm": 2.798299789428711, + "learning_rate": 3.0529120669965745e-06, + "loss": 1.4539, + "step": 1331 + }, + { + "epoch": 1.3682588597842835, + "grad_norm": 2.7995524406433105, + "learning_rate": 3.051008755234108e-06, + "loss": 1.3929, + "step": 1332 + }, + { + "epoch": 1.3692860811504879, + "grad_norm": 3.026519298553467, + "learning_rate": 3.049105443471641e-06, + "loss": 1.4, + "step": 1333 + }, + { + "epoch": 1.3703133025166925, + "grad_norm": 3.040426731109619, + "learning_rate": 3.0472021317091745e-06, + "loss": 1.5383, + "step": 1334 + }, + { + "epoch": 1.3713405238828968, + "grad_norm": 2.8738269805908203, + "learning_rate": 3.0452988199467075e-06, + "loss": 1.4581, + "step": 1335 + }, + { + "epoch": 1.3723677452491012, + "grad_norm": 2.895745038986206, + "learning_rate": 3.043395508184241e-06, + "loss": 1.4508, + "step": 1336 + }, + { + "epoch": 1.3733949666153056, + "grad_norm": 2.872739315032959, + "learning_rate": 3.041492196421774e-06, + "loss": 1.413, + "step": 1337 + }, + { + "epoch": 1.37442218798151, + "grad_norm": 2.717970132827759, + "learning_rate": 3.0395888846593076e-06, + "loss": 1.4433, + "step": 1338 + }, + { + "epoch": 1.3754494093477145, + "grad_norm": 2.799251079559326, + "learning_rate": 3.037685572896841e-06, + "loss": 1.4237, + "step": 1339 + }, + { + "epoch": 1.376476630713919, + "grad_norm": 2.8966126441955566, + "learning_rate": 3.035782261134374e-06, + "loss": 1.4386, + "step": 1340 + }, + { + "epoch": 1.376476630713919, + "eval_loss": 1.795745611190796, + "eval_runtime": 37.0421, + "eval_samples_per_second": 26.996, + "eval_steps_per_second": 1.134, + "step": 1340 + }, + { + "epoch": 1.3775038520801233, + "grad_norm": 3.0871410369873047, + "learning_rate": 3.0338789493719072e-06, + "loss": 1.4416, + "step": 1341 + }, + { + "epoch": 1.3785310734463276, + "grad_norm": 2.829460382461548, + "learning_rate": 3.0319756376094406e-06, + "loss": 1.4361, + "step": 1342 + }, + { + "epoch": 1.379558294812532, + "grad_norm": 3.0431859493255615, + "learning_rate": 3.030072325846974e-06, + "loss": 1.4431, + "step": 1343 + }, + { + "epoch": 1.3805855161787366, + "grad_norm": 2.966308116912842, + "learning_rate": 3.0281690140845073e-06, + "loss": 1.4022, + "step": 1344 + }, + { + "epoch": 1.381612737544941, + "grad_norm": 2.976332664489746, + "learning_rate": 3.0262657023220403e-06, + "loss": 1.4481, + "step": 1345 + }, + { + "epoch": 1.3826399589111453, + "grad_norm": 2.716144561767578, + "learning_rate": 3.0243623905595736e-06, + "loss": 1.4437, + "step": 1346 + }, + { + "epoch": 1.3836671802773497, + "grad_norm": 2.913194417953491, + "learning_rate": 3.022459078797107e-06, + "loss": 1.4414, + "step": 1347 + }, + { + "epoch": 1.384694401643554, + "grad_norm": 2.9889321327209473, + "learning_rate": 3.0205557670346404e-06, + "loss": 1.4333, + "step": 1348 + }, + { + "epoch": 1.3857216230097587, + "grad_norm": 2.657104730606079, + "learning_rate": 3.018652455272174e-06, + "loss": 1.3999, + "step": 1349 + }, + { + "epoch": 1.386748844375963, + "grad_norm": 2.988452196121216, + "learning_rate": 3.0167491435097075e-06, + "loss": 1.5192, + "step": 1350 + }, + { + "epoch": 1.386748844375963, + "eval_loss": 1.7963600158691406, + "eval_runtime": 37.062, + "eval_samples_per_second": 26.982, + "eval_steps_per_second": 1.133, + "step": 1350 + }, + { + "epoch": 1.3877760657421674, + "grad_norm": 2.9976272583007812, + "learning_rate": 3.014845831747241e-06, + "loss": 1.4288, + "step": 1351 + }, + { + "epoch": 1.388803287108372, + "grad_norm": 2.9255707263946533, + "learning_rate": 3.012942519984774e-06, + "loss": 1.4047, + "step": 1352 + }, + { + "epoch": 1.3898305084745763, + "grad_norm": 2.995906352996826, + "learning_rate": 3.011039208222307e-06, + "loss": 1.4151, + "step": 1353 + }, + { + "epoch": 1.3908577298407807, + "grad_norm": 2.745427370071411, + "learning_rate": 3.0091358964598405e-06, + "loss": 1.3784, + "step": 1354 + }, + { + "epoch": 1.391884951206985, + "grad_norm": 2.8206543922424316, + "learning_rate": 3.007232584697374e-06, + "loss": 1.4307, + "step": 1355 + }, + { + "epoch": 1.3929121725731894, + "grad_norm": 2.7975900173187256, + "learning_rate": 3.0053292729349073e-06, + "loss": 1.4363, + "step": 1356 + }, + { + "epoch": 1.393939393939394, + "grad_norm": 2.8240561485290527, + "learning_rate": 3.00342596117244e-06, + "loss": 1.4284, + "step": 1357 + }, + { + "epoch": 1.3949666153055984, + "grad_norm": 2.8556740283966064, + "learning_rate": 3.0015226494099736e-06, + "loss": 1.4252, + "step": 1358 + }, + { + "epoch": 1.3959938366718028, + "grad_norm": 2.828031539916992, + "learning_rate": 2.999619337647507e-06, + "loss": 1.4068, + "step": 1359 + }, + { + "epoch": 1.3970210580380071, + "grad_norm": 2.8133797645568848, + "learning_rate": 2.9977160258850403e-06, + "loss": 1.4114, + "step": 1360 + }, + { + "epoch": 1.3970210580380071, + "eval_loss": 1.7951425313949585, + "eval_runtime": 37.0796, + "eval_samples_per_second": 26.969, + "eval_steps_per_second": 1.133, + "step": 1360 + }, + { + "epoch": 1.3980482794042115, + "grad_norm": 2.8920187950134277, + "learning_rate": 2.9958127141225736e-06, + "loss": 1.3862, + "step": 1361 + }, + { + "epoch": 1.399075500770416, + "grad_norm": 2.936598777770996, + "learning_rate": 2.9939094023601066e-06, + "loss": 1.4565, + "step": 1362 + }, + { + "epoch": 1.4001027221366205, + "grad_norm": 2.8578219413757324, + "learning_rate": 2.99200609059764e-06, + "loss": 1.4291, + "step": 1363 + }, + { + "epoch": 1.4011299435028248, + "grad_norm": 2.8153412342071533, + "learning_rate": 2.9901027788351733e-06, + "loss": 1.3765, + "step": 1364 + }, + { + "epoch": 1.4021571648690292, + "grad_norm": 2.8321216106414795, + "learning_rate": 2.9881994670727067e-06, + "loss": 1.3914, + "step": 1365 + }, + { + "epoch": 1.4031843862352336, + "grad_norm": 2.7363576889038086, + "learning_rate": 2.98629615531024e-06, + "loss": 1.4091, + "step": 1366 + }, + { + "epoch": 1.4042116076014381, + "grad_norm": 2.8216500282287598, + "learning_rate": 2.984392843547773e-06, + "loss": 1.398, + "step": 1367 + }, + { + "epoch": 1.4052388289676425, + "grad_norm": 2.8419880867004395, + "learning_rate": 2.9824895317853063e-06, + "loss": 1.423, + "step": 1368 + }, + { + "epoch": 1.4062660503338469, + "grad_norm": 2.7820470333099365, + "learning_rate": 2.98058622002284e-06, + "loss": 1.4205, + "step": 1369 + }, + { + "epoch": 1.4072932717000515, + "grad_norm": 2.8953442573547363, + "learning_rate": 2.9786829082603735e-06, + "loss": 1.4667, + "step": 1370 + }, + { + "epoch": 1.4072932717000515, + "eval_loss": 1.7979086637496948, + "eval_runtime": 37.0854, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 1370 + }, + { + "epoch": 1.4083204930662558, + "grad_norm": 2.8208730220794678, + "learning_rate": 2.976779596497907e-06, + "loss": 1.4541, + "step": 1371 + }, + { + "epoch": 1.4093477144324602, + "grad_norm": 2.7347970008850098, + "learning_rate": 2.97487628473544e-06, + "loss": 1.4673, + "step": 1372 + }, + { + "epoch": 1.4103749357986646, + "grad_norm": 2.8528873920440674, + "learning_rate": 2.9729729729729736e-06, + "loss": 1.4267, + "step": 1373 + }, + { + "epoch": 1.411402157164869, + "grad_norm": 2.8298377990722656, + "learning_rate": 2.9710696612105065e-06, + "loss": 1.4713, + "step": 1374 + }, + { + "epoch": 1.4124293785310735, + "grad_norm": 2.847799301147461, + "learning_rate": 2.96916634944804e-06, + "loss": 1.4141, + "step": 1375 + }, + { + "epoch": 1.413456599897278, + "grad_norm": 2.8330819606781006, + "learning_rate": 2.9672630376855732e-06, + "loss": 1.4706, + "step": 1376 + }, + { + "epoch": 1.4144838212634823, + "grad_norm": 2.9275906085968018, + "learning_rate": 2.9653597259231066e-06, + "loss": 1.4371, + "step": 1377 + }, + { + "epoch": 1.4155110426296866, + "grad_norm": 2.7579028606414795, + "learning_rate": 2.96345641416064e-06, + "loss": 1.3921, + "step": 1378 + }, + { + "epoch": 1.416538263995891, + "grad_norm": 2.783045530319214, + "learning_rate": 2.961553102398173e-06, + "loss": 1.3295, + "step": 1379 + }, + { + "epoch": 1.4175654853620956, + "grad_norm": 2.946359157562256, + "learning_rate": 2.9596497906357063e-06, + "loss": 1.4655, + "step": 1380 + }, + { + "epoch": 1.4175654853620956, + "eval_loss": 1.7879260778427124, + "eval_runtime": 37.1105, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 1380 + }, + { + "epoch": 1.4185927067283, + "grad_norm": 3.1604318618774414, + "learning_rate": 2.9577464788732396e-06, + "loss": 1.4746, + "step": 1381 + }, + { + "epoch": 1.4196199280945043, + "grad_norm": 2.602548122406006, + "learning_rate": 2.955843167110773e-06, + "loss": 1.4024, + "step": 1382 + }, + { + "epoch": 1.420647149460709, + "grad_norm": 2.774137258529663, + "learning_rate": 2.9539398553483064e-06, + "loss": 1.4257, + "step": 1383 + }, + { + "epoch": 1.4216743708269133, + "grad_norm": 2.898256540298462, + "learning_rate": 2.9520365435858393e-06, + "loss": 1.4254, + "step": 1384 + }, + { + "epoch": 1.4227015921931176, + "grad_norm": 2.8015098571777344, + "learning_rate": 2.9501332318233727e-06, + "loss": 1.495, + "step": 1385 + }, + { + "epoch": 1.423728813559322, + "grad_norm": 2.7862651348114014, + "learning_rate": 2.948229920060906e-06, + "loss": 1.4355, + "step": 1386 + }, + { + "epoch": 1.4247560349255264, + "grad_norm": 2.7947680950164795, + "learning_rate": 2.9463266082984394e-06, + "loss": 1.4126, + "step": 1387 + }, + { + "epoch": 1.425783256291731, + "grad_norm": 2.8239989280700684, + "learning_rate": 2.9444232965359727e-06, + "loss": 1.417, + "step": 1388 + }, + { + "epoch": 1.4268104776579353, + "grad_norm": 2.845521926879883, + "learning_rate": 2.9425199847735057e-06, + "loss": 1.4406, + "step": 1389 + }, + { + "epoch": 1.4278376990241397, + "grad_norm": 2.9024460315704346, + "learning_rate": 2.94061667301104e-06, + "loss": 1.4083, + "step": 1390 + }, + { + "epoch": 1.4278376990241397, + "eval_loss": 1.7976808547973633, + "eval_runtime": 37.16, + "eval_samples_per_second": 26.911, + "eval_steps_per_second": 1.13, + "step": 1390 + }, + { + "epoch": 1.428864920390344, + "grad_norm": 2.7083208560943604, + "learning_rate": 2.938713361248573e-06, + "loss": 1.3428, + "step": 1391 + }, + { + "epoch": 1.4298921417565484, + "grad_norm": 2.7808196544647217, + "learning_rate": 2.936810049486106e-06, + "loss": 1.4262, + "step": 1392 + }, + { + "epoch": 1.430919363122753, + "grad_norm": 2.8307440280914307, + "learning_rate": 2.9349067377236396e-06, + "loss": 1.4328, + "step": 1393 + }, + { + "epoch": 1.4319465844889574, + "grad_norm": 2.8189167976379395, + "learning_rate": 2.933003425961173e-06, + "loss": 1.3878, + "step": 1394 + }, + { + "epoch": 1.4329738058551618, + "grad_norm": 2.7747530937194824, + "learning_rate": 2.9311001141987063e-06, + "loss": 1.404, + "step": 1395 + }, + { + "epoch": 1.4340010272213661, + "grad_norm": 2.747140884399414, + "learning_rate": 2.9291968024362392e-06, + "loss": 1.3954, + "step": 1396 + }, + { + "epoch": 1.4350282485875705, + "grad_norm": 2.665836811065674, + "learning_rate": 2.9272934906737726e-06, + "loss": 1.3868, + "step": 1397 + }, + { + "epoch": 1.436055469953775, + "grad_norm": 2.6918723583221436, + "learning_rate": 2.925390178911306e-06, + "loss": 1.3959, + "step": 1398 + }, + { + "epoch": 1.4370826913199795, + "grad_norm": 2.8271594047546387, + "learning_rate": 2.9234868671488393e-06, + "loss": 1.4114, + "step": 1399 + }, + { + "epoch": 1.4381099126861838, + "grad_norm": 3.037952184677124, + "learning_rate": 2.9215835553863727e-06, + "loss": 1.4267, + "step": 1400 + }, + { + "epoch": 1.4381099126861838, + "eval_loss": 1.7927908897399902, + "eval_runtime": 37.3671, + "eval_samples_per_second": 26.761, + "eval_steps_per_second": 1.124, + "step": 1400 + }, + { + "epoch": 1.4391371340523884, + "grad_norm": 3.0185632705688477, + "learning_rate": 2.9196802436239056e-06, + "loss": 1.4211, + "step": 1401 + }, + { + "epoch": 1.4401643554185928, + "grad_norm": 2.678468942642212, + "learning_rate": 2.917776931861439e-06, + "loss": 1.398, + "step": 1402 + }, + { + "epoch": 1.4411915767847971, + "grad_norm": 2.8228728771209717, + "learning_rate": 2.9158736200989723e-06, + "loss": 1.3946, + "step": 1403 + }, + { + "epoch": 1.4422187981510015, + "grad_norm": 2.6711578369140625, + "learning_rate": 2.9139703083365057e-06, + "loss": 1.3985, + "step": 1404 + }, + { + "epoch": 1.4432460195172059, + "grad_norm": 2.7891721725463867, + "learning_rate": 2.912066996574039e-06, + "loss": 1.4232, + "step": 1405 + }, + { + "epoch": 1.4442732408834105, + "grad_norm": 2.833613395690918, + "learning_rate": 2.910163684811572e-06, + "loss": 1.3629, + "step": 1406 + }, + { + "epoch": 1.4453004622496148, + "grad_norm": 3.0415570735931396, + "learning_rate": 2.9082603730491054e-06, + "loss": 1.4666, + "step": 1407 + }, + { + "epoch": 1.4463276836158192, + "grad_norm": 2.831063747406006, + "learning_rate": 2.9063570612866387e-06, + "loss": 1.4477, + "step": 1408 + }, + { + "epoch": 1.4473549049820236, + "grad_norm": 2.8253402709960938, + "learning_rate": 2.904453749524172e-06, + "loss": 1.3729, + "step": 1409 + }, + { + "epoch": 1.448382126348228, + "grad_norm": 2.9134843349456787, + "learning_rate": 2.902550437761706e-06, + "loss": 1.3958, + "step": 1410 + }, + { + "epoch": 1.448382126348228, + "eval_loss": 1.7833805084228516, + "eval_runtime": 36.9111, + "eval_samples_per_second": 27.092, + "eval_steps_per_second": 1.138, + "step": 1410 + }, + { + "epoch": 1.4494093477144325, + "grad_norm": 2.962355852127075, + "learning_rate": 2.9006471259992392e-06, + "loss": 1.4143, + "step": 1411 + }, + { + "epoch": 1.450436569080637, + "grad_norm": 2.738837242126465, + "learning_rate": 2.8987438142367726e-06, + "loss": 1.4051, + "step": 1412 + }, + { + "epoch": 1.4514637904468413, + "grad_norm": 2.766331195831299, + "learning_rate": 2.8968405024743055e-06, + "loss": 1.4628, + "step": 1413 + }, + { + "epoch": 1.4524910118130456, + "grad_norm": 2.6052863597869873, + "learning_rate": 2.894937190711839e-06, + "loss": 1.3662, + "step": 1414 + }, + { + "epoch": 1.45351823317925, + "grad_norm": 2.8454713821411133, + "learning_rate": 2.8930338789493723e-06, + "loss": 1.4395, + "step": 1415 + }, + { + "epoch": 1.4545454545454546, + "grad_norm": 2.683232545852661, + "learning_rate": 2.8911305671869056e-06, + "loss": 1.4243, + "step": 1416 + }, + { + "epoch": 1.455572675911659, + "grad_norm": 2.798797845840454, + "learning_rate": 2.889227255424439e-06, + "loss": 1.3627, + "step": 1417 + }, + { + "epoch": 1.4565998972778633, + "grad_norm": 2.821502923965454, + "learning_rate": 2.887323943661972e-06, + "loss": 1.4109, + "step": 1418 + }, + { + "epoch": 1.457627118644068, + "grad_norm": 2.8054518699645996, + "learning_rate": 2.8854206318995053e-06, + "loss": 1.3906, + "step": 1419 + }, + { + "epoch": 1.4586543400102723, + "grad_norm": 2.802000045776367, + "learning_rate": 2.8835173201370386e-06, + "loss": 1.4024, + "step": 1420 + }, + { + "epoch": 1.4586543400102723, + "eval_loss": 1.7884918451309204, + "eval_runtime": 37.0267, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 1420 + }, + { + "epoch": 1.4596815613764766, + "grad_norm": 3.097747564315796, + "learning_rate": 2.881614008374572e-06, + "loss": 1.4695, + "step": 1421 + }, + { + "epoch": 1.460708782742681, + "grad_norm": 3.01665997505188, + "learning_rate": 2.8797106966121054e-06, + "loss": 1.4321, + "step": 1422 + }, + { + "epoch": 1.4617360041088854, + "grad_norm": 2.9308114051818848, + "learning_rate": 2.8778073848496383e-06, + "loss": 1.4084, + "step": 1423 + }, + { + "epoch": 1.46276322547509, + "grad_norm": 2.8440439701080322, + "learning_rate": 2.8759040730871717e-06, + "loss": 1.4473, + "step": 1424 + }, + { + "epoch": 1.4637904468412943, + "grad_norm": 2.6690547466278076, + "learning_rate": 2.874000761324705e-06, + "loss": 1.407, + "step": 1425 + }, + { + "epoch": 1.4648176682074987, + "grad_norm": 2.7961244583129883, + "learning_rate": 2.8720974495622384e-06, + "loss": 1.4414, + "step": 1426 + }, + { + "epoch": 1.465844889573703, + "grad_norm": 2.891948938369751, + "learning_rate": 2.8701941377997718e-06, + "loss": 1.4185, + "step": 1427 + }, + { + "epoch": 1.4668721109399074, + "grad_norm": 2.790811538696289, + "learning_rate": 2.8682908260373047e-06, + "loss": 1.4528, + "step": 1428 + }, + { + "epoch": 1.467899332306112, + "grad_norm": 2.9391868114471436, + "learning_rate": 2.866387514274838e-06, + "loss": 1.4291, + "step": 1429 + }, + { + "epoch": 1.4689265536723164, + "grad_norm": 2.9476356506347656, + "learning_rate": 2.864484202512372e-06, + "loss": 1.4373, + "step": 1430 + }, + { + "epoch": 1.4689265536723164, + "eval_loss": 1.7926347255706787, + "eval_runtime": 37.0308, + "eval_samples_per_second": 27.005, + "eval_steps_per_second": 1.134, + "step": 1430 + }, + { + "epoch": 1.4699537750385208, + "grad_norm": 2.907172203063965, + "learning_rate": 2.8625808907499052e-06, + "loss": 1.3983, + "step": 1431 + }, + { + "epoch": 1.4709809964047253, + "grad_norm": 2.8310680389404297, + "learning_rate": 2.8606775789874386e-06, + "loss": 1.4281, + "step": 1432 + }, + { + "epoch": 1.4720082177709297, + "grad_norm": 2.8614001274108887, + "learning_rate": 2.858774267224972e-06, + "loss": 1.3975, + "step": 1433 + }, + { + "epoch": 1.473035439137134, + "grad_norm": 3.1007702350616455, + "learning_rate": 2.8568709554625053e-06, + "loss": 1.4252, + "step": 1434 + }, + { + "epoch": 1.4740626605033385, + "grad_norm": 2.879520893096924, + "learning_rate": 2.8549676437000382e-06, + "loss": 1.468, + "step": 1435 + }, + { + "epoch": 1.4750898818695428, + "grad_norm": 2.9735469818115234, + "learning_rate": 2.8530643319375716e-06, + "loss": 1.4341, + "step": 1436 + }, + { + "epoch": 1.4761171032357474, + "grad_norm": 3.0768795013427734, + "learning_rate": 2.851161020175105e-06, + "loss": 1.4122, + "step": 1437 + }, + { + "epoch": 1.4771443246019518, + "grad_norm": 2.930915355682373, + "learning_rate": 2.8492577084126383e-06, + "loss": 1.3904, + "step": 1438 + }, + { + "epoch": 1.4781715459681561, + "grad_norm": 2.659090280532837, + "learning_rate": 2.8473543966501717e-06, + "loss": 1.4261, + "step": 1439 + }, + { + "epoch": 1.4791987673343605, + "grad_norm": 3.020371913909912, + "learning_rate": 2.8454510848877046e-06, + "loss": 1.4636, + "step": 1440 + }, + { + "epoch": 1.4791987673343605, + "eval_loss": 1.8039511442184448, + "eval_runtime": 37.0373, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 1440 + }, + { + "epoch": 1.4802259887005649, + "grad_norm": 3.131986141204834, + "learning_rate": 2.843547773125238e-06, + "loss": 1.4178, + "step": 1441 + }, + { + "epoch": 1.4812532100667695, + "grad_norm": 2.8436167240142822, + "learning_rate": 2.8416444613627714e-06, + "loss": 1.4149, + "step": 1442 + }, + { + "epoch": 1.4822804314329738, + "grad_norm": 2.8628804683685303, + "learning_rate": 2.8397411496003047e-06, + "loss": 1.3751, + "step": 1443 + }, + { + "epoch": 1.4833076527991782, + "grad_norm": 2.917836904525757, + "learning_rate": 2.837837837837838e-06, + "loss": 1.4191, + "step": 1444 + }, + { + "epoch": 1.4843348741653826, + "grad_norm": 2.951035737991333, + "learning_rate": 2.835934526075371e-06, + "loss": 1.4352, + "step": 1445 + }, + { + "epoch": 1.485362095531587, + "grad_norm": 2.7296388149261475, + "learning_rate": 2.8340312143129044e-06, + "loss": 1.3856, + "step": 1446 + }, + { + "epoch": 1.4863893168977915, + "grad_norm": 2.7064192295074463, + "learning_rate": 2.8321279025504377e-06, + "loss": 1.3782, + "step": 1447 + }, + { + "epoch": 1.487416538263996, + "grad_norm": 2.9402806758880615, + "learning_rate": 2.830224590787971e-06, + "loss": 1.4829, + "step": 1448 + }, + { + "epoch": 1.4884437596302003, + "grad_norm": 2.9216394424438477, + "learning_rate": 2.8283212790255045e-06, + "loss": 1.4173, + "step": 1449 + }, + { + "epoch": 1.4894709809964048, + "grad_norm": 2.7382750511169434, + "learning_rate": 2.8264179672630383e-06, + "loss": 1.4431, + "step": 1450 + }, + { + "epoch": 1.4894709809964048, + "eval_loss": 1.798187255859375, + "eval_runtime": 37.0401, + "eval_samples_per_second": 26.998, + "eval_steps_per_second": 1.134, + "step": 1450 + }, + { + "epoch": 1.4904982023626092, + "grad_norm": 2.781414270401001, + "learning_rate": 2.8245146555005716e-06, + "loss": 1.4096, + "step": 1451 + }, + { + "epoch": 1.4915254237288136, + "grad_norm": 2.8284265995025635, + "learning_rate": 2.8226113437381046e-06, + "loss": 1.3561, + "step": 1452 + }, + { + "epoch": 1.492552645095018, + "grad_norm": 2.6621487140655518, + "learning_rate": 2.820708031975638e-06, + "loss": 1.4232, + "step": 1453 + }, + { + "epoch": 1.4935798664612223, + "grad_norm": 2.9116251468658447, + "learning_rate": 2.8188047202131713e-06, + "loss": 1.4431, + "step": 1454 + }, + { + "epoch": 1.494607087827427, + "grad_norm": 2.749554395675659, + "learning_rate": 2.8169014084507046e-06, + "loss": 1.4603, + "step": 1455 + }, + { + "epoch": 1.4956343091936313, + "grad_norm": 2.699585199356079, + "learning_rate": 2.814998096688238e-06, + "loss": 1.414, + "step": 1456 + }, + { + "epoch": 1.4966615305598356, + "grad_norm": 2.6940858364105225, + "learning_rate": 2.813094784925771e-06, + "loss": 1.3673, + "step": 1457 + }, + { + "epoch": 1.49768875192604, + "grad_norm": 2.859602689743042, + "learning_rate": 2.8111914731633043e-06, + "loss": 1.4664, + "step": 1458 + }, + { + "epoch": 1.4987159732922444, + "grad_norm": 2.8793201446533203, + "learning_rate": 2.8092881614008377e-06, + "loss": 1.4168, + "step": 1459 + }, + { + "epoch": 1.499743194658449, + "grad_norm": 2.8155760765075684, + "learning_rate": 2.807384849638371e-06, + "loss": 1.3917, + "step": 1460 + }, + { + "epoch": 1.499743194658449, + "eval_loss": 1.780331015586853, + "eval_runtime": 37.0489, + "eval_samples_per_second": 26.991, + "eval_steps_per_second": 1.134, + "step": 1460 + }, + { + "epoch": 1.5007704160246533, + "grad_norm": 2.7490365505218506, + "learning_rate": 2.8054815378759044e-06, + "loss": 1.4326, + "step": 1461 + }, + { + "epoch": 1.5017976373908577, + "grad_norm": 2.8913350105285645, + "learning_rate": 2.8035782261134378e-06, + "loss": 1.403, + "step": 1462 + }, + { + "epoch": 1.5028248587570623, + "grad_norm": 2.83650279045105, + "learning_rate": 2.8016749143509707e-06, + "loss": 1.4107, + "step": 1463 + }, + { + "epoch": 1.5038520801232664, + "grad_norm": 2.838649272918701, + "learning_rate": 2.799771602588504e-06, + "loss": 1.4131, + "step": 1464 + }, + { + "epoch": 1.504879301489471, + "grad_norm": 2.8424947261810303, + "learning_rate": 2.7978682908260374e-06, + "loss": 1.4322, + "step": 1465 + }, + { + "epoch": 1.5059065228556754, + "grad_norm": 2.8207895755767822, + "learning_rate": 2.795964979063571e-06, + "loss": 1.4464, + "step": 1466 + }, + { + "epoch": 1.5069337442218798, + "grad_norm": 2.897221088409424, + "learning_rate": 2.794061667301104e-06, + "loss": 1.4237, + "step": 1467 + }, + { + "epoch": 1.5079609655880843, + "grad_norm": 2.766166925430298, + "learning_rate": 2.792158355538637e-06, + "loss": 1.4375, + "step": 1468 + }, + { + "epoch": 1.5089881869542885, + "grad_norm": 2.8115463256835938, + "learning_rate": 2.7902550437761705e-06, + "loss": 1.4265, + "step": 1469 + }, + { + "epoch": 1.510015408320493, + "grad_norm": 2.826871156692505, + "learning_rate": 2.7883517320137042e-06, + "loss": 1.3933, + "step": 1470 + }, + { + "epoch": 1.510015408320493, + "eval_loss": 1.780141830444336, + "eval_runtime": 37.0572, + "eval_samples_per_second": 26.985, + "eval_steps_per_second": 1.133, + "step": 1470 + }, + { + "epoch": 1.5110426296866974, + "grad_norm": 2.737398147583008, + "learning_rate": 2.7864484202512376e-06, + "loss": 1.3898, + "step": 1471 + }, + { + "epoch": 1.5120698510529018, + "grad_norm": 2.9325804710388184, + "learning_rate": 2.784545108488771e-06, + "loss": 1.4139, + "step": 1472 + }, + { + "epoch": 1.5130970724191064, + "grad_norm": 2.7400333881378174, + "learning_rate": 2.7826417967263043e-06, + "loss": 1.4122, + "step": 1473 + }, + { + "epoch": 1.5141242937853108, + "grad_norm": 2.835869550704956, + "learning_rate": 2.7807384849638373e-06, + "loss": 1.4634, + "step": 1474 + }, + { + "epoch": 1.5151515151515151, + "grad_norm": 2.7512247562408447, + "learning_rate": 2.7788351732013706e-06, + "loss": 1.409, + "step": 1475 + }, + { + "epoch": 1.5161787365177197, + "grad_norm": 2.745054244995117, + "learning_rate": 2.776931861438904e-06, + "loss": 1.4004, + "step": 1476 + }, + { + "epoch": 1.5172059578839239, + "grad_norm": 2.7766499519348145, + "learning_rate": 2.7750285496764374e-06, + "loss": 1.4197, + "step": 1477 + }, + { + "epoch": 1.5182331792501285, + "grad_norm": 2.793663263320923, + "learning_rate": 2.7731252379139707e-06, + "loss": 1.4205, + "step": 1478 + }, + { + "epoch": 1.5192604006163328, + "grad_norm": 2.674372434616089, + "learning_rate": 2.771221926151504e-06, + "loss": 1.3648, + "step": 1479 + }, + { + "epoch": 1.5202876219825372, + "grad_norm": 2.7583112716674805, + "learning_rate": 2.769318614389037e-06, + "loss": 1.4296, + "step": 1480 + }, + { + "epoch": 1.5202876219825372, + "eval_loss": 1.7793656587600708, + "eval_runtime": 37.05, + "eval_samples_per_second": 26.991, + "eval_steps_per_second": 1.134, + "step": 1480 + }, + { + "epoch": 1.5213148433487418, + "grad_norm": 2.776137351989746, + "learning_rate": 2.7674153026265704e-06, + "loss": 1.4267, + "step": 1481 + }, + { + "epoch": 1.522342064714946, + "grad_norm": 2.8750247955322266, + "learning_rate": 2.7655119908641037e-06, + "loss": 1.4311, + "step": 1482 + }, + { + "epoch": 1.5233692860811505, + "grad_norm": 2.740621566772461, + "learning_rate": 2.763608679101637e-06, + "loss": 1.4198, + "step": 1483 + }, + { + "epoch": 1.5243965074473549, + "grad_norm": 2.875997304916382, + "learning_rate": 2.7617053673391705e-06, + "loss": 1.4557, + "step": 1484 + }, + { + "epoch": 1.5254237288135593, + "grad_norm": 2.8406331539154053, + "learning_rate": 2.7598020555767034e-06, + "loss": 1.4504, + "step": 1485 + }, + { + "epoch": 1.5264509501797638, + "grad_norm": 2.6013126373291016, + "learning_rate": 2.7578987438142368e-06, + "loss": 1.3829, + "step": 1486 + }, + { + "epoch": 1.5274781715459682, + "grad_norm": 2.7469065189361572, + "learning_rate": 2.75599543205177e-06, + "loss": 1.4314, + "step": 1487 + }, + { + "epoch": 1.5285053929121726, + "grad_norm": 2.8129167556762695, + "learning_rate": 2.7540921202893035e-06, + "loss": 1.4491, + "step": 1488 + }, + { + "epoch": 1.529532614278377, + "grad_norm": 2.9677329063415527, + "learning_rate": 2.752188808526837e-06, + "loss": 1.4292, + "step": 1489 + }, + { + "epoch": 1.5305598356445813, + "grad_norm": 2.7910637855529785, + "learning_rate": 2.75028549676437e-06, + "loss": 1.4157, + "step": 1490 + }, + { + "epoch": 1.5305598356445813, + "eval_loss": 1.7833755016326904, + "eval_runtime": 37.1104, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 1490 + }, + { + "epoch": 1.531587057010786, + "grad_norm": 2.7772984504699707, + "learning_rate": 2.7483821850019036e-06, + "loss": 1.4125, + "step": 1491 + }, + { + "epoch": 1.5326142783769903, + "grad_norm": 2.747514247894287, + "learning_rate": 2.746478873239437e-06, + "loss": 1.4057, + "step": 1492 + }, + { + "epoch": 1.5336414997431946, + "grad_norm": 2.7529642581939697, + "learning_rate": 2.7445755614769703e-06, + "loss": 1.4066, + "step": 1493 + }, + { + "epoch": 1.5346687211093992, + "grad_norm": 2.8375084400177, + "learning_rate": 2.7426722497145037e-06, + "loss": 1.3883, + "step": 1494 + }, + { + "epoch": 1.5356959424756034, + "grad_norm": 2.7483835220336914, + "learning_rate": 2.740768937952037e-06, + "loss": 1.3915, + "step": 1495 + }, + { + "epoch": 1.536723163841808, + "grad_norm": 2.837123394012451, + "learning_rate": 2.7388656261895704e-06, + "loss": 1.4082, + "step": 1496 + }, + { + "epoch": 1.5377503852080123, + "grad_norm": 2.909313440322876, + "learning_rate": 2.7369623144271033e-06, + "loss": 1.4815, + "step": 1497 + }, + { + "epoch": 1.5387776065742167, + "grad_norm": 2.931723117828369, + "learning_rate": 2.7350590026646367e-06, + "loss": 1.4099, + "step": 1498 + }, + { + "epoch": 1.5398048279404213, + "grad_norm": 2.9692134857177734, + "learning_rate": 2.73315569090217e-06, + "loss": 1.3637, + "step": 1499 + }, + { + "epoch": 1.5408320493066254, + "grad_norm": 2.974923610687256, + "learning_rate": 2.7312523791397034e-06, + "loss": 1.4505, + "step": 1500 + }, + { + "epoch": 1.5408320493066254, + "eval_loss": 1.7900456190109253, + "eval_runtime": 37.3025, + "eval_samples_per_second": 26.808, + "eval_steps_per_second": 1.126, + "step": 1500 + }, + { + "epoch": 1.54185927067283, + "grad_norm": 2.9976389408111572, + "learning_rate": 2.7293490673772368e-06, + "loss": 1.4726, + "step": 1501 + }, + { + "epoch": 1.5428864920390344, + "grad_norm": 2.8478479385375977, + "learning_rate": 2.7274457556147697e-06, + "loss": 1.4304, + "step": 1502 + }, + { + "epoch": 1.5439137134052388, + "grad_norm": 2.9100635051727295, + "learning_rate": 2.725542443852303e-06, + "loss": 1.4107, + "step": 1503 + }, + { + "epoch": 1.5449409347714433, + "grad_norm": 2.903132677078247, + "learning_rate": 2.7236391320898365e-06, + "loss": 1.4206, + "step": 1504 + }, + { + "epoch": 1.5459681561376477, + "grad_norm": 2.9177560806274414, + "learning_rate": 2.72173582032737e-06, + "loss": 1.3948, + "step": 1505 + }, + { + "epoch": 1.546995377503852, + "grad_norm": 2.7861175537109375, + "learning_rate": 2.719832508564903e-06, + "loss": 1.4123, + "step": 1506 + }, + { + "epoch": 1.5480225988700564, + "grad_norm": 2.787735939025879, + "learning_rate": 2.717929196802436e-06, + "loss": 1.4199, + "step": 1507 + }, + { + "epoch": 1.5490498202362608, + "grad_norm": 2.889784097671509, + "learning_rate": 2.7160258850399695e-06, + "loss": 1.3901, + "step": 1508 + }, + { + "epoch": 1.5500770416024654, + "grad_norm": 2.8383727073669434, + "learning_rate": 2.714122573277503e-06, + "loss": 1.4145, + "step": 1509 + }, + { + "epoch": 1.5511042629686698, + "grad_norm": 2.7701327800750732, + "learning_rate": 2.712219261515036e-06, + "loss": 1.4213, + "step": 1510 + }, + { + "epoch": 1.5511042629686698, + "eval_loss": 1.7863596677780151, + "eval_runtime": 36.8927, + "eval_samples_per_second": 27.106, + "eval_steps_per_second": 1.138, + "step": 1510 + }, + { + "epoch": 1.5521314843348741, + "grad_norm": 2.7612502574920654, + "learning_rate": 2.71031594975257e-06, + "loss": 1.4286, + "step": 1511 + }, + { + "epoch": 1.5531587057010787, + "grad_norm": 2.696126937866211, + "learning_rate": 2.7084126379901034e-06, + "loss": 1.4359, + "step": 1512 + }, + { + "epoch": 1.5541859270672829, + "grad_norm": 2.840223550796509, + "learning_rate": 2.7065093262276367e-06, + "loss": 1.4424, + "step": 1513 + }, + { + "epoch": 1.5552131484334875, + "grad_norm": 2.7309999465942383, + "learning_rate": 2.7046060144651697e-06, + "loss": 1.4567, + "step": 1514 + }, + { + "epoch": 1.5562403697996918, + "grad_norm": 2.8637242317199707, + "learning_rate": 2.702702702702703e-06, + "loss": 1.4699, + "step": 1515 + }, + { + "epoch": 1.5572675911658962, + "grad_norm": 2.653404474258423, + "learning_rate": 2.7007993909402364e-06, + "loss": 1.3445, + "step": 1516 + }, + { + "epoch": 1.5582948125321008, + "grad_norm": 2.9771206378936768, + "learning_rate": 2.6988960791777697e-06, + "loss": 1.4588, + "step": 1517 + }, + { + "epoch": 1.559322033898305, + "grad_norm": 2.888413667678833, + "learning_rate": 2.696992767415303e-06, + "loss": 1.425, + "step": 1518 + }, + { + "epoch": 1.5603492552645095, + "grad_norm": 2.8841593265533447, + "learning_rate": 2.695089455652836e-06, + "loss": 1.4101, + "step": 1519 + }, + { + "epoch": 1.5613764766307139, + "grad_norm": 2.990192413330078, + "learning_rate": 2.6931861438903694e-06, + "loss": 1.3897, + "step": 1520 + }, + { + "epoch": 1.5613764766307139, + "eval_loss": 1.7787587642669678, + "eval_runtime": 36.9999, + "eval_samples_per_second": 27.027, + "eval_steps_per_second": 1.135, + "step": 1520 + }, + { + "epoch": 1.5624036979969183, + "grad_norm": 2.8698887825012207, + "learning_rate": 2.6912828321279028e-06, + "loss": 1.4324, + "step": 1521 + }, + { + "epoch": 1.5634309193631228, + "grad_norm": 2.8924384117126465, + "learning_rate": 2.689379520365436e-06, + "loss": 1.4073, + "step": 1522 + }, + { + "epoch": 1.5644581407293272, + "grad_norm": 2.92336106300354, + "learning_rate": 2.6874762086029695e-06, + "loss": 1.4471, + "step": 1523 + }, + { + "epoch": 1.5654853620955316, + "grad_norm": 2.7212202548980713, + "learning_rate": 2.6855728968405024e-06, + "loss": 1.3902, + "step": 1524 + }, + { + "epoch": 1.5665125834617362, + "grad_norm": 2.711582660675049, + "learning_rate": 2.683669585078036e-06, + "loss": 1.4294, + "step": 1525 + }, + { + "epoch": 1.5675398048279403, + "grad_norm": 2.6806275844573975, + "learning_rate": 2.681766273315569e-06, + "loss": 1.3683, + "step": 1526 + }, + { + "epoch": 1.568567026194145, + "grad_norm": 2.920119524002075, + "learning_rate": 2.6798629615531025e-06, + "loss": 1.4194, + "step": 1527 + }, + { + "epoch": 1.5695942475603493, + "grad_norm": 2.7487800121307373, + "learning_rate": 2.677959649790636e-06, + "loss": 1.4091, + "step": 1528 + }, + { + "epoch": 1.5706214689265536, + "grad_norm": 2.7557766437530518, + "learning_rate": 2.676056338028169e-06, + "loss": 1.4453, + "step": 1529 + }, + { + "epoch": 1.5716486902927582, + "grad_norm": 2.7834057807922363, + "learning_rate": 2.674153026265702e-06, + "loss": 1.4381, + "step": 1530 + }, + { + "epoch": 1.5716486902927582, + "eval_loss": 1.7810354232788086, + "eval_runtime": 37.037, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 1530 + }, + { + "epoch": 1.5726759116589624, + "grad_norm": 3.107299327850342, + "learning_rate": 2.672249714503236e-06, + "loss": 1.4475, + "step": 1531 + }, + { + "epoch": 1.573703133025167, + "grad_norm": 2.813086748123169, + "learning_rate": 2.6703464027407693e-06, + "loss": 1.4062, + "step": 1532 + }, + { + "epoch": 1.5747303543913713, + "grad_norm": 2.781416416168213, + "learning_rate": 2.6684430909783027e-06, + "loss": 1.4268, + "step": 1533 + }, + { + "epoch": 1.5757575757575757, + "grad_norm": 2.681947708129883, + "learning_rate": 2.666539779215836e-06, + "loss": 1.4121, + "step": 1534 + }, + { + "epoch": 1.5767847971237803, + "grad_norm": 2.656013011932373, + "learning_rate": 2.6646364674533694e-06, + "loss": 1.4003, + "step": 1535 + }, + { + "epoch": 1.5778120184899846, + "grad_norm": 2.5161869525909424, + "learning_rate": 2.6627331556909024e-06, + "loss": 1.3127, + "step": 1536 + }, + { + "epoch": 1.578839239856189, + "grad_norm": 2.842507839202881, + "learning_rate": 2.6608298439284357e-06, + "loss": 1.4067, + "step": 1537 + }, + { + "epoch": 1.5798664612223934, + "grad_norm": 2.732778787612915, + "learning_rate": 2.658926532165969e-06, + "loss": 1.4071, + "step": 1538 + }, + { + "epoch": 1.5808936825885977, + "grad_norm": 2.802293300628662, + "learning_rate": 2.6570232204035024e-06, + "loss": 1.3637, + "step": 1539 + }, + { + "epoch": 1.5819209039548023, + "grad_norm": 2.927441358566284, + "learning_rate": 2.655119908641036e-06, + "loss": 1.405, + "step": 1540 + }, + { + "epoch": 1.5819209039548023, + "eval_loss": 1.7797114849090576, + "eval_runtime": 37.0514, + "eval_samples_per_second": 26.99, + "eval_steps_per_second": 1.134, + "step": 1540 + }, + { + "epoch": 1.5829481253210067, + "grad_norm": 2.959746837615967, + "learning_rate": 2.6532165968785687e-06, + "loss": 1.4123, + "step": 1541 + }, + { + "epoch": 1.583975346687211, + "grad_norm": 2.679551124572754, + "learning_rate": 2.651313285116102e-06, + "loss": 1.3742, + "step": 1542 + }, + { + "epoch": 1.5850025680534157, + "grad_norm": 2.7431442737579346, + "learning_rate": 2.6494099733536355e-06, + "loss": 1.3825, + "step": 1543 + }, + { + "epoch": 1.5860297894196198, + "grad_norm": 3.3630471229553223, + "learning_rate": 2.647506661591169e-06, + "loss": 1.5599, + "step": 1544 + }, + { + "epoch": 1.5870570107858244, + "grad_norm": 2.8524038791656494, + "learning_rate": 2.645603349828702e-06, + "loss": 1.4293, + "step": 1545 + }, + { + "epoch": 1.5880842321520288, + "grad_norm": 2.871917486190796, + "learning_rate": 2.643700038066235e-06, + "loss": 1.3716, + "step": 1546 + }, + { + "epoch": 1.5891114535182331, + "grad_norm": 2.7151851654052734, + "learning_rate": 2.6417967263037685e-06, + "loss": 1.43, + "step": 1547 + }, + { + "epoch": 1.5901386748844377, + "grad_norm": 2.7912635803222656, + "learning_rate": 2.639893414541302e-06, + "loss": 1.4401, + "step": 1548 + }, + { + "epoch": 1.5911658962506419, + "grad_norm": 2.9613499641418457, + "learning_rate": 2.6379901027788352e-06, + "loss": 1.4691, + "step": 1549 + }, + { + "epoch": 1.5921931176168465, + "grad_norm": 2.8831710815429688, + "learning_rate": 2.6360867910163686e-06, + "loss": 1.3975, + "step": 1550 + }, + { + "epoch": 1.5921931176168465, + "eval_loss": 1.7762643098831177, + "eval_runtime": 37.0485, + "eval_samples_per_second": 26.992, + "eval_steps_per_second": 1.134, + "step": 1550 + }, + { + "epoch": 1.5932203389830508, + "grad_norm": 2.6928372383117676, + "learning_rate": 2.6341834792539024e-06, + "loss": 1.4179, + "step": 1551 + }, + { + "epoch": 1.5942475603492552, + "grad_norm": 2.8088696002960205, + "learning_rate": 2.6322801674914357e-06, + "loss": 1.417, + "step": 1552 + }, + { + "epoch": 1.5952747817154598, + "grad_norm": 2.926135778427124, + "learning_rate": 2.6303768557289687e-06, + "loss": 1.4145, + "step": 1553 + }, + { + "epoch": 1.5963020030816641, + "grad_norm": 2.7283592224121094, + "learning_rate": 2.628473543966502e-06, + "loss": 1.3714, + "step": 1554 + }, + { + "epoch": 1.5973292244478685, + "grad_norm": 2.7970361709594727, + "learning_rate": 2.6265702322040354e-06, + "loss": 1.4393, + "step": 1555 + }, + { + "epoch": 1.5983564458140729, + "grad_norm": 2.717223644256592, + "learning_rate": 2.6246669204415688e-06, + "loss": 1.3857, + "step": 1556 + }, + { + "epoch": 1.5993836671802772, + "grad_norm": 2.817638635635376, + "learning_rate": 2.622763608679102e-06, + "loss": 1.3863, + "step": 1557 + }, + { + "epoch": 1.6004108885464818, + "grad_norm": 2.7856762409210205, + "learning_rate": 2.620860296916635e-06, + "loss": 1.4052, + "step": 1558 + }, + { + "epoch": 1.6014381099126862, + "grad_norm": 2.9281742572784424, + "learning_rate": 2.6189569851541684e-06, + "loss": 1.3417, + "step": 1559 + }, + { + "epoch": 1.6024653312788906, + "grad_norm": 2.8528225421905518, + "learning_rate": 2.617053673391702e-06, + "loss": 1.4486, + "step": 1560 + }, + { + "epoch": 1.6024653312788906, + "eval_loss": 1.771835207939148, + "eval_runtime": 37.0601, + "eval_samples_per_second": 26.983, + "eval_steps_per_second": 1.133, + "step": 1560 + }, + { + "epoch": 1.6034925526450952, + "grad_norm": 3.0976126194000244, + "learning_rate": 2.615150361629235e-06, + "loss": 1.4722, + "step": 1561 + }, + { + "epoch": 1.6045197740112993, + "grad_norm": 2.9993839263916016, + "learning_rate": 2.6132470498667685e-06, + "loss": 1.4265, + "step": 1562 + }, + { + "epoch": 1.605546995377504, + "grad_norm": 2.896975517272949, + "learning_rate": 2.6113437381043015e-06, + "loss": 1.4035, + "step": 1563 + }, + { + "epoch": 1.6065742167437083, + "grad_norm": 2.9382545948028564, + "learning_rate": 2.609440426341835e-06, + "loss": 1.3934, + "step": 1564 + }, + { + "epoch": 1.6076014381099126, + "grad_norm": 2.950702667236328, + "learning_rate": 2.607537114579368e-06, + "loss": 1.4272, + "step": 1565 + }, + { + "epoch": 1.6086286594761172, + "grad_norm": 2.888334274291992, + "learning_rate": 2.6056338028169015e-06, + "loss": 1.4088, + "step": 1566 + }, + { + "epoch": 1.6096558808423214, + "grad_norm": 2.6972687244415283, + "learning_rate": 2.603730491054435e-06, + "loss": 1.391, + "step": 1567 + }, + { + "epoch": 1.610683102208526, + "grad_norm": 2.834188461303711, + "learning_rate": 2.601827179291968e-06, + "loss": 1.4241, + "step": 1568 + }, + { + "epoch": 1.6117103235747303, + "grad_norm": 2.894808769226074, + "learning_rate": 2.599923867529501e-06, + "loss": 1.4297, + "step": 1569 + }, + { + "epoch": 1.6127375449409347, + "grad_norm": 2.9836509227752686, + "learning_rate": 2.5980205557670346e-06, + "loss": 1.4086, + "step": 1570 + }, + { + "epoch": 1.6127375449409347, + "eval_loss": 1.7795302867889404, + "eval_runtime": 37.0629, + "eval_samples_per_second": 26.981, + "eval_steps_per_second": 1.133, + "step": 1570 + }, + { + "epoch": 1.6137647663071393, + "grad_norm": 2.8716723918914795, + "learning_rate": 2.5961172440045684e-06, + "loss": 1.4425, + "step": 1571 + }, + { + "epoch": 1.6147919876733436, + "grad_norm": 2.8877105712890625, + "learning_rate": 2.5942139322421017e-06, + "loss": 1.4557, + "step": 1572 + }, + { + "epoch": 1.615819209039548, + "grad_norm": 2.8737246990203857, + "learning_rate": 2.592310620479635e-06, + "loss": 1.4015, + "step": 1573 + }, + { + "epoch": 1.6168464304057526, + "grad_norm": 2.7048048973083496, + "learning_rate": 2.5904073087171684e-06, + "loss": 1.361, + "step": 1574 + }, + { + "epoch": 1.6178736517719567, + "grad_norm": 2.79799222946167, + "learning_rate": 2.5885039969547014e-06, + "loss": 1.4325, + "step": 1575 + }, + { + "epoch": 1.6189008731381613, + "grad_norm": 2.788114309310913, + "learning_rate": 2.5866006851922347e-06, + "loss": 1.378, + "step": 1576 + }, + { + "epoch": 1.6199280945043657, + "grad_norm": 2.7423717975616455, + "learning_rate": 2.584697373429768e-06, + "loss": 1.3792, + "step": 1577 + }, + { + "epoch": 1.62095531587057, + "grad_norm": 2.777557849884033, + "learning_rate": 2.5827940616673015e-06, + "loss": 1.407, + "step": 1578 + }, + { + "epoch": 1.6219825372367747, + "grad_norm": 3.0097639560699463, + "learning_rate": 2.580890749904835e-06, + "loss": 1.4509, + "step": 1579 + }, + { + "epoch": 1.6230097586029788, + "grad_norm": 2.74116849899292, + "learning_rate": 2.5789874381423678e-06, + "loss": 1.3975, + "step": 1580 + }, + { + "epoch": 1.6230097586029788, + "eval_loss": 1.7757364511489868, + "eval_runtime": 37.0779, + "eval_samples_per_second": 26.97, + "eval_steps_per_second": 1.133, + "step": 1580 + }, + { + "epoch": 1.6240369799691834, + "grad_norm": 2.670633316040039, + "learning_rate": 2.577084126379901e-06, + "loss": 1.401, + "step": 1581 + }, + { + "epoch": 1.6250642013353878, + "grad_norm": 2.7129952907562256, + "learning_rate": 2.5751808146174345e-06, + "loss": 1.3663, + "step": 1582 + }, + { + "epoch": 1.6260914227015921, + "grad_norm": 2.839162826538086, + "learning_rate": 2.573277502854968e-06, + "loss": 1.4189, + "step": 1583 + }, + { + "epoch": 1.6271186440677967, + "grad_norm": 2.7310187816619873, + "learning_rate": 2.5713741910925012e-06, + "loss": 1.3899, + "step": 1584 + }, + { + "epoch": 1.6281458654340009, + "grad_norm": 2.6793200969696045, + "learning_rate": 2.569470879330034e-06, + "loss": 1.3926, + "step": 1585 + }, + { + "epoch": 1.6291730868002054, + "grad_norm": 2.946446418762207, + "learning_rate": 2.5675675675675675e-06, + "loss": 1.4539, + "step": 1586 + }, + { + "epoch": 1.6302003081664098, + "grad_norm": 2.683105945587158, + "learning_rate": 2.565664255805101e-06, + "loss": 1.399, + "step": 1587 + }, + { + "epoch": 1.6312275295326142, + "grad_norm": 2.8176918029785156, + "learning_rate": 2.5637609440426343e-06, + "loss": 1.427, + "step": 1588 + }, + { + "epoch": 1.6322547508988188, + "grad_norm": 2.646280288696289, + "learning_rate": 2.5618576322801676e-06, + "loss": 1.3738, + "step": 1589 + }, + { + "epoch": 1.6332819722650231, + "grad_norm": 2.9067888259887695, + "learning_rate": 2.5599543205177006e-06, + "loss": 1.4666, + "step": 1590 + }, + { + "epoch": 1.6332819722650231, + "eval_loss": 1.7719416618347168, + "eval_runtime": 37.1503, + "eval_samples_per_second": 26.918, + "eval_steps_per_second": 1.131, + "step": 1590 + }, + { + "epoch": 1.6343091936312275, + "grad_norm": 2.797149658203125, + "learning_rate": 2.558051008755234e-06, + "loss": 1.4162, + "step": 1591 + }, + { + "epoch": 1.635336414997432, + "grad_norm": 2.8294360637664795, + "learning_rate": 2.5561476969927677e-06, + "loss": 1.4658, + "step": 1592 + }, + { + "epoch": 1.6363636363636362, + "grad_norm": 2.738529682159424, + "learning_rate": 2.554244385230301e-06, + "loss": 1.3883, + "step": 1593 + }, + { + "epoch": 1.6373908577298408, + "grad_norm": 2.841029167175293, + "learning_rate": 2.5523410734678344e-06, + "loss": 1.4181, + "step": 1594 + }, + { + "epoch": 1.6384180790960452, + "grad_norm": 2.9257984161376953, + "learning_rate": 2.550437761705368e-06, + "loss": 1.3491, + "step": 1595 + }, + { + "epoch": 1.6394453004622496, + "grad_norm": 2.846874713897705, + "learning_rate": 2.548534449942901e-06, + "loss": 1.3832, + "step": 1596 + }, + { + "epoch": 1.6404725218284542, + "grad_norm": 2.9274935722351074, + "learning_rate": 2.546631138180434e-06, + "loss": 1.4215, + "step": 1597 + }, + { + "epoch": 1.6414997431946583, + "grad_norm": 2.9953150749206543, + "learning_rate": 2.5447278264179675e-06, + "loss": 1.4098, + "step": 1598 + }, + { + "epoch": 1.642526964560863, + "grad_norm": 2.7812561988830566, + "learning_rate": 2.542824514655501e-06, + "loss": 1.3806, + "step": 1599 + }, + { + "epoch": 1.6435541859270673, + "grad_norm": 2.9627180099487305, + "learning_rate": 2.540921202893034e-06, + "loss": 1.4623, + "step": 1600 + }, + { + "epoch": 1.6435541859270673, + "eval_loss": 1.7626967430114746, + "eval_runtime": 37.2795, + "eval_samples_per_second": 26.824, + "eval_steps_per_second": 1.127, + "step": 1600 + }, + { + "epoch": 1.6445814072932716, + "grad_norm": 2.8668668270111084, + "learning_rate": 2.5390178911305675e-06, + "loss": 1.4446, + "step": 1601 + }, + { + "epoch": 1.6456086286594762, + "grad_norm": 2.7962896823883057, + "learning_rate": 2.5371145793681005e-06, + "loss": 1.4155, + "step": 1602 + }, + { + "epoch": 1.6466358500256806, + "grad_norm": 2.76855731010437, + "learning_rate": 2.535211267605634e-06, + "loss": 1.4096, + "step": 1603 + }, + { + "epoch": 1.647663071391885, + "grad_norm": 2.7888379096984863, + "learning_rate": 2.533307955843167e-06, + "loss": 1.3544, + "step": 1604 + }, + { + "epoch": 1.6486902927580893, + "grad_norm": 2.9665539264678955, + "learning_rate": 2.5314046440807006e-06, + "loss": 1.4157, + "step": 1605 + }, + { + "epoch": 1.6497175141242937, + "grad_norm": 2.9601891040802, + "learning_rate": 2.529501332318234e-06, + "loss": 1.3963, + "step": 1606 + }, + { + "epoch": 1.6507447354904983, + "grad_norm": 2.897718667984009, + "learning_rate": 2.5275980205557673e-06, + "loss": 1.4498, + "step": 1607 + }, + { + "epoch": 1.6517719568567026, + "grad_norm": 2.9813809394836426, + "learning_rate": 2.5256947087933002e-06, + "loss": 1.4791, + "step": 1608 + }, + { + "epoch": 1.652799178222907, + "grad_norm": 3.025689125061035, + "learning_rate": 2.5237913970308336e-06, + "loss": 1.4616, + "step": 1609 + }, + { + "epoch": 1.6538263995891116, + "grad_norm": 2.6942813396453857, + "learning_rate": 2.521888085268367e-06, + "loss": 1.4112, + "step": 1610 + }, + { + "epoch": 1.6538263995891116, + "eval_loss": 1.775415301322937, + "eval_runtime": 36.8669, + "eval_samples_per_second": 27.125, + "eval_steps_per_second": 1.139, + "step": 1610 + }, + { + "epoch": 1.6548536209553157, + "grad_norm": 2.8190956115722656, + "learning_rate": 2.5199847735059003e-06, + "loss": 1.4559, + "step": 1611 + }, + { + "epoch": 1.6558808423215203, + "grad_norm": 2.7085635662078857, + "learning_rate": 2.518081461743434e-06, + "loss": 1.3905, + "step": 1612 + }, + { + "epoch": 1.6569080636877247, + "grad_norm": 2.856039524078369, + "learning_rate": 2.5161781499809675e-06, + "loss": 1.4223, + "step": 1613 + }, + { + "epoch": 1.657935285053929, + "grad_norm": 2.703293800354004, + "learning_rate": 2.5142748382185004e-06, + "loss": 1.3809, + "step": 1614 + }, + { + "epoch": 1.6589625064201337, + "grad_norm": 2.7698261737823486, + "learning_rate": 2.5123715264560338e-06, + "loss": 1.3732, + "step": 1615 + }, + { + "epoch": 1.6599897277863378, + "grad_norm": 2.8664398193359375, + "learning_rate": 2.510468214693567e-06, + "loss": 1.3994, + "step": 1616 + }, + { + "epoch": 1.6610169491525424, + "grad_norm": 2.843365430831909, + "learning_rate": 2.5085649029311005e-06, + "loss": 1.3466, + "step": 1617 + }, + { + "epoch": 1.6620441705187468, + "grad_norm": 2.7871944904327393, + "learning_rate": 2.506661591168634e-06, + "loss": 1.3096, + "step": 1618 + }, + { + "epoch": 1.6630713918849511, + "grad_norm": 2.9175424575805664, + "learning_rate": 2.504758279406167e-06, + "loss": 1.3931, + "step": 1619 + }, + { + "epoch": 1.6640986132511557, + "grad_norm": 2.926497220993042, + "learning_rate": 2.5028549676437e-06, + "loss": 1.4342, + "step": 1620 + }, + { + "epoch": 1.6640986132511557, + "eval_loss": 1.773324966430664, + "eval_runtime": 36.9745, + "eval_samples_per_second": 27.046, + "eval_steps_per_second": 1.136, + "step": 1620 + }, + { + "epoch": 1.66512583461736, + "grad_norm": 2.7389914989471436, + "learning_rate": 2.5009516558812335e-06, + "loss": 1.3808, + "step": 1621 + }, + { + "epoch": 1.6661530559835644, + "grad_norm": 2.7649383544921875, + "learning_rate": 2.499048344118767e-06, + "loss": 1.4307, + "step": 1622 + }, + { + "epoch": 1.667180277349769, + "grad_norm": 2.8508212566375732, + "learning_rate": 2.4971450323563002e-06, + "loss": 1.4699, + "step": 1623 + }, + { + "epoch": 1.6682074987159732, + "grad_norm": 2.8107385635375977, + "learning_rate": 2.4952417205938336e-06, + "loss": 1.3802, + "step": 1624 + }, + { + "epoch": 1.6692347200821778, + "grad_norm": 2.820322036743164, + "learning_rate": 2.4933384088313666e-06, + "loss": 1.3896, + "step": 1625 + }, + { + "epoch": 1.6702619414483821, + "grad_norm": 2.7912018299102783, + "learning_rate": 2.4914350970689e-06, + "loss": 1.3782, + "step": 1626 + }, + { + "epoch": 1.6712891628145865, + "grad_norm": 2.854904890060425, + "learning_rate": 2.4895317853064337e-06, + "loss": 1.386, + "step": 1627 + }, + { + "epoch": 1.672316384180791, + "grad_norm": 3.2605831623077393, + "learning_rate": 2.4876284735439666e-06, + "loss": 1.446, + "step": 1628 + }, + { + "epoch": 1.6733436055469952, + "grad_norm": 3.029327630996704, + "learning_rate": 2.4857251617815e-06, + "loss": 1.414, + "step": 1629 + }, + { + "epoch": 1.6743708269131998, + "grad_norm": 2.9044439792633057, + "learning_rate": 2.4838218500190334e-06, + "loss": 1.4219, + "step": 1630 + }, + { + "epoch": 1.6743708269131998, + "eval_loss": 1.778739094734192, + "eval_runtime": 36.9917, + "eval_samples_per_second": 27.033, + "eval_steps_per_second": 1.135, + "step": 1630 + }, + { + "epoch": 1.6753980482794042, + "grad_norm": 2.6776554584503174, + "learning_rate": 2.4819185382565667e-06, + "loss": 1.4001, + "step": 1631 + }, + { + "epoch": 1.6764252696456086, + "grad_norm": 2.841038227081299, + "learning_rate": 2.4800152264941e-06, + "loss": 1.406, + "step": 1632 + }, + { + "epoch": 1.6774524910118132, + "grad_norm": 2.848771095275879, + "learning_rate": 2.478111914731633e-06, + "loss": 1.3663, + "step": 1633 + }, + { + "epoch": 1.6784797123780173, + "grad_norm": 2.8403055667877197, + "learning_rate": 2.4762086029691664e-06, + "loss": 1.4065, + "step": 1634 + }, + { + "epoch": 1.6795069337442219, + "grad_norm": 3.0035014152526855, + "learning_rate": 2.4743052912066998e-06, + "loss": 1.5197, + "step": 1635 + }, + { + "epoch": 1.6805341551104263, + "grad_norm": 2.764261245727539, + "learning_rate": 2.472401979444233e-06, + "loss": 1.37, + "step": 1636 + }, + { + "epoch": 1.6815613764766306, + "grad_norm": 3.0496931076049805, + "learning_rate": 2.4704986676817665e-06, + "loss": 1.4657, + "step": 1637 + }, + { + "epoch": 1.6825885978428352, + "grad_norm": 2.8577051162719727, + "learning_rate": 2.4685953559193e-06, + "loss": 1.4019, + "step": 1638 + }, + { + "epoch": 1.6836158192090396, + "grad_norm": 2.8699984550476074, + "learning_rate": 2.466692044156833e-06, + "loss": 1.4508, + "step": 1639 + }, + { + "epoch": 1.684643040575244, + "grad_norm": 2.8038368225097656, + "learning_rate": 2.4647887323943666e-06, + "loss": 1.3859, + "step": 1640 + }, + { + "epoch": 1.684643040575244, + "eval_loss": 1.7762134075164795, + "eval_runtime": 37.0006, + "eval_samples_per_second": 27.027, + "eval_steps_per_second": 1.135, + "step": 1640 + }, + { + "epoch": 1.6856702619414485, + "grad_norm": 2.745168447494507, + "learning_rate": 2.4628854206319e-06, + "loss": 1.3581, + "step": 1641 + }, + { + "epoch": 1.6866974833076527, + "grad_norm": 2.8323488235473633, + "learning_rate": 2.460982108869433e-06, + "loss": 1.3661, + "step": 1642 + }, + { + "epoch": 1.6877247046738573, + "grad_norm": 2.8715322017669678, + "learning_rate": 2.4590787971069662e-06, + "loss": 1.4519, + "step": 1643 + }, + { + "epoch": 1.6887519260400616, + "grad_norm": 2.752286672592163, + "learning_rate": 2.4571754853444996e-06, + "loss": 1.3886, + "step": 1644 + }, + { + "epoch": 1.689779147406266, + "grad_norm": 2.7300102710723877, + "learning_rate": 2.455272173582033e-06, + "loss": 1.3767, + "step": 1645 + }, + { + "epoch": 1.6908063687724706, + "grad_norm": 2.803774356842041, + "learning_rate": 2.4533688618195663e-06, + "loss": 1.39, + "step": 1646 + }, + { + "epoch": 1.6918335901386747, + "grad_norm": 2.6127994060516357, + "learning_rate": 2.4514655500570997e-06, + "loss": 1.3186, + "step": 1647 + }, + { + "epoch": 1.6928608115048793, + "grad_norm": 2.718994140625, + "learning_rate": 2.449562238294633e-06, + "loss": 1.3693, + "step": 1648 + }, + { + "epoch": 1.6938880328710837, + "grad_norm": 2.635831117630005, + "learning_rate": 2.4476589265321664e-06, + "loss": 1.3342, + "step": 1649 + }, + { + "epoch": 1.694915254237288, + "grad_norm": 3.0353753566741943, + "learning_rate": 2.4457556147696993e-06, + "loss": 1.4433, + "step": 1650 + }, + { + "epoch": 1.694915254237288, + "eval_loss": 1.7774006128311157, + "eval_runtime": 37.0101, + "eval_samples_per_second": 27.02, + "eval_steps_per_second": 1.135, + "step": 1650 + }, + { + "epoch": 1.6959424756034926, + "grad_norm": 2.909377098083496, + "learning_rate": 2.4438523030072327e-06, + "loss": 1.4173, + "step": 1651 + }, + { + "epoch": 1.696969696969697, + "grad_norm": 2.9672727584838867, + "learning_rate": 2.441948991244766e-06, + "loss": 1.4274, + "step": 1652 + }, + { + "epoch": 1.6979969183359014, + "grad_norm": 2.822268486022949, + "learning_rate": 2.4400456794822994e-06, + "loss": 1.4012, + "step": 1653 + }, + { + "epoch": 1.6990241397021058, + "grad_norm": 2.8190255165100098, + "learning_rate": 2.438142367719833e-06, + "loss": 1.4251, + "step": 1654 + }, + { + "epoch": 1.7000513610683101, + "grad_norm": 2.8970980644226074, + "learning_rate": 2.4362390559573657e-06, + "loss": 1.3961, + "step": 1655 + }, + { + "epoch": 1.7010785824345147, + "grad_norm": 2.9281842708587646, + "learning_rate": 2.434335744194899e-06, + "loss": 1.3903, + "step": 1656 + }, + { + "epoch": 1.702105803800719, + "grad_norm": 3.0265564918518066, + "learning_rate": 2.432432432432433e-06, + "loss": 1.4319, + "step": 1657 + }, + { + "epoch": 1.7031330251669234, + "grad_norm": 2.785891056060791, + "learning_rate": 2.4305291206699662e-06, + "loss": 1.3561, + "step": 1658 + }, + { + "epoch": 1.704160246533128, + "grad_norm": 2.8406574726104736, + "learning_rate": 2.428625808907499e-06, + "loss": 1.4198, + "step": 1659 + }, + { + "epoch": 1.7051874678993322, + "grad_norm": 2.856170415878296, + "learning_rate": 2.4267224971450325e-06, + "loss": 1.446, + "step": 1660 + }, + { + "epoch": 1.7051874678993322, + "eval_loss": 1.768333077430725, + "eval_runtime": 37.0244, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 1660 + }, + { + "epoch": 1.7062146892655368, + "grad_norm": 2.7454917430877686, + "learning_rate": 2.424819185382566e-06, + "loss": 1.4046, + "step": 1661 + }, + { + "epoch": 1.7072419106317411, + "grad_norm": 2.9186973571777344, + "learning_rate": 2.4229158736200993e-06, + "loss": 1.3616, + "step": 1662 + }, + { + "epoch": 1.7082691319979455, + "grad_norm": 2.859410047531128, + "learning_rate": 2.4210125618576326e-06, + "loss": 1.3594, + "step": 1663 + }, + { + "epoch": 1.70929635336415, + "grad_norm": 2.8369641304016113, + "learning_rate": 2.4191092500951656e-06, + "loss": 1.3924, + "step": 1664 + }, + { + "epoch": 1.7103235747303542, + "grad_norm": 2.8959717750549316, + "learning_rate": 2.417205938332699e-06, + "loss": 1.4301, + "step": 1665 + }, + { + "epoch": 1.7113507960965588, + "grad_norm": 2.8268096446990967, + "learning_rate": 2.4153026265702323e-06, + "loss": 1.3397, + "step": 1666 + }, + { + "epoch": 1.7123780174627632, + "grad_norm": 2.717118740081787, + "learning_rate": 2.4133993148077657e-06, + "loss": 1.3706, + "step": 1667 + }, + { + "epoch": 1.7134052388289676, + "grad_norm": 2.865064859390259, + "learning_rate": 2.411496003045299e-06, + "loss": 1.4356, + "step": 1668 + }, + { + "epoch": 1.7144324601951721, + "grad_norm": 2.8731207847595215, + "learning_rate": 2.4095926912828324e-06, + "loss": 1.3786, + "step": 1669 + }, + { + "epoch": 1.7154596815613765, + "grad_norm": 2.7409722805023193, + "learning_rate": 2.4076893795203658e-06, + "loss": 1.3995, + "step": 1670 + }, + { + "epoch": 1.7154596815613765, + "eval_loss": 1.7796638011932373, + "eval_runtime": 37.0606, + "eval_samples_per_second": 26.983, + "eval_steps_per_second": 1.133, + "step": 1670 + }, + { + "epoch": 1.7164869029275809, + "grad_norm": 2.7279207706451416, + "learning_rate": 2.405786067757899e-06, + "loss": 1.3809, + "step": 1671 + }, + { + "epoch": 1.7175141242937855, + "grad_norm": 3.0839486122131348, + "learning_rate": 2.403882755995432e-06, + "loss": 1.472, + "step": 1672 + }, + { + "epoch": 1.7185413456599896, + "grad_norm": 2.810347557067871, + "learning_rate": 2.4019794442329654e-06, + "loss": 1.3999, + "step": 1673 + }, + { + "epoch": 1.7195685670261942, + "grad_norm": 2.843372106552124, + "learning_rate": 2.4000761324704988e-06, + "loss": 1.3667, + "step": 1674 + }, + { + "epoch": 1.7205957883923986, + "grad_norm": 2.8315467834472656, + "learning_rate": 2.398172820708032e-06, + "loss": 1.4193, + "step": 1675 + }, + { + "epoch": 1.721623009758603, + "grad_norm": 2.647193431854248, + "learning_rate": 2.3962695089455655e-06, + "loss": 1.3826, + "step": 1676 + }, + { + "epoch": 1.7226502311248075, + "grad_norm": 2.8349292278289795, + "learning_rate": 2.3943661971830984e-06, + "loss": 1.3776, + "step": 1677 + }, + { + "epoch": 1.7236774524910117, + "grad_norm": 2.723031520843506, + "learning_rate": 2.3924628854206322e-06, + "loss": 1.3851, + "step": 1678 + }, + { + "epoch": 1.7247046738572163, + "grad_norm": 2.7333242893218994, + "learning_rate": 2.3905595736581656e-06, + "loss": 1.3876, + "step": 1679 + }, + { + "epoch": 1.7257318952234206, + "grad_norm": 2.844578266143799, + "learning_rate": 2.388656261895699e-06, + "loss": 1.4353, + "step": 1680 + }, + { + "epoch": 1.7257318952234206, + "eval_loss": 1.7787370681762695, + "eval_runtime": 37.0536, + "eval_samples_per_second": 26.988, + "eval_steps_per_second": 1.133, + "step": 1680 + }, + { + "epoch": 1.726759116589625, + "grad_norm": 2.9545679092407227, + "learning_rate": 2.386752950133232e-06, + "loss": 1.4284, + "step": 1681 + }, + { + "epoch": 1.7277863379558296, + "grad_norm": 2.9255502223968506, + "learning_rate": 2.3848496383707653e-06, + "loss": 1.4102, + "step": 1682 + }, + { + "epoch": 1.7288135593220337, + "grad_norm": 3.032148838043213, + "learning_rate": 2.3829463266082986e-06, + "loss": 1.4044, + "step": 1683 + }, + { + "epoch": 1.7298407806882383, + "grad_norm": 2.960099935531616, + "learning_rate": 2.381043014845832e-06, + "loss": 1.4274, + "step": 1684 + }, + { + "epoch": 1.7308680020544427, + "grad_norm": 2.8444836139678955, + "learning_rate": 2.3791397030833653e-06, + "loss": 1.4059, + "step": 1685 + }, + { + "epoch": 1.731895223420647, + "grad_norm": 2.7982702255249023, + "learning_rate": 2.3772363913208983e-06, + "loss": 1.3865, + "step": 1686 + }, + { + "epoch": 1.7329224447868516, + "grad_norm": 2.762106418609619, + "learning_rate": 2.3753330795584316e-06, + "loss": 1.3728, + "step": 1687 + }, + { + "epoch": 1.733949666153056, + "grad_norm": 2.8059279918670654, + "learning_rate": 2.3734297677959654e-06, + "loss": 1.3924, + "step": 1688 + }, + { + "epoch": 1.7349768875192604, + "grad_norm": 2.913541316986084, + "learning_rate": 2.3715264560334984e-06, + "loss": 1.4347, + "step": 1689 + }, + { + "epoch": 1.736004108885465, + "grad_norm": 2.7966978549957275, + "learning_rate": 2.3696231442710317e-06, + "loss": 1.4246, + "step": 1690 + }, + { + "epoch": 1.736004108885465, + "eval_loss": 1.766000747680664, + "eval_runtime": 37.1692, + "eval_samples_per_second": 26.904, + "eval_steps_per_second": 1.13, + "step": 1690 + }, + { + "epoch": 1.7370313302516691, + "grad_norm": 2.795525550842285, + "learning_rate": 2.367719832508565e-06, + "loss": 1.4169, + "step": 1691 + }, + { + "epoch": 1.7380585516178737, + "grad_norm": 2.8152341842651367, + "learning_rate": 2.3658165207460985e-06, + "loss": 1.438, + "step": 1692 + }, + { + "epoch": 1.739085772984078, + "grad_norm": 2.6966261863708496, + "learning_rate": 2.363913208983632e-06, + "loss": 1.4164, + "step": 1693 + }, + { + "epoch": 1.7401129943502824, + "grad_norm": 2.875034809112549, + "learning_rate": 2.362009897221165e-06, + "loss": 1.4325, + "step": 1694 + }, + { + "epoch": 1.741140215716487, + "grad_norm": 2.9886977672576904, + "learning_rate": 2.360106585458698e-06, + "loss": 1.4277, + "step": 1695 + }, + { + "epoch": 1.7421674370826912, + "grad_norm": 2.735360860824585, + "learning_rate": 2.3582032736962315e-06, + "loss": 1.3688, + "step": 1696 + }, + { + "epoch": 1.7431946584488958, + "grad_norm": 2.823995351791382, + "learning_rate": 2.356299961933765e-06, + "loss": 1.4062, + "step": 1697 + }, + { + "epoch": 1.7442218798151001, + "grad_norm": 2.7787744998931885, + "learning_rate": 2.354396650171298e-06, + "loss": 1.3782, + "step": 1698 + }, + { + "epoch": 1.7452491011813045, + "grad_norm": 2.745983600616455, + "learning_rate": 2.3524933384088316e-06, + "loss": 1.3682, + "step": 1699 + }, + { + "epoch": 1.746276322547509, + "grad_norm": 2.853743553161621, + "learning_rate": 2.350590026646365e-06, + "loss": 1.4544, + "step": 1700 + }, + { + "epoch": 1.746276322547509, + "eval_loss": 1.7642710208892822, + "eval_runtime": 37.3227, + "eval_samples_per_second": 26.793, + "eval_steps_per_second": 1.125, + "step": 1700 + }, + { + "epoch": 1.7473035439137135, + "grad_norm": 2.8187296390533447, + "learning_rate": 2.3486867148838983e-06, + "loss": 1.3874, + "step": 1701 + }, + { + "epoch": 1.7483307652799178, + "grad_norm": 2.766657829284668, + "learning_rate": 2.3467834031214317e-06, + "loss": 1.37, + "step": 1702 + }, + { + "epoch": 1.7493579866461222, + "grad_norm": 2.796694755554199, + "learning_rate": 2.3448800913589646e-06, + "loss": 1.3909, + "step": 1703 + }, + { + "epoch": 1.7503852080123266, + "grad_norm": 2.8409624099731445, + "learning_rate": 2.342976779596498e-06, + "loss": 1.4091, + "step": 1704 + }, + { + "epoch": 1.7514124293785311, + "grad_norm": 2.7160699367523193, + "learning_rate": 2.3410734678340313e-06, + "loss": 1.3771, + "step": 1705 + }, + { + "epoch": 1.7524396507447355, + "grad_norm": 2.7986226081848145, + "learning_rate": 2.3391701560715647e-06, + "loss": 1.4323, + "step": 1706 + }, + { + "epoch": 1.7534668721109399, + "grad_norm": 2.814997911453247, + "learning_rate": 2.337266844309098e-06, + "loss": 1.3903, + "step": 1707 + }, + { + "epoch": 1.7544940934771445, + "grad_norm": 2.7355244159698486, + "learning_rate": 2.3353635325466314e-06, + "loss": 1.3955, + "step": 1708 + }, + { + "epoch": 1.7555213148433486, + "grad_norm": 2.8889238834381104, + "learning_rate": 2.3334602207841648e-06, + "loss": 1.4037, + "step": 1709 + }, + { + "epoch": 1.7565485362095532, + "grad_norm": 2.9253830909729004, + "learning_rate": 2.331556909021698e-06, + "loss": 1.3982, + "step": 1710 + }, + { + "epoch": 1.7565485362095532, + "eval_loss": 1.7730076313018799, + "eval_runtime": 36.931, + "eval_samples_per_second": 27.078, + "eval_steps_per_second": 1.137, + "step": 1710 + }, + { + "epoch": 1.7575757575757576, + "grad_norm": 2.8409945964813232, + "learning_rate": 2.3296535972592315e-06, + "loss": 1.4313, + "step": 1711 + }, + { + "epoch": 1.758602978941962, + "grad_norm": 2.7974541187286377, + "learning_rate": 2.3277502854967644e-06, + "loss": 1.375, + "step": 1712 + }, + { + "epoch": 1.7596302003081665, + "grad_norm": 2.828486204147339, + "learning_rate": 2.325846973734298e-06, + "loss": 1.4406, + "step": 1713 + }, + { + "epoch": 1.7606574216743707, + "grad_norm": 2.8244118690490723, + "learning_rate": 2.323943661971831e-06, + "loss": 1.3771, + "step": 1714 + }, + { + "epoch": 1.7616846430405753, + "grad_norm": 2.7830047607421875, + "learning_rate": 2.3220403502093645e-06, + "loss": 1.4408, + "step": 1715 + }, + { + "epoch": 1.7627118644067796, + "grad_norm": 2.889033079147339, + "learning_rate": 2.320137038446898e-06, + "loss": 1.438, + "step": 1716 + }, + { + "epoch": 1.763739085772984, + "grad_norm": 2.9460082054138184, + "learning_rate": 2.318233726684431e-06, + "loss": 1.4424, + "step": 1717 + }, + { + "epoch": 1.7647663071391886, + "grad_norm": 2.8492050170898438, + "learning_rate": 2.3163304149219646e-06, + "loss": 1.4354, + "step": 1718 + }, + { + "epoch": 1.765793528505393, + "grad_norm": 2.8115553855895996, + "learning_rate": 2.314427103159498e-06, + "loss": 1.4064, + "step": 1719 + }, + { + "epoch": 1.7668207498715973, + "grad_norm": 2.8483548164367676, + "learning_rate": 2.312523791397031e-06, + "loss": 1.3915, + "step": 1720 + }, + { + "epoch": 1.7668207498715973, + "eval_loss": 1.770641803741455, + "eval_runtime": 37.0317, + "eval_samples_per_second": 27.004, + "eval_steps_per_second": 1.134, + "step": 1720 + }, + { + "epoch": 1.767847971237802, + "grad_norm": 2.67576003074646, + "learning_rate": 2.3106204796345643e-06, + "loss": 1.3665, + "step": 1721 + }, + { + "epoch": 1.768875192604006, + "grad_norm": 2.815186023712158, + "learning_rate": 2.3087171678720976e-06, + "loss": 1.3975, + "step": 1722 + }, + { + "epoch": 1.7699024139702106, + "grad_norm": 2.9318761825561523, + "learning_rate": 2.306813856109631e-06, + "loss": 1.4506, + "step": 1723 + }, + { + "epoch": 1.770929635336415, + "grad_norm": 2.7512404918670654, + "learning_rate": 2.3049105443471644e-06, + "loss": 1.4047, + "step": 1724 + }, + { + "epoch": 1.7719568567026194, + "grad_norm": 2.8040406703948975, + "learning_rate": 2.3030072325846973e-06, + "loss": 1.3813, + "step": 1725 + }, + { + "epoch": 1.772984078068824, + "grad_norm": 2.7659668922424316, + "learning_rate": 2.3011039208222307e-06, + "loss": 1.373, + "step": 1726 + }, + { + "epoch": 1.774011299435028, + "grad_norm": 2.7986247539520264, + "learning_rate": 2.299200609059764e-06, + "loss": 1.3846, + "step": 1727 + }, + { + "epoch": 1.7750385208012327, + "grad_norm": 2.8309032917022705, + "learning_rate": 2.297297297297298e-06, + "loss": 1.3854, + "step": 1728 + }, + { + "epoch": 1.776065742167437, + "grad_norm": 2.7988357543945312, + "learning_rate": 2.2953939855348308e-06, + "loss": 1.3544, + "step": 1729 + }, + { + "epoch": 1.7770929635336414, + "grad_norm": 2.8682641983032227, + "learning_rate": 2.293490673772364e-06, + "loss": 1.4388, + "step": 1730 + }, + { + "epoch": 1.7770929635336414, + "eval_loss": 1.7706880569458008, + "eval_runtime": 37.0355, + "eval_samples_per_second": 27.001, + "eval_steps_per_second": 1.134, + "step": 1730 + }, + { + "epoch": 1.778120184899846, + "grad_norm": 2.8318183422088623, + "learning_rate": 2.2915873620098975e-06, + "loss": 1.4257, + "step": 1731 + }, + { + "epoch": 1.7791474062660502, + "grad_norm": 2.8488996028900146, + "learning_rate": 2.289684050247431e-06, + "loss": 1.4152, + "step": 1732 + }, + { + "epoch": 1.7801746276322548, + "grad_norm": 2.7117509841918945, + "learning_rate": 2.287780738484964e-06, + "loss": 1.3849, + "step": 1733 + }, + { + "epoch": 1.7812018489984591, + "grad_norm": 2.7878878116607666, + "learning_rate": 2.285877426722497e-06, + "loss": 1.4071, + "step": 1734 + }, + { + "epoch": 1.7822290703646635, + "grad_norm": 2.876696825027466, + "learning_rate": 2.2839741149600305e-06, + "loss": 1.4393, + "step": 1735 + }, + { + "epoch": 1.783256291730868, + "grad_norm": 2.8824517726898193, + "learning_rate": 2.282070803197564e-06, + "loss": 1.4225, + "step": 1736 + }, + { + "epoch": 1.7842835130970724, + "grad_norm": 2.8785459995269775, + "learning_rate": 2.2801674914350972e-06, + "loss": 1.4093, + "step": 1737 + }, + { + "epoch": 1.7853107344632768, + "grad_norm": 2.7482051849365234, + "learning_rate": 2.2782641796726306e-06, + "loss": 1.4004, + "step": 1738 + }, + { + "epoch": 1.7863379558294814, + "grad_norm": 2.812790632247925, + "learning_rate": 2.276360867910164e-06, + "loss": 1.4284, + "step": 1739 + }, + { + "epoch": 1.7873651771956856, + "grad_norm": 2.8116066455841064, + "learning_rate": 2.2744575561476973e-06, + "loss": 1.3601, + "step": 1740 + }, + { + "epoch": 1.7873651771956856, + "eval_loss": 1.780893325805664, + "eval_runtime": 37.0422, + "eval_samples_per_second": 26.996, + "eval_steps_per_second": 1.134, + "step": 1740 + }, + { + "epoch": 1.7883923985618901, + "grad_norm": 2.7929999828338623, + "learning_rate": 2.2725542443852307e-06, + "loss": 1.4034, + "step": 1741 + }, + { + "epoch": 1.7894196199280945, + "grad_norm": 2.8496506214141846, + "learning_rate": 2.2706509326227636e-06, + "loss": 1.4414, + "step": 1742 + }, + { + "epoch": 1.7904468412942989, + "grad_norm": 2.7631287574768066, + "learning_rate": 2.268747620860297e-06, + "loss": 1.4117, + "step": 1743 + }, + { + "epoch": 1.7914740626605035, + "grad_norm": 2.731952428817749, + "learning_rate": 2.2668443090978303e-06, + "loss": 1.4064, + "step": 1744 + }, + { + "epoch": 1.7925012840267076, + "grad_norm": 2.7945852279663086, + "learning_rate": 2.2649409973353637e-06, + "loss": 1.4329, + "step": 1745 + }, + { + "epoch": 1.7935285053929122, + "grad_norm": 2.7181437015533447, + "learning_rate": 2.263037685572897e-06, + "loss": 1.3821, + "step": 1746 + }, + { + "epoch": 1.7945557267591166, + "grad_norm": 2.792771816253662, + "learning_rate": 2.26113437381043e-06, + "loss": 1.3789, + "step": 1747 + }, + { + "epoch": 1.795582948125321, + "grad_norm": 2.9297096729278564, + "learning_rate": 2.259231062047964e-06, + "loss": 1.4049, + "step": 1748 + }, + { + "epoch": 1.7966101694915255, + "grad_norm": 2.9424288272857666, + "learning_rate": 2.257327750285497e-06, + "loss": 1.4275, + "step": 1749 + }, + { + "epoch": 1.7976373908577299, + "grad_norm": 2.880979537963867, + "learning_rate": 2.2554244385230305e-06, + "loss": 1.3911, + "step": 1750 + }, + { + "epoch": 1.7976373908577299, + "eval_loss": 1.7886379957199097, + "eval_runtime": 37.0413, + "eval_samples_per_second": 26.997, + "eval_steps_per_second": 1.134, + "step": 1750 + }, + { + "epoch": 1.7986646122239343, + "grad_norm": 2.9970686435699463, + "learning_rate": 2.2535211267605635e-06, + "loss": 1.3976, + "step": 1751 + }, + { + "epoch": 1.7996918335901386, + "grad_norm": 3.0055971145629883, + "learning_rate": 2.251617814998097e-06, + "loss": 1.4029, + "step": 1752 + }, + { + "epoch": 1.800719054956343, + "grad_norm": 2.8585352897644043, + "learning_rate": 2.24971450323563e-06, + "loss": 1.3279, + "step": 1753 + }, + { + "epoch": 1.8017462763225476, + "grad_norm": 2.8452632427215576, + "learning_rate": 2.2478111914731636e-06, + "loss": 1.4141, + "step": 1754 + }, + { + "epoch": 1.802773497688752, + "grad_norm": 2.642672538757324, + "learning_rate": 2.245907879710697e-06, + "loss": 1.2864, + "step": 1755 + }, + { + "epoch": 1.8038007190549563, + "grad_norm": 2.8887314796447754, + "learning_rate": 2.24400456794823e-06, + "loss": 1.4301, + "step": 1756 + }, + { + "epoch": 1.804827940421161, + "grad_norm": 2.735727071762085, + "learning_rate": 2.2421012561857632e-06, + "loss": 1.3908, + "step": 1757 + }, + { + "epoch": 1.805855161787365, + "grad_norm": 2.861366033554077, + "learning_rate": 2.240197944423297e-06, + "loss": 1.3757, + "step": 1758 + }, + { + "epoch": 1.8068823831535696, + "grad_norm": 2.7908987998962402, + "learning_rate": 2.23829463266083e-06, + "loss": 1.3758, + "step": 1759 + }, + { + "epoch": 1.807909604519774, + "grad_norm": 2.602125644683838, + "learning_rate": 2.2363913208983633e-06, + "loss": 1.369, + "step": 1760 + }, + { + "epoch": 1.807909604519774, + "eval_loss": 1.7980055809020996, + "eval_runtime": 37.0615, + "eval_samples_per_second": 26.982, + "eval_steps_per_second": 1.133, + "step": 1760 + }, + { + "epoch": 1.8089368258859784, + "grad_norm": 2.693953275680542, + "learning_rate": 2.2344880091358967e-06, + "loss": 1.4095, + "step": 1761 + }, + { + "epoch": 1.809964047252183, + "grad_norm": 2.772446632385254, + "learning_rate": 2.23258469737343e-06, + "loss": 1.386, + "step": 1762 + }, + { + "epoch": 1.810991268618387, + "grad_norm": 2.7845568656921387, + "learning_rate": 2.2306813856109634e-06, + "loss": 1.4293, + "step": 1763 + }, + { + "epoch": 1.8120184899845917, + "grad_norm": 2.786252498626709, + "learning_rate": 2.2287780738484963e-06, + "loss": 1.3805, + "step": 1764 + }, + { + "epoch": 1.813045711350796, + "grad_norm": 2.859053611755371, + "learning_rate": 2.2268747620860297e-06, + "loss": 1.4585, + "step": 1765 + }, + { + "epoch": 1.8140729327170004, + "grad_norm": 2.8749146461486816, + "learning_rate": 2.224971450323563e-06, + "loss": 1.393, + "step": 1766 + }, + { + "epoch": 1.815100154083205, + "grad_norm": 2.740787982940674, + "learning_rate": 2.2230681385610964e-06, + "loss": 1.3309, + "step": 1767 + }, + { + "epoch": 1.8161273754494094, + "grad_norm": 2.834662914276123, + "learning_rate": 2.2211648267986298e-06, + "loss": 1.3961, + "step": 1768 + }, + { + "epoch": 1.8171545968156138, + "grad_norm": 2.8033571243286133, + "learning_rate": 2.219261515036163e-06, + "loss": 1.3747, + "step": 1769 + }, + { + "epoch": 1.8181818181818183, + "grad_norm": 2.9068171977996826, + "learning_rate": 2.2173582032736965e-06, + "loss": 1.4091, + "step": 1770 + }, + { + "epoch": 1.8181818181818183, + "eval_loss": 1.7923928499221802, + "eval_runtime": 37.067, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 1770 + }, + { + "epoch": 1.8192090395480225, + "grad_norm": 2.8439955711364746, + "learning_rate": 2.21545489151123e-06, + "loss": 1.4281, + "step": 1771 + }, + { + "epoch": 1.820236260914227, + "grad_norm": 2.861231565475464, + "learning_rate": 2.2135515797487632e-06, + "loss": 1.3887, + "step": 1772 + }, + { + "epoch": 1.8212634822804314, + "grad_norm": 2.986788034439087, + "learning_rate": 2.211648267986296e-06, + "loss": 1.395, + "step": 1773 + }, + { + "epoch": 1.8222907036466358, + "grad_norm": 2.755415678024292, + "learning_rate": 2.2097449562238295e-06, + "loss": 1.3885, + "step": 1774 + }, + { + "epoch": 1.8233179250128404, + "grad_norm": 2.7685859203338623, + "learning_rate": 2.207841644461363e-06, + "loss": 1.3861, + "step": 1775 + }, + { + "epoch": 1.8243451463790445, + "grad_norm": 2.970322608947754, + "learning_rate": 2.2059383326988963e-06, + "loss": 1.4436, + "step": 1776 + }, + { + "epoch": 1.8253723677452491, + "grad_norm": 2.997922658920288, + "learning_rate": 2.2040350209364296e-06, + "loss": 1.3978, + "step": 1777 + }, + { + "epoch": 1.8263995891114535, + "grad_norm": 2.734039545059204, + "learning_rate": 2.2021317091739626e-06, + "loss": 1.3623, + "step": 1778 + }, + { + "epoch": 1.8274268104776579, + "grad_norm": 2.8555188179016113, + "learning_rate": 2.2002283974114963e-06, + "loss": 1.3998, + "step": 1779 + }, + { + "epoch": 1.8284540318438625, + "grad_norm": 2.8464975357055664, + "learning_rate": 2.1983250856490297e-06, + "loss": 1.3346, + "step": 1780 + }, + { + "epoch": 1.8284540318438625, + "eval_loss": 1.7876062393188477, + "eval_runtime": 37.0771, + "eval_samples_per_second": 26.971, + "eval_steps_per_second": 1.133, + "step": 1780 + }, + { + "epoch": 1.8294812532100666, + "grad_norm": 3.012516975402832, + "learning_rate": 2.1964217738865626e-06, + "loss": 1.3803, + "step": 1781 + }, + { + "epoch": 1.8305084745762712, + "grad_norm": 2.941331148147583, + "learning_rate": 2.194518462124096e-06, + "loss": 1.4033, + "step": 1782 + }, + { + "epoch": 1.8315356959424756, + "grad_norm": 3.090533494949341, + "learning_rate": 2.1926151503616294e-06, + "loss": 1.443, + "step": 1783 + }, + { + "epoch": 1.83256291730868, + "grad_norm": 2.890427827835083, + "learning_rate": 2.1907118385991627e-06, + "loss": 1.4089, + "step": 1784 + }, + { + "epoch": 1.8335901386748845, + "grad_norm": 2.8417627811431885, + "learning_rate": 2.188808526836696e-06, + "loss": 1.3722, + "step": 1785 + }, + { + "epoch": 1.8346173600410889, + "grad_norm": 2.7787647247314453, + "learning_rate": 2.1869052150742295e-06, + "loss": 1.4011, + "step": 1786 + }, + { + "epoch": 1.8356445814072933, + "grad_norm": 2.7441141605377197, + "learning_rate": 2.1850019033117624e-06, + "loss": 1.3262, + "step": 1787 + }, + { + "epoch": 1.8366718027734978, + "grad_norm": 2.873112678527832, + "learning_rate": 2.1830985915492958e-06, + "loss": 1.3896, + "step": 1788 + }, + { + "epoch": 1.837699024139702, + "grad_norm": 2.8022804260253906, + "learning_rate": 2.1811952797868295e-06, + "loss": 1.4035, + "step": 1789 + }, + { + "epoch": 1.8387262455059066, + "grad_norm": 2.8220114707946777, + "learning_rate": 2.1792919680243625e-06, + "loss": 1.3748, + "step": 1790 + }, + { + "epoch": 1.8387262455059066, + "eval_loss": 1.7839077711105347, + "eval_runtime": 37.0886, + "eval_samples_per_second": 26.962, + "eval_steps_per_second": 1.132, + "step": 1790 + }, + { + "epoch": 1.839753466872111, + "grad_norm": 2.769643783569336, + "learning_rate": 2.177388656261896e-06, + "loss": 1.4084, + "step": 1791 + }, + { + "epoch": 1.8407806882383153, + "grad_norm": 2.8555970191955566, + "learning_rate": 2.1754853444994292e-06, + "loss": 1.4772, + "step": 1792 + }, + { + "epoch": 1.84180790960452, + "grad_norm": 2.777965545654297, + "learning_rate": 2.1735820327369626e-06, + "loss": 1.3024, + "step": 1793 + }, + { + "epoch": 1.842835130970724, + "grad_norm": 2.8747634887695312, + "learning_rate": 2.171678720974496e-06, + "loss": 1.4195, + "step": 1794 + }, + { + "epoch": 1.8438623523369286, + "grad_norm": 2.8822247982025146, + "learning_rate": 2.169775409212029e-06, + "loss": 1.4151, + "step": 1795 + }, + { + "epoch": 1.844889573703133, + "grad_norm": 2.8203372955322266, + "learning_rate": 2.1678720974495622e-06, + "loss": 1.4223, + "step": 1796 + }, + { + "epoch": 1.8459167950693374, + "grad_norm": 2.632070779800415, + "learning_rate": 2.1659687856870956e-06, + "loss": 1.4141, + "step": 1797 + }, + { + "epoch": 1.846944016435542, + "grad_norm": 2.746530771255493, + "learning_rate": 2.164065473924629e-06, + "loss": 1.4351, + "step": 1798 + }, + { + "epoch": 1.8479712378017463, + "grad_norm": 2.7693254947662354, + "learning_rate": 2.1621621621621623e-06, + "loss": 1.3588, + "step": 1799 + }, + { + "epoch": 1.8489984591679507, + "grad_norm": 2.7944774627685547, + "learning_rate": 2.1602588503996957e-06, + "loss": 1.4181, + "step": 1800 + }, + { + "epoch": 1.8489984591679507, + "eval_loss": 1.780879259109497, + "eval_runtime": 37.2375, + "eval_samples_per_second": 26.855, + "eval_steps_per_second": 1.128, + "step": 1800 + }, + { + "epoch": 1.850025680534155, + "grad_norm": 2.827951669692993, + "learning_rate": 2.158355538637229e-06, + "loss": 1.3857, + "step": 1801 + }, + { + "epoch": 1.8510529019003594, + "grad_norm": 2.7634339332580566, + "learning_rate": 2.1564522268747624e-06, + "loss": 1.387, + "step": 1802 + }, + { + "epoch": 1.852080123266564, + "grad_norm": 3.014620542526245, + "learning_rate": 2.1545489151122958e-06, + "loss": 1.4595, + "step": 1803 + }, + { + "epoch": 1.8531073446327684, + "grad_norm": 2.7546815872192383, + "learning_rate": 2.1526456033498287e-06, + "loss": 1.4098, + "step": 1804 + }, + { + "epoch": 1.8541345659989728, + "grad_norm": 2.7237050533294678, + "learning_rate": 2.150742291587362e-06, + "loss": 1.4099, + "step": 1805 + }, + { + "epoch": 1.8551617873651773, + "grad_norm": 2.8354947566986084, + "learning_rate": 2.1488389798248954e-06, + "loss": 1.395, + "step": 1806 + }, + { + "epoch": 1.8561890087313815, + "grad_norm": 2.863032102584839, + "learning_rate": 2.146935668062429e-06, + "loss": 1.4275, + "step": 1807 + }, + { + "epoch": 1.857216230097586, + "grad_norm": 2.6963112354278564, + "learning_rate": 2.145032356299962e-06, + "loss": 1.3567, + "step": 1808 + }, + { + "epoch": 1.8582434514637904, + "grad_norm": 2.7628366947174072, + "learning_rate": 2.1431290445374955e-06, + "loss": 1.4036, + "step": 1809 + }, + { + "epoch": 1.8592706728299948, + "grad_norm": 2.790919065475464, + "learning_rate": 2.141225732775029e-06, + "loss": 1.3679, + "step": 1810 + }, + { + "epoch": 1.8592706728299948, + "eval_loss": 1.7804644107818604, + "eval_runtime": 36.9549, + "eval_samples_per_second": 27.06, + "eval_steps_per_second": 1.137, + "step": 1810 + }, + { + "epoch": 1.8602978941961994, + "grad_norm": 2.913525104522705, + "learning_rate": 2.1393224210125623e-06, + "loss": 1.4015, + "step": 1811 + }, + { + "epoch": 1.8613251155624035, + "grad_norm": 2.8357040882110596, + "learning_rate": 2.137419109250095e-06, + "loss": 1.3912, + "step": 1812 + }, + { + "epoch": 1.8623523369286081, + "grad_norm": 2.989529848098755, + "learning_rate": 2.1355157974876286e-06, + "loss": 1.4397, + "step": 1813 + }, + { + "epoch": 1.8633795582948125, + "grad_norm": 2.8843417167663574, + "learning_rate": 2.133612485725162e-06, + "loss": 1.4343, + "step": 1814 + }, + { + "epoch": 1.8644067796610169, + "grad_norm": 2.882571220397949, + "learning_rate": 2.1317091739626953e-06, + "loss": 1.342, + "step": 1815 + }, + { + "epoch": 1.8654340010272215, + "grad_norm": 2.8986611366271973, + "learning_rate": 2.1298058622002286e-06, + "loss": 1.3687, + "step": 1816 + }, + { + "epoch": 1.8664612223934258, + "grad_norm": 2.8779523372650146, + "learning_rate": 2.1279025504377616e-06, + "loss": 1.4005, + "step": 1817 + }, + { + "epoch": 1.8674884437596302, + "grad_norm": 2.7829904556274414, + "learning_rate": 2.125999238675295e-06, + "loss": 1.3843, + "step": 1818 + }, + { + "epoch": 1.8685156651258348, + "grad_norm": 3.0068447589874268, + "learning_rate": 2.1240959269128287e-06, + "loss": 1.4347, + "step": 1819 + }, + { + "epoch": 1.869542886492039, + "grad_norm": 2.945225715637207, + "learning_rate": 2.122192615150362e-06, + "loss": 1.3934, + "step": 1820 + }, + { + "epoch": 1.869542886492039, + "eval_loss": 1.7796295881271362, + "eval_runtime": 37.0423, + "eval_samples_per_second": 26.996, + "eval_steps_per_second": 1.134, + "step": 1820 + }, + { + "epoch": 1.8705701078582435, + "grad_norm": 2.79140305519104, + "learning_rate": 2.120289303387895e-06, + "loss": 1.3387, + "step": 1821 + }, + { + "epoch": 1.8715973292244479, + "grad_norm": 2.897294282913208, + "learning_rate": 2.1183859916254284e-06, + "loss": 1.3905, + "step": 1822 + }, + { + "epoch": 1.8726245505906522, + "grad_norm": 2.8181400299072266, + "learning_rate": 2.1164826798629618e-06, + "loss": 1.4219, + "step": 1823 + }, + { + "epoch": 1.8736517719568568, + "grad_norm": 2.695333480834961, + "learning_rate": 2.114579368100495e-06, + "loss": 1.38, + "step": 1824 + }, + { + "epoch": 1.874678993323061, + "grad_norm": 2.7668285369873047, + "learning_rate": 2.1126760563380285e-06, + "loss": 1.3613, + "step": 1825 + }, + { + "epoch": 1.8757062146892656, + "grad_norm": 2.723405599594116, + "learning_rate": 2.1107727445755614e-06, + "loss": 1.3344, + "step": 1826 + }, + { + "epoch": 1.87673343605547, + "grad_norm": 3.01149320602417, + "learning_rate": 2.1088694328130948e-06, + "loss": 1.4672, + "step": 1827 + }, + { + "epoch": 1.8777606574216743, + "grad_norm": 2.7828991413116455, + "learning_rate": 2.106966121050628e-06, + "loss": 1.3671, + "step": 1828 + }, + { + "epoch": 1.878787878787879, + "grad_norm": 2.792518138885498, + "learning_rate": 2.1050628092881615e-06, + "loss": 1.4182, + "step": 1829 + }, + { + "epoch": 1.879815100154083, + "grad_norm": 2.935760736465454, + "learning_rate": 2.103159497525695e-06, + "loss": 1.4623, + "step": 1830 + }, + { + "epoch": 1.879815100154083, + "eval_loss": 1.7818807363510132, + "eval_runtime": 37.0809, + "eval_samples_per_second": 26.968, + "eval_steps_per_second": 1.133, + "step": 1830 + }, + { + "epoch": 1.8808423215202876, + "grad_norm": 2.849071741104126, + "learning_rate": 2.1012561857632282e-06, + "loss": 1.3777, + "step": 1831 + }, + { + "epoch": 1.881869542886492, + "grad_norm": 2.978045701980591, + "learning_rate": 2.0993528740007616e-06, + "loss": 1.4325, + "step": 1832 + }, + { + "epoch": 1.8828967642526964, + "grad_norm": 2.8763439655303955, + "learning_rate": 2.097449562238295e-06, + "loss": 1.4417, + "step": 1833 + }, + { + "epoch": 1.883923985618901, + "grad_norm": 2.891221523284912, + "learning_rate": 2.095546250475828e-06, + "loss": 1.3881, + "step": 1834 + }, + { + "epoch": 1.8849512069851053, + "grad_norm": 2.9743025302886963, + "learning_rate": 2.0936429387133613e-06, + "loss": 1.3956, + "step": 1835 + }, + { + "epoch": 1.8859784283513097, + "grad_norm": 2.7557995319366455, + "learning_rate": 2.0917396269508946e-06, + "loss": 1.4511, + "step": 1836 + }, + { + "epoch": 1.8870056497175143, + "grad_norm": 2.878021240234375, + "learning_rate": 2.089836315188428e-06, + "loss": 1.4206, + "step": 1837 + }, + { + "epoch": 1.8880328710837184, + "grad_norm": 2.9022324085235596, + "learning_rate": 2.0879330034259614e-06, + "loss": 1.4256, + "step": 1838 + }, + { + "epoch": 1.889060092449923, + "grad_norm": 2.8805174827575684, + "learning_rate": 2.0860296916634947e-06, + "loss": 1.3752, + "step": 1839 + }, + { + "epoch": 1.8900873138161274, + "grad_norm": 2.887437105178833, + "learning_rate": 2.084126379901028e-06, + "loss": 1.4636, + "step": 1840 + }, + { + "epoch": 1.8900873138161274, + "eval_loss": 1.775463581085205, + "eval_runtime": 37.0928, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 1840 + }, + { + "epoch": 1.8911145351823317, + "grad_norm": 2.8477346897125244, + "learning_rate": 2.0822230681385614e-06, + "loss": 1.3653, + "step": 1841 + }, + { + "epoch": 1.8921417565485363, + "grad_norm": 2.939516067504883, + "learning_rate": 2.080319756376095e-06, + "loss": 1.4216, + "step": 1842 + }, + { + "epoch": 1.8931689779147405, + "grad_norm": 2.869184970855713, + "learning_rate": 2.0784164446136277e-06, + "loss": 1.4138, + "step": 1843 + }, + { + "epoch": 1.894196199280945, + "grad_norm": 2.7639729976654053, + "learning_rate": 2.076513132851161e-06, + "loss": 1.3837, + "step": 1844 + }, + { + "epoch": 1.8952234206471494, + "grad_norm": 2.9656383991241455, + "learning_rate": 2.0746098210886945e-06, + "loss": 1.4168, + "step": 1845 + }, + { + "epoch": 1.8962506420133538, + "grad_norm": 2.80155873298645, + "learning_rate": 2.072706509326228e-06, + "loss": 1.3879, + "step": 1846 + }, + { + "epoch": 1.8972778633795584, + "grad_norm": 2.742859363555908, + "learning_rate": 2.070803197563761e-06, + "loss": 1.397, + "step": 1847 + }, + { + "epoch": 1.8983050847457628, + "grad_norm": 2.790229082107544, + "learning_rate": 2.068899885801294e-06, + "loss": 1.3894, + "step": 1848 + }, + { + "epoch": 1.8993323061119671, + "grad_norm": 2.87921142578125, + "learning_rate": 2.066996574038828e-06, + "loss": 1.4122, + "step": 1849 + }, + { + "epoch": 1.9003595274781715, + "grad_norm": 2.937082052230835, + "learning_rate": 2.0650932622763613e-06, + "loss": 1.435, + "step": 1850 + }, + { + "epoch": 1.9003595274781715, + "eval_loss": 1.7634403705596924, + "eval_runtime": 37.0377, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 1850 + }, + { + "epoch": 1.9013867488443759, + "grad_norm": 2.830686569213867, + "learning_rate": 2.0631899505138942e-06, + "loss": 1.3248, + "step": 1851 + }, + { + "epoch": 1.9024139702105805, + "grad_norm": 2.878131151199341, + "learning_rate": 2.0612866387514276e-06, + "loss": 1.3778, + "step": 1852 + }, + { + "epoch": 1.9034411915767848, + "grad_norm": 2.894076108932495, + "learning_rate": 2.059383326988961e-06, + "loss": 1.4073, + "step": 1853 + }, + { + "epoch": 1.9044684129429892, + "grad_norm": 2.816612958908081, + "learning_rate": 2.0574800152264943e-06, + "loss": 1.3873, + "step": 1854 + }, + { + "epoch": 1.9054956343091938, + "grad_norm": 2.905581474304199, + "learning_rate": 2.0555767034640277e-06, + "loss": 1.3644, + "step": 1855 + }, + { + "epoch": 1.906522855675398, + "grad_norm": 2.7126173973083496, + "learning_rate": 2.053673391701561e-06, + "loss": 1.3402, + "step": 1856 + }, + { + "epoch": 1.9075500770416025, + "grad_norm": 2.8092000484466553, + "learning_rate": 2.051770079939094e-06, + "loss": 1.4095, + "step": 1857 + }, + { + "epoch": 1.9085772984078069, + "grad_norm": 2.7571353912353516, + "learning_rate": 2.0498667681766273e-06, + "loss": 1.3657, + "step": 1858 + }, + { + "epoch": 1.9096045197740112, + "grad_norm": 2.8373515605926514, + "learning_rate": 2.047963456414161e-06, + "loss": 1.3643, + "step": 1859 + }, + { + "epoch": 1.9106317411402158, + "grad_norm": 2.825209379196167, + "learning_rate": 2.046060144651694e-06, + "loss": 1.3859, + "step": 1860 + }, + { + "epoch": 1.9106317411402158, + "eval_loss": 1.7664082050323486, + "eval_runtime": 37.1245, + "eval_samples_per_second": 26.936, + "eval_steps_per_second": 1.131, + "step": 1860 + }, + { + "epoch": 1.91165896250642, + "grad_norm": 2.7947168350219727, + "learning_rate": 2.0441568328892274e-06, + "loss": 1.4002, + "step": 1861 + }, + { + "epoch": 1.9126861838726246, + "grad_norm": 2.8314290046691895, + "learning_rate": 2.0422535211267608e-06, + "loss": 1.4017, + "step": 1862 + }, + { + "epoch": 1.913713405238829, + "grad_norm": 2.7938215732574463, + "learning_rate": 2.040350209364294e-06, + "loss": 1.379, + "step": 1863 + }, + { + "epoch": 1.9147406266050333, + "grad_norm": 2.7426047325134277, + "learning_rate": 2.0384468976018275e-06, + "loss": 1.3615, + "step": 1864 + }, + { + "epoch": 1.915767847971238, + "grad_norm": 2.916954278945923, + "learning_rate": 2.0365435858393604e-06, + "loss": 1.3756, + "step": 1865 + }, + { + "epoch": 1.9167950693374423, + "grad_norm": 2.7564570903778076, + "learning_rate": 2.034640274076894e-06, + "loss": 1.3716, + "step": 1866 + }, + { + "epoch": 1.9178222907036466, + "grad_norm": 2.911621332168579, + "learning_rate": 2.032736962314427e-06, + "loss": 1.4378, + "step": 1867 + }, + { + "epoch": 1.9188495120698512, + "grad_norm": 2.805112361907959, + "learning_rate": 2.0308336505519605e-06, + "loss": 1.3823, + "step": 1868 + }, + { + "epoch": 1.9198767334360554, + "grad_norm": 3.079866647720337, + "learning_rate": 2.028930338789494e-06, + "loss": 1.4388, + "step": 1869 + }, + { + "epoch": 1.92090395480226, + "grad_norm": 2.9153823852539062, + "learning_rate": 2.0270270270270273e-06, + "loss": 1.359, + "step": 1870 + }, + { + "epoch": 1.92090395480226, + "eval_loss": 1.7608962059020996, + "eval_runtime": 37.1265, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 1870 + }, + { + "epoch": 1.9219311761684643, + "grad_norm": 2.8992526531219482, + "learning_rate": 2.0251237152645606e-06, + "loss": 1.3814, + "step": 1871 + }, + { + "epoch": 1.9229583975346687, + "grad_norm": 2.8838071823120117, + "learning_rate": 2.023220403502094e-06, + "loss": 1.3642, + "step": 1872 + }, + { + "epoch": 1.9239856189008733, + "grad_norm": 2.8734290599823, + "learning_rate": 2.0213170917396273e-06, + "loss": 1.3366, + "step": 1873 + }, + { + "epoch": 1.9250128402670774, + "grad_norm": 2.929142951965332, + "learning_rate": 2.0194137799771603e-06, + "loss": 1.3725, + "step": 1874 + }, + { + "epoch": 1.926040061633282, + "grad_norm": 2.984711170196533, + "learning_rate": 2.0175104682146937e-06, + "loss": 1.3825, + "step": 1875 + }, + { + "epoch": 1.9270672829994864, + "grad_norm": 2.916740894317627, + "learning_rate": 2.015607156452227e-06, + "loss": 1.3261, + "step": 1876 + }, + { + "epoch": 1.9280945043656907, + "grad_norm": 2.8688554763793945, + "learning_rate": 2.0137038446897604e-06, + "loss": 1.365, + "step": 1877 + }, + { + "epoch": 1.9291217257318953, + "grad_norm": 2.932525396347046, + "learning_rate": 2.0118005329272937e-06, + "loss": 1.3654, + "step": 1878 + }, + { + "epoch": 1.9301489470980995, + "grad_norm": 2.8874945640563965, + "learning_rate": 2.0098972211648267e-06, + "loss": 1.3745, + "step": 1879 + }, + { + "epoch": 1.931176168464304, + "grad_norm": 2.7608230113983154, + "learning_rate": 2.0079939094023605e-06, + "loss": 1.3718, + "step": 1880 + }, + { + "epoch": 1.931176168464304, + "eval_loss": 1.7662792205810547, + "eval_runtime": 37.1426, + "eval_samples_per_second": 26.923, + "eval_steps_per_second": 1.131, + "step": 1880 + }, + { + "epoch": 1.9322033898305084, + "grad_norm": 2.7790932655334473, + "learning_rate": 2.006090597639894e-06, + "loss": 1.4227, + "step": 1881 + }, + { + "epoch": 1.9332306111967128, + "grad_norm": 2.9143898487091064, + "learning_rate": 2.0041872858774268e-06, + "loss": 1.4396, + "step": 1882 + }, + { + "epoch": 1.9342578325629174, + "grad_norm": 2.744253158569336, + "learning_rate": 2.00228397411496e-06, + "loss": 1.4123, + "step": 1883 + }, + { + "epoch": 1.9352850539291218, + "grad_norm": 2.7352776527404785, + "learning_rate": 2.0003806623524935e-06, + "loss": 1.3936, + "step": 1884 + }, + { + "epoch": 1.9363122752953261, + "grad_norm": 2.7074668407440186, + "learning_rate": 1.998477350590027e-06, + "loss": 1.3201, + "step": 1885 + }, + { + "epoch": 1.9373394966615307, + "grad_norm": 2.848219871520996, + "learning_rate": 1.9965740388275602e-06, + "loss": 1.4048, + "step": 1886 + }, + { + "epoch": 1.9383667180277349, + "grad_norm": 2.7833354473114014, + "learning_rate": 1.994670727065093e-06, + "loss": 1.4007, + "step": 1887 + }, + { + "epoch": 1.9393939393939394, + "grad_norm": 2.83203387260437, + "learning_rate": 1.9927674153026265e-06, + "loss": 1.3471, + "step": 1888 + }, + { + "epoch": 1.9404211607601438, + "grad_norm": 2.99426007270813, + "learning_rate": 1.99086410354016e-06, + "loss": 1.407, + "step": 1889 + }, + { + "epoch": 1.9414483821263482, + "grad_norm": 2.8964831829071045, + "learning_rate": 1.9889607917776937e-06, + "loss": 1.3784, + "step": 1890 + }, + { + "epoch": 1.9414483821263482, + "eval_loss": 1.7859200239181519, + "eval_runtime": 37.16, + "eval_samples_per_second": 26.911, + "eval_steps_per_second": 1.13, + "step": 1890 + }, + { + "epoch": 1.9424756034925528, + "grad_norm": 2.938148021697998, + "learning_rate": 1.9870574800152266e-06, + "loss": 1.3932, + "step": 1891 + }, + { + "epoch": 1.943502824858757, + "grad_norm": 2.9096012115478516, + "learning_rate": 1.98515416825276e-06, + "loss": 1.4051, + "step": 1892 + }, + { + "epoch": 1.9445300462249615, + "grad_norm": 2.8964526653289795, + "learning_rate": 1.9832508564902933e-06, + "loss": 1.3961, + "step": 1893 + }, + { + "epoch": 1.9455572675911659, + "grad_norm": 2.681792974472046, + "learning_rate": 1.9813475447278267e-06, + "loss": 1.3175, + "step": 1894 + }, + { + "epoch": 1.9465844889573702, + "grad_norm": 2.8340206146240234, + "learning_rate": 1.97944423296536e-06, + "loss": 1.3955, + "step": 1895 + }, + { + "epoch": 1.9476117103235748, + "grad_norm": 2.9196391105651855, + "learning_rate": 1.977540921202893e-06, + "loss": 1.3833, + "step": 1896 + }, + { + "epoch": 1.9486389316897792, + "grad_norm": 2.7774484157562256, + "learning_rate": 1.9756376094404264e-06, + "loss": 1.3645, + "step": 1897 + }, + { + "epoch": 1.9496661530559836, + "grad_norm": 2.7222559452056885, + "learning_rate": 1.9737342976779597e-06, + "loss": 1.342, + "step": 1898 + }, + { + "epoch": 1.950693374422188, + "grad_norm": 2.816084623336792, + "learning_rate": 1.971830985915493e-06, + "loss": 1.4255, + "step": 1899 + }, + { + "epoch": 1.9517205957883923, + "grad_norm": 2.7530157566070557, + "learning_rate": 1.9699276741530264e-06, + "loss": 1.3792, + "step": 1900 + }, + { + "epoch": 1.9517205957883923, + "eval_loss": 1.7724789381027222, + "eval_runtime": 37.3499, + "eval_samples_per_second": 26.774, + "eval_steps_per_second": 1.125, + "step": 1900 + }, + { + "epoch": 1.9527478171545969, + "grad_norm": 2.7674543857574463, + "learning_rate": 1.96802436239056e-06, + "loss": 1.3888, + "step": 1901 + }, + { + "epoch": 1.9537750385208013, + "grad_norm": 2.9137115478515625, + "learning_rate": 1.966121050628093e-06, + "loss": 1.3656, + "step": 1902 + }, + { + "epoch": 1.9548022598870056, + "grad_norm": 2.8127288818359375, + "learning_rate": 1.9642177388656265e-06, + "loss": 1.4109, + "step": 1903 + }, + { + "epoch": 1.9558294812532102, + "grad_norm": 2.906191110610962, + "learning_rate": 1.9623144271031595e-06, + "loss": 1.4258, + "step": 1904 + }, + { + "epoch": 1.9568567026194144, + "grad_norm": 2.9624578952789307, + "learning_rate": 1.960411115340693e-06, + "loss": 1.4328, + "step": 1905 + }, + { + "epoch": 1.957883923985619, + "grad_norm": 2.8785276412963867, + "learning_rate": 1.958507803578226e-06, + "loss": 1.3912, + "step": 1906 + }, + { + "epoch": 1.9589111453518233, + "grad_norm": 2.848287343978882, + "learning_rate": 1.9566044918157596e-06, + "loss": 1.4052, + "step": 1907 + }, + { + "epoch": 1.9599383667180277, + "grad_norm": 2.766186237335205, + "learning_rate": 1.954701180053293e-06, + "loss": 1.4064, + "step": 1908 + }, + { + "epoch": 1.9609655880842323, + "grad_norm": 2.8182034492492676, + "learning_rate": 1.952797868290826e-06, + "loss": 1.3724, + "step": 1909 + }, + { + "epoch": 1.9619928094504364, + "grad_norm": 2.826488494873047, + "learning_rate": 1.9508945565283596e-06, + "loss": 1.421, + "step": 1910 + }, + { + "epoch": 1.9619928094504364, + "eval_loss": 1.780763864517212, + "eval_runtime": 36.9249, + "eval_samples_per_second": 27.082, + "eval_steps_per_second": 1.137, + "step": 1910 + }, + { + "epoch": 1.963020030816641, + "grad_norm": 2.876213312149048, + "learning_rate": 1.948991244765893e-06, + "loss": 1.3885, + "step": 1911 + }, + { + "epoch": 1.9640472521828454, + "grad_norm": 2.897611141204834, + "learning_rate": 1.9470879330034264e-06, + "loss": 1.4126, + "step": 1912 + }, + { + "epoch": 1.9650744735490497, + "grad_norm": 2.806605100631714, + "learning_rate": 1.9451846212409593e-06, + "loss": 1.3506, + "step": 1913 + }, + { + "epoch": 1.9661016949152543, + "grad_norm": 2.879645824432373, + "learning_rate": 1.9432813094784927e-06, + "loss": 1.409, + "step": 1914 + }, + { + "epoch": 1.9671289162814587, + "grad_norm": 2.871046304702759, + "learning_rate": 1.941377997716026e-06, + "loss": 1.3417, + "step": 1915 + }, + { + "epoch": 1.968156137647663, + "grad_norm": 2.8778395652770996, + "learning_rate": 1.9394746859535594e-06, + "loss": 1.3779, + "step": 1916 + }, + { + "epoch": 1.9691833590138677, + "grad_norm": 2.8736093044281006, + "learning_rate": 1.9375713741910928e-06, + "loss": 1.3355, + "step": 1917 + }, + { + "epoch": 1.9702105803800718, + "grad_norm": 2.822600841522217, + "learning_rate": 1.9356680624286257e-06, + "loss": 1.3779, + "step": 1918 + }, + { + "epoch": 1.9712378017462764, + "grad_norm": 2.779498338699341, + "learning_rate": 1.933764750666159e-06, + "loss": 1.3328, + "step": 1919 + }, + { + "epoch": 1.9722650231124808, + "grad_norm": 2.9568161964416504, + "learning_rate": 1.931861438903693e-06, + "loss": 1.3902, + "step": 1920 + }, + { + "epoch": 1.9722650231124808, + "eval_loss": 1.769534707069397, + "eval_runtime": 37.0243, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 1920 + }, + { + "epoch": 1.9732922444786851, + "grad_norm": 2.7917582988739014, + "learning_rate": 1.929958127141226e-06, + "loss": 1.3517, + "step": 1921 + }, + { + "epoch": 1.9743194658448897, + "grad_norm": 2.8624138832092285, + "learning_rate": 1.928054815378759e-06, + "loss": 1.3765, + "step": 1922 + }, + { + "epoch": 1.9753466872110939, + "grad_norm": 2.874619960784912, + "learning_rate": 1.9261515036162925e-06, + "loss": 1.4548, + "step": 1923 + }, + { + "epoch": 1.9763739085772984, + "grad_norm": 2.921558141708374, + "learning_rate": 1.924248191853826e-06, + "loss": 1.4122, + "step": 1924 + }, + { + "epoch": 1.9774011299435028, + "grad_norm": 2.853748321533203, + "learning_rate": 1.9223448800913592e-06, + "loss": 1.3909, + "step": 1925 + }, + { + "epoch": 1.9784283513097072, + "grad_norm": 2.828735828399658, + "learning_rate": 1.9204415683288926e-06, + "loss": 1.3748, + "step": 1926 + }, + { + "epoch": 1.9794555726759118, + "grad_norm": 2.8109824657440186, + "learning_rate": 1.9185382565664255e-06, + "loss": 1.4051, + "step": 1927 + }, + { + "epoch": 1.980482794042116, + "grad_norm": 3.035722494125366, + "learning_rate": 1.916634944803959e-06, + "loss": 1.42, + "step": 1928 + }, + { + "epoch": 1.9815100154083205, + "grad_norm": 2.9289391040802, + "learning_rate": 1.9147316330414923e-06, + "loss": 1.3965, + "step": 1929 + }, + { + "epoch": 1.9825372367745249, + "grad_norm": 2.853398561477661, + "learning_rate": 1.9128283212790256e-06, + "loss": 1.3869, + "step": 1930 + }, + { + "epoch": 1.9825372367745249, + "eval_loss": 1.7659740447998047, + "eval_runtime": 37.0479, + "eval_samples_per_second": 26.992, + "eval_steps_per_second": 1.134, + "step": 1930 + }, + { + "epoch": 1.9835644581407292, + "grad_norm": 2.9305245876312256, + "learning_rate": 1.910925009516559e-06, + "loss": 1.4265, + "step": 1931 + }, + { + "epoch": 1.9845916795069338, + "grad_norm": 2.8339171409606934, + "learning_rate": 1.9090216977540924e-06, + "loss": 1.3934, + "step": 1932 + }, + { + "epoch": 1.9856189008731382, + "grad_norm": 2.929807424545288, + "learning_rate": 1.9071183859916257e-06, + "loss": 1.4339, + "step": 1933 + }, + { + "epoch": 1.9866461222393426, + "grad_norm": 2.9070687294006348, + "learning_rate": 1.9052150742291589e-06, + "loss": 1.4004, + "step": 1934 + }, + { + "epoch": 1.9876733436055471, + "grad_norm": 2.9756429195404053, + "learning_rate": 1.9033117624666922e-06, + "loss": 1.404, + "step": 1935 + }, + { + "epoch": 1.9887005649717513, + "grad_norm": 2.831841468811035, + "learning_rate": 1.9014084507042254e-06, + "loss": 1.3668, + "step": 1936 + }, + { + "epoch": 1.9897277863379559, + "grad_norm": 2.8649723529815674, + "learning_rate": 1.8995051389417587e-06, + "loss": 1.3625, + "step": 1937 + }, + { + "epoch": 1.9907550077041603, + "grad_norm": 2.7570254802703857, + "learning_rate": 1.8976018271792921e-06, + "loss": 1.3427, + "step": 1938 + }, + { + "epoch": 1.9917822290703646, + "grad_norm": 2.9655392169952393, + "learning_rate": 1.8956985154168253e-06, + "loss": 1.4039, + "step": 1939 + }, + { + "epoch": 1.9928094504365692, + "grad_norm": 2.840308427810669, + "learning_rate": 1.8937952036543588e-06, + "loss": 1.3678, + "step": 1940 + }, + { + "epoch": 1.9928094504365692, + "eval_loss": 1.7668676376342773, + "eval_runtime": 37.0279, + "eval_samples_per_second": 27.007, + "eval_steps_per_second": 1.134, + "step": 1940 + }, + { + "epoch": 1.9938366718027734, + "grad_norm": 2.7608296871185303, + "learning_rate": 1.8918918918918922e-06, + "loss": 1.3353, + "step": 1941 + }, + { + "epoch": 1.994863893168978, + "grad_norm": 2.861088275909424, + "learning_rate": 1.8899885801294253e-06, + "loss": 1.3866, + "step": 1942 + }, + { + "epoch": 1.9958911145351823, + "grad_norm": 2.8978993892669678, + "learning_rate": 1.8880852683669587e-06, + "loss": 1.4054, + "step": 1943 + }, + { + "epoch": 1.9969183359013867, + "grad_norm": 2.850844144821167, + "learning_rate": 1.886181956604492e-06, + "loss": 1.3692, + "step": 1944 + }, + { + "epoch": 1.9979455572675913, + "grad_norm": 2.8984055519104004, + "learning_rate": 1.8842786448420252e-06, + "loss": 1.3681, + "step": 1945 + }, + { + "epoch": 1.9989727786337956, + "grad_norm": 2.8746683597564697, + "learning_rate": 1.8823753330795586e-06, + "loss": 1.4047, + "step": 1946 + }, + { + "epoch": 2.0, + "grad_norm": 2.7244315147399902, + "learning_rate": 1.8804720213170917e-06, + "loss": 1.3003, + "step": 1947 + }, + { + "epoch": 2.0010272213662046, + "grad_norm": 3.10217022895813, + "learning_rate": 1.878568709554625e-06, + "loss": 1.3375, + "step": 1948 + }, + { + "epoch": 2.0020544427324087, + "grad_norm": 3.1777870655059814, + "learning_rate": 1.8766653977921585e-06, + "loss": 1.3775, + "step": 1949 + }, + { + "epoch": 2.0030816640986133, + "grad_norm": 3.008085250854492, + "learning_rate": 1.874762086029692e-06, + "loss": 1.3435, + "step": 1950 + }, + { + "epoch": 2.0030816640986133, + "eval_loss": 1.7830486297607422, + "eval_runtime": 37.0192, + "eval_samples_per_second": 27.013, + "eval_steps_per_second": 1.135, + "step": 1950 + }, + { + "epoch": 2.0041088854648175, + "grad_norm": 3.228102445602417, + "learning_rate": 1.8728587742672252e-06, + "loss": 1.3697, + "step": 1951 + }, + { + "epoch": 2.005136106831022, + "grad_norm": 3.392749071121216, + "learning_rate": 1.8709554625047585e-06, + "loss": 1.3813, + "step": 1952 + }, + { + "epoch": 2.0061633281972266, + "grad_norm": 3.264998197555542, + "learning_rate": 1.8690521507422917e-06, + "loss": 1.3203, + "step": 1953 + }, + { + "epoch": 2.007190549563431, + "grad_norm": 3.344672679901123, + "learning_rate": 1.867148838979825e-06, + "loss": 1.3413, + "step": 1954 + }, + { + "epoch": 2.0082177709296354, + "grad_norm": 3.138148307800293, + "learning_rate": 1.8652455272173584e-06, + "loss": 1.3606, + "step": 1955 + }, + { + "epoch": 2.00924499229584, + "grad_norm": 3.1569597721099854, + "learning_rate": 1.8633422154548916e-06, + "loss": 1.3846, + "step": 1956 + }, + { + "epoch": 2.010272213662044, + "grad_norm": 3.0843050479888916, + "learning_rate": 1.861438903692425e-06, + "loss": 1.282, + "step": 1957 + }, + { + "epoch": 2.0112994350282487, + "grad_norm": 3.3988428115844727, + "learning_rate": 1.859535591929958e-06, + "loss": 1.3183, + "step": 1958 + }, + { + "epoch": 2.012326656394453, + "grad_norm": 3.3206543922424316, + "learning_rate": 1.8576322801674915e-06, + "loss": 1.3874, + "step": 1959 + }, + { + "epoch": 2.0133538777606574, + "grad_norm": 3.1254804134368896, + "learning_rate": 1.855728968405025e-06, + "loss": 1.3335, + "step": 1960 + }, + { + "epoch": 2.0133538777606574, + "eval_loss": 1.780193567276001, + "eval_runtime": 37.0457, + "eval_samples_per_second": 26.994, + "eval_steps_per_second": 1.134, + "step": 1960 + }, + { + "epoch": 2.014381099126862, + "grad_norm": 3.2736592292785645, + "learning_rate": 1.8538256566425584e-06, + "loss": 1.3489, + "step": 1961 + }, + { + "epoch": 2.015408320493066, + "grad_norm": 3.1620168685913086, + "learning_rate": 1.8519223448800915e-06, + "loss": 1.3898, + "step": 1962 + }, + { + "epoch": 2.0164355418592708, + "grad_norm": 3.3050506114959717, + "learning_rate": 1.850019033117625e-06, + "loss": 1.3292, + "step": 1963 + }, + { + "epoch": 2.017462763225475, + "grad_norm": 3.192668914794922, + "learning_rate": 1.848115721355158e-06, + "loss": 1.3319, + "step": 1964 + }, + { + "epoch": 2.0184899845916795, + "grad_norm": 3.1518166065216064, + "learning_rate": 1.8462124095926914e-06, + "loss": 1.3433, + "step": 1965 + }, + { + "epoch": 2.019517205957884, + "grad_norm": 3.365323305130005, + "learning_rate": 1.8443090978302248e-06, + "loss": 1.362, + "step": 1966 + }, + { + "epoch": 2.0205444273240882, + "grad_norm": 3.4271507263183594, + "learning_rate": 1.842405786067758e-06, + "loss": 1.3867, + "step": 1967 + }, + { + "epoch": 2.021571648690293, + "grad_norm": 3.167879581451416, + "learning_rate": 1.8405024743052913e-06, + "loss": 1.3341, + "step": 1968 + }, + { + "epoch": 2.022598870056497, + "grad_norm": 3.1220595836639404, + "learning_rate": 1.8385991625428247e-06, + "loss": 1.379, + "step": 1969 + }, + { + "epoch": 2.0236260914227016, + "grad_norm": 3.03648042678833, + "learning_rate": 1.8366958507803578e-06, + "loss": 1.3095, + "step": 1970 + }, + { + "epoch": 2.0236260914227016, + "eval_loss": 1.7782403230667114, + "eval_runtime": 37.0535, + "eval_samples_per_second": 26.988, + "eval_steps_per_second": 1.133, + "step": 1970 + }, + { + "epoch": 2.024653312788906, + "grad_norm": 3.254167318344116, + "learning_rate": 1.8347925390178914e-06, + "loss": 1.3625, + "step": 1971 + }, + { + "epoch": 2.0256805341551103, + "grad_norm": 3.0616867542266846, + "learning_rate": 1.8328892272554247e-06, + "loss": 1.3059, + "step": 1972 + }, + { + "epoch": 2.026707755521315, + "grad_norm": 3.236483097076416, + "learning_rate": 1.8309859154929579e-06, + "loss": 1.3737, + "step": 1973 + }, + { + "epoch": 2.0277349768875195, + "grad_norm": 3.1941466331481934, + "learning_rate": 1.8290826037304913e-06, + "loss": 1.35, + "step": 1974 + }, + { + "epoch": 2.0287621982537236, + "grad_norm": 3.0594732761383057, + "learning_rate": 1.8271792919680246e-06, + "loss": 1.3261, + "step": 1975 + }, + { + "epoch": 2.029789419619928, + "grad_norm": 3.157635450363159, + "learning_rate": 1.8252759802055578e-06, + "loss": 1.3224, + "step": 1976 + }, + { + "epoch": 2.0308166409861323, + "grad_norm": 3.380322217941284, + "learning_rate": 1.8233726684430911e-06, + "loss": 1.3742, + "step": 1977 + }, + { + "epoch": 2.031843862352337, + "grad_norm": 3.397022247314453, + "learning_rate": 1.8214693566806243e-06, + "loss": 1.324, + "step": 1978 + }, + { + "epoch": 2.0328710837185415, + "grad_norm": 3.187364101409912, + "learning_rate": 1.8195660449181576e-06, + "loss": 1.3435, + "step": 1979 + }, + { + "epoch": 2.0338983050847457, + "grad_norm": 3.1330549716949463, + "learning_rate": 1.817662733155691e-06, + "loss": 1.3472, + "step": 1980 + }, + { + "epoch": 2.0338983050847457, + "eval_loss": 1.7825566530227661, + "eval_runtime": 37.0315, + "eval_samples_per_second": 27.004, + "eval_steps_per_second": 1.134, + "step": 1980 + }, + { + "epoch": 2.0349255264509503, + "grad_norm": 3.1697194576263428, + "learning_rate": 1.8157594213932244e-06, + "loss": 1.347, + "step": 1981 + }, + { + "epoch": 2.0359527478171544, + "grad_norm": 3.1682229042053223, + "learning_rate": 1.8138561096307577e-06, + "loss": 1.3588, + "step": 1982 + }, + { + "epoch": 2.036979969183359, + "grad_norm": 3.1456754207611084, + "learning_rate": 1.811952797868291e-06, + "loss": 1.2981, + "step": 1983 + }, + { + "epoch": 2.0380071905495636, + "grad_norm": 3.1854794025421143, + "learning_rate": 1.8100494861058242e-06, + "loss": 1.3429, + "step": 1984 + }, + { + "epoch": 2.0390344119157677, + "grad_norm": 3.2500100135803223, + "learning_rate": 1.8081461743433576e-06, + "loss": 1.3572, + "step": 1985 + }, + { + "epoch": 2.0400616332819723, + "grad_norm": 3.0852582454681396, + "learning_rate": 1.806242862580891e-06, + "loss": 1.3081, + "step": 1986 + }, + { + "epoch": 2.0410888546481765, + "grad_norm": 3.3680732250213623, + "learning_rate": 1.8043395508184241e-06, + "loss": 1.3584, + "step": 1987 + }, + { + "epoch": 2.042116076014381, + "grad_norm": 3.1346354484558105, + "learning_rate": 1.8024362390559575e-06, + "loss": 1.3437, + "step": 1988 + }, + { + "epoch": 2.0431432973805856, + "grad_norm": 3.1831793785095215, + "learning_rate": 1.8005329272934906e-06, + "loss": 1.3206, + "step": 1989 + }, + { + "epoch": 2.04417051874679, + "grad_norm": 3.3129804134368896, + "learning_rate": 1.798629615531024e-06, + "loss": 1.3992, + "step": 1990 + }, + { + "epoch": 2.04417051874679, + "eval_loss": 1.791827917098999, + "eval_runtime": 37.1302, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 1990 + }, + { + "epoch": 2.0451977401129944, + "grad_norm": 3.2966983318328857, + "learning_rate": 1.7967263037685576e-06, + "loss": 1.3374, + "step": 1991 + }, + { + "epoch": 2.046224961479199, + "grad_norm": 3.2134904861450195, + "learning_rate": 1.794822992006091e-06, + "loss": 1.3304, + "step": 1992 + }, + { + "epoch": 2.047252182845403, + "grad_norm": 3.3718373775482178, + "learning_rate": 1.792919680243624e-06, + "loss": 1.3898, + "step": 1993 + }, + { + "epoch": 2.0482794042116077, + "grad_norm": 3.1467645168304443, + "learning_rate": 1.7910163684811574e-06, + "loss": 1.3089, + "step": 1994 + }, + { + "epoch": 2.049306625577812, + "grad_norm": 3.1504383087158203, + "learning_rate": 1.7891130567186906e-06, + "loss": 1.3398, + "step": 1995 + }, + { + "epoch": 2.0503338469440164, + "grad_norm": 3.2479350566864014, + "learning_rate": 1.787209744956224e-06, + "loss": 1.3099, + "step": 1996 + }, + { + "epoch": 2.051361068310221, + "grad_norm": 3.311795234680176, + "learning_rate": 1.7853064331937573e-06, + "loss": 1.3369, + "step": 1997 + }, + { + "epoch": 2.052388289676425, + "grad_norm": 3.1674633026123047, + "learning_rate": 1.7834031214312905e-06, + "loss": 1.3732, + "step": 1998 + }, + { + "epoch": 2.0534155110426298, + "grad_norm": 3.1175575256347656, + "learning_rate": 1.7814998096688238e-06, + "loss": 1.3159, + "step": 1999 + }, + { + "epoch": 2.054442732408834, + "grad_norm": 3.216402530670166, + "learning_rate": 1.779596497906357e-06, + "loss": 1.3479, + "step": 2000 + }, + { + "epoch": 2.054442732408834, + "eval_loss": 1.7845674753189087, + "eval_runtime": 37.2755, + "eval_samples_per_second": 26.827, + "eval_steps_per_second": 1.127, + "step": 2000 + }, + { + "epoch": 2.0554699537750385, + "grad_norm": 3.2530016899108887, + "learning_rate": 1.7776931861438906e-06, + "loss": 1.3924, + "step": 2001 + }, + { + "epoch": 2.056497175141243, + "grad_norm": 3.222146511077881, + "learning_rate": 1.775789874381424e-06, + "loss": 1.3826, + "step": 2002 + }, + { + "epoch": 2.0575243965074472, + "grad_norm": 3.3821890354156494, + "learning_rate": 1.7738865626189573e-06, + "loss": 1.3664, + "step": 2003 + }, + { + "epoch": 2.058551617873652, + "grad_norm": 3.2899715900421143, + "learning_rate": 1.7719832508564904e-06, + "loss": 1.3, + "step": 2004 + }, + { + "epoch": 2.0595788392398564, + "grad_norm": 3.2912356853485107, + "learning_rate": 1.7700799390940238e-06, + "loss": 1.3297, + "step": 2005 + }, + { + "epoch": 2.0606060606060606, + "grad_norm": 3.397883892059326, + "learning_rate": 1.768176627331557e-06, + "loss": 1.3622, + "step": 2006 + }, + { + "epoch": 2.061633281972265, + "grad_norm": 3.230801820755005, + "learning_rate": 1.7662733155690903e-06, + "loss": 1.353, + "step": 2007 + }, + { + "epoch": 2.0626605033384693, + "grad_norm": 3.2354795932769775, + "learning_rate": 1.7643700038066237e-06, + "loss": 1.3777, + "step": 2008 + }, + { + "epoch": 2.063687724704674, + "grad_norm": 3.086163282394409, + "learning_rate": 1.7624666920441568e-06, + "loss": 1.3087, + "step": 2009 + }, + { + "epoch": 2.0647149460708785, + "grad_norm": 3.1875600814819336, + "learning_rate": 1.7605633802816902e-06, + "loss": 1.3045, + "step": 2010 + }, + { + "epoch": 2.0647149460708785, + "eval_loss": 1.7808178663253784, + "eval_runtime": 36.9437, + "eval_samples_per_second": 27.068, + "eval_steps_per_second": 1.137, + "step": 2010 + }, + { + "epoch": 2.0657421674370826, + "grad_norm": 3.33947491645813, + "learning_rate": 1.7586600685192238e-06, + "loss": 1.3992, + "step": 2011 + }, + { + "epoch": 2.066769388803287, + "grad_norm": 3.314852476119995, + "learning_rate": 1.756756756756757e-06, + "loss": 1.4015, + "step": 2012 + }, + { + "epoch": 2.0677966101694913, + "grad_norm": 3.1022720336914062, + "learning_rate": 1.7548534449942903e-06, + "loss": 1.3664, + "step": 2013 + }, + { + "epoch": 2.068823831535696, + "grad_norm": 3.2397265434265137, + "learning_rate": 1.7529501332318236e-06, + "loss": 1.3429, + "step": 2014 + }, + { + "epoch": 2.0698510529019005, + "grad_norm": 3.1528069972991943, + "learning_rate": 1.7510468214693568e-06, + "loss": 1.3176, + "step": 2015 + }, + { + "epoch": 2.0708782742681047, + "grad_norm": 3.157923460006714, + "learning_rate": 1.7491435097068902e-06, + "loss": 1.3156, + "step": 2016 + }, + { + "epoch": 2.0719054956343093, + "grad_norm": 3.094817876815796, + "learning_rate": 1.7472401979444233e-06, + "loss": 1.3552, + "step": 2017 + }, + { + "epoch": 2.0729327170005134, + "grad_norm": 3.0657732486724854, + "learning_rate": 1.7453368861819567e-06, + "loss": 1.2914, + "step": 2018 + }, + { + "epoch": 2.073959938366718, + "grad_norm": 3.0663399696350098, + "learning_rate": 1.74343357441949e-06, + "loss": 1.344, + "step": 2019 + }, + { + "epoch": 2.0749871597329226, + "grad_norm": 3.1718454360961914, + "learning_rate": 1.7415302626570232e-06, + "loss": 1.3987, + "step": 2020 + }, + { + "epoch": 2.0749871597329226, + "eval_loss": 1.7751562595367432, + "eval_runtime": 37.0572, + "eval_samples_per_second": 26.985, + "eval_steps_per_second": 1.133, + "step": 2020 + }, + { + "epoch": 2.0760143810991267, + "grad_norm": 3.0701513290405273, + "learning_rate": 1.7396269508945568e-06, + "loss": 1.3327, + "step": 2021 + }, + { + "epoch": 2.0770416024653313, + "grad_norm": 3.1184256076812744, + "learning_rate": 1.7377236391320901e-06, + "loss": 1.3111, + "step": 2022 + }, + { + "epoch": 2.078068823831536, + "grad_norm": 3.2951152324676514, + "learning_rate": 1.7358203273696233e-06, + "loss": 1.3774, + "step": 2023 + }, + { + "epoch": 2.07909604519774, + "grad_norm": 3.2582433223724365, + "learning_rate": 1.7339170156071566e-06, + "loss": 1.3332, + "step": 2024 + }, + { + "epoch": 2.0801232665639446, + "grad_norm": 3.1642305850982666, + "learning_rate": 1.73201370384469e-06, + "loss": 1.3294, + "step": 2025 + }, + { + "epoch": 2.081150487930149, + "grad_norm": 3.2165608406066895, + "learning_rate": 1.7301103920822231e-06, + "loss": 1.344, + "step": 2026 + }, + { + "epoch": 2.0821777092963534, + "grad_norm": 3.247485637664795, + "learning_rate": 1.7282070803197565e-06, + "loss": 1.3496, + "step": 2027 + }, + { + "epoch": 2.083204930662558, + "grad_norm": 3.192835569381714, + "learning_rate": 1.7263037685572897e-06, + "loss": 1.3345, + "step": 2028 + }, + { + "epoch": 2.084232152028762, + "grad_norm": 3.2298715114593506, + "learning_rate": 1.724400456794823e-06, + "loss": 1.3666, + "step": 2029 + }, + { + "epoch": 2.0852593733949667, + "grad_norm": 3.091574192047119, + "learning_rate": 1.7224971450323564e-06, + "loss": 1.3321, + "step": 2030 + }, + { + "epoch": 2.0852593733949667, + "eval_loss": 1.7775636911392212, + "eval_runtime": 37.0648, + "eval_samples_per_second": 26.98, + "eval_steps_per_second": 1.133, + "step": 2030 + }, + { + "epoch": 2.086286594761171, + "grad_norm": 3.235821008682251, + "learning_rate": 1.72059383326989e-06, + "loss": 1.3372, + "step": 2031 + }, + { + "epoch": 2.0873138161273754, + "grad_norm": 3.333073377609253, + "learning_rate": 1.7186905215074231e-06, + "loss": 1.3926, + "step": 2032 + }, + { + "epoch": 2.08834103749358, + "grad_norm": 3.235518217086792, + "learning_rate": 1.7167872097449565e-06, + "loss": 1.3179, + "step": 2033 + }, + { + "epoch": 2.089368258859784, + "grad_norm": 3.1183924674987793, + "learning_rate": 1.7148838979824896e-06, + "loss": 1.3402, + "step": 2034 + }, + { + "epoch": 2.0903954802259888, + "grad_norm": 3.245620012283325, + "learning_rate": 1.712980586220023e-06, + "loss": 1.3351, + "step": 2035 + }, + { + "epoch": 2.091422701592193, + "grad_norm": 3.413208246231079, + "learning_rate": 1.7110772744575563e-06, + "loss": 1.3199, + "step": 2036 + }, + { + "epoch": 2.0924499229583975, + "grad_norm": 3.183093786239624, + "learning_rate": 1.7091739626950895e-06, + "loss": 1.3295, + "step": 2037 + }, + { + "epoch": 2.093477144324602, + "grad_norm": 3.2691562175750732, + "learning_rate": 1.7072706509326229e-06, + "loss": 1.3623, + "step": 2038 + }, + { + "epoch": 2.0945043656908062, + "grad_norm": 3.2356081008911133, + "learning_rate": 1.7053673391701562e-06, + "loss": 1.3547, + "step": 2039 + }, + { + "epoch": 2.095531587057011, + "grad_norm": 3.138723134994507, + "learning_rate": 1.7034640274076894e-06, + "loss": 1.3189, + "step": 2040 + }, + { + "epoch": 2.095531587057011, + "eval_loss": 1.7775871753692627, + "eval_runtime": 37.0654, + "eval_samples_per_second": 26.979, + "eval_steps_per_second": 1.133, + "step": 2040 + }, + { + "epoch": 2.0965588084232154, + "grad_norm": 3.2420573234558105, + "learning_rate": 1.701560715645223e-06, + "loss": 1.3862, + "step": 2041 + }, + { + "epoch": 2.0975860297894195, + "grad_norm": 3.137073278427124, + "learning_rate": 1.6996574038827563e-06, + "loss": 1.3076, + "step": 2042 + }, + { + "epoch": 2.098613251155624, + "grad_norm": 3.122952461242676, + "learning_rate": 1.6977540921202895e-06, + "loss": 1.3493, + "step": 2043 + }, + { + "epoch": 2.0996404725218283, + "grad_norm": 3.1556475162506104, + "learning_rate": 1.6958507803578228e-06, + "loss": 1.3337, + "step": 2044 + }, + { + "epoch": 2.100667693888033, + "grad_norm": 3.2341697216033936, + "learning_rate": 1.693947468595356e-06, + "loss": 1.3575, + "step": 2045 + }, + { + "epoch": 2.1016949152542375, + "grad_norm": 3.411771774291992, + "learning_rate": 1.6920441568328893e-06, + "loss": 1.3754, + "step": 2046 + }, + { + "epoch": 2.1027221366204416, + "grad_norm": 3.0538063049316406, + "learning_rate": 1.6901408450704227e-06, + "loss": 1.2938, + "step": 2047 + }, + { + "epoch": 2.103749357986646, + "grad_norm": 3.2879412174224854, + "learning_rate": 1.6882375333079559e-06, + "loss": 1.3313, + "step": 2048 + }, + { + "epoch": 2.1047765793528503, + "grad_norm": 3.1848442554473877, + "learning_rate": 1.6863342215454892e-06, + "loss": 1.3835, + "step": 2049 + }, + { + "epoch": 2.105803800719055, + "grad_norm": 3.1774306297302246, + "learning_rate": 1.6844309097830226e-06, + "loss": 1.3741, + "step": 2050 + }, + { + "epoch": 2.105803800719055, + "eval_loss": 1.77629554271698, + "eval_runtime": 37.0728, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 2050 + }, + { + "epoch": 2.1068310220852595, + "grad_norm": 3.218514919281006, + "learning_rate": 1.682527598020556e-06, + "loss": 1.3592, + "step": 2051 + }, + { + "epoch": 2.1078582434514637, + "grad_norm": 3.2372989654541016, + "learning_rate": 1.6806242862580893e-06, + "loss": 1.376, + "step": 2052 + }, + { + "epoch": 2.1088854648176683, + "grad_norm": 3.115527629852295, + "learning_rate": 1.6787209744956227e-06, + "loss": 1.3192, + "step": 2053 + }, + { + "epoch": 2.109912686183873, + "grad_norm": 3.06538987159729, + "learning_rate": 1.6768176627331558e-06, + "loss": 1.303, + "step": 2054 + }, + { + "epoch": 2.110939907550077, + "grad_norm": 3.075857162475586, + "learning_rate": 1.6749143509706892e-06, + "loss": 1.3504, + "step": 2055 + }, + { + "epoch": 2.1119671289162816, + "grad_norm": 3.2284908294677734, + "learning_rate": 1.6730110392082225e-06, + "loss": 1.3284, + "step": 2056 + }, + { + "epoch": 2.1129943502824857, + "grad_norm": 3.3034918308258057, + "learning_rate": 1.6711077274457557e-06, + "loss": 1.349, + "step": 2057 + }, + { + "epoch": 2.1140215716486903, + "grad_norm": 3.1239676475524902, + "learning_rate": 1.669204415683289e-06, + "loss": 1.2957, + "step": 2058 + }, + { + "epoch": 2.115048793014895, + "grad_norm": 3.4618728160858154, + "learning_rate": 1.6673011039208222e-06, + "loss": 1.3626, + "step": 2059 + }, + { + "epoch": 2.116076014381099, + "grad_norm": 3.4596455097198486, + "learning_rate": 1.6653977921583556e-06, + "loss": 1.3436, + "step": 2060 + }, + { + "epoch": 2.116076014381099, + "eval_loss": 1.7774609327316284, + "eval_runtime": 37.12, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 2060 + }, + { + "epoch": 2.1171032357473036, + "grad_norm": 3.6342878341674805, + "learning_rate": 1.663494480395889e-06, + "loss": 1.3618, + "step": 2061 + }, + { + "epoch": 2.118130457113508, + "grad_norm": 3.329286575317383, + "learning_rate": 1.6615911686334225e-06, + "loss": 1.3187, + "step": 2062 + }, + { + "epoch": 2.1191576784797124, + "grad_norm": 3.2518863677978516, + "learning_rate": 1.6596878568709557e-06, + "loss": 1.3412, + "step": 2063 + }, + { + "epoch": 2.120184899845917, + "grad_norm": 3.1159286499023438, + "learning_rate": 1.657784545108489e-06, + "loss": 1.331, + "step": 2064 + }, + { + "epoch": 2.121212121212121, + "grad_norm": 3.352390766143799, + "learning_rate": 1.6558812333460222e-06, + "loss": 1.3668, + "step": 2065 + }, + { + "epoch": 2.1222393425783257, + "grad_norm": 3.1647236347198486, + "learning_rate": 1.6539779215835555e-06, + "loss": 1.3574, + "step": 2066 + }, + { + "epoch": 2.12326656394453, + "grad_norm": 3.268143653869629, + "learning_rate": 1.652074609821089e-06, + "loss": 1.3247, + "step": 2067 + }, + { + "epoch": 2.1242937853107344, + "grad_norm": 3.081690788269043, + "learning_rate": 1.650171298058622e-06, + "loss": 1.3269, + "step": 2068 + }, + { + "epoch": 2.125321006676939, + "grad_norm": 3.1442677974700928, + "learning_rate": 1.6482679862961554e-06, + "loss": 1.329, + "step": 2069 + }, + { + "epoch": 2.126348228043143, + "grad_norm": 3.434091091156006, + "learning_rate": 1.6463646745336886e-06, + "loss": 1.3249, + "step": 2070 + }, + { + "epoch": 2.126348228043143, + "eval_loss": 1.7745627164840698, + "eval_runtime": 37.115, + "eval_samples_per_second": 26.943, + "eval_steps_per_second": 1.132, + "step": 2070 + }, + { + "epoch": 2.1273754494093478, + "grad_norm": 3.312300443649292, + "learning_rate": 1.644461362771222e-06, + "loss": 1.3091, + "step": 2071 + }, + { + "epoch": 2.128402670775552, + "grad_norm": 3.319504976272583, + "learning_rate": 1.6425580510087555e-06, + "loss": 1.3439, + "step": 2072 + }, + { + "epoch": 2.1294298921417565, + "grad_norm": 3.2242252826690674, + "learning_rate": 1.6406547392462889e-06, + "loss": 1.3068, + "step": 2073 + }, + { + "epoch": 2.130457113507961, + "grad_norm": 3.1730830669403076, + "learning_rate": 1.638751427483822e-06, + "loss": 1.3079, + "step": 2074 + }, + { + "epoch": 2.131484334874165, + "grad_norm": 3.2944724559783936, + "learning_rate": 1.6368481157213554e-06, + "loss": 1.3502, + "step": 2075 + }, + { + "epoch": 2.13251155624037, + "grad_norm": 3.1474711894989014, + "learning_rate": 1.6349448039588885e-06, + "loss": 1.3308, + "step": 2076 + }, + { + "epoch": 2.1335387776065744, + "grad_norm": 3.291567802429199, + "learning_rate": 1.6330414921964219e-06, + "loss": 1.3834, + "step": 2077 + }, + { + "epoch": 2.1345659989727785, + "grad_norm": 3.2176096439361572, + "learning_rate": 1.6311381804339553e-06, + "loss": 1.3134, + "step": 2078 + }, + { + "epoch": 2.135593220338983, + "grad_norm": 3.2517168521881104, + "learning_rate": 1.6292348686714884e-06, + "loss": 1.357, + "step": 2079 + }, + { + "epoch": 2.1366204417051873, + "grad_norm": 3.1429731845855713, + "learning_rate": 1.6273315569090218e-06, + "loss": 1.3327, + "step": 2080 + }, + { + "epoch": 2.1366204417051873, + "eval_loss": 1.7769756317138672, + "eval_runtime": 37.1282, + "eval_samples_per_second": 26.934, + "eval_steps_per_second": 1.131, + "step": 2080 + }, + { + "epoch": 2.137647663071392, + "grad_norm": 3.3201053142547607, + "learning_rate": 1.625428245146555e-06, + "loss": 1.4052, + "step": 2081 + }, + { + "epoch": 2.1386748844375965, + "grad_norm": 3.2188704013824463, + "learning_rate": 1.6235249333840885e-06, + "loss": 1.3829, + "step": 2082 + }, + { + "epoch": 2.1397021058038006, + "grad_norm": 3.2554514408111572, + "learning_rate": 1.6216216216216219e-06, + "loss": 1.3356, + "step": 2083 + }, + { + "epoch": 2.140729327170005, + "grad_norm": 3.286280870437622, + "learning_rate": 1.6197183098591552e-06, + "loss": 1.3627, + "step": 2084 + }, + { + "epoch": 2.1417565485362093, + "grad_norm": 3.4556727409362793, + "learning_rate": 1.6178149980966884e-06, + "loss": 1.35, + "step": 2085 + }, + { + "epoch": 2.142783769902414, + "grad_norm": 3.186457395553589, + "learning_rate": 1.6159116863342217e-06, + "loss": 1.3218, + "step": 2086 + }, + { + "epoch": 2.1438109912686185, + "grad_norm": 3.360995054244995, + "learning_rate": 1.6140083745717549e-06, + "loss": 1.3895, + "step": 2087 + }, + { + "epoch": 2.1448382126348227, + "grad_norm": 3.1155898571014404, + "learning_rate": 1.6121050628092882e-06, + "loss": 1.3113, + "step": 2088 + }, + { + "epoch": 2.1458654340010272, + "grad_norm": 3.300079584121704, + "learning_rate": 1.6102017510468216e-06, + "loss": 1.379, + "step": 2089 + }, + { + "epoch": 2.146892655367232, + "grad_norm": 3.1357364654541016, + "learning_rate": 1.6082984392843548e-06, + "loss": 1.2912, + "step": 2090 + }, + { + "epoch": 2.146892655367232, + "eval_loss": 1.7777626514434814, + "eval_runtime": 37.1337, + "eval_samples_per_second": 26.93, + "eval_steps_per_second": 1.131, + "step": 2090 + }, + { + "epoch": 2.147919876733436, + "grad_norm": 3.267508029937744, + "learning_rate": 1.6063951275218881e-06, + "loss": 1.3489, + "step": 2091 + }, + { + "epoch": 2.1489470980996406, + "grad_norm": 3.3493881225585938, + "learning_rate": 1.6044918157594217e-06, + "loss": 1.395, + "step": 2092 + }, + { + "epoch": 2.1499743194658447, + "grad_norm": 3.3308827877044678, + "learning_rate": 1.6025885039969548e-06, + "loss": 1.3225, + "step": 2093 + }, + { + "epoch": 2.1510015408320493, + "grad_norm": 3.253615617752075, + "learning_rate": 1.6006851922344882e-06, + "loss": 1.3537, + "step": 2094 + }, + { + "epoch": 2.152028762198254, + "grad_norm": 3.2310163974761963, + "learning_rate": 1.5987818804720216e-06, + "loss": 1.3581, + "step": 2095 + }, + { + "epoch": 2.153055983564458, + "grad_norm": 3.228498935699463, + "learning_rate": 1.5968785687095547e-06, + "loss": 1.3591, + "step": 2096 + }, + { + "epoch": 2.1540832049306626, + "grad_norm": 3.145155191421509, + "learning_rate": 1.594975256947088e-06, + "loss": 1.3068, + "step": 2097 + }, + { + "epoch": 2.1551104262968668, + "grad_norm": 3.19036865234375, + "learning_rate": 1.5930719451846212e-06, + "loss": 1.3645, + "step": 2098 + }, + { + "epoch": 2.1561376476630714, + "grad_norm": 3.1367108821868896, + "learning_rate": 1.5911686334221546e-06, + "loss": 1.3138, + "step": 2099 + }, + { + "epoch": 2.157164869029276, + "grad_norm": 3.268315076828003, + "learning_rate": 1.589265321659688e-06, + "loss": 1.3677, + "step": 2100 + }, + { + "epoch": 2.157164869029276, + "eval_loss": 1.7752312421798706, + "eval_runtime": 37.3358, + "eval_samples_per_second": 26.784, + "eval_steps_per_second": 1.125, + "step": 2100 + }, + { + "epoch": 2.15819209039548, + "grad_norm": 3.0902626514434814, + "learning_rate": 1.5873620098972211e-06, + "loss": 1.2737, + "step": 2101 + }, + { + "epoch": 2.1592193117616847, + "grad_norm": 3.183992385864258, + "learning_rate": 1.5854586981347547e-06, + "loss": 1.318, + "step": 2102 + }, + { + "epoch": 2.1602465331278893, + "grad_norm": 3.348939895629883, + "learning_rate": 1.583555386372288e-06, + "loss": 1.3618, + "step": 2103 + }, + { + "epoch": 2.1612737544940934, + "grad_norm": 3.395472526550293, + "learning_rate": 1.5816520746098212e-06, + "loss": 1.378, + "step": 2104 + }, + { + "epoch": 2.162300975860298, + "grad_norm": 3.367633819580078, + "learning_rate": 1.5797487628473546e-06, + "loss": 1.3318, + "step": 2105 + }, + { + "epoch": 2.163328197226502, + "grad_norm": 3.1082911491394043, + "learning_rate": 1.577845451084888e-06, + "loss": 1.3049, + "step": 2106 + }, + { + "epoch": 2.1643554185927067, + "grad_norm": 3.3265581130981445, + "learning_rate": 1.575942139322421e-06, + "loss": 1.3368, + "step": 2107 + }, + { + "epoch": 2.1653826399589113, + "grad_norm": 3.167185068130493, + "learning_rate": 1.5740388275599544e-06, + "loss": 1.2669, + "step": 2108 + }, + { + "epoch": 2.1664098613251155, + "grad_norm": 3.3141398429870605, + "learning_rate": 1.5721355157974876e-06, + "loss": 1.3532, + "step": 2109 + }, + { + "epoch": 2.16743708269132, + "grad_norm": 3.2805464267730713, + "learning_rate": 1.570232204035021e-06, + "loss": 1.3291, + "step": 2110 + }, + { + "epoch": 2.16743708269132, + "eval_loss": 1.7720998525619507, + "eval_runtime": 36.915, + "eval_samples_per_second": 27.089, + "eval_steps_per_second": 1.138, + "step": 2110 + }, + { + "epoch": 2.168464304057524, + "grad_norm": 3.2257862091064453, + "learning_rate": 1.5683288922725543e-06, + "loss": 1.3417, + "step": 2111 + }, + { + "epoch": 2.169491525423729, + "grad_norm": 3.1542367935180664, + "learning_rate": 1.5664255805100879e-06, + "loss": 1.3475, + "step": 2112 + }, + { + "epoch": 2.1705187467899334, + "grad_norm": 3.1991381645202637, + "learning_rate": 1.564522268747621e-06, + "loss": 1.3259, + "step": 2113 + }, + { + "epoch": 2.1715459681561375, + "grad_norm": 3.0904557704925537, + "learning_rate": 1.5626189569851544e-06, + "loss": 1.3489, + "step": 2114 + }, + { + "epoch": 2.172573189522342, + "grad_norm": 3.1544106006622314, + "learning_rate": 1.5607156452226875e-06, + "loss": 1.2626, + "step": 2115 + }, + { + "epoch": 2.1736004108885463, + "grad_norm": 3.2527761459350586, + "learning_rate": 1.558812333460221e-06, + "loss": 1.3497, + "step": 2116 + }, + { + "epoch": 2.174627632254751, + "grad_norm": 3.3398337364196777, + "learning_rate": 1.5569090216977543e-06, + "loss": 1.3492, + "step": 2117 + }, + { + "epoch": 2.1756548536209555, + "grad_norm": 3.246671199798584, + "learning_rate": 1.5550057099352874e-06, + "loss": 1.3495, + "step": 2118 + }, + { + "epoch": 2.1766820749871596, + "grad_norm": 3.4486782550811768, + "learning_rate": 1.5531023981728208e-06, + "loss": 1.3541, + "step": 2119 + }, + { + "epoch": 2.177709296353364, + "grad_norm": 3.209132432937622, + "learning_rate": 1.5511990864103542e-06, + "loss": 1.3134, + "step": 2120 + }, + { + "epoch": 2.177709296353364, + "eval_loss": 1.7801611423492432, + "eval_runtime": 36.9888, + "eval_samples_per_second": 27.035, + "eval_steps_per_second": 1.135, + "step": 2120 + }, + { + "epoch": 2.1787365177195683, + "grad_norm": 3.225480794906616, + "learning_rate": 1.5492957746478873e-06, + "loss": 1.3486, + "step": 2121 + }, + { + "epoch": 2.179763739085773, + "grad_norm": 3.2712321281433105, + "learning_rate": 1.5473924628854209e-06, + "loss": 1.3659, + "step": 2122 + }, + { + "epoch": 2.1807909604519775, + "grad_norm": 3.233598470687866, + "learning_rate": 1.5454891511229542e-06, + "loss": 1.3295, + "step": 2123 + }, + { + "epoch": 2.1818181818181817, + "grad_norm": 3.167868137359619, + "learning_rate": 1.5435858393604874e-06, + "loss": 1.3372, + "step": 2124 + }, + { + "epoch": 2.1828454031843862, + "grad_norm": 3.364309310913086, + "learning_rate": 1.5416825275980208e-06, + "loss": 1.356, + "step": 2125 + }, + { + "epoch": 2.183872624550591, + "grad_norm": 3.125401020050049, + "learning_rate": 1.539779215835554e-06, + "loss": 1.3306, + "step": 2126 + }, + { + "epoch": 2.184899845916795, + "grad_norm": 3.256223201751709, + "learning_rate": 1.5378759040730873e-06, + "loss": 1.3256, + "step": 2127 + }, + { + "epoch": 2.1859270672829996, + "grad_norm": 3.309673547744751, + "learning_rate": 1.5359725923106206e-06, + "loss": 1.3753, + "step": 2128 + }, + { + "epoch": 2.1869542886492037, + "grad_norm": 3.238278388977051, + "learning_rate": 1.5340692805481538e-06, + "loss": 1.3459, + "step": 2129 + }, + { + "epoch": 2.1879815100154083, + "grad_norm": 3.16030216217041, + "learning_rate": 1.5321659687856871e-06, + "loss": 1.3108, + "step": 2130 + }, + { + "epoch": 2.1879815100154083, + "eval_loss": 1.7832409143447876, + "eval_runtime": 37.0244, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 2130 + }, + { + "epoch": 2.189008731381613, + "grad_norm": 3.124526262283325, + "learning_rate": 1.5302626570232205e-06, + "loss": 1.3097, + "step": 2131 + }, + { + "epoch": 2.190035952747817, + "grad_norm": 3.112478256225586, + "learning_rate": 1.5283593452607539e-06, + "loss": 1.278, + "step": 2132 + }, + { + "epoch": 2.1910631741140216, + "grad_norm": 3.4559900760650635, + "learning_rate": 1.5264560334982872e-06, + "loss": 1.385, + "step": 2133 + }, + { + "epoch": 2.1920903954802258, + "grad_norm": 3.1703872680664062, + "learning_rate": 1.5245527217358206e-06, + "loss": 1.3299, + "step": 2134 + }, + { + "epoch": 2.1931176168464304, + "grad_norm": 3.314497947692871, + "learning_rate": 1.5226494099733537e-06, + "loss": 1.3474, + "step": 2135 + }, + { + "epoch": 2.194144838212635, + "grad_norm": 3.0944998264312744, + "learning_rate": 1.520746098210887e-06, + "loss": 1.3214, + "step": 2136 + }, + { + "epoch": 2.195172059578839, + "grad_norm": 3.2073676586151123, + "learning_rate": 1.5188427864484205e-06, + "loss": 1.3766, + "step": 2137 + }, + { + "epoch": 2.1961992809450437, + "grad_norm": 3.0689923763275146, + "learning_rate": 1.5169394746859536e-06, + "loss": 1.319, + "step": 2138 + }, + { + "epoch": 2.1972265023112483, + "grad_norm": 3.1764934062957764, + "learning_rate": 1.515036162923487e-06, + "loss": 1.3342, + "step": 2139 + }, + { + "epoch": 2.1982537236774524, + "grad_norm": 3.1994433403015137, + "learning_rate": 1.5131328511610201e-06, + "loss": 1.3462, + "step": 2140 + }, + { + "epoch": 2.1982537236774524, + "eval_loss": 1.7849186658859253, + "eval_runtime": 37.0183, + "eval_samples_per_second": 27.014, + "eval_steps_per_second": 1.135, + "step": 2140 + }, + { + "epoch": 2.199280945043657, + "grad_norm": 3.37168025970459, + "learning_rate": 1.5112295393985535e-06, + "loss": 1.3805, + "step": 2141 + }, + { + "epoch": 2.200308166409861, + "grad_norm": 3.1596593856811523, + "learning_rate": 1.509326227636087e-06, + "loss": 1.3543, + "step": 2142 + }, + { + "epoch": 2.2013353877760657, + "grad_norm": 3.086030960083008, + "learning_rate": 1.5074229158736204e-06, + "loss": 1.2834, + "step": 2143 + }, + { + "epoch": 2.2023626091422703, + "grad_norm": 3.2379868030548096, + "learning_rate": 1.5055196041111536e-06, + "loss": 1.342, + "step": 2144 + }, + { + "epoch": 2.2033898305084745, + "grad_norm": 3.233773708343506, + "learning_rate": 1.503616292348687e-06, + "loss": 1.3455, + "step": 2145 + }, + { + "epoch": 2.204417051874679, + "grad_norm": 3.4700264930725098, + "learning_rate": 1.50171298058622e-06, + "loss": 1.3741, + "step": 2146 + }, + { + "epoch": 2.205444273240883, + "grad_norm": 3.1174497604370117, + "learning_rate": 1.4998096688237535e-06, + "loss": 1.2824, + "step": 2147 + }, + { + "epoch": 2.206471494607088, + "grad_norm": 3.1174497604370117, + "learning_rate": 1.4998096688237535e-06, + "loss": 1.356, + "step": 2148 + }, + { + "epoch": 2.2074987159732924, + "grad_norm": 3.3671200275421143, + "learning_rate": 1.4979063570612868e-06, + "loss": 1.3421, + "step": 2149 + }, + { + "epoch": 2.2085259373394965, + "grad_norm": 3.2470571994781494, + "learning_rate": 1.49600304529882e-06, + "loss": 1.3181, + "step": 2150 + }, + { + "epoch": 2.2085259373394965, + "eval_loss": 1.7855594158172607, + "eval_runtime": 37.0613, + "eval_samples_per_second": 26.982, + "eval_steps_per_second": 1.133, + "step": 2150 + }, + { + "epoch": 2.209553158705701, + "grad_norm": 3.3434650897979736, + "learning_rate": 1.4940997335363533e-06, + "loss": 1.3418, + "step": 2151 + }, + { + "epoch": 2.2105803800719057, + "grad_norm": 3.169682502746582, + "learning_rate": 1.4921964217738865e-06, + "loss": 1.3279, + "step": 2152 + }, + { + "epoch": 2.21160760143811, + "grad_norm": 3.3032784461975098, + "learning_rate": 1.49029311001142e-06, + "loss": 1.3582, + "step": 2153 + }, + { + "epoch": 2.2126348228043144, + "grad_norm": 3.2480883598327637, + "learning_rate": 1.4883897982489534e-06, + "loss": 1.362, + "step": 2154 + }, + { + "epoch": 2.2136620441705186, + "grad_norm": 3.2822015285491943, + "learning_rate": 1.4864864864864868e-06, + "loss": 1.3299, + "step": 2155 + }, + { + "epoch": 2.214689265536723, + "grad_norm": 3.2832584381103516, + "learning_rate": 1.48458317472402e-06, + "loss": 1.3374, + "step": 2156 + }, + { + "epoch": 2.2157164869029278, + "grad_norm": 3.057344913482666, + "learning_rate": 1.4826798629615533e-06, + "loss": 1.2969, + "step": 2157 + }, + { + "epoch": 2.216743708269132, + "grad_norm": 3.279812812805176, + "learning_rate": 1.4807765511990864e-06, + "loss": 1.3083, + "step": 2158 + }, + { + "epoch": 2.2177709296353365, + "grad_norm": 3.273972272872925, + "learning_rate": 1.4788732394366198e-06, + "loss": 1.3285, + "step": 2159 + }, + { + "epoch": 2.2187981510015407, + "grad_norm": 3.231541633605957, + "learning_rate": 1.4769699276741532e-06, + "loss": 1.3178, + "step": 2160 + }, + { + "epoch": 2.2187981510015407, + "eval_loss": 1.7819184064865112, + "eval_runtime": 37.0481, + "eval_samples_per_second": 26.992, + "eval_steps_per_second": 1.134, + "step": 2160 + }, + { + "epoch": 2.2198253723677452, + "grad_norm": 3.125030755996704, + "learning_rate": 1.4750666159116863e-06, + "loss": 1.2916, + "step": 2161 + }, + { + "epoch": 2.22085259373395, + "grad_norm": 3.172628164291382, + "learning_rate": 1.4731633041492197e-06, + "loss": 1.3554, + "step": 2162 + }, + { + "epoch": 2.221879815100154, + "grad_norm": 3.4434096813201904, + "learning_rate": 1.4712599923867528e-06, + "loss": 1.3769, + "step": 2163 + }, + { + "epoch": 2.2229070364663586, + "grad_norm": 3.1751928329467773, + "learning_rate": 1.4693566806242864e-06, + "loss": 1.2994, + "step": 2164 + }, + { + "epoch": 2.2239342578325627, + "grad_norm": 3.153481960296631, + "learning_rate": 1.4674533688618198e-06, + "loss": 1.3245, + "step": 2165 + }, + { + "epoch": 2.2249614791987673, + "grad_norm": 3.2735719680786133, + "learning_rate": 1.4655500570993531e-06, + "loss": 1.3312, + "step": 2166 + }, + { + "epoch": 2.225988700564972, + "grad_norm": 3.34511137008667, + "learning_rate": 1.4636467453368863e-06, + "loss": 1.3805, + "step": 2167 + }, + { + "epoch": 2.227015921931176, + "grad_norm": 3.2341556549072266, + "learning_rate": 1.4617434335744197e-06, + "loss": 1.3694, + "step": 2168 + }, + { + "epoch": 2.2280431432973806, + "grad_norm": 3.2977921962738037, + "learning_rate": 1.4598401218119528e-06, + "loss": 1.3546, + "step": 2169 + }, + { + "epoch": 2.2290703646635848, + "grad_norm": 3.2472712993621826, + "learning_rate": 1.4579368100494862e-06, + "loss": 1.3382, + "step": 2170 + }, + { + "epoch": 2.2290703646635848, + "eval_loss": 1.7886496782302856, + "eval_runtime": 37.0389, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 2170 + }, + { + "epoch": 2.2300975860297894, + "grad_norm": 3.219923257827759, + "learning_rate": 1.4560334982870195e-06, + "loss": 1.3479, + "step": 2171 + }, + { + "epoch": 2.231124807395994, + "grad_norm": 3.297858476638794, + "learning_rate": 1.4541301865245527e-06, + "loss": 1.3507, + "step": 2172 + }, + { + "epoch": 2.232152028762198, + "grad_norm": 3.2059547901153564, + "learning_rate": 1.452226874762086e-06, + "loss": 1.3111, + "step": 2173 + }, + { + "epoch": 2.2331792501284027, + "grad_norm": 3.320009231567383, + "learning_rate": 1.4503235629996196e-06, + "loss": 1.3853, + "step": 2174 + }, + { + "epoch": 2.2342064714946073, + "grad_norm": 3.116455554962158, + "learning_rate": 1.4484202512371528e-06, + "loss": 1.2811, + "step": 2175 + }, + { + "epoch": 2.2352336928608114, + "grad_norm": 3.2926032543182373, + "learning_rate": 1.4465169394746861e-06, + "loss": 1.3899, + "step": 2176 + }, + { + "epoch": 2.236260914227016, + "grad_norm": 3.2553927898406982, + "learning_rate": 1.4446136277122195e-06, + "loss": 1.289, + "step": 2177 + }, + { + "epoch": 2.23728813559322, + "grad_norm": 3.3934881687164307, + "learning_rate": 1.4427103159497526e-06, + "loss": 1.314, + "step": 2178 + }, + { + "epoch": 2.2383153569594247, + "grad_norm": 3.380460262298584, + "learning_rate": 1.440807004187286e-06, + "loss": 1.3433, + "step": 2179 + }, + { + "epoch": 2.2393425783256293, + "grad_norm": 3.6436474323272705, + "learning_rate": 1.4389036924248192e-06, + "loss": 1.366, + "step": 2180 + }, + { + "epoch": 2.2393425783256293, + "eval_loss": 1.7869333028793335, + "eval_runtime": 37.0766, + "eval_samples_per_second": 26.971, + "eval_steps_per_second": 1.133, + "step": 2180 + }, + { + "epoch": 2.2403697996918335, + "grad_norm": 3.26050066947937, + "learning_rate": 1.4370003806623525e-06, + "loss": 1.3365, + "step": 2181 + }, + { + "epoch": 2.241397021058038, + "grad_norm": 3.31622576713562, + "learning_rate": 1.4350970688998859e-06, + "loss": 1.3857, + "step": 2182 + }, + { + "epoch": 2.242424242424242, + "grad_norm": 3.2586722373962402, + "learning_rate": 1.433193757137419e-06, + "loss": 1.3216, + "step": 2183 + }, + { + "epoch": 2.243451463790447, + "grad_norm": 3.00472092628479, + "learning_rate": 1.4312904453749526e-06, + "loss": 1.2854, + "step": 2184 + }, + { + "epoch": 2.2444786851566514, + "grad_norm": 3.261831760406494, + "learning_rate": 1.429387133612486e-06, + "loss": 1.3284, + "step": 2185 + }, + { + "epoch": 2.2455059065228555, + "grad_norm": 3.1558244228363037, + "learning_rate": 1.4274838218500191e-06, + "loss": 1.3478, + "step": 2186 + }, + { + "epoch": 2.24653312788906, + "grad_norm": 3.3475098609924316, + "learning_rate": 1.4255805100875525e-06, + "loss": 1.3973, + "step": 2187 + }, + { + "epoch": 2.2475603492552647, + "grad_norm": 3.2336196899414062, + "learning_rate": 1.4236771983250858e-06, + "loss": 1.3334, + "step": 2188 + }, + { + "epoch": 2.248587570621469, + "grad_norm": 3.107243537902832, + "learning_rate": 1.421773886562619e-06, + "loss": 1.3104, + "step": 2189 + }, + { + "epoch": 2.2496147919876734, + "grad_norm": 3.0853683948516846, + "learning_rate": 1.4198705748001524e-06, + "loss": 1.326, + "step": 2190 + }, + { + "epoch": 2.2496147919876734, + "eval_loss": 1.7847614288330078, + "eval_runtime": 37.1233, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 2190 + }, + { + "epoch": 2.2506420133538776, + "grad_norm": 3.3024446964263916, + "learning_rate": 1.4179672630376855e-06, + "loss": 1.3083, + "step": 2191 + }, + { + "epoch": 2.251669234720082, + "grad_norm": 3.2713446617126465, + "learning_rate": 1.4160639512752189e-06, + "loss": 1.3272, + "step": 2192 + }, + { + "epoch": 2.2526964560862868, + "grad_norm": 3.2788100242614746, + "learning_rate": 1.4141606395127522e-06, + "loss": 1.2954, + "step": 2193 + }, + { + "epoch": 2.253723677452491, + "grad_norm": 3.1132969856262207, + "learning_rate": 1.4122573277502858e-06, + "loss": 1.3539, + "step": 2194 + }, + { + "epoch": 2.2547508988186955, + "grad_norm": 3.1661081314086914, + "learning_rate": 1.410354015987819e-06, + "loss": 1.3275, + "step": 2195 + }, + { + "epoch": 2.2557781201848996, + "grad_norm": 3.156219959259033, + "learning_rate": 1.4084507042253523e-06, + "loss": 1.3608, + "step": 2196 + }, + { + "epoch": 2.2568053415511042, + "grad_norm": 3.0919189453125, + "learning_rate": 1.4065473924628855e-06, + "loss": 1.3308, + "step": 2197 + }, + { + "epoch": 2.257832562917309, + "grad_norm": 3.244119882583618, + "learning_rate": 1.4046440807004188e-06, + "loss": 1.3188, + "step": 2198 + }, + { + "epoch": 2.258859784283513, + "grad_norm": 3.2616937160491943, + "learning_rate": 1.4027407689379522e-06, + "loss": 1.3105, + "step": 2199 + }, + { + "epoch": 2.2598870056497176, + "grad_norm": 3.2117950916290283, + "learning_rate": 1.4008374571754854e-06, + "loss": 1.3396, + "step": 2200 + }, + { + "epoch": 2.2598870056497176, + "eval_loss": 1.7757095098495483, + "eval_runtime": 37.2922, + "eval_samples_per_second": 26.815, + "eval_steps_per_second": 1.126, + "step": 2200 + } + ], + "logging_steps": 1, + "max_steps": 2919, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 6.858080700596224e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2200/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/config.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/config.json @@ -0,0 +1,36 @@ +{ + "_name_or_path": "meta-llama/Llama-3.2-3B", + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128001, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 3072, + "initializer_range": 0.02, + "intermediate_size": 8192, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 24, + "num_hidden_layers": 28, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 32.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": true, + "torch_dtype": "float16", + "transformers_version": "4.45.1", + "use_cache": true, + "vocab_size": 128256 +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/generation_config.json @@ -0,0 +1,9 @@ +{ + "_from_model_config": true, + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": 128001, + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.45.1" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/latest b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/latest new file mode 100644 index 0000000000000000000000000000000000000000..98f8bed9a5485ee900d9931cc06950de69499848 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/latest @@ -0,0 +1 @@ +global_step2500 \ No newline at end of file diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/model.safetensors.index.json @@ -0,0 +1,262 @@ +{ + "metadata": { + "total_size": 7213504512 + }, + "weight_map": { + "lm_head.weight": "model-00002-of-00002.safetensors", + "model.embed_tokens.weight": "model-00001-of-00002.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors", + "model.norm.weight": "model-00002-of-00002.safetensors" + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..0b782d4602b361f7577abb52b710a8c0e088588d --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/tokenizer_config.json @@ -0,0 +1,2078 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128256": { + "content": "[PAD]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128257": { + "content": "🅁", + "lstrip": true, + "normalized": true, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "bos_token": "<|begin_of_text|>", + "clean_up_tokenization_spaces": true, + "eos_token": "<|end_of_text|>", + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "[PAD]", + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..017742abe1d8fcc78b159b9d7489682ecbd20607 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/trainer_state.json @@ -0,0 +1,19533 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.5680534155110424, + "eval_steps": 10, + "global_step": 2500, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001027221366204417, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7525, + "step": 1 + }, + { + "epoch": 0.002054442732408834, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7605, + "step": 2 + }, + { + "epoch": 0.0030816640986132513, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8008, + "step": 3 + }, + { + "epoch": 0.004108885464817668, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.809, + "step": 4 + }, + { + "epoch": 0.005136106831022085, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7841, + "step": 5 + }, + { + "epoch": 0.0061633281972265025, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7924, + "step": 6 + }, + { + "epoch": 0.007190549563430919, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7732, + "step": 7 + }, + { + "epoch": 0.008217770929635337, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7953, + "step": 8 + }, + { + "epoch": 0.009244992295839754, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7747, + "step": 9 + }, + { + "epoch": 0.01027221366204417, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7909, + "step": 10 + }, + { + "epoch": 0.01027221366204417, + "eval_loss": 2.811410665512085, + "eval_runtime": 37.9307, + "eval_samples_per_second": 26.364, + "eval_steps_per_second": 1.107, + "step": 10 + }, + { + "epoch": 0.011299435028248588, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7592, + "step": 11 + }, + { + "epoch": 0.012326656394453005, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7394, + "step": 12 + }, + { + "epoch": 0.01335387776065742, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.8111, + "step": 13 + }, + { + "epoch": 0.014381099126861838, + "grad_norm": 0.0, + "learning_rate": 0.0, + "loss": 2.7786, + "step": 14 + }, + { + "epoch": 0.015408320493066256, + "grad_norm": 5.814355373382568, + "learning_rate": 1.7123287671232876e-08, + "loss": 2.7711, + "step": 15 + }, + { + "epoch": 0.016435541859270673, + "grad_norm": 6.2984514236450195, + "learning_rate": 3.424657534246575e-08, + "loss": 2.7517, + "step": 16 + }, + { + "epoch": 0.01746276322547509, + "grad_norm": 6.235473155975342, + "learning_rate": 5.136986301369863e-08, + "loss": 2.7191, + "step": 17 + }, + { + "epoch": 0.01848998459167951, + "grad_norm": 6.235473155975342, + "learning_rate": 5.136986301369863e-08, + "loss": 2.806, + "step": 18 + }, + { + "epoch": 0.019517205957883924, + "grad_norm": 6.621068000793457, + "learning_rate": 6.84931506849315e-08, + "loss": 2.7192, + "step": 19 + }, + { + "epoch": 0.02054442732408834, + "grad_norm": 6.781496524810791, + "learning_rate": 8.561643835616439e-08, + "loss": 2.7862, + "step": 20 + }, + { + "epoch": 0.02054442732408834, + "eval_loss": 2.8109023571014404, + "eval_runtime": 37.527, + "eval_samples_per_second": 26.647, + "eval_steps_per_second": 1.119, + "step": 20 + }, + { + "epoch": 0.02157164869029276, + "grad_norm": 6.336451530456543, + "learning_rate": 1.0273972602739726e-07, + "loss": 2.7997, + "step": 21 + }, + { + "epoch": 0.022598870056497175, + "grad_norm": 6.184426307678223, + "learning_rate": 1.1986301369863014e-07, + "loss": 2.7938, + "step": 22 + }, + { + "epoch": 0.02362609142270159, + "grad_norm": 6.443874359130859, + "learning_rate": 1.36986301369863e-07, + "loss": 2.7485, + "step": 23 + }, + { + "epoch": 0.02465331278890601, + "grad_norm": 6.17667293548584, + "learning_rate": 1.541095890410959e-07, + "loss": 2.7779, + "step": 24 + }, + { + "epoch": 0.025680534155110426, + "grad_norm": 5.850281715393066, + "learning_rate": 1.7123287671232878e-07, + "loss": 2.7598, + "step": 25 + }, + { + "epoch": 0.02670775552131484, + "grad_norm": 5.788112640380859, + "learning_rate": 1.8835616438356165e-07, + "loss": 2.7837, + "step": 26 + }, + { + "epoch": 0.02773497688751926, + "grad_norm": 5.17655086517334, + "learning_rate": 2.0547945205479452e-07, + "loss": 2.7513, + "step": 27 + }, + { + "epoch": 0.028762198253723677, + "grad_norm": 4.698268890380859, + "learning_rate": 2.226027397260274e-07, + "loss": 2.7714, + "step": 28 + }, + { + "epoch": 0.029789419619928096, + "grad_norm": 4.562674522399902, + "learning_rate": 2.397260273972603e-07, + "loss": 2.7759, + "step": 29 + }, + { + "epoch": 0.030816640986132512, + "grad_norm": 4.138142108917236, + "learning_rate": 2.568493150684932e-07, + "loss": 2.7656, + "step": 30 + }, + { + "epoch": 0.030816640986132512, + "eval_loss": 2.784187078475952, + "eval_runtime": 36.9691, + "eval_samples_per_second": 27.05, + "eval_steps_per_second": 1.136, + "step": 30 + }, + { + "epoch": 0.03184386235233693, + "grad_norm": 4.39779806137085, + "learning_rate": 2.73972602739726e-07, + "loss": 2.7377, + "step": 31 + }, + { + "epoch": 0.03287108371854135, + "grad_norm": 3.9160122871398926, + "learning_rate": 2.910958904109589e-07, + "loss": 2.7753, + "step": 32 + }, + { + "epoch": 0.03389830508474576, + "grad_norm": 4.828811168670654, + "learning_rate": 3.082191780821918e-07, + "loss": 2.7311, + "step": 33 + }, + { + "epoch": 0.03492552645095018, + "grad_norm": 4.6300435066223145, + "learning_rate": 3.2534246575342466e-07, + "loss": 2.7287, + "step": 34 + }, + { + "epoch": 0.035952747817154594, + "grad_norm": 4.47259521484375, + "learning_rate": 3.4246575342465755e-07, + "loss": 2.7117, + "step": 35 + }, + { + "epoch": 0.03697996918335902, + "grad_norm": 4.557999610900879, + "learning_rate": 3.595890410958904e-07, + "loss": 2.6291, + "step": 36 + }, + { + "epoch": 0.03800719054956343, + "grad_norm": 4.3974761962890625, + "learning_rate": 3.767123287671233e-07, + "loss": 2.7421, + "step": 37 + }, + { + "epoch": 0.03903441191576785, + "grad_norm": 3.6151702404022217, + "learning_rate": 3.938356164383562e-07, + "loss": 2.6744, + "step": 38 + }, + { + "epoch": 0.040061633281972264, + "grad_norm": 3.165523052215576, + "learning_rate": 4.1095890410958903e-07, + "loss": 2.6694, + "step": 39 + }, + { + "epoch": 0.04108885464817668, + "grad_norm": 3.2827563285827637, + "learning_rate": 4.2808219178082193e-07, + "loss": 2.7177, + "step": 40 + }, + { + "epoch": 0.04108885464817668, + "eval_loss": 2.705791711807251, + "eval_runtime": 36.995, + "eval_samples_per_second": 27.031, + "eval_steps_per_second": 1.135, + "step": 40 + }, + { + "epoch": 0.042116076014381096, + "grad_norm": 3.241729736328125, + "learning_rate": 4.452054794520548e-07, + "loss": 2.6703, + "step": 41 + }, + { + "epoch": 0.04314329738058552, + "grad_norm": 3.401049852371216, + "learning_rate": 4.6232876712328767e-07, + "loss": 2.6036, + "step": 42 + }, + { + "epoch": 0.044170518746789934, + "grad_norm": 3.2498552799224854, + "learning_rate": 4.794520547945206e-07, + "loss": 2.6666, + "step": 43 + }, + { + "epoch": 0.04519774011299435, + "grad_norm": 3.1376283168792725, + "learning_rate": 4.965753424657534e-07, + "loss": 2.6331, + "step": 44 + }, + { + "epoch": 0.046224961479198766, + "grad_norm": 2.885286569595337, + "learning_rate": 5.136986301369864e-07, + "loss": 2.6609, + "step": 45 + }, + { + "epoch": 0.04725218284540318, + "grad_norm": 2.636420249938965, + "learning_rate": 5.308219178082192e-07, + "loss": 2.6462, + "step": 46 + }, + { + "epoch": 0.048279404211607604, + "grad_norm": 2.5609490871429443, + "learning_rate": 5.47945205479452e-07, + "loss": 2.6821, + "step": 47 + }, + { + "epoch": 0.04930662557781202, + "grad_norm": 2.7683653831481934, + "learning_rate": 5.65068493150685e-07, + "loss": 2.61, + "step": 48 + }, + { + "epoch": 0.050333846944016436, + "grad_norm": 3.3999135494232178, + "learning_rate": 5.821917808219178e-07, + "loss": 2.6166, + "step": 49 + }, + { + "epoch": 0.05136106831022085, + "grad_norm": 2.922145128250122, + "learning_rate": 5.993150684931507e-07, + "loss": 2.5657, + "step": 50 + }, + { + "epoch": 0.05136106831022085, + "eval_loss": 2.6353814601898193, + "eval_runtime": 37.0591, + "eval_samples_per_second": 26.984, + "eval_steps_per_second": 1.133, + "step": 50 + }, + { + "epoch": 0.05238828967642527, + "grad_norm": 2.5821900367736816, + "learning_rate": 6.164383561643836e-07, + "loss": 2.5512, + "step": 51 + }, + { + "epoch": 0.05341551104262968, + "grad_norm": 2.2299063205718994, + "learning_rate": 6.335616438356165e-07, + "loss": 2.6082, + "step": 52 + }, + { + "epoch": 0.054442732408834106, + "grad_norm": 2.0938310623168945, + "learning_rate": 6.506849315068493e-07, + "loss": 2.5301, + "step": 53 + }, + { + "epoch": 0.05546995377503852, + "grad_norm": 2.186629295349121, + "learning_rate": 6.678082191780823e-07, + "loss": 2.5755, + "step": 54 + }, + { + "epoch": 0.05649717514124294, + "grad_norm": 2.1747493743896484, + "learning_rate": 6.849315068493151e-07, + "loss": 2.5451, + "step": 55 + }, + { + "epoch": 0.05752439650744735, + "grad_norm": 2.5470354557037354, + "learning_rate": 7.020547945205481e-07, + "loss": 2.5159, + "step": 56 + }, + { + "epoch": 0.05855161787365177, + "grad_norm": 2.5056023597717285, + "learning_rate": 7.191780821917808e-07, + "loss": 2.5431, + "step": 57 + }, + { + "epoch": 0.05957883923985619, + "grad_norm": 2.6059749126434326, + "learning_rate": 7.363013698630137e-07, + "loss": 2.4684, + "step": 58 + }, + { + "epoch": 0.06060606060606061, + "grad_norm": 1.87150239944458, + "learning_rate": 7.534246575342466e-07, + "loss": 2.5269, + "step": 59 + }, + { + "epoch": 0.061633281972265024, + "grad_norm": 1.8777480125427246, + "learning_rate": 7.705479452054795e-07, + "loss": 2.5253, + "step": 60 + }, + { + "epoch": 0.061633281972265024, + "eval_loss": 2.5705313682556152, + "eval_runtime": 37.0929, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 60 + }, + { + "epoch": 0.06266050333846944, + "grad_norm": 1.8500003814697266, + "learning_rate": 7.876712328767124e-07, + "loss": 2.4987, + "step": 61 + }, + { + "epoch": 0.06368772470467386, + "grad_norm": 1.9473347663879395, + "learning_rate": 8.047945205479453e-07, + "loss": 2.4866, + "step": 62 + }, + { + "epoch": 0.06471494607087827, + "grad_norm": 2.2350873947143555, + "learning_rate": 8.219178082191781e-07, + "loss": 2.5343, + "step": 63 + }, + { + "epoch": 0.0657421674370827, + "grad_norm": 2.1791350841522217, + "learning_rate": 8.39041095890411e-07, + "loss": 2.5185, + "step": 64 + }, + { + "epoch": 0.0667693888032871, + "grad_norm": 1.8400903940200806, + "learning_rate": 8.561643835616439e-07, + "loss": 2.3897, + "step": 65 + }, + { + "epoch": 0.06779661016949153, + "grad_norm": 1.7193644046783447, + "learning_rate": 8.732876712328768e-07, + "loss": 2.485, + "step": 66 + }, + { + "epoch": 0.06882383153569595, + "grad_norm": 1.8478909730911255, + "learning_rate": 8.904109589041097e-07, + "loss": 2.4404, + "step": 67 + }, + { + "epoch": 0.06985105290190036, + "grad_norm": 1.8803719282150269, + "learning_rate": 9.075342465753426e-07, + "loss": 2.4933, + "step": 68 + }, + { + "epoch": 0.07087827426810478, + "grad_norm": 1.6866114139556885, + "learning_rate": 9.246575342465753e-07, + "loss": 2.4441, + "step": 69 + }, + { + "epoch": 0.07190549563430919, + "grad_norm": 1.807783842086792, + "learning_rate": 9.417808219178083e-07, + "loss": 2.5477, + "step": 70 + }, + { + "epoch": 0.07190549563430919, + "eval_loss": 2.5140035152435303, + "eval_runtime": 37.1384, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 70 + }, + { + "epoch": 0.07293271700051361, + "grad_norm": 1.7754000425338745, + "learning_rate": 9.589041095890411e-07, + "loss": 2.44, + "step": 71 + }, + { + "epoch": 0.07395993836671803, + "grad_norm": 1.7819652557373047, + "learning_rate": 9.76027397260274e-07, + "loss": 2.402, + "step": 72 + }, + { + "epoch": 0.07498715973292244, + "grad_norm": 1.6666278839111328, + "learning_rate": 9.931506849315068e-07, + "loss": 2.4511, + "step": 73 + }, + { + "epoch": 0.07601438109912687, + "grad_norm": 2.0625369548797607, + "learning_rate": 1.0102739726027399e-06, + "loss": 2.424, + "step": 74 + }, + { + "epoch": 0.07704160246533127, + "grad_norm": 1.9408000707626343, + "learning_rate": 1.0273972602739727e-06, + "loss": 2.4286, + "step": 75 + }, + { + "epoch": 0.0780688238315357, + "grad_norm": 1.7670842409133911, + "learning_rate": 1.0445205479452056e-06, + "loss": 2.407, + "step": 76 + }, + { + "epoch": 0.07909604519774012, + "grad_norm": 3.045717716217041, + "learning_rate": 1.0616438356164384e-06, + "loss": 2.425, + "step": 77 + }, + { + "epoch": 0.08012326656394453, + "grad_norm": 1.9278619289398193, + "learning_rate": 1.0787671232876712e-06, + "loss": 2.4306, + "step": 78 + }, + { + "epoch": 0.08115048793014895, + "grad_norm": 2.405146837234497, + "learning_rate": 1.095890410958904e-06, + "loss": 2.3499, + "step": 79 + }, + { + "epoch": 0.08217770929635336, + "grad_norm": 2.1083860397338867, + "learning_rate": 1.1130136986301371e-06, + "loss": 2.4115, + "step": 80 + }, + { + "epoch": 0.08217770929635336, + "eval_loss": 2.4501075744628906, + "eval_runtime": 37.1247, + "eval_samples_per_second": 26.936, + "eval_steps_per_second": 1.131, + "step": 80 + }, + { + "epoch": 0.08320493066255778, + "grad_norm": 2.181162118911743, + "learning_rate": 1.13013698630137e-06, + "loss": 2.3517, + "step": 81 + }, + { + "epoch": 0.08423215202876219, + "grad_norm": 2.1833999156951904, + "learning_rate": 1.1472602739726028e-06, + "loss": 2.38, + "step": 82 + }, + { + "epoch": 0.08525937339496661, + "grad_norm": 2.670384645462036, + "learning_rate": 1.1643835616438357e-06, + "loss": 2.3833, + "step": 83 + }, + { + "epoch": 0.08628659476117104, + "grad_norm": 2.438676118850708, + "learning_rate": 1.1815068493150685e-06, + "loss": 2.3399, + "step": 84 + }, + { + "epoch": 0.08731381612737545, + "grad_norm": 2.7584919929504395, + "learning_rate": 1.1986301369863014e-06, + "loss": 2.31, + "step": 85 + }, + { + "epoch": 0.08834103749357987, + "grad_norm": 2.4129064083099365, + "learning_rate": 1.2157534246575344e-06, + "loss": 2.3758, + "step": 86 + }, + { + "epoch": 0.08936825885978428, + "grad_norm": 2.366154193878174, + "learning_rate": 1.2328767123287673e-06, + "loss": 2.3154, + "step": 87 + }, + { + "epoch": 0.0903954802259887, + "grad_norm": 2.4916300773620605, + "learning_rate": 1.25e-06, + "loss": 2.3022, + "step": 88 + }, + { + "epoch": 0.09142270159219312, + "grad_norm": 2.4606640338897705, + "learning_rate": 1.267123287671233e-06, + "loss": 2.3103, + "step": 89 + }, + { + "epoch": 0.09244992295839753, + "grad_norm": 2.4134910106658936, + "learning_rate": 1.284246575342466e-06, + "loss": 2.2672, + "step": 90 + }, + { + "epoch": 0.09244992295839753, + "eval_loss": 2.3742051124572754, + "eval_runtime": 37.1258, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 90 + }, + { + "epoch": 0.09347714432460195, + "grad_norm": 2.302776336669922, + "learning_rate": 1.3013698630136986e-06, + "loss": 2.2798, + "step": 91 + }, + { + "epoch": 0.09450436569080636, + "grad_norm": 3.397819757461548, + "learning_rate": 1.3184931506849317e-06, + "loss": 2.2947, + "step": 92 + }, + { + "epoch": 0.09553158705701079, + "grad_norm": 2.6030948162078857, + "learning_rate": 1.3356164383561645e-06, + "loss": 2.2734, + "step": 93 + }, + { + "epoch": 0.09655880842321521, + "grad_norm": 2.9380528926849365, + "learning_rate": 1.3527397260273976e-06, + "loss": 2.2635, + "step": 94 + }, + { + "epoch": 0.09758602978941962, + "grad_norm": 2.9196345806121826, + "learning_rate": 1.3698630136986302e-06, + "loss": 2.3161, + "step": 95 + }, + { + "epoch": 0.09861325115562404, + "grad_norm": 2.2380659580230713, + "learning_rate": 1.386986301369863e-06, + "loss": 2.21, + "step": 96 + }, + { + "epoch": 0.09964047252182845, + "grad_norm": 2.730910539627075, + "learning_rate": 1.4041095890410961e-06, + "loss": 2.2562, + "step": 97 + }, + { + "epoch": 0.10066769388803287, + "grad_norm": 2.994748830795288, + "learning_rate": 1.421232876712329e-06, + "loss": 2.1597, + "step": 98 + }, + { + "epoch": 0.1016949152542373, + "grad_norm": 2.2228171825408936, + "learning_rate": 1.4383561643835616e-06, + "loss": 2.1835, + "step": 99 + }, + { + "epoch": 0.1027221366204417, + "grad_norm": 2.715527296066284, + "learning_rate": 1.4554794520547946e-06, + "loss": 2.2074, + "step": 100 + }, + { + "epoch": 0.1027221366204417, + "eval_loss": 2.3061234951019287, + "eval_runtime": 37.1393, + "eval_samples_per_second": 26.926, + "eval_steps_per_second": 1.131, + "step": 100 + }, + { + "epoch": 0.10374935798664613, + "grad_norm": 2.7054452896118164, + "learning_rate": 1.4726027397260275e-06, + "loss": 2.2089, + "step": 101 + }, + { + "epoch": 0.10477657935285054, + "grad_norm": 2.795694589614868, + "learning_rate": 1.4897260273972605e-06, + "loss": 2.1802, + "step": 102 + }, + { + "epoch": 0.10580380071905496, + "grad_norm": 2.634822368621826, + "learning_rate": 1.5068493150684932e-06, + "loss": 2.1428, + "step": 103 + }, + { + "epoch": 0.10683102208525937, + "grad_norm": 2.2418437004089355, + "learning_rate": 1.5239726027397262e-06, + "loss": 2.1706, + "step": 104 + }, + { + "epoch": 0.10785824345146379, + "grad_norm": 2.471231698989868, + "learning_rate": 1.541095890410959e-06, + "loss": 2.2248, + "step": 105 + }, + { + "epoch": 0.10888546481766821, + "grad_norm": 2.4154577255249023, + "learning_rate": 1.5582191780821921e-06, + "loss": 2.1873, + "step": 106 + }, + { + "epoch": 0.10991268618387262, + "grad_norm": 2.2823996543884277, + "learning_rate": 1.5753424657534248e-06, + "loss": 2.2237, + "step": 107 + }, + { + "epoch": 0.11093990755007704, + "grad_norm": 2.8695461750030518, + "learning_rate": 1.5924657534246576e-06, + "loss": 2.1966, + "step": 108 + }, + { + "epoch": 0.11196712891628145, + "grad_norm": 3.319650650024414, + "learning_rate": 1.6095890410958907e-06, + "loss": 2.066, + "step": 109 + }, + { + "epoch": 0.11299435028248588, + "grad_norm": 2.815196990966797, + "learning_rate": 1.6267123287671235e-06, + "loss": 2.1355, + "step": 110 + }, + { + "epoch": 0.11299435028248588, + "eval_loss": 2.250187635421753, + "eval_runtime": 37.1645, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 110 + }, + { + "epoch": 0.1140215716486903, + "grad_norm": 2.9195759296417236, + "learning_rate": 1.6438356164383561e-06, + "loss": 2.0713, + "step": 111 + }, + { + "epoch": 0.1150487930148947, + "grad_norm": 2.5600695610046387, + "learning_rate": 1.6609589041095892e-06, + "loss": 2.1573, + "step": 112 + }, + { + "epoch": 0.11607601438109913, + "grad_norm": 2.7566795349121094, + "learning_rate": 1.678082191780822e-06, + "loss": 2.0909, + "step": 113 + }, + { + "epoch": 0.11710323574730354, + "grad_norm": 2.5593106746673584, + "learning_rate": 1.695205479452055e-06, + "loss": 2.144, + "step": 114 + }, + { + "epoch": 0.11813045711350796, + "grad_norm": 2.594931125640869, + "learning_rate": 1.7123287671232877e-06, + "loss": 2.0606, + "step": 115 + }, + { + "epoch": 0.11915767847971238, + "grad_norm": 2.769916534423828, + "learning_rate": 1.7294520547945206e-06, + "loss": 2.1737, + "step": 116 + }, + { + "epoch": 0.12018489984591679, + "grad_norm": 2.523179054260254, + "learning_rate": 1.7465753424657536e-06, + "loss": 2.1247, + "step": 117 + }, + { + "epoch": 0.12121212121212122, + "grad_norm": 2.8153676986694336, + "learning_rate": 1.7636986301369865e-06, + "loss": 2.1257, + "step": 118 + }, + { + "epoch": 0.12223934257832562, + "grad_norm": 2.5389721393585205, + "learning_rate": 1.7808219178082193e-06, + "loss": 2.1049, + "step": 119 + }, + { + "epoch": 0.12326656394453005, + "grad_norm": 2.9753856658935547, + "learning_rate": 1.7979452054794521e-06, + "loss": 2.1871, + "step": 120 + }, + { + "epoch": 0.12326656394453005, + "eval_loss": 2.2089297771453857, + "eval_runtime": 37.897, + "eval_samples_per_second": 26.387, + "eval_steps_per_second": 1.108, + "step": 120 + }, + { + "epoch": 0.12429378531073447, + "grad_norm": 2.737180471420288, + "learning_rate": 1.8150684931506852e-06, + "loss": 2.1166, + "step": 121 + }, + { + "epoch": 0.12532100667693888, + "grad_norm": 2.783801794052124, + "learning_rate": 1.832191780821918e-06, + "loss": 2.0932, + "step": 122 + }, + { + "epoch": 0.1263482280431433, + "grad_norm": 2.9615402221679688, + "learning_rate": 1.8493150684931507e-06, + "loss": 2.0327, + "step": 123 + }, + { + "epoch": 0.12737544940934772, + "grad_norm": 3.053622007369995, + "learning_rate": 1.8664383561643837e-06, + "loss": 2.0665, + "step": 124 + }, + { + "epoch": 0.12840267077555212, + "grad_norm": 3.265355348587036, + "learning_rate": 1.8835616438356166e-06, + "loss": 2.0243, + "step": 125 + }, + { + "epoch": 0.12942989214175654, + "grad_norm": 2.4237143993377686, + "learning_rate": 1.9006849315068496e-06, + "loss": 2.0902, + "step": 126 + }, + { + "epoch": 0.13045711350796096, + "grad_norm": 2.926950693130493, + "learning_rate": 1.9178082191780823e-06, + "loss": 2.0057, + "step": 127 + }, + { + "epoch": 0.1314843348741654, + "grad_norm": 2.5317165851593018, + "learning_rate": 1.9349315068493153e-06, + "loss": 2.0719, + "step": 128 + }, + { + "epoch": 0.1325115562403698, + "grad_norm": 2.4999701976776123, + "learning_rate": 1.952054794520548e-06, + "loss": 2.0969, + "step": 129 + }, + { + "epoch": 0.1335387776065742, + "grad_norm": 2.7063920497894287, + "learning_rate": 1.969178082191781e-06, + "loss": 2.0902, + "step": 130 + }, + { + "epoch": 0.1335387776065742, + "eval_loss": 2.175224542617798, + "eval_runtime": 37.4519, + "eval_samples_per_second": 26.701, + "eval_steps_per_second": 1.121, + "step": 130 + }, + { + "epoch": 0.13456599897277863, + "grad_norm": 2.650860071182251, + "learning_rate": 1.9863013698630136e-06, + "loss": 2.0191, + "step": 131 + }, + { + "epoch": 0.13559322033898305, + "grad_norm": 2.79182505607605, + "learning_rate": 2.0034246575342467e-06, + "loss": 2.0856, + "step": 132 + }, + { + "epoch": 0.13662044170518747, + "grad_norm": 2.453162431716919, + "learning_rate": 2.0205479452054797e-06, + "loss": 2.0679, + "step": 133 + }, + { + "epoch": 0.1376476630713919, + "grad_norm": 2.9630799293518066, + "learning_rate": 2.037671232876713e-06, + "loss": 2.0038, + "step": 134 + }, + { + "epoch": 0.1386748844375963, + "grad_norm": 3.0807266235351562, + "learning_rate": 2.0547945205479454e-06, + "loss": 2.0367, + "step": 135 + }, + { + "epoch": 0.1397021058038007, + "grad_norm": 2.9712910652160645, + "learning_rate": 2.071917808219178e-06, + "loss": 2.0571, + "step": 136 + }, + { + "epoch": 0.14072932717000514, + "grad_norm": 3.0804853439331055, + "learning_rate": 2.089041095890411e-06, + "loss": 1.9725, + "step": 137 + }, + { + "epoch": 0.14175654853620956, + "grad_norm": 2.659608840942383, + "learning_rate": 2.106164383561644e-06, + "loss": 2.0572, + "step": 138 + }, + { + "epoch": 0.14278376990241398, + "grad_norm": 2.8093109130859375, + "learning_rate": 2.123287671232877e-06, + "loss": 2.0178, + "step": 139 + }, + { + "epoch": 0.14381099126861838, + "grad_norm": 2.8163881301879883, + "learning_rate": 2.14041095890411e-06, + "loss": 2.0085, + "step": 140 + }, + { + "epoch": 0.14381099126861838, + "eval_loss": 2.147092342376709, + "eval_runtime": 37.2453, + "eval_samples_per_second": 26.849, + "eval_steps_per_second": 1.128, + "step": 140 + }, + { + "epoch": 0.1448382126348228, + "grad_norm": 3.795241594314575, + "learning_rate": 2.1575342465753425e-06, + "loss": 2.0149, + "step": 141 + }, + { + "epoch": 0.14586543400102722, + "grad_norm": 3.3287174701690674, + "learning_rate": 2.1746575342465755e-06, + "loss": 1.9658, + "step": 142 + }, + { + "epoch": 0.14689265536723164, + "grad_norm": 2.803213596343994, + "learning_rate": 2.191780821917808e-06, + "loss": 2.0983, + "step": 143 + }, + { + "epoch": 0.14791987673343607, + "grad_norm": 3.2640199661254883, + "learning_rate": 2.2089041095890412e-06, + "loss": 1.9865, + "step": 144 + }, + { + "epoch": 0.14894709809964046, + "grad_norm": 2.95979380607605, + "learning_rate": 2.2260273972602743e-06, + "loss": 1.9669, + "step": 145 + }, + { + "epoch": 0.14997431946584489, + "grad_norm": 3.737175226211548, + "learning_rate": 2.243150684931507e-06, + "loss": 1.9691, + "step": 146 + }, + { + "epoch": 0.1510015408320493, + "grad_norm": 2.7248430252075195, + "learning_rate": 2.26027397260274e-06, + "loss": 1.9764, + "step": 147 + }, + { + "epoch": 0.15202876219825373, + "grad_norm": 2.929856061935425, + "learning_rate": 2.2773972602739726e-06, + "loss": 1.9967, + "step": 148 + }, + { + "epoch": 0.15305598356445815, + "grad_norm": 2.8998641967773438, + "learning_rate": 2.2945205479452057e-06, + "loss": 2.0842, + "step": 149 + }, + { + "epoch": 0.15408320493066255, + "grad_norm": 2.9672694206237793, + "learning_rate": 2.3116438356164387e-06, + "loss": 1.9915, + "step": 150 + }, + { + "epoch": 0.15408320493066255, + "eval_loss": 2.122715473175049, + "eval_runtime": 37.1938, + "eval_samples_per_second": 26.886, + "eval_steps_per_second": 1.129, + "step": 150 + }, + { + "epoch": 0.15511042629686697, + "grad_norm": 2.8529887199401855, + "learning_rate": 2.3287671232876713e-06, + "loss": 1.9022, + "step": 151 + }, + { + "epoch": 0.1561376476630714, + "grad_norm": 2.914121389389038, + "learning_rate": 2.3458904109589044e-06, + "loss": 1.956, + "step": 152 + }, + { + "epoch": 0.15716486902927582, + "grad_norm": 3.6286494731903076, + "learning_rate": 2.363013698630137e-06, + "loss": 2.0118, + "step": 153 + }, + { + "epoch": 0.15819209039548024, + "grad_norm": 2.789562702178955, + "learning_rate": 2.38013698630137e-06, + "loss": 1.9935, + "step": 154 + }, + { + "epoch": 0.15921931176168463, + "grad_norm": 3.367119550704956, + "learning_rate": 2.3972602739726027e-06, + "loss": 1.9519, + "step": 155 + }, + { + "epoch": 0.16024653312788906, + "grad_norm": 2.9345803260803223, + "learning_rate": 2.4143835616438358e-06, + "loss": 1.9017, + "step": 156 + }, + { + "epoch": 0.16127375449409348, + "grad_norm": 2.8674590587615967, + "learning_rate": 2.431506849315069e-06, + "loss": 1.9731, + "step": 157 + }, + { + "epoch": 0.1623009758602979, + "grad_norm": 3.144766092300415, + "learning_rate": 2.4486301369863015e-06, + "loss": 1.9811, + "step": 158 + }, + { + "epoch": 0.1633281972265023, + "grad_norm": 3.845395088195801, + "learning_rate": 2.4657534246575345e-06, + "loss": 1.9523, + "step": 159 + }, + { + "epoch": 0.16435541859270672, + "grad_norm": 2.776700973510742, + "learning_rate": 2.482876712328767e-06, + "loss": 1.9687, + "step": 160 + }, + { + "epoch": 0.16435541859270672, + "eval_loss": 2.0929579734802246, + "eval_runtime": 37.1917, + "eval_samples_per_second": 26.888, + "eval_steps_per_second": 1.129, + "step": 160 + }, + { + "epoch": 0.16538263995891114, + "grad_norm": 3.4260904788970947, + "learning_rate": 2.5e-06, + "loss": 1.9665, + "step": 161 + }, + { + "epoch": 0.16640986132511557, + "grad_norm": 3.1217620372772217, + "learning_rate": 2.5171232876712333e-06, + "loss": 1.9361, + "step": 162 + }, + { + "epoch": 0.16743708269132, + "grad_norm": 3.342068910598755, + "learning_rate": 2.534246575342466e-06, + "loss": 1.9084, + "step": 163 + }, + { + "epoch": 0.16846430405752438, + "grad_norm": 3.6485838890075684, + "learning_rate": 2.551369863013699e-06, + "loss": 1.9542, + "step": 164 + }, + { + "epoch": 0.1694915254237288, + "grad_norm": 3.1247780323028564, + "learning_rate": 2.568493150684932e-06, + "loss": 1.9717, + "step": 165 + }, + { + "epoch": 0.17051874678993323, + "grad_norm": 2.9856860637664795, + "learning_rate": 2.585616438356164e-06, + "loss": 1.9574, + "step": 166 + }, + { + "epoch": 0.17154596815613765, + "grad_norm": 3.337902069091797, + "learning_rate": 2.6027397260273973e-06, + "loss": 1.8884, + "step": 167 + }, + { + "epoch": 0.17257318952234207, + "grad_norm": 2.844534158706665, + "learning_rate": 2.6198630136986303e-06, + "loss": 1.9545, + "step": 168 + }, + { + "epoch": 0.17360041088854647, + "grad_norm": 3.75907564163208, + "learning_rate": 2.6369863013698634e-06, + "loss": 1.9907, + "step": 169 + }, + { + "epoch": 0.1746276322547509, + "grad_norm": 2.9321961402893066, + "learning_rate": 2.654109589041096e-06, + "loss": 1.9663, + "step": 170 + }, + { + "epoch": 0.1746276322547509, + "eval_loss": 2.0753843784332275, + "eval_runtime": 37.2389, + "eval_samples_per_second": 26.854, + "eval_steps_per_second": 1.128, + "step": 170 + }, + { + "epoch": 0.17565485362095531, + "grad_norm": 2.952523946762085, + "learning_rate": 2.671232876712329e-06, + "loss": 1.9065, + "step": 171 + }, + { + "epoch": 0.17668207498715974, + "grad_norm": 2.952463150024414, + "learning_rate": 2.688356164383562e-06, + "loss": 1.8994, + "step": 172 + }, + { + "epoch": 0.17770929635336416, + "grad_norm": 3.1352880001068115, + "learning_rate": 2.705479452054795e-06, + "loss": 1.8902, + "step": 173 + }, + { + "epoch": 0.17873651771956856, + "grad_norm": 3.5727322101593018, + "learning_rate": 2.7226027397260274e-06, + "loss": 1.8454, + "step": 174 + }, + { + "epoch": 0.17976373908577298, + "grad_norm": 3.0562493801116943, + "learning_rate": 2.7397260273972604e-06, + "loss": 1.984, + "step": 175 + }, + { + "epoch": 0.1807909604519774, + "grad_norm": 3.367924690246582, + "learning_rate": 2.7568493150684935e-06, + "loss": 1.9398, + "step": 176 + }, + { + "epoch": 0.18181818181818182, + "grad_norm": 3.1982340812683105, + "learning_rate": 2.773972602739726e-06, + "loss": 1.8566, + "step": 177 + }, + { + "epoch": 0.18284540318438625, + "grad_norm": 3.8962080478668213, + "learning_rate": 2.791095890410959e-06, + "loss": 1.9459, + "step": 178 + }, + { + "epoch": 0.18387262455059064, + "grad_norm": 3.3503036499023438, + "learning_rate": 2.8082191780821922e-06, + "loss": 1.8348, + "step": 179 + }, + { + "epoch": 0.18489984591679506, + "grad_norm": 3.276139497756958, + "learning_rate": 2.8253424657534253e-06, + "loss": 1.9849, + "step": 180 + }, + { + "epoch": 0.18489984591679506, + "eval_loss": 2.0605413913726807, + "eval_runtime": 37.2276, + "eval_samples_per_second": 26.862, + "eval_steps_per_second": 1.128, + "step": 180 + }, + { + "epoch": 0.1859270672829995, + "grad_norm": 3.3787975311279297, + "learning_rate": 2.842465753424658e-06, + "loss": 1.9517, + "step": 181 + }, + { + "epoch": 0.1869542886492039, + "grad_norm": 3.054290294647217, + "learning_rate": 2.8595890410958905e-06, + "loss": 1.9572, + "step": 182 + }, + { + "epoch": 0.18798151001540833, + "grad_norm": 3.0737054347991943, + "learning_rate": 2.876712328767123e-06, + "loss": 1.9674, + "step": 183 + }, + { + "epoch": 0.18900873138161273, + "grad_norm": 3.2833855152130127, + "learning_rate": 2.8938356164383562e-06, + "loss": 1.8185, + "step": 184 + }, + { + "epoch": 0.19003595274781715, + "grad_norm": 2.70452880859375, + "learning_rate": 2.9109589041095893e-06, + "loss": 1.8552, + "step": 185 + }, + { + "epoch": 0.19106317411402157, + "grad_norm": 3.2723655700683594, + "learning_rate": 2.9280821917808223e-06, + "loss": 1.8433, + "step": 186 + }, + { + "epoch": 0.192090395480226, + "grad_norm": 2.844536066055298, + "learning_rate": 2.945205479452055e-06, + "loss": 1.8488, + "step": 187 + }, + { + "epoch": 0.19311761684643042, + "grad_norm": 2.944545269012451, + "learning_rate": 2.962328767123288e-06, + "loss": 1.9131, + "step": 188 + }, + { + "epoch": 0.1941448382126348, + "grad_norm": 2.812204122543335, + "learning_rate": 2.979452054794521e-06, + "loss": 1.8977, + "step": 189 + }, + { + "epoch": 0.19517205957883924, + "grad_norm": 2.8999950885772705, + "learning_rate": 2.9965753424657533e-06, + "loss": 1.8477, + "step": 190 + }, + { + "epoch": 0.19517205957883924, + "eval_loss": 2.051161527633667, + "eval_runtime": 37.2363, + "eval_samples_per_second": 26.856, + "eval_steps_per_second": 1.128, + "step": 190 + }, + { + "epoch": 0.19619928094504366, + "grad_norm": 2.7814741134643555, + "learning_rate": 3.0136986301369864e-06, + "loss": 1.8989, + "step": 191 + }, + { + "epoch": 0.19722650231124808, + "grad_norm": 2.93346905708313, + "learning_rate": 3.0308219178082194e-06, + "loss": 1.8748, + "step": 192 + }, + { + "epoch": 0.1982537236774525, + "grad_norm": 3.00400447845459, + "learning_rate": 3.0479452054794525e-06, + "loss": 1.838, + "step": 193 + }, + { + "epoch": 0.1992809450436569, + "grad_norm": 2.949695110321045, + "learning_rate": 3.065068493150685e-06, + "loss": 1.8817, + "step": 194 + }, + { + "epoch": 0.20030816640986132, + "grad_norm": 2.8858301639556885, + "learning_rate": 3.082191780821918e-06, + "loss": 1.8333, + "step": 195 + }, + { + "epoch": 0.20133538777606574, + "grad_norm": 3.123706102371216, + "learning_rate": 3.099315068493151e-06, + "loss": 1.8533, + "step": 196 + }, + { + "epoch": 0.20236260914227017, + "grad_norm": 3.0130505561828613, + "learning_rate": 3.1164383561643843e-06, + "loss": 1.842, + "step": 197 + }, + { + "epoch": 0.2033898305084746, + "grad_norm": 2.8823280334472656, + "learning_rate": 3.1335616438356165e-06, + "loss": 1.8025, + "step": 198 + }, + { + "epoch": 0.20441705187467898, + "grad_norm": 2.920884132385254, + "learning_rate": 3.1506849315068495e-06, + "loss": 1.9488, + "step": 199 + }, + { + "epoch": 0.2054442732408834, + "grad_norm": 2.9152207374572754, + "learning_rate": 3.167808219178082e-06, + "loss": 1.8163, + "step": 200 + }, + { + "epoch": 0.2054442732408834, + "eval_loss": 2.033583641052246, + "eval_runtime": 37.127, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 200 + }, + { + "epoch": 0.20647149460708783, + "grad_norm": 2.904230833053589, + "learning_rate": 3.184931506849315e-06, + "loss": 1.8207, + "step": 201 + }, + { + "epoch": 0.20749871597329225, + "grad_norm": 3.128448486328125, + "learning_rate": 3.2020547945205483e-06, + "loss": 1.8051, + "step": 202 + }, + { + "epoch": 0.20852593733949668, + "grad_norm": 3.0049452781677246, + "learning_rate": 3.2191780821917813e-06, + "loss": 1.7823, + "step": 203 + }, + { + "epoch": 0.20955315870570107, + "grad_norm": 2.7612497806549072, + "learning_rate": 3.236301369863014e-06, + "loss": 1.7703, + "step": 204 + }, + { + "epoch": 0.2105803800719055, + "grad_norm": 2.9178879261016846, + "learning_rate": 3.253424657534247e-06, + "loss": 1.7427, + "step": 205 + }, + { + "epoch": 0.21160760143810992, + "grad_norm": 2.836967945098877, + "learning_rate": 3.2705479452054796e-06, + "loss": 1.8347, + "step": 206 + }, + { + "epoch": 0.21263482280431434, + "grad_norm": 2.8437204360961914, + "learning_rate": 3.2876712328767123e-06, + "loss": 1.8147, + "step": 207 + }, + { + "epoch": 0.21366204417051873, + "grad_norm": 3.0728225708007812, + "learning_rate": 3.3047945205479453e-06, + "loss": 1.7823, + "step": 208 + }, + { + "epoch": 0.21468926553672316, + "grad_norm": 3.357631206512451, + "learning_rate": 3.3219178082191784e-06, + "loss": 1.8013, + "step": 209 + }, + { + "epoch": 0.21571648690292758, + "grad_norm": 2.7996842861175537, + "learning_rate": 3.3390410958904114e-06, + "loss": 1.8279, + "step": 210 + }, + { + "epoch": 0.21571648690292758, + "eval_loss": 2.024500846862793, + "eval_runtime": 37.0122, + "eval_samples_per_second": 27.018, + "eval_steps_per_second": 1.135, + "step": 210 + }, + { + "epoch": 0.216743708269132, + "grad_norm": 3.3635380268096924, + "learning_rate": 3.356164383561644e-06, + "loss": 1.8509, + "step": 211 + }, + { + "epoch": 0.21777092963533642, + "grad_norm": 2.703932523727417, + "learning_rate": 3.373287671232877e-06, + "loss": 1.8984, + "step": 212 + }, + { + "epoch": 0.21879815100154082, + "grad_norm": 2.9341695308685303, + "learning_rate": 3.39041095890411e-06, + "loss": 1.8381, + "step": 213 + }, + { + "epoch": 0.21982537236774524, + "grad_norm": 3.072448253631592, + "learning_rate": 3.4075342465753424e-06, + "loss": 1.8779, + "step": 214 + }, + { + "epoch": 0.22085259373394966, + "grad_norm": 2.850555419921875, + "learning_rate": 3.4246575342465754e-06, + "loss": 1.7656, + "step": 215 + }, + { + "epoch": 0.2218798151001541, + "grad_norm": 3.0218873023986816, + "learning_rate": 3.4417808219178085e-06, + "loss": 1.829, + "step": 216 + }, + { + "epoch": 0.2229070364663585, + "grad_norm": 2.8475091457366943, + "learning_rate": 3.458904109589041e-06, + "loss": 1.8121, + "step": 217 + }, + { + "epoch": 0.2239342578325629, + "grad_norm": 2.8760111331939697, + "learning_rate": 3.476027397260274e-06, + "loss": 1.8281, + "step": 218 + }, + { + "epoch": 0.22496147919876733, + "grad_norm": 2.976670265197754, + "learning_rate": 3.4931506849315072e-06, + "loss": 1.8274, + "step": 219 + }, + { + "epoch": 0.22598870056497175, + "grad_norm": 2.912505865097046, + "learning_rate": 3.5102739726027403e-06, + "loss": 1.8316, + "step": 220 + }, + { + "epoch": 0.22598870056497175, + "eval_loss": 2.0203635692596436, + "eval_runtime": 37.0433, + "eval_samples_per_second": 26.995, + "eval_steps_per_second": 1.134, + "step": 220 + }, + { + "epoch": 0.22701592193117617, + "grad_norm": 3.1133334636688232, + "learning_rate": 3.527397260273973e-06, + "loss": 1.8277, + "step": 221 + }, + { + "epoch": 0.2280431432973806, + "grad_norm": 3.289062261581421, + "learning_rate": 3.5445205479452056e-06, + "loss": 1.7893, + "step": 222 + }, + { + "epoch": 0.229070364663585, + "grad_norm": 2.814558267593384, + "learning_rate": 3.5616438356164386e-06, + "loss": 1.8095, + "step": 223 + }, + { + "epoch": 0.2300975860297894, + "grad_norm": 3.48117733001709, + "learning_rate": 3.5787671232876712e-06, + "loss": 1.8701, + "step": 224 + }, + { + "epoch": 0.23112480739599384, + "grad_norm": 2.922144651412964, + "learning_rate": 3.5958904109589043e-06, + "loss": 1.8479, + "step": 225 + }, + { + "epoch": 0.23215202876219826, + "grad_norm": 3.552428960800171, + "learning_rate": 3.6130136986301373e-06, + "loss": 1.7822, + "step": 226 + }, + { + "epoch": 0.23317925012840268, + "grad_norm": 3.058274269104004, + "learning_rate": 3.6301369863013704e-06, + "loss": 1.8228, + "step": 227 + }, + { + "epoch": 0.23420647149460708, + "grad_norm": 3.365823268890381, + "learning_rate": 3.647260273972603e-06, + "loss": 1.8415, + "step": 228 + }, + { + "epoch": 0.2352336928608115, + "grad_norm": 2.87333607673645, + "learning_rate": 3.664383561643836e-06, + "loss": 1.8474, + "step": 229 + }, + { + "epoch": 0.23626091422701592, + "grad_norm": 3.185807228088379, + "learning_rate": 3.6815068493150687e-06, + "loss": 1.7854, + "step": 230 + }, + { + "epoch": 0.23626091422701592, + "eval_loss": 2.0132367610931396, + "eval_runtime": 37.3246, + "eval_samples_per_second": 26.792, + "eval_steps_per_second": 1.125, + "step": 230 + }, + { + "epoch": 0.23728813559322035, + "grad_norm": 3.02864933013916, + "learning_rate": 3.6986301369863014e-06, + "loss": 1.8081, + "step": 231 + }, + { + "epoch": 0.23831535695942477, + "grad_norm": 2.892451286315918, + "learning_rate": 3.7157534246575344e-06, + "loss": 1.7619, + "step": 232 + }, + { + "epoch": 0.23934257832562916, + "grad_norm": 3.24947452545166, + "learning_rate": 3.7328767123287675e-06, + "loss": 1.8729, + "step": 233 + }, + { + "epoch": 0.24036979969183359, + "grad_norm": 3.13212513923645, + "learning_rate": 3.7500000000000005e-06, + "loss": 1.7849, + "step": 234 + }, + { + "epoch": 0.241397021058038, + "grad_norm": 3.004814863204956, + "learning_rate": 3.767123287671233e-06, + "loss": 1.8217, + "step": 235 + }, + { + "epoch": 0.24242424242424243, + "grad_norm": 2.845555067062378, + "learning_rate": 3.784246575342466e-06, + "loss": 1.8438, + "step": 236 + }, + { + "epoch": 0.24345146379044685, + "grad_norm": 3.1796751022338867, + "learning_rate": 3.8013698630136993e-06, + "loss": 1.7807, + "step": 237 + }, + { + "epoch": 0.24447868515665125, + "grad_norm": 2.7663002014160156, + "learning_rate": 3.818493150684932e-06, + "loss": 1.7981, + "step": 238 + }, + { + "epoch": 0.24550590652285567, + "grad_norm": 3.0547749996185303, + "learning_rate": 3.8356164383561645e-06, + "loss": 1.7543, + "step": 239 + }, + { + "epoch": 0.2465331278890601, + "grad_norm": 2.6994550228118896, + "learning_rate": 3.852739726027397e-06, + "loss": 1.8612, + "step": 240 + }, + { + "epoch": 0.2465331278890601, + "eval_loss": 2.0076475143432617, + "eval_runtime": 37.3223, + "eval_samples_per_second": 26.794, + "eval_steps_per_second": 1.125, + "step": 240 + }, + { + "epoch": 0.24756034925526452, + "grad_norm": 2.769717216491699, + "learning_rate": 3.869863013698631e-06, + "loss": 1.8352, + "step": 241 + }, + { + "epoch": 0.24858757062146894, + "grad_norm": 2.7526769638061523, + "learning_rate": 3.886986301369863e-06, + "loss": 1.7775, + "step": 242 + }, + { + "epoch": 0.24961479198767333, + "grad_norm": 3.0213067531585693, + "learning_rate": 3.904109589041096e-06, + "loss": 1.7901, + "step": 243 + }, + { + "epoch": 0.25064201335387776, + "grad_norm": 2.8562498092651367, + "learning_rate": 3.921232876712329e-06, + "loss": 1.7483, + "step": 244 + }, + { + "epoch": 0.2516692347200822, + "grad_norm": 2.812774896621704, + "learning_rate": 3.938356164383562e-06, + "loss": 1.7643, + "step": 245 + }, + { + "epoch": 0.2526964560862866, + "grad_norm": 3.2509777545928955, + "learning_rate": 3.9554794520547955e-06, + "loss": 1.729, + "step": 246 + }, + { + "epoch": 0.253723677452491, + "grad_norm": 3.6394171714782715, + "learning_rate": 3.972602739726027e-06, + "loss": 1.8253, + "step": 247 + }, + { + "epoch": 0.25475089881869545, + "grad_norm": 3.1234781742095947, + "learning_rate": 3.989726027397261e-06, + "loss": 1.7457, + "step": 248 + }, + { + "epoch": 0.25577812018489987, + "grad_norm": 2.8587522506713867, + "learning_rate": 4.006849315068493e-06, + "loss": 1.804, + "step": 249 + }, + { + "epoch": 0.25680534155110424, + "grad_norm": 2.9718286991119385, + "learning_rate": 4.023972602739726e-06, + "loss": 1.8051, + "step": 250 + }, + { + "epoch": 0.25680534155110424, + "eval_loss": 2.0024874210357666, + "eval_runtime": 37.2884, + "eval_samples_per_second": 26.818, + "eval_steps_per_second": 1.126, + "step": 250 + }, + { + "epoch": 0.25783256291730866, + "grad_norm": 3.007124423980713, + "learning_rate": 4.0410958904109595e-06, + "loss": 1.7946, + "step": 251 + }, + { + "epoch": 0.2588597842835131, + "grad_norm": 3.1715426445007324, + "learning_rate": 4.058219178082192e-06, + "loss": 1.7811, + "step": 252 + }, + { + "epoch": 0.2598870056497175, + "grad_norm": 2.814269542694092, + "learning_rate": 4.075342465753426e-06, + "loss": 1.8024, + "step": 253 + }, + { + "epoch": 0.26091422701592193, + "grad_norm": 2.841212034225464, + "learning_rate": 4.092465753424658e-06, + "loss": 1.7713, + "step": 254 + }, + { + "epoch": 0.26194144838212635, + "grad_norm": 3.0878522396087646, + "learning_rate": 4.109589041095891e-06, + "loss": 1.7579, + "step": 255 + }, + { + "epoch": 0.2629686697483308, + "grad_norm": 2.946709394454956, + "learning_rate": 4.1267123287671235e-06, + "loss": 1.823, + "step": 256 + }, + { + "epoch": 0.2639958911145352, + "grad_norm": 2.8136606216430664, + "learning_rate": 4.143835616438356e-06, + "loss": 1.7731, + "step": 257 + }, + { + "epoch": 0.2650231124807396, + "grad_norm": 2.745762825012207, + "learning_rate": 4.16095890410959e-06, + "loss": 1.8445, + "step": 258 + }, + { + "epoch": 0.26605033384694404, + "grad_norm": 3.0986413955688477, + "learning_rate": 4.178082191780822e-06, + "loss": 1.8231, + "step": 259 + }, + { + "epoch": 0.2670775552131484, + "grad_norm": 2.6990394592285156, + "learning_rate": 4.195205479452055e-06, + "loss": 1.7635, + "step": 260 + }, + { + "epoch": 0.2670775552131484, + "eval_loss": 1.9839800596237183, + "eval_runtime": 37.1612, + "eval_samples_per_second": 26.91, + "eval_steps_per_second": 1.13, + "step": 260 + }, + { + "epoch": 0.26810477657935283, + "grad_norm": 3.281421184539795, + "learning_rate": 4.212328767123288e-06, + "loss": 1.7587, + "step": 261 + }, + { + "epoch": 0.26913199794555726, + "grad_norm": 3.1233623027801514, + "learning_rate": 4.229452054794521e-06, + "loss": 1.7673, + "step": 262 + }, + { + "epoch": 0.2701592193117617, + "grad_norm": 3.3131625652313232, + "learning_rate": 4.246575342465754e-06, + "loss": 1.7755, + "step": 263 + }, + { + "epoch": 0.2711864406779661, + "grad_norm": 3.3002350330352783, + "learning_rate": 4.263698630136986e-06, + "loss": 1.7575, + "step": 264 + }, + { + "epoch": 0.2722136620441705, + "grad_norm": 2.965397357940674, + "learning_rate": 4.28082191780822e-06, + "loss": 1.7162, + "step": 265 + }, + { + "epoch": 0.27324088341037495, + "grad_norm": 2.9904608726501465, + "learning_rate": 4.297945205479452e-06, + "loss": 1.8794, + "step": 266 + }, + { + "epoch": 0.27426810477657937, + "grad_norm": 2.707996129989624, + "learning_rate": 4.315068493150685e-06, + "loss": 1.7646, + "step": 267 + }, + { + "epoch": 0.2752953261427838, + "grad_norm": 2.8878190517425537, + "learning_rate": 4.3321917808219185e-06, + "loss": 1.7559, + "step": 268 + }, + { + "epoch": 0.2763225475089882, + "grad_norm": 3.1676602363586426, + "learning_rate": 4.349315068493151e-06, + "loss": 1.7806, + "step": 269 + }, + { + "epoch": 0.2773497688751926, + "grad_norm": 2.9179935455322266, + "learning_rate": 4.3664383561643846e-06, + "loss": 1.7493, + "step": 270 + }, + { + "epoch": 0.2773497688751926, + "eval_loss": 1.9736579656600952, + "eval_runtime": 37.0983, + "eval_samples_per_second": 26.955, + "eval_steps_per_second": 1.132, + "step": 270 + }, + { + "epoch": 0.278376990241397, + "grad_norm": 3.0081794261932373, + "learning_rate": 4.383561643835616e-06, + "loss": 1.7405, + "step": 271 + }, + { + "epoch": 0.2794042116076014, + "grad_norm": 3.240462064743042, + "learning_rate": 4.40068493150685e-06, + "loss": 1.8326, + "step": 272 + }, + { + "epoch": 0.28043143297380585, + "grad_norm": 3.324059247970581, + "learning_rate": 4.4178082191780825e-06, + "loss": 1.7819, + "step": 273 + }, + { + "epoch": 0.2814586543400103, + "grad_norm": 2.8438684940338135, + "learning_rate": 4.434931506849315e-06, + "loss": 1.7435, + "step": 274 + }, + { + "epoch": 0.2824858757062147, + "grad_norm": 2.9525997638702393, + "learning_rate": 4.4520547945205486e-06, + "loss": 1.7123, + "step": 275 + }, + { + "epoch": 0.2835130970724191, + "grad_norm": 2.972714900970459, + "learning_rate": 4.469178082191781e-06, + "loss": 1.7693, + "step": 276 + }, + { + "epoch": 0.28454031843862354, + "grad_norm": 3.4248580932617188, + "learning_rate": 4.486301369863014e-06, + "loss": 1.7502, + "step": 277 + }, + { + "epoch": 0.28556753980482796, + "grad_norm": 2.9928481578826904, + "learning_rate": 4.503424657534247e-06, + "loss": 1.8064, + "step": 278 + }, + { + "epoch": 0.28659476117103233, + "grad_norm": 2.944093942642212, + "learning_rate": 4.52054794520548e-06, + "loss": 1.7075, + "step": 279 + }, + { + "epoch": 0.28762198253723675, + "grad_norm": 3.2028305530548096, + "learning_rate": 4.537671232876713e-06, + "loss": 1.7366, + "step": 280 + }, + { + "epoch": 0.28762198253723675, + "eval_loss": 1.9768972396850586, + "eval_runtime": 37.1078, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 280 + }, + { + "epoch": 0.2886492039034412, + "grad_norm": 2.8642148971557617, + "learning_rate": 4.554794520547945e-06, + "loss": 1.7909, + "step": 281 + }, + { + "epoch": 0.2896764252696456, + "grad_norm": 3.2716004848480225, + "learning_rate": 4.571917808219179e-06, + "loss": 1.7906, + "step": 282 + }, + { + "epoch": 0.29070364663585, + "grad_norm": 2.9601588249206543, + "learning_rate": 4.589041095890411e-06, + "loss": 1.7731, + "step": 283 + }, + { + "epoch": 0.29173086800205444, + "grad_norm": 3.1850085258483887, + "learning_rate": 4.606164383561644e-06, + "loss": 1.7335, + "step": 284 + }, + { + "epoch": 0.29275808936825887, + "grad_norm": 3.128908634185791, + "learning_rate": 4.6232876712328774e-06, + "loss": 1.748, + "step": 285 + }, + { + "epoch": 0.2937853107344633, + "grad_norm": 3.063014507293701, + "learning_rate": 4.64041095890411e-06, + "loss": 1.7632, + "step": 286 + }, + { + "epoch": 0.2948125321006677, + "grad_norm": 3.467817783355713, + "learning_rate": 4.657534246575343e-06, + "loss": 1.7295, + "step": 287 + }, + { + "epoch": 0.29583975346687214, + "grad_norm": 2.7431066036224365, + "learning_rate": 4.674657534246575e-06, + "loss": 1.7728, + "step": 288 + }, + { + "epoch": 0.2968669748330765, + "grad_norm": 3.271421432495117, + "learning_rate": 4.691780821917809e-06, + "loss": 1.7403, + "step": 289 + }, + { + "epoch": 0.2978941961992809, + "grad_norm": 2.908351421356201, + "learning_rate": 4.7089041095890414e-06, + "loss": 1.7983, + "step": 290 + }, + { + "epoch": 0.2978941961992809, + "eval_loss": 1.955810308456421, + "eval_runtime": 37.0914, + "eval_samples_per_second": 26.96, + "eval_steps_per_second": 1.132, + "step": 290 + }, + { + "epoch": 0.29892141756548535, + "grad_norm": 3.046114683151245, + "learning_rate": 4.726027397260274e-06, + "loss": 1.7612, + "step": 291 + }, + { + "epoch": 0.29994863893168977, + "grad_norm": 3.275486707687378, + "learning_rate": 4.7431506849315075e-06, + "loss": 1.7548, + "step": 292 + }, + { + "epoch": 0.3009758602978942, + "grad_norm": 2.819552183151245, + "learning_rate": 4.76027397260274e-06, + "loss": 1.7645, + "step": 293 + }, + { + "epoch": 0.3020030816640986, + "grad_norm": 3.234227418899536, + "learning_rate": 4.777397260273973e-06, + "loss": 1.715, + "step": 294 + }, + { + "epoch": 0.30303030303030304, + "grad_norm": 2.988713502883911, + "learning_rate": 4.7945205479452054e-06, + "loss": 1.7343, + "step": 295 + }, + { + "epoch": 0.30405752439650746, + "grad_norm": 3.195387601852417, + "learning_rate": 4.811643835616439e-06, + "loss": 1.7421, + "step": 296 + }, + { + "epoch": 0.3050847457627119, + "grad_norm": 3.617765426635742, + "learning_rate": 4.8287671232876716e-06, + "loss": 1.6661, + "step": 297 + }, + { + "epoch": 0.3061119671289163, + "grad_norm": 2.741631269454956, + "learning_rate": 4.845890410958904e-06, + "loss": 1.7188, + "step": 298 + }, + { + "epoch": 0.3071391884951207, + "grad_norm": 3.3230361938476562, + "learning_rate": 4.863013698630138e-06, + "loss": 1.721, + "step": 299 + }, + { + "epoch": 0.3081664098613251, + "grad_norm": 2.696352005004883, + "learning_rate": 4.88013698630137e-06, + "loss": 1.7137, + "step": 300 + }, + { + "epoch": 0.3081664098613251, + "eval_loss": 1.9677459001541138, + "eval_runtime": 37.1101, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 300 + }, + { + "epoch": 0.3091936312275295, + "grad_norm": 3.162966251373291, + "learning_rate": 4.897260273972603e-06, + "loss": 1.6956, + "step": 301 + }, + { + "epoch": 0.31022085259373394, + "grad_norm": 2.6055824756622314, + "learning_rate": 4.914383561643836e-06, + "loss": 1.8174, + "step": 302 + }, + { + "epoch": 0.31124807395993837, + "grad_norm": 2.816882610321045, + "learning_rate": 4.931506849315069e-06, + "loss": 1.7444, + "step": 303 + }, + { + "epoch": 0.3122752953261428, + "grad_norm": 2.9913320541381836, + "learning_rate": 4.948630136986302e-06, + "loss": 1.7744, + "step": 304 + }, + { + "epoch": 0.3133025166923472, + "grad_norm": 2.992614507675171, + "learning_rate": 4.965753424657534e-06, + "loss": 1.8155, + "step": 305 + }, + { + "epoch": 0.31432973805855163, + "grad_norm": 3.2482688426971436, + "learning_rate": 4.982876712328768e-06, + "loss": 1.6903, + "step": 306 + }, + { + "epoch": 0.31535695942475606, + "grad_norm": 2.7701141834259033, + "learning_rate": 5e-06, + "loss": 1.7228, + "step": 307 + }, + { + "epoch": 0.3163841807909605, + "grad_norm": 3.179245948791504, + "learning_rate": 4.998096688237534e-06, + "loss": 1.6644, + "step": 308 + }, + { + "epoch": 0.31741140215716485, + "grad_norm": 2.845163106918335, + "learning_rate": 4.996193376475067e-06, + "loss": 1.772, + "step": 309 + }, + { + "epoch": 0.31843862352336927, + "grad_norm": 3.1509721279144287, + "learning_rate": 4.9942900647126005e-06, + "loss": 1.6662, + "step": 310 + }, + { + "epoch": 0.31843862352336927, + "eval_loss": 1.9684854745864868, + "eval_runtime": 37.0507, + "eval_samples_per_second": 26.99, + "eval_steps_per_second": 1.134, + "step": 310 + }, + { + "epoch": 0.3194658448895737, + "grad_norm": 2.8594536781311035, + "learning_rate": 4.992386752950134e-06, + "loss": 1.6587, + "step": 311 + }, + { + "epoch": 0.3204930662557781, + "grad_norm": 3.358464002609253, + "learning_rate": 4.990483441187667e-06, + "loss": 1.7021, + "step": 312 + }, + { + "epoch": 0.32152028762198254, + "grad_norm": 2.6954872608184814, + "learning_rate": 4.9885801294252e-06, + "loss": 1.7301, + "step": 313 + }, + { + "epoch": 0.32254750898818696, + "grad_norm": 2.9256083965301514, + "learning_rate": 4.986676817662733e-06, + "loss": 1.7316, + "step": 314 + }, + { + "epoch": 0.3235747303543914, + "grad_norm": 2.840461254119873, + "learning_rate": 4.9847735059002665e-06, + "loss": 1.6848, + "step": 315 + }, + { + "epoch": 0.3246019517205958, + "grad_norm": 2.9138107299804688, + "learning_rate": 4.9828701941378e-06, + "loss": 1.765, + "step": 316 + }, + { + "epoch": 0.3256291730868002, + "grad_norm": 2.858368396759033, + "learning_rate": 4.980966882375333e-06, + "loss": 1.7078, + "step": 317 + }, + { + "epoch": 0.3266563944530046, + "grad_norm": 2.905238628387451, + "learning_rate": 4.979063570612867e-06, + "loss": 1.6765, + "step": 318 + }, + { + "epoch": 0.327683615819209, + "grad_norm": 3.117845296859741, + "learning_rate": 4.9771602588504e-06, + "loss": 1.6419, + "step": 319 + }, + { + "epoch": 0.32871083718541344, + "grad_norm": 2.5665884017944336, + "learning_rate": 4.975256947087933e-06, + "loss": 1.7424, + "step": 320 + }, + { + "epoch": 0.32871083718541344, + "eval_loss": 1.9528932571411133, + "eval_runtime": 37.2179, + "eval_samples_per_second": 26.869, + "eval_steps_per_second": 1.128, + "step": 320 + }, + { + "epoch": 0.32973805855161786, + "grad_norm": 2.499392032623291, + "learning_rate": 4.973353635325467e-06, + "loss": 1.6544, + "step": 321 + }, + { + "epoch": 0.3307652799178223, + "grad_norm": 2.8056774139404297, + "learning_rate": 4.971450323563e-06, + "loss": 1.7085, + "step": 322 + }, + { + "epoch": 0.3317925012840267, + "grad_norm": 2.850602388381958, + "learning_rate": 4.969547011800533e-06, + "loss": 1.6489, + "step": 323 + }, + { + "epoch": 0.33281972265023113, + "grad_norm": 2.783855438232422, + "learning_rate": 4.967643700038067e-06, + "loss": 1.7153, + "step": 324 + }, + { + "epoch": 0.33384694401643555, + "grad_norm": 2.7178030014038086, + "learning_rate": 4.9657403882756e-06, + "loss": 1.7104, + "step": 325 + }, + { + "epoch": 0.33487416538264, + "grad_norm": 2.7497682571411133, + "learning_rate": 4.9638370765131335e-06, + "loss": 1.6778, + "step": 326 + }, + { + "epoch": 0.3359013867488444, + "grad_norm": 2.688120126724243, + "learning_rate": 4.961933764750667e-06, + "loss": 1.6673, + "step": 327 + }, + { + "epoch": 0.33692860811504877, + "grad_norm": 2.661578416824341, + "learning_rate": 4.9600304529882e-06, + "loss": 1.7353, + "step": 328 + }, + { + "epoch": 0.3379558294812532, + "grad_norm": 2.759965419769287, + "learning_rate": 4.9581271412257335e-06, + "loss": 1.7506, + "step": 329 + }, + { + "epoch": 0.3389830508474576, + "grad_norm": 2.972900390625, + "learning_rate": 4.956223829463266e-06, + "loss": 1.7114, + "step": 330 + }, + { + "epoch": 0.3389830508474576, + "eval_loss": 1.9369672536849976, + "eval_runtime": 37.2876, + "eval_samples_per_second": 26.819, + "eval_steps_per_second": 1.126, + "step": 330 + }, + { + "epoch": 0.34001027221366203, + "grad_norm": 2.7194814682006836, + "learning_rate": 4.954320517700799e-06, + "loss": 1.621, + "step": 331 + }, + { + "epoch": 0.34103749357986646, + "grad_norm": 2.822052001953125, + "learning_rate": 4.952417205938333e-06, + "loss": 1.7215, + "step": 332 + }, + { + "epoch": 0.3420647149460709, + "grad_norm": 2.778348207473755, + "learning_rate": 4.950513894175866e-06, + "loss": 1.7049, + "step": 333 + }, + { + "epoch": 0.3430919363122753, + "grad_norm": 2.8513247966766357, + "learning_rate": 4.9486105824133995e-06, + "loss": 1.684, + "step": 334 + }, + { + "epoch": 0.3441191576784797, + "grad_norm": 2.5596683025360107, + "learning_rate": 4.946707270650933e-06, + "loss": 1.6992, + "step": 335 + }, + { + "epoch": 0.34514637904468415, + "grad_norm": 2.850728750228882, + "learning_rate": 4.944803958888466e-06, + "loss": 1.7095, + "step": 336 + }, + { + "epoch": 0.34617360041088857, + "grad_norm": 2.8497488498687744, + "learning_rate": 4.942900647126e-06, + "loss": 1.7361, + "step": 337 + }, + { + "epoch": 0.34720082177709294, + "grad_norm": 2.9133079051971436, + "learning_rate": 4.940997335363533e-06, + "loss": 1.6783, + "step": 338 + }, + { + "epoch": 0.34822804314329736, + "grad_norm": 2.7611589431762695, + "learning_rate": 4.939094023601066e-06, + "loss": 1.6407, + "step": 339 + }, + { + "epoch": 0.3492552645095018, + "grad_norm": 2.7190754413604736, + "learning_rate": 4.9371907118386e-06, + "loss": 1.6886, + "step": 340 + }, + { + "epoch": 0.3492552645095018, + "eval_loss": 1.9296809434890747, + "eval_runtime": 37.4339, + "eval_samples_per_second": 26.714, + "eval_steps_per_second": 1.122, + "step": 340 + }, + { + "epoch": 0.3502824858757062, + "grad_norm": 2.863070487976074, + "learning_rate": 4.935287400076133e-06, + "loss": 1.6912, + "step": 341 + }, + { + "epoch": 0.35130970724191063, + "grad_norm": 2.5978336334228516, + "learning_rate": 4.933384088313666e-06, + "loss": 1.7107, + "step": 342 + }, + { + "epoch": 0.35233692860811505, + "grad_norm": 2.8323776721954346, + "learning_rate": 4.9314807765512e-06, + "loss": 1.6723, + "step": 343 + }, + { + "epoch": 0.3533641499743195, + "grad_norm": 2.7128350734710693, + "learning_rate": 4.929577464788733e-06, + "loss": 1.5993, + "step": 344 + }, + { + "epoch": 0.3543913713405239, + "grad_norm": 2.618936061859131, + "learning_rate": 4.9276741530262665e-06, + "loss": 1.6769, + "step": 345 + }, + { + "epoch": 0.3554185927067283, + "grad_norm": 2.9575355052948, + "learning_rate": 4.9257708412638e-06, + "loss": 1.6743, + "step": 346 + }, + { + "epoch": 0.35644581407293274, + "grad_norm": 2.756464719772339, + "learning_rate": 4.923867529501332e-06, + "loss": 1.6858, + "step": 347 + }, + { + "epoch": 0.3574730354391371, + "grad_norm": 2.6581053733825684, + "learning_rate": 4.921964217738866e-06, + "loss": 1.6515, + "step": 348 + }, + { + "epoch": 0.35850025680534153, + "grad_norm": 2.890632390975952, + "learning_rate": 4.920060905976399e-06, + "loss": 1.7378, + "step": 349 + }, + { + "epoch": 0.35952747817154596, + "grad_norm": 2.7895255088806152, + "learning_rate": 4.9181575942139325e-06, + "loss": 1.7131, + "step": 350 + }, + { + "epoch": 0.35952747817154596, + "eval_loss": 1.9226924180984497, + "eval_runtime": 37.5186, + "eval_samples_per_second": 26.653, + "eval_steps_per_second": 1.119, + "step": 350 + }, + { + "epoch": 0.3605546995377504, + "grad_norm": 2.7698557376861572, + "learning_rate": 4.916254282451466e-06, + "loss": 1.71, + "step": 351 + }, + { + "epoch": 0.3615819209039548, + "grad_norm": 2.956592321395874, + "learning_rate": 4.914350970688999e-06, + "loss": 1.7487, + "step": 352 + }, + { + "epoch": 0.3626091422701592, + "grad_norm": 2.6062943935394287, + "learning_rate": 4.9124476589265325e-06, + "loss": 1.6486, + "step": 353 + }, + { + "epoch": 0.36363636363636365, + "grad_norm": 2.588418483734131, + "learning_rate": 4.910544347164066e-06, + "loss": 1.6468, + "step": 354 + }, + { + "epoch": 0.36466358500256807, + "grad_norm": 2.4665889739990234, + "learning_rate": 4.908641035401599e-06, + "loss": 1.7264, + "step": 355 + }, + { + "epoch": 0.3656908063687725, + "grad_norm": 2.701957941055298, + "learning_rate": 4.906737723639133e-06, + "loss": 1.7914, + "step": 356 + }, + { + "epoch": 0.3667180277349769, + "grad_norm": 2.68169903755188, + "learning_rate": 4.904834411876665e-06, + "loss": 1.6609, + "step": 357 + }, + { + "epoch": 0.3677452491011813, + "grad_norm": 2.633120536804199, + "learning_rate": 4.902931100114199e-06, + "loss": 1.6998, + "step": 358 + }, + { + "epoch": 0.3687724704673857, + "grad_norm": 2.549065113067627, + "learning_rate": 4.901027788351733e-06, + "loss": 1.6227, + "step": 359 + }, + { + "epoch": 0.3697996918335901, + "grad_norm": 2.6537365913391113, + "learning_rate": 4.899124476589266e-06, + "loss": 1.6678, + "step": 360 + }, + { + "epoch": 0.3697996918335901, + "eval_loss": 1.9240167140960693, + "eval_runtime": 37.3212, + "eval_samples_per_second": 26.794, + "eval_steps_per_second": 1.125, + "step": 360 + }, + { + "epoch": 0.37082691319979455, + "grad_norm": 2.8710854053497314, + "learning_rate": 4.8972211648267994e-06, + "loss": 1.6377, + "step": 361 + }, + { + "epoch": 0.371854134565999, + "grad_norm": 2.8168065547943115, + "learning_rate": 4.895317853064333e-06, + "loss": 1.6888, + "step": 362 + }, + { + "epoch": 0.3728813559322034, + "grad_norm": 2.7674975395202637, + "learning_rate": 4.893414541301866e-06, + "loss": 1.6532, + "step": 363 + }, + { + "epoch": 0.3739085772984078, + "grad_norm": 2.5604195594787598, + "learning_rate": 4.891511229539399e-06, + "loss": 1.6513, + "step": 364 + }, + { + "epoch": 0.37493579866461224, + "grad_norm": 2.5669636726379395, + "learning_rate": 4.889607917776932e-06, + "loss": 1.6832, + "step": 365 + }, + { + "epoch": 0.37596302003081666, + "grad_norm": 2.650104522705078, + "learning_rate": 4.887704606014465e-06, + "loss": 1.6836, + "step": 366 + }, + { + "epoch": 0.37699024139702103, + "grad_norm": 2.702894926071167, + "learning_rate": 4.885801294251999e-06, + "loss": 1.6916, + "step": 367 + }, + { + "epoch": 0.37801746276322545, + "grad_norm": 2.991715431213379, + "learning_rate": 4.883897982489532e-06, + "loss": 1.6146, + "step": 368 + }, + { + "epoch": 0.3790446841294299, + "grad_norm": 2.725093126296997, + "learning_rate": 4.8819946707270655e-06, + "loss": 1.7124, + "step": 369 + }, + { + "epoch": 0.3800719054956343, + "grad_norm": 2.8419744968414307, + "learning_rate": 4.880091358964599e-06, + "loss": 1.7002, + "step": 370 + }, + { + "epoch": 0.3800719054956343, + "eval_loss": 1.9024126529693604, + "eval_runtime": 37.231, + "eval_samples_per_second": 26.859, + "eval_steps_per_second": 1.128, + "step": 370 + }, + { + "epoch": 0.3810991268618387, + "grad_norm": 2.3915929794311523, + "learning_rate": 4.878188047202132e-06, + "loss": 1.6096, + "step": 371 + }, + { + "epoch": 0.38212634822804314, + "grad_norm": 2.577662944793701, + "learning_rate": 4.876284735439666e-06, + "loss": 1.772, + "step": 372 + }, + { + "epoch": 0.38315356959424757, + "grad_norm": 2.6686151027679443, + "learning_rate": 4.874381423677199e-06, + "loss": 1.7375, + "step": 373 + }, + { + "epoch": 0.384180790960452, + "grad_norm": 2.4534966945648193, + "learning_rate": 4.8724781119147315e-06, + "loss": 1.6603, + "step": 374 + }, + { + "epoch": 0.3852080123266564, + "grad_norm": 2.7308576107025146, + "learning_rate": 4.870574800152265e-06, + "loss": 1.6999, + "step": 375 + }, + { + "epoch": 0.38623523369286084, + "grad_norm": 2.551203727722168, + "learning_rate": 4.868671488389798e-06, + "loss": 1.6193, + "step": 376 + }, + { + "epoch": 0.3872624550590652, + "grad_norm": 2.78924822807312, + "learning_rate": 4.8667681766273316e-06, + "loss": 1.6359, + "step": 377 + }, + { + "epoch": 0.3882896764252696, + "grad_norm": 2.7972333431243896, + "learning_rate": 4.864864864864866e-06, + "loss": 1.6057, + "step": 378 + }, + { + "epoch": 0.38931689779147405, + "grad_norm": 2.721069097518921, + "learning_rate": 4.862961553102399e-06, + "loss": 1.6661, + "step": 379 + }, + { + "epoch": 0.39034411915767847, + "grad_norm": 2.817729949951172, + "learning_rate": 4.8610582413399325e-06, + "loss": 1.7053, + "step": 380 + }, + { + "epoch": 0.39034411915767847, + "eval_loss": 1.9177663326263428, + "eval_runtime": 37.2158, + "eval_samples_per_second": 26.87, + "eval_steps_per_second": 1.129, + "step": 380 + }, + { + "epoch": 0.3913713405238829, + "grad_norm": 2.800873279571533, + "learning_rate": 4.859154929577465e-06, + "loss": 1.6584, + "step": 381 + }, + { + "epoch": 0.3923985618900873, + "grad_norm": 2.6885769367218018, + "learning_rate": 4.857251617814998e-06, + "loss": 1.661, + "step": 382 + }, + { + "epoch": 0.39342578325629174, + "grad_norm": 2.56797456741333, + "learning_rate": 4.855348306052532e-06, + "loss": 1.7058, + "step": 383 + }, + { + "epoch": 0.39445300462249616, + "grad_norm": 2.5235650539398193, + "learning_rate": 4.853444994290065e-06, + "loss": 1.6977, + "step": 384 + }, + { + "epoch": 0.3954802259887006, + "grad_norm": 2.713711738586426, + "learning_rate": 4.8515416825275985e-06, + "loss": 1.6449, + "step": 385 + }, + { + "epoch": 0.396507447354905, + "grad_norm": 2.7639107704162598, + "learning_rate": 4.849638370765132e-06, + "loss": 1.7169, + "step": 386 + }, + { + "epoch": 0.3975346687211094, + "grad_norm": 2.4775755405426025, + "learning_rate": 4.847735059002665e-06, + "loss": 1.6423, + "step": 387 + }, + { + "epoch": 0.3985618900873138, + "grad_norm": 2.574907064437866, + "learning_rate": 4.8458317472401985e-06, + "loss": 1.5911, + "step": 388 + }, + { + "epoch": 0.3995891114535182, + "grad_norm": 2.667907953262329, + "learning_rate": 4.843928435477732e-06, + "loss": 1.6883, + "step": 389 + }, + { + "epoch": 0.40061633281972264, + "grad_norm": 2.592353343963623, + "learning_rate": 4.842025123715265e-06, + "loss": 1.639, + "step": 390 + }, + { + "epoch": 0.40061633281972264, + "eval_loss": 1.9100708961486816, + "eval_runtime": 37.2379, + "eval_samples_per_second": 26.854, + "eval_steps_per_second": 1.128, + "step": 390 + }, + { + "epoch": 0.40164355418592707, + "grad_norm": 2.8238866329193115, + "learning_rate": 4.840121811952798e-06, + "loss": 1.6835, + "step": 391 + }, + { + "epoch": 0.4026707755521315, + "grad_norm": 2.769272804260254, + "learning_rate": 4.838218500190331e-06, + "loss": 1.6651, + "step": 392 + }, + { + "epoch": 0.4036979969183359, + "grad_norm": 2.749093532562256, + "learning_rate": 4.8363151884278645e-06, + "loss": 1.6815, + "step": 393 + }, + { + "epoch": 0.40472521828454033, + "grad_norm": 2.7952616214752197, + "learning_rate": 4.834411876665398e-06, + "loss": 1.6622, + "step": 394 + }, + { + "epoch": 0.40575243965074476, + "grad_norm": 2.5004849433898926, + "learning_rate": 4.832508564902931e-06, + "loss": 1.6768, + "step": 395 + }, + { + "epoch": 0.4067796610169492, + "grad_norm": 2.938015937805176, + "learning_rate": 4.830605253140465e-06, + "loss": 1.6171, + "step": 396 + }, + { + "epoch": 0.40780688238315355, + "grad_norm": 2.737666130065918, + "learning_rate": 4.828701941377998e-06, + "loss": 1.6572, + "step": 397 + }, + { + "epoch": 0.40883410374935797, + "grad_norm": 2.6246249675750732, + "learning_rate": 4.826798629615531e-06, + "loss": 1.609, + "step": 398 + }, + { + "epoch": 0.4098613251155624, + "grad_norm": 2.6501305103302, + "learning_rate": 4.824895317853065e-06, + "loss": 1.6235, + "step": 399 + }, + { + "epoch": 0.4108885464817668, + "grad_norm": 2.89320707321167, + "learning_rate": 4.822992006090598e-06, + "loss": 1.6494, + "step": 400 + }, + { + "epoch": 0.4108885464817668, + "eval_loss": 1.8916113376617432, + "eval_runtime": 37.1849, + "eval_samples_per_second": 26.893, + "eval_steps_per_second": 1.129, + "step": 400 + }, + { + "epoch": 0.41191576784797124, + "grad_norm": 2.4857001304626465, + "learning_rate": 4.821088694328131e-06, + "loss": 1.594, + "step": 401 + }, + { + "epoch": 0.41294298921417566, + "grad_norm": 2.903836250305176, + "learning_rate": 4.819185382565665e-06, + "loss": 1.6319, + "step": 402 + }, + { + "epoch": 0.4139702105803801, + "grad_norm": 2.69596266746521, + "learning_rate": 4.817282070803198e-06, + "loss": 1.6843, + "step": 403 + }, + { + "epoch": 0.4149974319465845, + "grad_norm": 2.5105295181274414, + "learning_rate": 4.8153787590407315e-06, + "loss": 1.6725, + "step": 404 + }, + { + "epoch": 0.41602465331278893, + "grad_norm": 2.7997353076934814, + "learning_rate": 4.813475447278265e-06, + "loss": 1.6017, + "step": 405 + }, + { + "epoch": 0.41705187467899335, + "grad_norm": 2.6215803623199463, + "learning_rate": 4.811572135515798e-06, + "loss": 1.6824, + "step": 406 + }, + { + "epoch": 0.4180790960451977, + "grad_norm": 2.76534104347229, + "learning_rate": 4.809668823753332e-06, + "loss": 1.6111, + "step": 407 + }, + { + "epoch": 0.41910631741140214, + "grad_norm": 2.6977133750915527, + "learning_rate": 4.807765511990864e-06, + "loss": 1.6165, + "step": 408 + }, + { + "epoch": 0.42013353877760656, + "grad_norm": 2.6771748065948486, + "learning_rate": 4.8058622002283975e-06, + "loss": 1.648, + "step": 409 + }, + { + "epoch": 0.421160760143811, + "grad_norm": 2.746731996536255, + "learning_rate": 4.803958888465931e-06, + "loss": 1.6946, + "step": 410 + }, + { + "epoch": 0.421160760143811, + "eval_loss": 1.878269076347351, + "eval_runtime": 37.0935, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 410 + }, + { + "epoch": 0.4221879815100154, + "grad_norm": 2.688074827194214, + "learning_rate": 4.802055576703464e-06, + "loss": 1.6545, + "step": 411 + }, + { + "epoch": 0.42321520287621983, + "grad_norm": 2.7336084842681885, + "learning_rate": 4.8001522649409976e-06, + "loss": 1.7305, + "step": 412 + }, + { + "epoch": 0.42424242424242425, + "grad_norm": 2.6902427673339844, + "learning_rate": 4.798248953178531e-06, + "loss": 1.612, + "step": 413 + }, + { + "epoch": 0.4252696456086287, + "grad_norm": 2.5339174270629883, + "learning_rate": 4.796345641416064e-06, + "loss": 1.5947, + "step": 414 + }, + { + "epoch": 0.4262968669748331, + "grad_norm": 2.6804542541503906, + "learning_rate": 4.794442329653598e-06, + "loss": 1.634, + "step": 415 + }, + { + "epoch": 0.42732408834103747, + "grad_norm": 2.5764853954315186, + "learning_rate": 4.792539017891131e-06, + "loss": 1.6947, + "step": 416 + }, + { + "epoch": 0.4283513097072419, + "grad_norm": 2.4067094326019287, + "learning_rate": 4.790635706128664e-06, + "loss": 1.616, + "step": 417 + }, + { + "epoch": 0.4293785310734463, + "grad_norm": 2.576206684112549, + "learning_rate": 4.788732394366197e-06, + "loss": 1.6529, + "step": 418 + }, + { + "epoch": 0.43040575243965074, + "grad_norm": 2.57826828956604, + "learning_rate": 4.786829082603731e-06, + "loss": 1.5771, + "step": 419 + }, + { + "epoch": 0.43143297380585516, + "grad_norm": 2.5002858638763428, + "learning_rate": 4.7849257708412645e-06, + "loss": 1.5814, + "step": 420 + }, + { + "epoch": 0.43143297380585516, + "eval_loss": 1.8940727710723877, + "eval_runtime": 37.0932, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 420 + }, + { + "epoch": 0.4324601951720596, + "grad_norm": 2.618067502975464, + "learning_rate": 4.783022459078798e-06, + "loss": 1.6645, + "step": 421 + }, + { + "epoch": 0.433487416538264, + "grad_norm": 2.5745837688446045, + "learning_rate": 4.781119147316331e-06, + "loss": 1.602, + "step": 422 + }, + { + "epoch": 0.4345146379044684, + "grad_norm": 2.8009719848632812, + "learning_rate": 4.7792158355538645e-06, + "loss": 1.6269, + "step": 423 + }, + { + "epoch": 0.43554185927067285, + "grad_norm": 2.8486008644104004, + "learning_rate": 4.777312523791398e-06, + "loss": 1.6402, + "step": 424 + }, + { + "epoch": 0.43656908063687727, + "grad_norm": 2.5588459968566895, + "learning_rate": 4.77540921202893e-06, + "loss": 1.6122, + "step": 425 + }, + { + "epoch": 0.43759630200308164, + "grad_norm": 3.1455588340759277, + "learning_rate": 4.773505900266464e-06, + "loss": 1.6542, + "step": 426 + }, + { + "epoch": 0.43862352336928606, + "grad_norm": 2.6221370697021484, + "learning_rate": 4.771602588503997e-06, + "loss": 1.6339, + "step": 427 + }, + { + "epoch": 0.4396507447354905, + "grad_norm": 3.006816864013672, + "learning_rate": 4.7696992767415305e-06, + "loss": 1.6381, + "step": 428 + }, + { + "epoch": 0.4406779661016949, + "grad_norm": 2.7317590713500977, + "learning_rate": 4.767795964979064e-06, + "loss": 1.6689, + "step": 429 + }, + { + "epoch": 0.44170518746789933, + "grad_norm": 2.9259109497070312, + "learning_rate": 4.765892653216597e-06, + "loss": 1.6793, + "step": 430 + }, + { + "epoch": 0.44170518746789933, + "eval_loss": 1.8887399435043335, + "eval_runtime": 37.1334, + "eval_samples_per_second": 26.93, + "eval_steps_per_second": 1.131, + "step": 430 + }, + { + "epoch": 0.44273240883410375, + "grad_norm": 2.8546087741851807, + "learning_rate": 4.763989341454131e-06, + "loss": 1.6706, + "step": 431 + }, + { + "epoch": 0.4437596302003082, + "grad_norm": 2.6085422039031982, + "learning_rate": 4.762086029691664e-06, + "loss": 1.5855, + "step": 432 + }, + { + "epoch": 0.4447868515665126, + "grad_norm": 2.622277021408081, + "learning_rate": 4.760182717929197e-06, + "loss": 1.6068, + "step": 433 + }, + { + "epoch": 0.445814072932717, + "grad_norm": 2.5327494144439697, + "learning_rate": 4.758279406166731e-06, + "loss": 1.6303, + "step": 434 + }, + { + "epoch": 0.44684129429892144, + "grad_norm": 2.7626633644104004, + "learning_rate": 4.756376094404263e-06, + "loss": 1.5567, + "step": 435 + }, + { + "epoch": 0.4478685156651258, + "grad_norm": 2.613476514816284, + "learning_rate": 4.7544727826417966e-06, + "loss": 1.6345, + "step": 436 + }, + { + "epoch": 0.44889573703133023, + "grad_norm": 2.58027720451355, + "learning_rate": 4.75256947087933e-06, + "loss": 1.675, + "step": 437 + }, + { + "epoch": 0.44992295839753466, + "grad_norm": 2.6709048748016357, + "learning_rate": 4.750666159116863e-06, + "loss": 1.6335, + "step": 438 + }, + { + "epoch": 0.4509501797637391, + "grad_norm": 2.7493093013763428, + "learning_rate": 4.7487628473543975e-06, + "loss": 1.6061, + "step": 439 + }, + { + "epoch": 0.4519774011299435, + "grad_norm": 2.725672960281372, + "learning_rate": 4.746859535591931e-06, + "loss": 1.5515, + "step": 440 + }, + { + "epoch": 0.4519774011299435, + "eval_loss": 1.8779852390289307, + "eval_runtime": 37.3135, + "eval_samples_per_second": 26.8, + "eval_steps_per_second": 1.126, + "step": 440 + }, + { + "epoch": 0.4530046224961479, + "grad_norm": 2.784477472305298, + "learning_rate": 4.744956223829464e-06, + "loss": 1.6146, + "step": 441 + }, + { + "epoch": 0.45403184386235235, + "grad_norm": 2.758119821548462, + "learning_rate": 4.743052912066997e-06, + "loss": 1.5945, + "step": 442 + }, + { + "epoch": 0.45505906522855677, + "grad_norm": 2.9006264209747314, + "learning_rate": 4.74114960030453e-06, + "loss": 1.6461, + "step": 443 + }, + { + "epoch": 0.4560862865947612, + "grad_norm": 2.8473031520843506, + "learning_rate": 4.7392462885420635e-06, + "loss": 1.6565, + "step": 444 + }, + { + "epoch": 0.4571135079609656, + "grad_norm": 2.715806722640991, + "learning_rate": 4.737342976779597e-06, + "loss": 1.5414, + "step": 445 + }, + { + "epoch": 0.45814072932717, + "grad_norm": 2.6588337421417236, + "learning_rate": 4.73543966501713e-06, + "loss": 1.6284, + "step": 446 + }, + { + "epoch": 0.4591679506933744, + "grad_norm": 3.1152243614196777, + "learning_rate": 4.7335363532546636e-06, + "loss": 1.6237, + "step": 447 + }, + { + "epoch": 0.4601951720595788, + "grad_norm": 2.545490264892578, + "learning_rate": 4.731633041492197e-06, + "loss": 1.6149, + "step": 448 + }, + { + "epoch": 0.46122239342578325, + "grad_norm": 3.0324277877807617, + "learning_rate": 4.72972972972973e-06, + "loss": 1.6277, + "step": 449 + }, + { + "epoch": 0.4622496147919877, + "grad_norm": 2.885970115661621, + "learning_rate": 4.727826417967264e-06, + "loss": 1.6625, + "step": 450 + }, + { + "epoch": 0.4622496147919877, + "eval_loss": 1.8651624917984009, + "eval_runtime": 37.3486, + "eval_samples_per_second": 26.775, + "eval_steps_per_second": 1.125, + "step": 450 + }, + { + "epoch": 0.4632768361581921, + "grad_norm": 2.7258739471435547, + "learning_rate": 4.725923106204797e-06, + "loss": 1.6263, + "step": 451 + }, + { + "epoch": 0.4643040575243965, + "grad_norm": 2.631122589111328, + "learning_rate": 4.72401979444233e-06, + "loss": 1.6081, + "step": 452 + }, + { + "epoch": 0.46533127889060094, + "grad_norm": 2.5882441997528076, + "learning_rate": 4.722116482679863e-06, + "loss": 1.6237, + "step": 453 + }, + { + "epoch": 0.46635850025680536, + "grad_norm": 2.66310715675354, + "learning_rate": 4.720213170917396e-06, + "loss": 1.6306, + "step": 454 + }, + { + "epoch": 0.46738572162300973, + "grad_norm": 2.6906657218933105, + "learning_rate": 4.71830985915493e-06, + "loss": 1.6492, + "step": 455 + }, + { + "epoch": 0.46841294298921415, + "grad_norm": 2.6957409381866455, + "learning_rate": 4.716406547392463e-06, + "loss": 1.5993, + "step": 456 + }, + { + "epoch": 0.4694401643554186, + "grad_norm": 2.904311418533325, + "learning_rate": 4.714503235629996e-06, + "loss": 1.6633, + "step": 457 + }, + { + "epoch": 0.470467385721623, + "grad_norm": 2.6853256225585938, + "learning_rate": 4.71259992386753e-06, + "loss": 1.5824, + "step": 458 + }, + { + "epoch": 0.4714946070878274, + "grad_norm": 2.4767656326293945, + "learning_rate": 4.710696612105063e-06, + "loss": 1.6145, + "step": 459 + }, + { + "epoch": 0.47252182845403184, + "grad_norm": 2.5569417476654053, + "learning_rate": 4.708793300342596e-06, + "loss": 1.6116, + "step": 460 + }, + { + "epoch": 0.47252182845403184, + "eval_loss": 1.8785003423690796, + "eval_runtime": 37.401, + "eval_samples_per_second": 26.737, + "eval_steps_per_second": 1.123, + "step": 460 + }, + { + "epoch": 0.47354904982023627, + "grad_norm": 2.6568074226379395, + "learning_rate": 4.70688998858013e-06, + "loss": 1.6097, + "step": 461 + }, + { + "epoch": 0.4745762711864407, + "grad_norm": 2.6002280712127686, + "learning_rate": 4.704986676817663e-06, + "loss": 1.6025, + "step": 462 + }, + { + "epoch": 0.4756034925526451, + "grad_norm": 2.6055355072021484, + "learning_rate": 4.7030833650551965e-06, + "loss": 1.6322, + "step": 463 + }, + { + "epoch": 0.47663071391884954, + "grad_norm": 2.6288487911224365, + "learning_rate": 4.70118005329273e-06, + "loss": 1.594, + "step": 464 + }, + { + "epoch": 0.4776579352850539, + "grad_norm": 2.6981899738311768, + "learning_rate": 4.699276741530263e-06, + "loss": 1.6707, + "step": 465 + }, + { + "epoch": 0.4786851566512583, + "grad_norm": 2.6832222938537598, + "learning_rate": 4.697373429767797e-06, + "loss": 1.6435, + "step": 466 + }, + { + "epoch": 0.47971237801746275, + "grad_norm": 2.563159227371216, + "learning_rate": 4.69547011800533e-06, + "loss": 1.6249, + "step": 467 + }, + { + "epoch": 0.48073959938366717, + "grad_norm": 2.545778751373291, + "learning_rate": 4.693566806242863e-06, + "loss": 1.6309, + "step": 468 + }, + { + "epoch": 0.4817668207498716, + "grad_norm": 2.555227518081665, + "learning_rate": 4.691663494480397e-06, + "loss": 1.6128, + "step": 469 + }, + { + "epoch": 0.482794042116076, + "grad_norm": 2.612576723098755, + "learning_rate": 4.689760182717929e-06, + "loss": 1.6484, + "step": 470 + }, + { + "epoch": 0.482794042116076, + "eval_loss": 1.8684974908828735, + "eval_runtime": 37.2607, + "eval_samples_per_second": 26.838, + "eval_steps_per_second": 1.127, + "step": 470 + }, + { + "epoch": 0.48382126348228044, + "grad_norm": 2.4981396198272705, + "learning_rate": 4.6878568709554626e-06, + "loss": 1.5393, + "step": 471 + }, + { + "epoch": 0.48484848484848486, + "grad_norm": 2.793700695037842, + "learning_rate": 4.685953559192996e-06, + "loss": 1.608, + "step": 472 + }, + { + "epoch": 0.4858757062146893, + "grad_norm": 2.7605607509613037, + "learning_rate": 4.684050247430529e-06, + "loss": 1.6236, + "step": 473 + }, + { + "epoch": 0.4869029275808937, + "grad_norm": 2.413719892501831, + "learning_rate": 4.682146935668063e-06, + "loss": 1.5555, + "step": 474 + }, + { + "epoch": 0.4879301489470981, + "grad_norm": 2.9032249450683594, + "learning_rate": 4.680243623905596e-06, + "loss": 1.6042, + "step": 475 + }, + { + "epoch": 0.4889573703133025, + "grad_norm": 2.5445175170898438, + "learning_rate": 4.678340312143129e-06, + "loss": 1.6265, + "step": 476 + }, + { + "epoch": 0.4899845916795069, + "grad_norm": 2.6955716609954834, + "learning_rate": 4.676437000380663e-06, + "loss": 1.5879, + "step": 477 + }, + { + "epoch": 0.49101181304571134, + "grad_norm": 2.8399500846862793, + "learning_rate": 4.674533688618196e-06, + "loss": 1.7009, + "step": 478 + }, + { + "epoch": 0.49203903441191577, + "grad_norm": 2.635035991668701, + "learning_rate": 4.6726303768557295e-06, + "loss": 1.6381, + "step": 479 + }, + { + "epoch": 0.4930662557781202, + "grad_norm": 2.7609052658081055, + "learning_rate": 4.670727065093263e-06, + "loss": 1.5492, + "step": 480 + }, + { + "epoch": 0.4930662557781202, + "eval_loss": 1.8574503660202026, + "eval_runtime": 37.2173, + "eval_samples_per_second": 26.869, + "eval_steps_per_second": 1.129, + "step": 480 + }, + { + "epoch": 0.4940934771443246, + "grad_norm": 2.593766927719116, + "learning_rate": 4.668823753330796e-06, + "loss": 1.5693, + "step": 481 + }, + { + "epoch": 0.49512069851052903, + "grad_norm": 2.8000242710113525, + "learning_rate": 4.6669204415683295e-06, + "loss": 1.5998, + "step": 482 + }, + { + "epoch": 0.49614791987673346, + "grad_norm": 2.522688627243042, + "learning_rate": 4.665017129805863e-06, + "loss": 1.6653, + "step": 483 + }, + { + "epoch": 0.4971751412429379, + "grad_norm": 2.854534864425659, + "learning_rate": 4.663113818043396e-06, + "loss": 1.6393, + "step": 484 + }, + { + "epoch": 0.49820236260914225, + "grad_norm": 2.5333967208862305, + "learning_rate": 4.66121050628093e-06, + "loss": 1.606, + "step": 485 + }, + { + "epoch": 0.49922958397534667, + "grad_norm": 2.565260171890259, + "learning_rate": 4.659307194518463e-06, + "loss": 1.6471, + "step": 486 + }, + { + "epoch": 0.5002568053415511, + "grad_norm": 2.570220947265625, + "learning_rate": 4.6574038827559955e-06, + "loss": 1.6023, + "step": 487 + }, + { + "epoch": 0.5012840267077555, + "grad_norm": 2.460740089416504, + "learning_rate": 4.655500570993529e-06, + "loss": 1.6208, + "step": 488 + }, + { + "epoch": 0.50231124807396, + "grad_norm": 2.6339426040649414, + "learning_rate": 4.653597259231062e-06, + "loss": 1.6533, + "step": 489 + }, + { + "epoch": 0.5033384694401644, + "grad_norm": 2.681044578552246, + "learning_rate": 4.651693947468596e-06, + "loss": 1.5437, + "step": 490 + }, + { + "epoch": 0.5033384694401644, + "eval_loss": 1.8701337575912476, + "eval_runtime": 37.1714, + "eval_samples_per_second": 26.902, + "eval_steps_per_second": 1.13, + "step": 490 + }, + { + "epoch": 0.5043656908063687, + "grad_norm": 2.6900649070739746, + "learning_rate": 4.649790635706129e-06, + "loss": 1.6816, + "step": 491 + }, + { + "epoch": 0.5053929121725732, + "grad_norm": 2.6106340885162354, + "learning_rate": 4.647887323943662e-06, + "loss": 1.5784, + "step": 492 + }, + { + "epoch": 0.5064201335387776, + "grad_norm": 3.0704469680786133, + "learning_rate": 4.645984012181196e-06, + "loss": 1.6065, + "step": 493 + }, + { + "epoch": 0.507447354904982, + "grad_norm": 2.648425817489624, + "learning_rate": 4.644080700418729e-06, + "loss": 1.5134, + "step": 494 + }, + { + "epoch": 0.5084745762711864, + "grad_norm": 2.509066343307495, + "learning_rate": 4.642177388656262e-06, + "loss": 1.6324, + "step": 495 + }, + { + "epoch": 0.5095017976373909, + "grad_norm": 2.8628857135772705, + "learning_rate": 4.640274076893796e-06, + "loss": 1.5957, + "step": 496 + }, + { + "epoch": 0.5105290190035953, + "grad_norm": 2.83579683303833, + "learning_rate": 4.638370765131328e-06, + "loss": 1.6556, + "step": 497 + }, + { + "epoch": 0.5115562403697997, + "grad_norm": 2.6618151664733887, + "learning_rate": 4.636467453368862e-06, + "loss": 1.5738, + "step": 498 + }, + { + "epoch": 0.5125834617360041, + "grad_norm": 2.936943769454956, + "learning_rate": 4.634564141606396e-06, + "loss": 1.5597, + "step": 499 + }, + { + "epoch": 0.5136106831022085, + "grad_norm": 2.8520913124084473, + "learning_rate": 4.632660829843929e-06, + "loss": 1.6316, + "step": 500 + }, + { + "epoch": 0.5136106831022085, + "eval_loss": 1.8548446893692017, + "eval_runtime": 37.1878, + "eval_samples_per_second": 26.891, + "eval_steps_per_second": 1.129, + "step": 500 + }, + { + "epoch": 0.514637904468413, + "grad_norm": 2.446373462677002, + "learning_rate": 4.630757518081463e-06, + "loss": 1.5533, + "step": 501 + }, + { + "epoch": 0.5156651258346173, + "grad_norm": 2.8339035511016846, + "learning_rate": 4.628854206318996e-06, + "loss": 1.6904, + "step": 502 + }, + { + "epoch": 0.5166923472008218, + "grad_norm": 2.664249897003174, + "learning_rate": 4.626950894556529e-06, + "loss": 1.5926, + "step": 503 + }, + { + "epoch": 0.5177195685670262, + "grad_norm": 2.681185483932495, + "learning_rate": 4.625047582794062e-06, + "loss": 1.6348, + "step": 504 + }, + { + "epoch": 0.5187467899332306, + "grad_norm": 2.468263864517212, + "learning_rate": 4.623144271031595e-06, + "loss": 1.5477, + "step": 505 + }, + { + "epoch": 0.519774011299435, + "grad_norm": 2.539066791534424, + "learning_rate": 4.6212409592691286e-06, + "loss": 1.5615, + "step": 506 + }, + { + "epoch": 0.5208012326656395, + "grad_norm": 2.822235345840454, + "learning_rate": 4.619337647506662e-06, + "loss": 1.6398, + "step": 507 + }, + { + "epoch": 0.5218284540318439, + "grad_norm": 2.6063523292541504, + "learning_rate": 4.617434335744195e-06, + "loss": 1.5992, + "step": 508 + }, + { + "epoch": 0.5228556753980482, + "grad_norm": 2.624810218811035, + "learning_rate": 4.615531023981729e-06, + "loss": 1.5708, + "step": 509 + }, + { + "epoch": 0.5238828967642527, + "grad_norm": 2.6324198246002197, + "learning_rate": 4.613627712219262e-06, + "loss": 1.5865, + "step": 510 + }, + { + "epoch": 0.5238828967642527, + "eval_loss": 1.8442509174346924, + "eval_runtime": 36.9899, + "eval_samples_per_second": 27.034, + "eval_steps_per_second": 1.135, + "step": 510 + }, + { + "epoch": 0.5249101181304571, + "grad_norm": 2.5496973991394043, + "learning_rate": 4.611724400456795e-06, + "loss": 1.6048, + "step": 511 + }, + { + "epoch": 0.5259373394966615, + "grad_norm": 2.4150168895721436, + "learning_rate": 4.609821088694329e-06, + "loss": 1.6009, + "step": 512 + }, + { + "epoch": 0.5269645608628659, + "grad_norm": 2.8799378871917725, + "learning_rate": 4.607917776931862e-06, + "loss": 1.5761, + "step": 513 + }, + { + "epoch": 0.5279917822290704, + "grad_norm": 2.749016761779785, + "learning_rate": 4.606014465169395e-06, + "loss": 1.6202, + "step": 514 + }, + { + "epoch": 0.5290190035952748, + "grad_norm": 2.6993141174316406, + "learning_rate": 4.604111153406928e-06, + "loss": 1.632, + "step": 515 + }, + { + "epoch": 0.5300462249614792, + "grad_norm": 2.9027445316314697, + "learning_rate": 4.602207841644461e-06, + "loss": 1.5352, + "step": 516 + }, + { + "epoch": 0.5310734463276836, + "grad_norm": 2.874267816543579, + "learning_rate": 4.600304529881995e-06, + "loss": 1.642, + "step": 517 + }, + { + "epoch": 0.5321006676938881, + "grad_norm": 2.6057357788085938, + "learning_rate": 4.598401218119528e-06, + "loss": 1.6179, + "step": 518 + }, + { + "epoch": 0.5331278890600925, + "grad_norm": 2.658812999725342, + "learning_rate": 4.5964979063570614e-06, + "loss": 1.6285, + "step": 519 + }, + { + "epoch": 0.5341551104262968, + "grad_norm": 2.667215585708618, + "learning_rate": 4.594594594594596e-06, + "loss": 1.5694, + "step": 520 + }, + { + "epoch": 0.5341551104262968, + "eval_loss": 1.8490455150604248, + "eval_runtime": 37.0914, + "eval_samples_per_second": 26.96, + "eval_steps_per_second": 1.132, + "step": 520 + }, + { + "epoch": 0.5351823317925013, + "grad_norm": 2.5979607105255127, + "learning_rate": 4.592691282832128e-06, + "loss": 1.6091, + "step": 521 + }, + { + "epoch": 0.5362095531587057, + "grad_norm": 2.552816152572632, + "learning_rate": 4.5907879710696615e-06, + "loss": 1.6145, + "step": 522 + }, + { + "epoch": 0.5372367745249101, + "grad_norm": 2.575045347213745, + "learning_rate": 4.588884659307195e-06, + "loss": 1.5444, + "step": 523 + }, + { + "epoch": 0.5382639958911145, + "grad_norm": 2.650202751159668, + "learning_rate": 4.586981347544728e-06, + "loss": 1.6181, + "step": 524 + }, + { + "epoch": 0.539291217257319, + "grad_norm": 2.7869739532470703, + "learning_rate": 4.585078035782262e-06, + "loss": 1.563, + "step": 525 + }, + { + "epoch": 0.5403184386235234, + "grad_norm": 2.6828651428222656, + "learning_rate": 4.583174724019795e-06, + "loss": 1.6004, + "step": 526 + }, + { + "epoch": 0.5413456599897278, + "grad_norm": 2.552387237548828, + "learning_rate": 4.581271412257328e-06, + "loss": 1.5351, + "step": 527 + }, + { + "epoch": 0.5423728813559322, + "grad_norm": 2.785661458969116, + "learning_rate": 4.579368100494862e-06, + "loss": 1.5264, + "step": 528 + }, + { + "epoch": 0.5434001027221366, + "grad_norm": 2.832939863204956, + "learning_rate": 4.577464788732395e-06, + "loss": 1.6261, + "step": 529 + }, + { + "epoch": 0.544427324088341, + "grad_norm": 2.641766309738159, + "learning_rate": 4.575561476969928e-06, + "loss": 1.6019, + "step": 530 + }, + { + "epoch": 0.544427324088341, + "eval_loss": 1.8477164506912231, + "eval_runtime": 37.1444, + "eval_samples_per_second": 26.922, + "eval_steps_per_second": 1.131, + "step": 530 + }, + { + "epoch": 0.5454545454545454, + "grad_norm": 3.0502524375915527, + "learning_rate": 4.573658165207461e-06, + "loss": 1.6436, + "step": 531 + }, + { + "epoch": 0.5464817668207499, + "grad_norm": 2.64772891998291, + "learning_rate": 4.571754853444994e-06, + "loss": 1.6085, + "step": 532 + }, + { + "epoch": 0.5475089881869543, + "grad_norm": 2.549692392349243, + "learning_rate": 4.569851541682528e-06, + "loss": 1.5977, + "step": 533 + }, + { + "epoch": 0.5485362095531587, + "grad_norm": 2.719322681427002, + "learning_rate": 4.567948229920061e-06, + "loss": 1.5659, + "step": 534 + }, + { + "epoch": 0.5495634309193631, + "grad_norm": 2.5005698204040527, + "learning_rate": 4.566044918157594e-06, + "loss": 1.5301, + "step": 535 + }, + { + "epoch": 0.5505906522855676, + "grad_norm": 2.7379472255706787, + "learning_rate": 4.564141606395128e-06, + "loss": 1.5721, + "step": 536 + }, + { + "epoch": 0.551617873651772, + "grad_norm": 2.6673905849456787, + "learning_rate": 4.562238294632661e-06, + "loss": 1.6105, + "step": 537 + }, + { + "epoch": 0.5526450950179764, + "grad_norm": 2.6745669841766357, + "learning_rate": 4.5603349828701945e-06, + "loss": 1.577, + "step": 538 + }, + { + "epoch": 0.5536723163841808, + "grad_norm": 2.593475341796875, + "learning_rate": 4.558431671107728e-06, + "loss": 1.6084, + "step": 539 + }, + { + "epoch": 0.5546995377503852, + "grad_norm": 2.6577041149139404, + "learning_rate": 4.556528359345261e-06, + "loss": 1.6417, + "step": 540 + }, + { + "epoch": 0.5546995377503852, + "eval_loss": 1.839761734008789, + "eval_runtime": 37.1765, + "eval_samples_per_second": 26.899, + "eval_steps_per_second": 1.13, + "step": 540 + }, + { + "epoch": 0.5557267591165896, + "grad_norm": 2.6834933757781982, + "learning_rate": 4.5546250475827946e-06, + "loss": 1.5803, + "step": 541 + }, + { + "epoch": 0.556753980482794, + "grad_norm": 2.464665174484253, + "learning_rate": 4.552721735820328e-06, + "loss": 1.5324, + "step": 542 + }, + { + "epoch": 0.5577812018489985, + "grad_norm": 2.4488964080810547, + "learning_rate": 4.550818424057861e-06, + "loss": 1.5767, + "step": 543 + }, + { + "epoch": 0.5588084232152029, + "grad_norm": 2.763124465942383, + "learning_rate": 4.548915112295395e-06, + "loss": 1.5903, + "step": 544 + }, + { + "epoch": 0.5598356445814073, + "grad_norm": 2.7473502159118652, + "learning_rate": 4.547011800532928e-06, + "loss": 1.5966, + "step": 545 + }, + { + "epoch": 0.5608628659476117, + "grad_norm": 3.050267219543457, + "learning_rate": 4.545108488770461e-06, + "loss": 1.5492, + "step": 546 + }, + { + "epoch": 0.5618900873138162, + "grad_norm": 2.910439968109131, + "learning_rate": 4.543205177007995e-06, + "loss": 1.6357, + "step": 547 + }, + { + "epoch": 0.5629173086800205, + "grad_norm": 2.5621109008789062, + "learning_rate": 4.541301865245527e-06, + "loss": 1.5693, + "step": 548 + }, + { + "epoch": 0.5639445300462249, + "grad_norm": 2.7492010593414307, + "learning_rate": 4.539398553483061e-06, + "loss": 1.615, + "step": 549 + }, + { + "epoch": 0.5649717514124294, + "grad_norm": 2.6651856899261475, + "learning_rate": 4.537495241720594e-06, + "loss": 1.5804, + "step": 550 + }, + { + "epoch": 0.5649717514124294, + "eval_loss": 1.8344007730484009, + "eval_runtime": 37.1218, + "eval_samples_per_second": 26.938, + "eval_steps_per_second": 1.131, + "step": 550 + }, + { + "epoch": 0.5659989727786338, + "grad_norm": 2.5933356285095215, + "learning_rate": 4.535591929958127e-06, + "loss": 1.6108, + "step": 551 + }, + { + "epoch": 0.5670261941448382, + "grad_norm": 2.660522222518921, + "learning_rate": 4.533688618195661e-06, + "loss": 1.5977, + "step": 552 + }, + { + "epoch": 0.5680534155110426, + "grad_norm": 2.5636332035064697, + "learning_rate": 4.531785306433194e-06, + "loss": 1.5802, + "step": 553 + }, + { + "epoch": 0.5690806368772471, + "grad_norm": 2.759533166885376, + "learning_rate": 4.529881994670727e-06, + "loss": 1.5656, + "step": 554 + }, + { + "epoch": 0.5701078582434514, + "grad_norm": 2.6530680656433105, + "learning_rate": 4.527978682908261e-06, + "loss": 1.5693, + "step": 555 + }, + { + "epoch": 0.5711350796096559, + "grad_norm": 2.688427448272705, + "learning_rate": 4.526075371145794e-06, + "loss": 1.5769, + "step": 556 + }, + { + "epoch": 0.5721623009758603, + "grad_norm": 2.5163164138793945, + "learning_rate": 4.5241720593833275e-06, + "loss": 1.593, + "step": 557 + }, + { + "epoch": 0.5731895223420647, + "grad_norm": 2.612612724304199, + "learning_rate": 4.52226874762086e-06, + "loss": 1.5854, + "step": 558 + }, + { + "epoch": 0.5742167437082691, + "grad_norm": 2.647853374481201, + "learning_rate": 4.520365435858393e-06, + "loss": 1.6085, + "step": 559 + }, + { + "epoch": 0.5752439650744735, + "grad_norm": 2.7160937786102295, + "learning_rate": 4.518462124095928e-06, + "loss": 1.6353, + "step": 560 + }, + { + "epoch": 0.5752439650744735, + "eval_loss": 1.8585809469223022, + "eval_runtime": 37.328, + "eval_samples_per_second": 26.79, + "eval_steps_per_second": 1.125, + "step": 560 + }, + { + "epoch": 0.576271186440678, + "grad_norm": 2.9025261402130127, + "learning_rate": 4.516558812333461e-06, + "loss": 1.5903, + "step": 561 + }, + { + "epoch": 0.5772984078068824, + "grad_norm": 2.5665547847747803, + "learning_rate": 4.514655500570994e-06, + "loss": 1.5775, + "step": 562 + }, + { + "epoch": 0.5783256291730868, + "grad_norm": 2.591381311416626, + "learning_rate": 4.512752188808528e-06, + "loss": 1.6387, + "step": 563 + }, + { + "epoch": 0.5793528505392912, + "grad_norm": 2.7540411949157715, + "learning_rate": 4.510848877046061e-06, + "loss": 1.6432, + "step": 564 + }, + { + "epoch": 0.5803800719054957, + "grad_norm": 2.6911778450012207, + "learning_rate": 4.5089455652835936e-06, + "loss": 1.6078, + "step": 565 + }, + { + "epoch": 0.5814072932717, + "grad_norm": 2.609619140625, + "learning_rate": 4.507042253521127e-06, + "loss": 1.5709, + "step": 566 + }, + { + "epoch": 0.5824345146379045, + "grad_norm": 2.4930672645568848, + "learning_rate": 4.50513894175866e-06, + "loss": 1.5305, + "step": 567 + }, + { + "epoch": 0.5834617360041089, + "grad_norm": 2.490992307662964, + "learning_rate": 4.503235629996194e-06, + "loss": 1.61, + "step": 568 + }, + { + "epoch": 0.5844889573703133, + "grad_norm": 2.7229626178741455, + "learning_rate": 4.501332318233727e-06, + "loss": 1.5996, + "step": 569 + }, + { + "epoch": 0.5855161787365177, + "grad_norm": 2.5289995670318604, + "learning_rate": 4.49942900647126e-06, + "loss": 1.5624, + "step": 570 + }, + { + "epoch": 0.5855161787365177, + "eval_loss": 1.867126226425171, + "eval_runtime": 37.4117, + "eval_samples_per_second": 26.73, + "eval_steps_per_second": 1.123, + "step": 570 + }, + { + "epoch": 0.5865434001027221, + "grad_norm": 2.6244945526123047, + "learning_rate": 4.497525694708794e-06, + "loss": 1.6154, + "step": 571 + }, + { + "epoch": 0.5875706214689266, + "grad_norm": 2.4343698024749756, + "learning_rate": 4.495622382946327e-06, + "loss": 1.5967, + "step": 572 + }, + { + "epoch": 0.588597842835131, + "grad_norm": 2.5899314880371094, + "learning_rate": 4.4937190711838605e-06, + "loss": 1.5601, + "step": 573 + }, + { + "epoch": 0.5896250642013354, + "grad_norm": 2.829028367996216, + "learning_rate": 4.491815759421394e-06, + "loss": 1.6315, + "step": 574 + }, + { + "epoch": 0.5906522855675398, + "grad_norm": 2.748849868774414, + "learning_rate": 4.489912447658926e-06, + "loss": 1.5591, + "step": 575 + }, + { + "epoch": 0.5916795069337443, + "grad_norm": 2.7016541957855225, + "learning_rate": 4.48800913589646e-06, + "loss": 1.5575, + "step": 576 + }, + { + "epoch": 0.5927067282999486, + "grad_norm": 2.6260931491851807, + "learning_rate": 4.486105824133993e-06, + "loss": 1.5823, + "step": 577 + }, + { + "epoch": 0.593733949666153, + "grad_norm": 2.797224998474121, + "learning_rate": 4.4842025123715264e-06, + "loss": 1.5902, + "step": 578 + }, + { + "epoch": 0.5947611710323575, + "grad_norm": 2.671342611312866, + "learning_rate": 4.48229920060906e-06, + "loss": 1.551, + "step": 579 + }, + { + "epoch": 0.5957883923985619, + "grad_norm": 2.626988172531128, + "learning_rate": 4.480395888846594e-06, + "loss": 1.5312, + "step": 580 + }, + { + "epoch": 0.5957883923985619, + "eval_loss": 1.8685871362686157, + "eval_runtime": 37.388, + "eval_samples_per_second": 26.747, + "eval_steps_per_second": 1.123, + "step": 580 + }, + { + "epoch": 0.5968156137647663, + "grad_norm": 2.8422539234161377, + "learning_rate": 4.478492577084127e-06, + "loss": 1.6393, + "step": 581 + }, + { + "epoch": 0.5978428351309707, + "grad_norm": 2.6877822875976562, + "learning_rate": 4.47658926532166e-06, + "loss": 1.5376, + "step": 582 + }, + { + "epoch": 0.5988700564971752, + "grad_norm": 2.736452341079712, + "learning_rate": 4.474685953559193e-06, + "loss": 1.5798, + "step": 583 + }, + { + "epoch": 0.5998972778633795, + "grad_norm": 2.732513904571533, + "learning_rate": 4.472782641796727e-06, + "loss": 1.6897, + "step": 584 + }, + { + "epoch": 0.600924499229584, + "grad_norm": 2.8732001781463623, + "learning_rate": 4.47087933003426e-06, + "loss": 1.533, + "step": 585 + }, + { + "epoch": 0.6019517205957884, + "grad_norm": 2.6099846363067627, + "learning_rate": 4.468976018271793e-06, + "loss": 1.5481, + "step": 586 + }, + { + "epoch": 0.6029789419619929, + "grad_norm": 2.754856586456299, + "learning_rate": 4.467072706509327e-06, + "loss": 1.5497, + "step": 587 + }, + { + "epoch": 0.6040061633281972, + "grad_norm": 2.6755611896514893, + "learning_rate": 4.46516939474686e-06, + "loss": 1.5882, + "step": 588 + }, + { + "epoch": 0.6050333846944016, + "grad_norm": 2.632248640060425, + "learning_rate": 4.463266082984393e-06, + "loss": 1.5481, + "step": 589 + }, + { + "epoch": 0.6060606060606061, + "grad_norm": 2.6425793170928955, + "learning_rate": 4.461362771221927e-06, + "loss": 1.6607, + "step": 590 + }, + { + "epoch": 0.6060606060606061, + "eval_loss": 1.863004446029663, + "eval_runtime": 37.2064, + "eval_samples_per_second": 26.877, + "eval_steps_per_second": 1.129, + "step": 590 + }, + { + "epoch": 0.6070878274268104, + "grad_norm": 2.6340153217315674, + "learning_rate": 4.45945945945946e-06, + "loss": 1.5474, + "step": 591 + }, + { + "epoch": 0.6081150487930149, + "grad_norm": 2.7601606845855713, + "learning_rate": 4.457556147696993e-06, + "loss": 1.5569, + "step": 592 + }, + { + "epoch": 0.6091422701592193, + "grad_norm": 2.8151822090148926, + "learning_rate": 4.455652835934526e-06, + "loss": 1.6441, + "step": 593 + }, + { + "epoch": 0.6101694915254238, + "grad_norm": 2.757577419281006, + "learning_rate": 4.453749524172059e-06, + "loss": 1.6339, + "step": 594 + }, + { + "epoch": 0.6111967128916281, + "grad_norm": 2.6828997135162354, + "learning_rate": 4.451846212409593e-06, + "loss": 1.5885, + "step": 595 + }, + { + "epoch": 0.6122239342578326, + "grad_norm": 2.6645700931549072, + "learning_rate": 4.449942900647126e-06, + "loss": 1.5866, + "step": 596 + }, + { + "epoch": 0.613251155624037, + "grad_norm": 2.572789430618286, + "learning_rate": 4.4480395888846595e-06, + "loss": 1.5255, + "step": 597 + }, + { + "epoch": 0.6142783769902413, + "grad_norm": 2.6281981468200684, + "learning_rate": 4.446136277122193e-06, + "loss": 1.5905, + "step": 598 + }, + { + "epoch": 0.6153055983564458, + "grad_norm": 2.6164042949676514, + "learning_rate": 4.444232965359726e-06, + "loss": 1.5274, + "step": 599 + }, + { + "epoch": 0.6163328197226502, + "grad_norm": 2.606343984603882, + "learning_rate": 4.4423296535972596e-06, + "loss": 1.5846, + "step": 600 + }, + { + "epoch": 0.6163328197226502, + "eval_loss": 1.8634170293807983, + "eval_runtime": 37.1568, + "eval_samples_per_second": 26.913, + "eval_steps_per_second": 1.13, + "step": 600 + }, + { + "epoch": 0.6173600410888547, + "grad_norm": 2.8743226528167725, + "learning_rate": 4.440426341834793e-06, + "loss": 1.538, + "step": 601 + }, + { + "epoch": 0.618387262455059, + "grad_norm": 2.5392212867736816, + "learning_rate": 4.438523030072326e-06, + "loss": 1.585, + "step": 602 + }, + { + "epoch": 0.6194144838212635, + "grad_norm": 2.715684413909912, + "learning_rate": 4.43661971830986e-06, + "loss": 1.6028, + "step": 603 + }, + { + "epoch": 0.6204417051874679, + "grad_norm": 2.514934778213501, + "learning_rate": 4.434716406547393e-06, + "loss": 1.5525, + "step": 604 + }, + { + "epoch": 0.6214689265536724, + "grad_norm": 2.685807943344116, + "learning_rate": 4.432813094784926e-06, + "loss": 1.5456, + "step": 605 + }, + { + "epoch": 0.6224961479198767, + "grad_norm": 2.848557949066162, + "learning_rate": 4.43090978302246e-06, + "loss": 1.6017, + "step": 606 + }, + { + "epoch": 0.6235233692860811, + "grad_norm": 2.6820950508117676, + "learning_rate": 4.429006471259993e-06, + "loss": 1.573, + "step": 607 + }, + { + "epoch": 0.6245505906522856, + "grad_norm": 2.493544340133667, + "learning_rate": 4.4271031594975265e-06, + "loss": 1.5576, + "step": 608 + }, + { + "epoch": 0.6255778120184899, + "grad_norm": 2.6240968704223633, + "learning_rate": 4.425199847735059e-06, + "loss": 1.5312, + "step": 609 + }, + { + "epoch": 0.6266050333846944, + "grad_norm": 2.516113042831421, + "learning_rate": 4.423296535972592e-06, + "loss": 1.5284, + "step": 610 + }, + { + "epoch": 0.6266050333846944, + "eval_loss": 1.8370482921600342, + "eval_runtime": 36.9401, + "eval_samples_per_second": 27.071, + "eval_steps_per_second": 1.137, + "step": 610 + }, + { + "epoch": 0.6276322547508988, + "grad_norm": 2.8341944217681885, + "learning_rate": 4.421393224210126e-06, + "loss": 1.5609, + "step": 611 + }, + { + "epoch": 0.6286594761171033, + "grad_norm": 2.5433242321014404, + "learning_rate": 4.419489912447659e-06, + "loss": 1.5253, + "step": 612 + }, + { + "epoch": 0.6296866974833076, + "grad_norm": 2.6761138439178467, + "learning_rate": 4.4175866006851924e-06, + "loss": 1.6171, + "step": 613 + }, + { + "epoch": 0.6307139188495121, + "grad_norm": 2.570707321166992, + "learning_rate": 4.415683288922726e-06, + "loss": 1.5766, + "step": 614 + }, + { + "epoch": 0.6317411402157165, + "grad_norm": 2.5421109199523926, + "learning_rate": 4.413779977160259e-06, + "loss": 1.5616, + "step": 615 + }, + { + "epoch": 0.632768361581921, + "grad_norm": 2.6767091751098633, + "learning_rate": 4.4118766653977925e-06, + "loss": 1.5531, + "step": 616 + }, + { + "epoch": 0.6337955829481253, + "grad_norm": 2.7730000019073486, + "learning_rate": 4.409973353635326e-06, + "loss": 1.6224, + "step": 617 + }, + { + "epoch": 0.6348228043143297, + "grad_norm": 2.7248494625091553, + "learning_rate": 4.408070041872859e-06, + "loss": 1.5752, + "step": 618 + }, + { + "epoch": 0.6358500256805342, + "grad_norm": 2.5071568489074707, + "learning_rate": 4.406166730110393e-06, + "loss": 1.5178, + "step": 619 + }, + { + "epoch": 0.6368772470467385, + "grad_norm": 2.5583243370056152, + "learning_rate": 4.404263418347925e-06, + "loss": 1.5298, + "step": 620 + }, + { + "epoch": 0.6368772470467385, + "eval_loss": 1.833156943321228, + "eval_runtime": 37.1108, + "eval_samples_per_second": 26.946, + "eval_steps_per_second": 1.132, + "step": 620 + }, + { + "epoch": 0.637904468412943, + "grad_norm": 2.492332696914673, + "learning_rate": 4.402360106585459e-06, + "loss": 1.5443, + "step": 621 + }, + { + "epoch": 0.6389316897791474, + "grad_norm": 2.5734798908233643, + "learning_rate": 4.400456794822993e-06, + "loss": 1.5362, + "step": 622 + }, + { + "epoch": 0.6399589111453519, + "grad_norm": 2.7597742080688477, + "learning_rate": 4.398553483060526e-06, + "loss": 1.56, + "step": 623 + }, + { + "epoch": 0.6409861325115562, + "grad_norm": 2.6911630630493164, + "learning_rate": 4.396650171298059e-06, + "loss": 1.5138, + "step": 624 + }, + { + "epoch": 0.6420133538777607, + "grad_norm": 2.596557378768921, + "learning_rate": 4.394746859535593e-06, + "loss": 1.5381, + "step": 625 + }, + { + "epoch": 0.6430405752439651, + "grad_norm": 2.5787081718444824, + "learning_rate": 4.392843547773125e-06, + "loss": 1.5284, + "step": 626 + }, + { + "epoch": 0.6440677966101694, + "grad_norm": 2.5975289344787598, + "learning_rate": 4.390940236010659e-06, + "loss": 1.5359, + "step": 627 + }, + { + "epoch": 0.6450950179763739, + "grad_norm": 2.5175442695617676, + "learning_rate": 4.389036924248192e-06, + "loss": 1.5608, + "step": 628 + }, + { + "epoch": 0.6461222393425783, + "grad_norm": 2.535139322280884, + "learning_rate": 4.387133612485725e-06, + "loss": 1.5856, + "step": 629 + }, + { + "epoch": 0.6471494607087828, + "grad_norm": 2.5596396923065186, + "learning_rate": 4.385230300723259e-06, + "loss": 1.5467, + "step": 630 + }, + { + "epoch": 0.6471494607087828, + "eval_loss": 1.835809588432312, + "eval_runtime": 37.1073, + "eval_samples_per_second": 26.949, + "eval_steps_per_second": 1.132, + "step": 630 + }, + { + "epoch": 0.6481766820749871, + "grad_norm": 2.691441297531128, + "learning_rate": 4.383326988960792e-06, + "loss": 1.5516, + "step": 631 + }, + { + "epoch": 0.6492039034411916, + "grad_norm": 2.534902811050415, + "learning_rate": 4.3814236771983255e-06, + "loss": 1.5215, + "step": 632 + }, + { + "epoch": 0.650231124807396, + "grad_norm": 2.626211643218994, + "learning_rate": 4.379520365435859e-06, + "loss": 1.5811, + "step": 633 + }, + { + "epoch": 0.6512583461736005, + "grad_norm": 2.6646480560302734, + "learning_rate": 4.377617053673392e-06, + "loss": 1.5674, + "step": 634 + }, + { + "epoch": 0.6522855675398048, + "grad_norm": 2.5728375911712646, + "learning_rate": 4.3757137419109256e-06, + "loss": 1.5655, + "step": 635 + }, + { + "epoch": 0.6533127889060092, + "grad_norm": 2.72817325592041, + "learning_rate": 4.373810430148459e-06, + "loss": 1.6283, + "step": 636 + }, + { + "epoch": 0.6543400102722137, + "grad_norm": 2.5395050048828125, + "learning_rate": 4.3719071183859914e-06, + "loss": 1.5191, + "step": 637 + }, + { + "epoch": 0.655367231638418, + "grad_norm": 2.964163064956665, + "learning_rate": 4.370003806623525e-06, + "loss": 1.5839, + "step": 638 + }, + { + "epoch": 0.6563944530046225, + "grad_norm": 2.517214298248291, + "learning_rate": 4.368100494861058e-06, + "loss": 1.5278, + "step": 639 + }, + { + "epoch": 0.6574216743708269, + "grad_norm": 2.765145778656006, + "learning_rate": 4.3661971830985915e-06, + "loss": 1.5718, + "step": 640 + }, + { + "epoch": 0.6574216743708269, + "eval_loss": 1.8324981927871704, + "eval_runtime": 37.0883, + "eval_samples_per_second": 26.963, + "eval_steps_per_second": 1.132, + "step": 640 + }, + { + "epoch": 0.6584488957370314, + "grad_norm": 2.463439464569092, + "learning_rate": 4.364293871336126e-06, + "loss": 1.5378, + "step": 641 + }, + { + "epoch": 0.6594761171032357, + "grad_norm": 2.7756192684173584, + "learning_rate": 4.362390559573659e-06, + "loss": 1.5601, + "step": 642 + }, + { + "epoch": 0.6605033384694402, + "grad_norm": 2.4134342670440674, + "learning_rate": 4.3604872478111925e-06, + "loss": 1.5252, + "step": 643 + }, + { + "epoch": 0.6615305598356446, + "grad_norm": 2.4549670219421387, + "learning_rate": 4.358583936048725e-06, + "loss": 1.4602, + "step": 644 + }, + { + "epoch": 0.662557781201849, + "grad_norm": 2.6815311908721924, + "learning_rate": 4.356680624286258e-06, + "loss": 1.5865, + "step": 645 + }, + { + "epoch": 0.6635850025680534, + "grad_norm": 2.5216286182403564, + "learning_rate": 4.354777312523792e-06, + "loss": 1.5432, + "step": 646 + }, + { + "epoch": 0.6646122239342578, + "grad_norm": 2.5547335147857666, + "learning_rate": 4.352874000761325e-06, + "loss": 1.59, + "step": 647 + }, + { + "epoch": 0.6656394453004623, + "grad_norm": 2.587646484375, + "learning_rate": 4.3509706889988584e-06, + "loss": 1.5351, + "step": 648 + }, + { + "epoch": 0.6666666666666666, + "grad_norm": 2.530768871307373, + "learning_rate": 4.349067377236392e-06, + "loss": 1.603, + "step": 649 + }, + { + "epoch": 0.6676938880328711, + "grad_norm": 2.727213144302368, + "learning_rate": 4.347164065473925e-06, + "loss": 1.6032, + "step": 650 + }, + { + "epoch": 0.6676938880328711, + "eval_loss": 1.8441001176834106, + "eval_runtime": 37.0797, + "eval_samples_per_second": 26.969, + "eval_steps_per_second": 1.133, + "step": 650 + }, + { + "epoch": 0.6687211093990755, + "grad_norm": 2.750441312789917, + "learning_rate": 4.3452607537114585e-06, + "loss": 1.6172, + "step": 651 + }, + { + "epoch": 0.66974833076528, + "grad_norm": 2.571692705154419, + "learning_rate": 4.343357441948992e-06, + "loss": 1.528, + "step": 652 + }, + { + "epoch": 0.6707755521314843, + "grad_norm": 2.5415565967559814, + "learning_rate": 4.341454130186525e-06, + "loss": 1.5629, + "step": 653 + }, + { + "epoch": 0.6718027734976888, + "grad_norm": 2.4970591068267822, + "learning_rate": 4.339550818424058e-06, + "loss": 1.5299, + "step": 654 + }, + { + "epoch": 0.6728299948638932, + "grad_norm": 2.5283010005950928, + "learning_rate": 4.337647506661591e-06, + "loss": 1.5149, + "step": 655 + }, + { + "epoch": 0.6738572162300975, + "grad_norm": 2.5205166339874268, + "learning_rate": 4.3357441948991245e-06, + "loss": 1.5486, + "step": 656 + }, + { + "epoch": 0.674884437596302, + "grad_norm": 2.4621481895446777, + "learning_rate": 4.333840883136658e-06, + "loss": 1.5769, + "step": 657 + }, + { + "epoch": 0.6759116589625064, + "grad_norm": 2.5954174995422363, + "learning_rate": 4.331937571374191e-06, + "loss": 1.5242, + "step": 658 + }, + { + "epoch": 0.6769388803287109, + "grad_norm": 2.553523063659668, + "learning_rate": 4.3300342596117246e-06, + "loss": 1.5339, + "step": 659 + }, + { + "epoch": 0.6779661016949152, + "grad_norm": 2.5289323329925537, + "learning_rate": 4.328130947849258e-06, + "loss": 1.5371, + "step": 660 + }, + { + "epoch": 0.6779661016949152, + "eval_loss": 1.8562791347503662, + "eval_runtime": 37.1299, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 660 + }, + { + "epoch": 0.6789933230611197, + "grad_norm": 2.537010908126831, + "learning_rate": 4.326227636086791e-06, + "loss": 1.5005, + "step": 661 + }, + { + "epoch": 0.6800205444273241, + "grad_norm": 2.6716437339782715, + "learning_rate": 4.324324324324325e-06, + "loss": 1.5501, + "step": 662 + }, + { + "epoch": 0.6810477657935285, + "grad_norm": 2.4969863891601562, + "learning_rate": 4.322421012561858e-06, + "loss": 1.5474, + "step": 663 + }, + { + "epoch": 0.6820749871597329, + "grad_norm": 2.6803722381591797, + "learning_rate": 4.320517700799391e-06, + "loss": 1.5397, + "step": 664 + }, + { + "epoch": 0.6831022085259374, + "grad_norm": 2.6370725631713867, + "learning_rate": 4.318614389036925e-06, + "loss": 1.6274, + "step": 665 + }, + { + "epoch": 0.6841294298921418, + "grad_norm": 2.6329426765441895, + "learning_rate": 4.316711077274458e-06, + "loss": 1.6015, + "step": 666 + }, + { + "epoch": 0.6851566512583461, + "grad_norm": 2.6113548278808594, + "learning_rate": 4.3148077655119915e-06, + "loss": 1.5015, + "step": 667 + }, + { + "epoch": 0.6861838726245506, + "grad_norm": 2.718801498413086, + "learning_rate": 4.312904453749525e-06, + "loss": 1.597, + "step": 668 + }, + { + "epoch": 0.687211093990755, + "grad_norm": 2.638871908187866, + "learning_rate": 4.311001141987058e-06, + "loss": 1.5528, + "step": 669 + }, + { + "epoch": 0.6882383153569595, + "grad_norm": 2.74849534034729, + "learning_rate": 4.3090978302245916e-06, + "loss": 1.5627, + "step": 670 + }, + { + "epoch": 0.6882383153569595, + "eval_loss": 1.8493062257766724, + "eval_runtime": 37.2964, + "eval_samples_per_second": 26.812, + "eval_steps_per_second": 1.126, + "step": 670 + }, + { + "epoch": 0.6892655367231638, + "grad_norm": 2.6839728355407715, + "learning_rate": 4.307194518462124e-06, + "loss": 1.552, + "step": 671 + }, + { + "epoch": 0.6902927580893683, + "grad_norm": 2.844496488571167, + "learning_rate": 4.3052912066996574e-06, + "loss": 1.6218, + "step": 672 + }, + { + "epoch": 0.6913199794555727, + "grad_norm": 2.4792141914367676, + "learning_rate": 4.303387894937191e-06, + "loss": 1.592, + "step": 673 + }, + { + "epoch": 0.6923472008217771, + "grad_norm": 2.4852263927459717, + "learning_rate": 4.301484583174724e-06, + "loss": 1.5165, + "step": 674 + }, + { + "epoch": 0.6933744221879815, + "grad_norm": 2.5438406467437744, + "learning_rate": 4.2995812714122575e-06, + "loss": 1.5109, + "step": 675 + }, + { + "epoch": 0.6944016435541859, + "grad_norm": 2.5431931018829346, + "learning_rate": 4.297677959649791e-06, + "loss": 1.5345, + "step": 676 + }, + { + "epoch": 0.6954288649203904, + "grad_norm": 2.650897979736328, + "learning_rate": 4.295774647887324e-06, + "loss": 1.5607, + "step": 677 + }, + { + "epoch": 0.6964560862865947, + "grad_norm": 2.721167802810669, + "learning_rate": 4.293871336124858e-06, + "loss": 1.5764, + "step": 678 + }, + { + "epoch": 0.6974833076527992, + "grad_norm": 2.5402960777282715, + "learning_rate": 4.291968024362391e-06, + "loss": 1.5505, + "step": 679 + }, + { + "epoch": 0.6985105290190036, + "grad_norm": 2.6878864765167236, + "learning_rate": 4.290064712599924e-06, + "loss": 1.5821, + "step": 680 + }, + { + "epoch": 0.6985105290190036, + "eval_loss": 1.8400572538375854, + "eval_runtime": 37.3885, + "eval_samples_per_second": 26.746, + "eval_steps_per_second": 1.123, + "step": 680 + }, + { + "epoch": 0.699537750385208, + "grad_norm": 2.6703765392303467, + "learning_rate": 4.288161400837458e-06, + "loss": 1.5794, + "step": 681 + }, + { + "epoch": 0.7005649717514124, + "grad_norm": 2.626725196838379, + "learning_rate": 4.286258089074991e-06, + "loss": 1.4986, + "step": 682 + }, + { + "epoch": 0.7015921931176169, + "grad_norm": 2.6062862873077393, + "learning_rate": 4.284354777312524e-06, + "loss": 1.5606, + "step": 683 + }, + { + "epoch": 0.7026194144838213, + "grad_norm": 2.6359503269195557, + "learning_rate": 4.282451465550058e-06, + "loss": 1.5174, + "step": 684 + }, + { + "epoch": 0.7036466358500256, + "grad_norm": 2.8453221321105957, + "learning_rate": 4.280548153787591e-06, + "loss": 1.543, + "step": 685 + }, + { + "epoch": 0.7046738572162301, + "grad_norm": 2.5195140838623047, + "learning_rate": 4.2786448420251245e-06, + "loss": 1.5462, + "step": 686 + }, + { + "epoch": 0.7057010785824345, + "grad_norm": 2.757230281829834, + "learning_rate": 4.276741530262658e-06, + "loss": 1.5033, + "step": 687 + }, + { + "epoch": 0.706728299948639, + "grad_norm": 2.708148241043091, + "learning_rate": 4.27483821850019e-06, + "loss": 1.5286, + "step": 688 + }, + { + "epoch": 0.7077555213148433, + "grad_norm": 2.6836555004119873, + "learning_rate": 4.272934906737724e-06, + "loss": 1.5973, + "step": 689 + }, + { + "epoch": 0.7087827426810478, + "grad_norm": 2.6970176696777344, + "learning_rate": 4.271031594975257e-06, + "loss": 1.5844, + "step": 690 + }, + { + "epoch": 0.7087827426810478, + "eval_loss": 1.8414137363433838, + "eval_runtime": 37.3847, + "eval_samples_per_second": 26.749, + "eval_steps_per_second": 1.123, + "step": 690 + }, + { + "epoch": 0.7098099640472522, + "grad_norm": 2.69212007522583, + "learning_rate": 4.2691282832127905e-06, + "loss": 1.5817, + "step": 691 + }, + { + "epoch": 0.7108371854134566, + "grad_norm": 2.6203925609588623, + "learning_rate": 4.267224971450324e-06, + "loss": 1.5411, + "step": 692 + }, + { + "epoch": 0.711864406779661, + "grad_norm": 2.418916940689087, + "learning_rate": 4.265321659687857e-06, + "loss": 1.483, + "step": 693 + }, + { + "epoch": 0.7128916281458655, + "grad_norm": 2.784196615219116, + "learning_rate": 4.2634183479253906e-06, + "loss": 1.5286, + "step": 694 + }, + { + "epoch": 0.7139188495120699, + "grad_norm": 2.675043821334839, + "learning_rate": 4.261515036162924e-06, + "loss": 1.5067, + "step": 695 + }, + { + "epoch": 0.7149460708782742, + "grad_norm": 2.7075769901275635, + "learning_rate": 4.259611724400457e-06, + "loss": 1.5316, + "step": 696 + }, + { + "epoch": 0.7159732922444787, + "grad_norm": 2.911898612976074, + "learning_rate": 4.257708412637991e-06, + "loss": 1.5203, + "step": 697 + }, + { + "epoch": 0.7170005136106831, + "grad_norm": 2.748126983642578, + "learning_rate": 4.255805100875523e-06, + "loss": 1.465, + "step": 698 + }, + { + "epoch": 0.7180277349768875, + "grad_norm": 2.6017746925354004, + "learning_rate": 4.2539017891130565e-06, + "loss": 1.4956, + "step": 699 + }, + { + "epoch": 0.7190549563430919, + "grad_norm": 3.0877811908721924, + "learning_rate": 4.25199847735059e-06, + "loss": 1.587, + "step": 700 + }, + { + "epoch": 0.7190549563430919, + "eval_loss": 1.8209773302078247, + "eval_runtime": 37.1796, + "eval_samples_per_second": 26.896, + "eval_steps_per_second": 1.13, + "step": 700 + }, + { + "epoch": 0.7200821777092964, + "grad_norm": 2.6982834339141846, + "learning_rate": 4.250095165588123e-06, + "loss": 1.4984, + "step": 701 + }, + { + "epoch": 0.7211093990755008, + "grad_norm": 2.7425835132598877, + "learning_rate": 4.2481918538256575e-06, + "loss": 1.5596, + "step": 702 + }, + { + "epoch": 0.7221366204417052, + "grad_norm": 2.838890314102173, + "learning_rate": 4.246288542063191e-06, + "loss": 1.5207, + "step": 703 + }, + { + "epoch": 0.7231638418079096, + "grad_norm": 2.694871664047241, + "learning_rate": 4.244385230300724e-06, + "loss": 1.5592, + "step": 704 + }, + { + "epoch": 0.724191063174114, + "grad_norm": 2.630206823348999, + "learning_rate": 4.242481918538257e-06, + "loss": 1.5005, + "step": 705 + }, + { + "epoch": 0.7252182845403184, + "grad_norm": 2.9697742462158203, + "learning_rate": 4.24057860677579e-06, + "loss": 1.5329, + "step": 706 + }, + { + "epoch": 0.7262455059065228, + "grad_norm": 2.8384499549865723, + "learning_rate": 4.2386752950133234e-06, + "loss": 1.5731, + "step": 707 + }, + { + "epoch": 0.7272727272727273, + "grad_norm": 2.530134916305542, + "learning_rate": 4.236771983250857e-06, + "loss": 1.5187, + "step": 708 + }, + { + "epoch": 0.7282999486389317, + "grad_norm": 2.601384162902832, + "learning_rate": 4.23486867148839e-06, + "loss": 1.5069, + "step": 709 + }, + { + "epoch": 0.7293271700051361, + "grad_norm": 2.7746055126190186, + "learning_rate": 4.2329653597259235e-06, + "loss": 1.5377, + "step": 710 + }, + { + "epoch": 0.7293271700051361, + "eval_loss": 1.8144561052322388, + "eval_runtime": 36.9983, + "eval_samples_per_second": 27.028, + "eval_steps_per_second": 1.135, + "step": 710 + }, + { + "epoch": 0.7303543913713405, + "grad_norm": 2.860414505004883, + "learning_rate": 4.231062047963457e-06, + "loss": 1.5604, + "step": 711 + }, + { + "epoch": 0.731381612737545, + "grad_norm": 2.7479686737060547, + "learning_rate": 4.22915873620099e-06, + "loss": 1.578, + "step": 712 + }, + { + "epoch": 0.7324088341037494, + "grad_norm": 3.093319892883301, + "learning_rate": 4.227255424438524e-06, + "loss": 1.5609, + "step": 713 + }, + { + "epoch": 0.7334360554699538, + "grad_norm": 2.611778497695923, + "learning_rate": 4.225352112676057e-06, + "loss": 1.5775, + "step": 714 + }, + { + "epoch": 0.7344632768361582, + "grad_norm": 2.5740373134613037, + "learning_rate": 4.2234488009135895e-06, + "loss": 1.5059, + "step": 715 + }, + { + "epoch": 0.7354904982023626, + "grad_norm": 2.5564701557159424, + "learning_rate": 4.221545489151123e-06, + "loss": 1.4687, + "step": 716 + }, + { + "epoch": 0.736517719568567, + "grad_norm": 2.566359758377075, + "learning_rate": 4.219642177388656e-06, + "loss": 1.5084, + "step": 717 + }, + { + "epoch": 0.7375449409347714, + "grad_norm": 2.632016897201538, + "learning_rate": 4.2177388656261896e-06, + "loss": 1.5449, + "step": 718 + }, + { + "epoch": 0.7385721623009759, + "grad_norm": 2.525616407394409, + "learning_rate": 4.215835553863723e-06, + "loss": 1.4882, + "step": 719 + }, + { + "epoch": 0.7395993836671803, + "grad_norm": 2.674612522125244, + "learning_rate": 4.213932242101256e-06, + "loss": 1.4576, + "step": 720 + }, + { + "epoch": 0.7395993836671803, + "eval_loss": 1.8278547525405884, + "eval_runtime": 37.0938, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 720 + }, + { + "epoch": 0.7406266050333847, + "grad_norm": 2.8636388778686523, + "learning_rate": 4.21202893033879e-06, + "loss": 1.5149, + "step": 721 + }, + { + "epoch": 0.7416538263995891, + "grad_norm": 2.7943782806396484, + "learning_rate": 4.210125618576323e-06, + "loss": 1.5409, + "step": 722 + }, + { + "epoch": 0.7426810477657936, + "grad_norm": 2.5911595821380615, + "learning_rate": 4.208222306813856e-06, + "loss": 1.5324, + "step": 723 + }, + { + "epoch": 0.743708269131998, + "grad_norm": 2.7497329711914062, + "learning_rate": 4.20631899505139e-06, + "loss": 1.5842, + "step": 724 + }, + { + "epoch": 0.7447354904982023, + "grad_norm": 2.649540424346924, + "learning_rate": 4.204415683288923e-06, + "loss": 1.438, + "step": 725 + }, + { + "epoch": 0.7457627118644068, + "grad_norm": 2.696852684020996, + "learning_rate": 4.2025123715264565e-06, + "loss": 1.5538, + "step": 726 + }, + { + "epoch": 0.7467899332306112, + "grad_norm": 2.553107738494873, + "learning_rate": 4.20060905976399e-06, + "loss": 1.5115, + "step": 727 + }, + { + "epoch": 0.7478171545968156, + "grad_norm": 2.559718370437622, + "learning_rate": 4.198705748001523e-06, + "loss": 1.4613, + "step": 728 + }, + { + "epoch": 0.74884437596302, + "grad_norm": 2.615692377090454, + "learning_rate": 4.1968024362390566e-06, + "loss": 1.5655, + "step": 729 + }, + { + "epoch": 0.7498715973292245, + "grad_norm": 2.432598114013672, + "learning_rate": 4.19489912447659e-06, + "loss": 1.5192, + "step": 730 + }, + { + "epoch": 0.7498715973292245, + "eval_loss": 1.8231704235076904, + "eval_runtime": 37.1648, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 730 + }, + { + "epoch": 0.7508988186954288, + "grad_norm": 2.63838529586792, + "learning_rate": 4.192995812714123e-06, + "loss": 1.5888, + "step": 731 + }, + { + "epoch": 0.7519260400616333, + "grad_norm": 2.711975574493408, + "learning_rate": 4.191092500951656e-06, + "loss": 1.5309, + "step": 732 + }, + { + "epoch": 0.7529532614278377, + "grad_norm": 2.459561347961426, + "learning_rate": 4.189189189189189e-06, + "loss": 1.5141, + "step": 733 + }, + { + "epoch": 0.7539804827940421, + "grad_norm": 2.7000269889831543, + "learning_rate": 4.1872858774267225e-06, + "loss": 1.5148, + "step": 734 + }, + { + "epoch": 0.7550077041602465, + "grad_norm": 2.8983139991760254, + "learning_rate": 4.185382565664256e-06, + "loss": 1.5766, + "step": 735 + }, + { + "epoch": 0.7560349255264509, + "grad_norm": 2.600132942199707, + "learning_rate": 4.183479253901789e-06, + "loss": 1.4987, + "step": 736 + }, + { + "epoch": 0.7570621468926554, + "grad_norm": 2.579481363296509, + "learning_rate": 4.181575942139323e-06, + "loss": 1.4761, + "step": 737 + }, + { + "epoch": 0.7580893682588598, + "grad_norm": 2.708799123764038, + "learning_rate": 4.179672630376856e-06, + "loss": 1.5278, + "step": 738 + }, + { + "epoch": 0.7591165896250642, + "grad_norm": 2.4836950302124023, + "learning_rate": 4.177769318614389e-06, + "loss": 1.5056, + "step": 739 + }, + { + "epoch": 0.7601438109912686, + "grad_norm": 2.4556121826171875, + "learning_rate": 4.175866006851923e-06, + "loss": 1.5637, + "step": 740 + }, + { + "epoch": 0.7601438109912686, + "eval_loss": 1.8386361598968506, + "eval_runtime": 37.1304, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 740 + }, + { + "epoch": 0.7611710323574731, + "grad_norm": 2.504653215408325, + "learning_rate": 4.173962695089456e-06, + "loss": 1.5519, + "step": 741 + }, + { + "epoch": 0.7621982537236774, + "grad_norm": 2.50435733795166, + "learning_rate": 4.1720593833269894e-06, + "loss": 1.5379, + "step": 742 + }, + { + "epoch": 0.7632254750898819, + "grad_norm": 2.5818099975585938, + "learning_rate": 4.170156071564523e-06, + "loss": 1.5462, + "step": 743 + }, + { + "epoch": 0.7642526964560863, + "grad_norm": 2.6530778408050537, + "learning_rate": 4.168252759802056e-06, + "loss": 1.6112, + "step": 744 + }, + { + "epoch": 0.7652799178222907, + "grad_norm": 2.463993549346924, + "learning_rate": 4.1663494480395895e-06, + "loss": 1.5174, + "step": 745 + }, + { + "epoch": 0.7663071391884951, + "grad_norm": 2.6928186416625977, + "learning_rate": 4.164446136277123e-06, + "loss": 1.5639, + "step": 746 + }, + { + "epoch": 0.7673343605546995, + "grad_norm": 2.5539324283599854, + "learning_rate": 4.162542824514656e-06, + "loss": 1.5557, + "step": 747 + }, + { + "epoch": 0.768361581920904, + "grad_norm": 2.6068854331970215, + "learning_rate": 4.16063951275219e-06, + "loss": 1.5309, + "step": 748 + }, + { + "epoch": 0.7693888032871083, + "grad_norm": 2.5830016136169434, + "learning_rate": 4.158736200989722e-06, + "loss": 1.4876, + "step": 749 + }, + { + "epoch": 0.7704160246533128, + "grad_norm": 2.6156973838806152, + "learning_rate": 4.1568328892272555e-06, + "loss": 1.5284, + "step": 750 + }, + { + "epoch": 0.7704160246533128, + "eval_loss": 1.8401216268539429, + "eval_runtime": 37.1657, + "eval_samples_per_second": 26.907, + "eval_steps_per_second": 1.13, + "step": 750 + }, + { + "epoch": 0.7714432460195172, + "grad_norm": 2.6343140602111816, + "learning_rate": 4.154929577464789e-06, + "loss": 1.5348, + "step": 751 + }, + { + "epoch": 0.7724704673857217, + "grad_norm": 2.483319044113159, + "learning_rate": 4.153026265702322e-06, + "loss": 1.5344, + "step": 752 + }, + { + "epoch": 0.773497688751926, + "grad_norm": 2.755953550338745, + "learning_rate": 4.1511229539398556e-06, + "loss": 1.4715, + "step": 753 + }, + { + "epoch": 0.7745249101181304, + "grad_norm": 2.5264573097229004, + "learning_rate": 4.149219642177389e-06, + "loss": 1.5236, + "step": 754 + }, + { + "epoch": 0.7755521314843349, + "grad_norm": 2.793328046798706, + "learning_rate": 4.147316330414922e-06, + "loss": 1.5316, + "step": 755 + }, + { + "epoch": 0.7765793528505393, + "grad_norm": 2.7607345581054688, + "learning_rate": 4.145413018652456e-06, + "loss": 1.4518, + "step": 756 + }, + { + "epoch": 0.7776065742167437, + "grad_norm": 2.5960099697113037, + "learning_rate": 4.143509706889989e-06, + "loss": 1.5499, + "step": 757 + }, + { + "epoch": 0.7786337955829481, + "grad_norm": 2.752039909362793, + "learning_rate": 4.141606395127522e-06, + "loss": 1.5433, + "step": 758 + }, + { + "epoch": 0.7796610169491526, + "grad_norm": 2.724228620529175, + "learning_rate": 4.139703083365056e-06, + "loss": 1.5242, + "step": 759 + }, + { + "epoch": 0.7806882383153569, + "grad_norm": 2.668635368347168, + "learning_rate": 4.137799771602588e-06, + "loss": 1.563, + "step": 760 + }, + { + "epoch": 0.7806882383153569, + "eval_loss": 1.8386328220367432, + "eval_runtime": 37.1498, + "eval_samples_per_second": 26.918, + "eval_steps_per_second": 1.131, + "step": 760 + }, + { + "epoch": 0.7817154596815614, + "grad_norm": 2.5619916915893555, + "learning_rate": 4.135896459840122e-06, + "loss": 1.5398, + "step": 761 + }, + { + "epoch": 0.7827426810477658, + "grad_norm": 2.642075538635254, + "learning_rate": 4.133993148077656e-06, + "loss": 1.4938, + "step": 762 + }, + { + "epoch": 0.7837699024139703, + "grad_norm": 2.700350522994995, + "learning_rate": 4.132089836315189e-06, + "loss": 1.5196, + "step": 763 + }, + { + "epoch": 0.7847971237801746, + "grad_norm": 2.9511470794677734, + "learning_rate": 4.1301865245527226e-06, + "loss": 1.531, + "step": 764 + }, + { + "epoch": 0.785824345146379, + "grad_norm": 2.681643486022949, + "learning_rate": 4.128283212790256e-06, + "loss": 1.4747, + "step": 765 + }, + { + "epoch": 0.7868515665125835, + "grad_norm": 2.5386908054351807, + "learning_rate": 4.1263799010277884e-06, + "loss": 1.4424, + "step": 766 + }, + { + "epoch": 0.7878787878787878, + "grad_norm": 2.8831303119659424, + "learning_rate": 4.124476589265322e-06, + "loss": 1.5677, + "step": 767 + }, + { + "epoch": 0.7889060092449923, + "grad_norm": 2.593442678451538, + "learning_rate": 4.122573277502855e-06, + "loss": 1.5043, + "step": 768 + }, + { + "epoch": 0.7899332306111967, + "grad_norm": 2.6778204441070557, + "learning_rate": 4.1206699657403885e-06, + "loss": 1.5053, + "step": 769 + }, + { + "epoch": 0.7909604519774012, + "grad_norm": 2.6590428352355957, + "learning_rate": 4.118766653977922e-06, + "loss": 1.5338, + "step": 770 + }, + { + "epoch": 0.7909604519774012, + "eval_loss": 1.8266831636428833, + "eval_runtime": 37.1428, + "eval_samples_per_second": 26.923, + "eval_steps_per_second": 1.131, + "step": 770 + }, + { + "epoch": 0.7919876733436055, + "grad_norm": 2.6895735263824463, + "learning_rate": 4.116863342215455e-06, + "loss": 1.5067, + "step": 771 + }, + { + "epoch": 0.79301489470981, + "grad_norm": 2.5284390449523926, + "learning_rate": 4.114960030452989e-06, + "loss": 1.4775, + "step": 772 + }, + { + "epoch": 0.7940421160760144, + "grad_norm": 2.610442876815796, + "learning_rate": 4.113056718690522e-06, + "loss": 1.519, + "step": 773 + }, + { + "epoch": 0.7950693374422187, + "grad_norm": 2.5558695793151855, + "learning_rate": 4.111153406928055e-06, + "loss": 1.4565, + "step": 774 + }, + { + "epoch": 0.7960965588084232, + "grad_norm": 2.6313698291778564, + "learning_rate": 4.109250095165589e-06, + "loss": 1.5626, + "step": 775 + }, + { + "epoch": 0.7971237801746276, + "grad_norm": 2.553133726119995, + "learning_rate": 4.107346783403122e-06, + "loss": 1.5591, + "step": 776 + }, + { + "epoch": 0.7981510015408321, + "grad_norm": 2.6599066257476807, + "learning_rate": 4.105443471640655e-06, + "loss": 1.5064, + "step": 777 + }, + { + "epoch": 0.7991782229070364, + "grad_norm": 2.5582847595214844, + "learning_rate": 4.103540159878188e-06, + "loss": 1.4824, + "step": 778 + }, + { + "epoch": 0.8002054442732409, + "grad_norm": 2.658006429672241, + "learning_rate": 4.101636848115721e-06, + "loss": 1.5187, + "step": 779 + }, + { + "epoch": 0.8012326656394453, + "grad_norm": 2.5531060695648193, + "learning_rate": 4.099733536353255e-06, + "loss": 1.4862, + "step": 780 + }, + { + "epoch": 0.8012326656394453, + "eval_loss": 1.8293416500091553, + "eval_runtime": 37.2532, + "eval_samples_per_second": 26.843, + "eval_steps_per_second": 1.127, + "step": 780 + }, + { + "epoch": 0.8022598870056498, + "grad_norm": 2.7279789447784424, + "learning_rate": 4.097830224590788e-06, + "loss": 1.5299, + "step": 781 + }, + { + "epoch": 0.8032871083718541, + "grad_norm": 2.7395598888397217, + "learning_rate": 4.095926912828322e-06, + "loss": 1.6016, + "step": 782 + }, + { + "epoch": 0.8043143297380585, + "grad_norm": 2.6175215244293213, + "learning_rate": 4.094023601065855e-06, + "loss": 1.5657, + "step": 783 + }, + { + "epoch": 0.805341551104263, + "grad_norm": 2.4423458576202393, + "learning_rate": 4.092120289303388e-06, + "loss": 1.5085, + "step": 784 + }, + { + "epoch": 0.8063687724704673, + "grad_norm": 2.569363594055176, + "learning_rate": 4.0902169775409215e-06, + "loss": 1.5272, + "step": 785 + }, + { + "epoch": 0.8073959938366718, + "grad_norm": 2.5281481742858887, + "learning_rate": 4.088313665778455e-06, + "loss": 1.473, + "step": 786 + }, + { + "epoch": 0.8084232152028762, + "grad_norm": 2.527724504470825, + "learning_rate": 4.086410354015988e-06, + "loss": 1.4978, + "step": 787 + }, + { + "epoch": 0.8094504365690807, + "grad_norm": 2.524348497390747, + "learning_rate": 4.0845070422535216e-06, + "loss": 1.4818, + "step": 788 + }, + { + "epoch": 0.810477657935285, + "grad_norm": 2.5176353454589844, + "learning_rate": 4.082603730491055e-06, + "loss": 1.4946, + "step": 789 + }, + { + "epoch": 0.8115048793014895, + "grad_norm": 2.63879132270813, + "learning_rate": 4.080700418728588e-06, + "loss": 1.5411, + "step": 790 + }, + { + "epoch": 0.8115048793014895, + "eval_loss": 1.8180036544799805, + "eval_runtime": 37.4282, + "eval_samples_per_second": 26.718, + "eval_steps_per_second": 1.122, + "step": 790 + }, + { + "epoch": 0.8125321006676939, + "grad_norm": 2.5492401123046875, + "learning_rate": 4.078797106966122e-06, + "loss": 1.5185, + "step": 791 + }, + { + "epoch": 0.8135593220338984, + "grad_norm": 2.668663501739502, + "learning_rate": 4.076893795203655e-06, + "loss": 1.4187, + "step": 792 + }, + { + "epoch": 0.8145865434001027, + "grad_norm": 2.500401496887207, + "learning_rate": 4.074990483441188e-06, + "loss": 1.4902, + "step": 793 + }, + { + "epoch": 0.8156137647663071, + "grad_norm": 2.7383482456207275, + "learning_rate": 4.073087171678721e-06, + "loss": 1.5454, + "step": 794 + }, + { + "epoch": 0.8166409861325116, + "grad_norm": 2.7144806385040283, + "learning_rate": 4.071183859916254e-06, + "loss": 1.5336, + "step": 795 + }, + { + "epoch": 0.8176682074987159, + "grad_norm": 2.8088390827178955, + "learning_rate": 4.069280548153788e-06, + "loss": 1.5107, + "step": 796 + }, + { + "epoch": 0.8186954288649204, + "grad_norm": 2.786731004714966, + "learning_rate": 4.067377236391321e-06, + "loss": 1.4762, + "step": 797 + }, + { + "epoch": 0.8197226502311248, + "grad_norm": 2.7185654640197754, + "learning_rate": 4.065473924628854e-06, + "loss": 1.5129, + "step": 798 + }, + { + "epoch": 0.8207498715973293, + "grad_norm": 2.5182769298553467, + "learning_rate": 4.063570612866388e-06, + "loss": 1.5392, + "step": 799 + }, + { + "epoch": 0.8217770929635336, + "grad_norm": 2.594088554382324, + "learning_rate": 4.061667301103921e-06, + "loss": 1.5389, + "step": 800 + }, + { + "epoch": 0.8217770929635336, + "eval_loss": 1.8078373670578003, + "eval_runtime": 37.4191, + "eval_samples_per_second": 26.724, + "eval_steps_per_second": 1.122, + "step": 800 + }, + { + "epoch": 0.8228043143297381, + "grad_norm": 2.5099830627441406, + "learning_rate": 4.0597639893414544e-06, + "loss": 1.5033, + "step": 801 + }, + { + "epoch": 0.8238315356959425, + "grad_norm": 2.514415740966797, + "learning_rate": 4.057860677578988e-06, + "loss": 1.5152, + "step": 802 + }, + { + "epoch": 0.8248587570621468, + "grad_norm": 2.738563060760498, + "learning_rate": 4.055957365816521e-06, + "loss": 1.5302, + "step": 803 + }, + { + "epoch": 0.8258859784283513, + "grad_norm": 2.633472204208374, + "learning_rate": 4.0540540540540545e-06, + "loss": 1.4735, + "step": 804 + }, + { + "epoch": 0.8269131997945557, + "grad_norm": 2.6635546684265137, + "learning_rate": 4.052150742291588e-06, + "loss": 1.4675, + "step": 805 + }, + { + "epoch": 0.8279404211607602, + "grad_norm": 2.515171527862549, + "learning_rate": 4.050247430529121e-06, + "loss": 1.4682, + "step": 806 + }, + { + "epoch": 0.8289676425269645, + "grad_norm": 2.689464569091797, + "learning_rate": 4.048344118766655e-06, + "loss": 1.5499, + "step": 807 + }, + { + "epoch": 0.829994863893169, + "grad_norm": 2.639822244644165, + "learning_rate": 4.046440807004188e-06, + "loss": 1.5297, + "step": 808 + }, + { + "epoch": 0.8310220852593734, + "grad_norm": 2.6470797061920166, + "learning_rate": 4.044537495241721e-06, + "loss": 1.5215, + "step": 809 + }, + { + "epoch": 0.8320493066255779, + "grad_norm": 2.4769506454467773, + "learning_rate": 4.042634183479255e-06, + "loss": 1.4859, + "step": 810 + }, + { + "epoch": 0.8320493066255779, + "eval_loss": 1.8010139465332031, + "eval_runtime": 36.974, + "eval_samples_per_second": 27.046, + "eval_steps_per_second": 1.136, + "step": 810 + }, + { + "epoch": 0.8330765279917822, + "grad_norm": 2.72485613822937, + "learning_rate": 4.040730871716787e-06, + "loss": 1.5301, + "step": 811 + }, + { + "epoch": 0.8341037493579867, + "grad_norm": 2.522390365600586, + "learning_rate": 4.038827559954321e-06, + "loss": 1.5199, + "step": 812 + }, + { + "epoch": 0.8351309707241911, + "grad_norm": 2.6865367889404297, + "learning_rate": 4.036924248191854e-06, + "loss": 1.5605, + "step": 813 + }, + { + "epoch": 0.8361581920903954, + "grad_norm": 2.6478004455566406, + "learning_rate": 4.035020936429387e-06, + "loss": 1.5291, + "step": 814 + }, + { + "epoch": 0.8371854134565999, + "grad_norm": 2.632477045059204, + "learning_rate": 4.033117624666921e-06, + "loss": 1.5134, + "step": 815 + }, + { + "epoch": 0.8382126348228043, + "grad_norm": 2.683198928833008, + "learning_rate": 4.031214312904454e-06, + "loss": 1.4787, + "step": 816 + }, + { + "epoch": 0.8392398561890088, + "grad_norm": 2.536470651626587, + "learning_rate": 4.029311001141987e-06, + "loss": 1.4717, + "step": 817 + }, + { + "epoch": 0.8402670775552131, + "grad_norm": 2.5733802318573, + "learning_rate": 4.027407689379521e-06, + "loss": 1.4507, + "step": 818 + }, + { + "epoch": 0.8412942989214176, + "grad_norm": 2.779752016067505, + "learning_rate": 4.025504377617054e-06, + "loss": 1.503, + "step": 819 + }, + { + "epoch": 0.842321520287622, + "grad_norm": 2.6932976245880127, + "learning_rate": 4.0236010658545875e-06, + "loss": 1.4815, + "step": 820 + }, + { + "epoch": 0.842321520287622, + "eval_loss": 1.7960872650146484, + "eval_runtime": 37.0481, + "eval_samples_per_second": 26.992, + "eval_steps_per_second": 1.134, + "step": 820 + }, + { + "epoch": 0.8433487416538265, + "grad_norm": 2.7531607151031494, + "learning_rate": 4.02169775409212e-06, + "loss": 1.5286, + "step": 821 + }, + { + "epoch": 0.8443759630200308, + "grad_norm": 2.8817851543426514, + "learning_rate": 4.019794442329653e-06, + "loss": 1.5765, + "step": 822 + }, + { + "epoch": 0.8454031843862352, + "grad_norm": 2.544451951980591, + "learning_rate": 4.0178911305671876e-06, + "loss": 1.452, + "step": 823 + }, + { + "epoch": 0.8464304057524397, + "grad_norm": 2.8142952919006348, + "learning_rate": 4.015987818804721e-06, + "loss": 1.5374, + "step": 824 + }, + { + "epoch": 0.847457627118644, + "grad_norm": 2.5938656330108643, + "learning_rate": 4.014084507042254e-06, + "loss": 1.4756, + "step": 825 + }, + { + "epoch": 0.8484848484848485, + "grad_norm": 2.7007956504821777, + "learning_rate": 4.012181195279788e-06, + "loss": 1.445, + "step": 826 + }, + { + "epoch": 0.8495120698510529, + "grad_norm": 2.9885144233703613, + "learning_rate": 4.010277883517321e-06, + "loss": 1.5516, + "step": 827 + }, + { + "epoch": 0.8505392912172574, + "grad_norm": 2.814607620239258, + "learning_rate": 4.0083745717548535e-06, + "loss": 1.5375, + "step": 828 + }, + { + "epoch": 0.8515665125834617, + "grad_norm": 2.9187042713165283, + "learning_rate": 4.006471259992387e-06, + "loss": 1.496, + "step": 829 + }, + { + "epoch": 0.8525937339496662, + "grad_norm": 2.7408974170684814, + "learning_rate": 4.00456794822992e-06, + "loss": 1.5195, + "step": 830 + }, + { + "epoch": 0.8525937339496662, + "eval_loss": 1.795318365097046, + "eval_runtime": 37.1058, + "eval_samples_per_second": 26.95, + "eval_steps_per_second": 1.132, + "step": 830 + }, + { + "epoch": 0.8536209553158706, + "grad_norm": 2.6824872493743896, + "learning_rate": 4.002664636467454e-06, + "loss": 1.5568, + "step": 831 + }, + { + "epoch": 0.8546481766820749, + "grad_norm": 2.7598671913146973, + "learning_rate": 4.000761324704987e-06, + "loss": 1.5134, + "step": 832 + }, + { + "epoch": 0.8556753980482794, + "grad_norm": 2.728449583053589, + "learning_rate": 3.99885801294252e-06, + "loss": 1.4675, + "step": 833 + }, + { + "epoch": 0.8567026194144838, + "grad_norm": 2.900674343109131, + "learning_rate": 3.996954701180054e-06, + "loss": 1.5259, + "step": 834 + }, + { + "epoch": 0.8577298407806883, + "grad_norm": 2.818105697631836, + "learning_rate": 3.995051389417587e-06, + "loss": 1.4533, + "step": 835 + }, + { + "epoch": 0.8587570621468926, + "grad_norm": 2.791207790374756, + "learning_rate": 3.9931480776551204e-06, + "loss": 1.4692, + "step": 836 + }, + { + "epoch": 0.8597842835130971, + "grad_norm": 2.584268808364868, + "learning_rate": 3.991244765892654e-06, + "loss": 1.554, + "step": 837 + }, + { + "epoch": 0.8608115048793015, + "grad_norm": 2.5646660327911377, + "learning_rate": 3.989341454130186e-06, + "loss": 1.4782, + "step": 838 + }, + { + "epoch": 0.861838726245506, + "grad_norm": 2.4912099838256836, + "learning_rate": 3.98743814236772e-06, + "loss": 1.4494, + "step": 839 + }, + { + "epoch": 0.8628659476117103, + "grad_norm": 2.4632041454315186, + "learning_rate": 3.985534830605253e-06, + "loss": 1.4911, + "step": 840 + }, + { + "epoch": 0.8628659476117103, + "eval_loss": 1.7955427169799805, + "eval_runtime": 37.0188, + "eval_samples_per_second": 27.013, + "eval_steps_per_second": 1.135, + "step": 840 + }, + { + "epoch": 0.8638931689779148, + "grad_norm": 2.6388566493988037, + "learning_rate": 3.983631518842786e-06, + "loss": 1.5135, + "step": 841 + }, + { + "epoch": 0.8649203903441192, + "grad_norm": 2.7619292736053467, + "learning_rate": 3.98172820708032e-06, + "loss": 1.5279, + "step": 842 + }, + { + "epoch": 0.8659476117103235, + "grad_norm": 2.510636806488037, + "learning_rate": 3.979824895317854e-06, + "loss": 1.4815, + "step": 843 + }, + { + "epoch": 0.866974833076528, + "grad_norm": 2.6847214698791504, + "learning_rate": 3.977921583555387e-06, + "loss": 1.4556, + "step": 844 + }, + { + "epoch": 0.8680020544427324, + "grad_norm": 2.5366628170013428, + "learning_rate": 3.97601827179292e-06, + "loss": 1.5297, + "step": 845 + }, + { + "epoch": 0.8690292758089369, + "grad_norm": 2.6018271446228027, + "learning_rate": 3.974114960030453e-06, + "loss": 1.4642, + "step": 846 + }, + { + "epoch": 0.8700564971751412, + "grad_norm": 2.612107276916504, + "learning_rate": 3.9722116482679866e-06, + "loss": 1.5065, + "step": 847 + }, + { + "epoch": 0.8710837185413457, + "grad_norm": 2.6163978576660156, + "learning_rate": 3.97030833650552e-06, + "loss": 1.4913, + "step": 848 + }, + { + "epoch": 0.8721109399075501, + "grad_norm": 2.516697645187378, + "learning_rate": 3.968405024743053e-06, + "loss": 1.5006, + "step": 849 + }, + { + "epoch": 0.8731381612737545, + "grad_norm": 2.6012990474700928, + "learning_rate": 3.966501712980587e-06, + "loss": 1.4276, + "step": 850 + }, + { + "epoch": 0.8731381612737545, + "eval_loss": 1.807686448097229, + "eval_runtime": 37.0357, + "eval_samples_per_second": 27.001, + "eval_steps_per_second": 1.134, + "step": 850 + }, + { + "epoch": 0.8741653826399589, + "grad_norm": 2.6806578636169434, + "learning_rate": 3.96459840121812e-06, + "loss": 1.4849, + "step": 851 + }, + { + "epoch": 0.8751926040061633, + "grad_norm": 2.5139150619506836, + "learning_rate": 3.962695089455653e-06, + "loss": 1.4919, + "step": 852 + }, + { + "epoch": 0.8762198253723678, + "grad_norm": 2.779398202896118, + "learning_rate": 3.960791777693187e-06, + "loss": 1.5446, + "step": 853 + }, + { + "epoch": 0.8772470467385721, + "grad_norm": 2.7080724239349365, + "learning_rate": 3.95888846593072e-06, + "loss": 1.5215, + "step": 854 + }, + { + "epoch": 0.8782742681047766, + "grad_norm": 2.658066749572754, + "learning_rate": 3.956985154168253e-06, + "loss": 1.5315, + "step": 855 + }, + { + "epoch": 0.879301489470981, + "grad_norm": 2.6813952922821045, + "learning_rate": 3.955081842405786e-06, + "loss": 1.549, + "step": 856 + }, + { + "epoch": 0.8803287108371854, + "grad_norm": 2.744917869567871, + "learning_rate": 3.953178530643319e-06, + "loss": 1.5717, + "step": 857 + }, + { + "epoch": 0.8813559322033898, + "grad_norm": 2.869288921356201, + "learning_rate": 3.951275218880853e-06, + "loss": 1.5882, + "step": 858 + }, + { + "epoch": 0.8823831535695943, + "grad_norm": 2.633666515350342, + "learning_rate": 3.949371907118386e-06, + "loss": 1.4332, + "step": 859 + }, + { + "epoch": 0.8834103749357987, + "grad_norm": 2.7296335697174072, + "learning_rate": 3.9474685953559194e-06, + "loss": 1.5349, + "step": 860 + }, + { + "epoch": 0.8834103749357987, + "eval_loss": 1.80648934841156, + "eval_runtime": 37.0387, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 860 + }, + { + "epoch": 0.884437596302003, + "grad_norm": 3.0827677249908447, + "learning_rate": 3.945565283593453e-06, + "loss": 1.5108, + "step": 861 + }, + { + "epoch": 0.8854648176682075, + "grad_norm": 2.81630539894104, + "learning_rate": 3.943661971830986e-06, + "loss": 1.493, + "step": 862 + }, + { + "epoch": 0.8864920390344119, + "grad_norm": 2.804215908050537, + "learning_rate": 3.9417586600685195e-06, + "loss": 1.5432, + "step": 863 + }, + { + "epoch": 0.8875192604006163, + "grad_norm": 2.9585464000701904, + "learning_rate": 3.939855348306053e-06, + "loss": 1.5202, + "step": 864 + }, + { + "epoch": 0.8885464817668207, + "grad_norm": 2.717324733734131, + "learning_rate": 3.937952036543586e-06, + "loss": 1.5172, + "step": 865 + }, + { + "epoch": 0.8895737031330252, + "grad_norm": 2.62406325340271, + "learning_rate": 3.93604872478112e-06, + "loss": 1.5064, + "step": 866 + }, + { + "epoch": 0.8906009244992296, + "grad_norm": 3.2518985271453857, + "learning_rate": 3.934145413018653e-06, + "loss": 1.5159, + "step": 867 + }, + { + "epoch": 0.891628145865434, + "grad_norm": 2.6304733753204346, + "learning_rate": 3.932242101256186e-06, + "loss": 1.4723, + "step": 868 + }, + { + "epoch": 0.8926553672316384, + "grad_norm": 2.6788930892944336, + "learning_rate": 3.93033878949372e-06, + "loss": 1.4759, + "step": 869 + }, + { + "epoch": 0.8936825885978429, + "grad_norm": 2.8033249378204346, + "learning_rate": 3.928435477731253e-06, + "loss": 1.4435, + "step": 870 + }, + { + "epoch": 0.8936825885978429, + "eval_loss": 1.8105196952819824, + "eval_runtime": 37.062, + "eval_samples_per_second": 26.982, + "eval_steps_per_second": 1.133, + "step": 870 + }, + { + "epoch": 0.8947098099640473, + "grad_norm": 2.8569960594177246, + "learning_rate": 3.9265321659687864e-06, + "loss": 1.5424, + "step": 871 + }, + { + "epoch": 0.8957370313302516, + "grad_norm": 2.8037121295928955, + "learning_rate": 3.924628854206319e-06, + "loss": 1.5368, + "step": 872 + }, + { + "epoch": 0.8967642526964561, + "grad_norm": 2.7444164752960205, + "learning_rate": 3.922725542443852e-06, + "loss": 1.5102, + "step": 873 + }, + { + "epoch": 0.8977914740626605, + "grad_norm": 2.8689560890197754, + "learning_rate": 3.920822230681386e-06, + "loss": 1.4649, + "step": 874 + }, + { + "epoch": 0.8988186954288649, + "grad_norm": 2.677800178527832, + "learning_rate": 3.918918918918919e-06, + "loss": 1.4815, + "step": 875 + }, + { + "epoch": 0.8998459167950693, + "grad_norm": 2.8248631954193115, + "learning_rate": 3.917015607156452e-06, + "loss": 1.5193, + "step": 876 + }, + { + "epoch": 0.9008731381612738, + "grad_norm": 2.8272030353546143, + "learning_rate": 3.915112295393986e-06, + "loss": 1.4968, + "step": 877 + }, + { + "epoch": 0.9019003595274782, + "grad_norm": 3.0890679359436035, + "learning_rate": 3.913208983631519e-06, + "loss": 1.5705, + "step": 878 + }, + { + "epoch": 0.9029275808936826, + "grad_norm": 2.7617108821868896, + "learning_rate": 3.9113056718690525e-06, + "loss": 1.5598, + "step": 879 + }, + { + "epoch": 0.903954802259887, + "grad_norm": 2.7723312377929688, + "learning_rate": 3.909402360106586e-06, + "loss": 1.4781, + "step": 880 + }, + { + "epoch": 0.903954802259887, + "eval_loss": 1.8097552061080933, + "eval_runtime": 37.0572, + "eval_samples_per_second": 26.985, + "eval_steps_per_second": 1.133, + "step": 880 + }, + { + "epoch": 0.9049820236260914, + "grad_norm": 2.82560133934021, + "learning_rate": 3.907499048344119e-06, + "loss": 1.5224, + "step": 881 + }, + { + "epoch": 0.9060092449922958, + "grad_norm": 2.881899118423462, + "learning_rate": 3.905595736581652e-06, + "loss": 1.5169, + "step": 882 + }, + { + "epoch": 0.9070364663585002, + "grad_norm": 2.8723788261413574, + "learning_rate": 3.903692424819186e-06, + "loss": 1.5225, + "step": 883 + }, + { + "epoch": 0.9080636877247047, + "grad_norm": 2.5932493209838867, + "learning_rate": 3.901789113056719e-06, + "loss": 1.4796, + "step": 884 + }, + { + "epoch": 0.9090909090909091, + "grad_norm": 2.6493358612060547, + "learning_rate": 3.899885801294253e-06, + "loss": 1.5324, + "step": 885 + }, + { + "epoch": 0.9101181304571135, + "grad_norm": 2.630830764770508, + "learning_rate": 3.897982489531786e-06, + "loss": 1.5113, + "step": 886 + }, + { + "epoch": 0.9111453518233179, + "grad_norm": 2.8183815479278564, + "learning_rate": 3.896079177769319e-06, + "loss": 1.5677, + "step": 887 + }, + { + "epoch": 0.9121725731895224, + "grad_norm": 2.811241865158081, + "learning_rate": 3.894175866006853e-06, + "loss": 1.5357, + "step": 888 + }, + { + "epoch": 0.9131997945557268, + "grad_norm": 2.6256868839263916, + "learning_rate": 3.892272554244385e-06, + "loss": 1.5603, + "step": 889 + }, + { + "epoch": 0.9142270159219312, + "grad_norm": 2.677891254425049, + "learning_rate": 3.890369242481919e-06, + "loss": 1.4748, + "step": 890 + }, + { + "epoch": 0.9142270159219312, + "eval_loss": 1.7937747240066528, + "eval_runtime": 37.3281, + "eval_samples_per_second": 26.789, + "eval_steps_per_second": 1.125, + "step": 890 + }, + { + "epoch": 0.9152542372881356, + "grad_norm": 2.639749765396118, + "learning_rate": 3.888465930719452e-06, + "loss": 1.5477, + "step": 891 + }, + { + "epoch": 0.91628145865434, + "grad_norm": 2.622887134552002, + "learning_rate": 3.886562618956985e-06, + "loss": 1.481, + "step": 892 + }, + { + "epoch": 0.9173086800205444, + "grad_norm": 2.659916639328003, + "learning_rate": 3.884659307194519e-06, + "loss": 1.5044, + "step": 893 + }, + { + "epoch": 0.9183359013867488, + "grad_norm": 2.524515151977539, + "learning_rate": 3.882755995432052e-06, + "loss": 1.4378, + "step": 894 + }, + { + "epoch": 0.9193631227529533, + "grad_norm": 2.5722432136535645, + "learning_rate": 3.8808526836695854e-06, + "loss": 1.5173, + "step": 895 + }, + { + "epoch": 0.9203903441191577, + "grad_norm": 2.6036641597747803, + "learning_rate": 3.878949371907119e-06, + "loss": 1.4823, + "step": 896 + }, + { + "epoch": 0.9214175654853621, + "grad_norm": 2.671830415725708, + "learning_rate": 3.877046060144652e-06, + "loss": 1.4679, + "step": 897 + }, + { + "epoch": 0.9224447868515665, + "grad_norm": 2.649601936340332, + "learning_rate": 3.8751427483821855e-06, + "loss": 1.4651, + "step": 898 + }, + { + "epoch": 0.923472008217771, + "grad_norm": 2.6315200328826904, + "learning_rate": 3.873239436619718e-06, + "loss": 1.5029, + "step": 899 + }, + { + "epoch": 0.9244992295839753, + "grad_norm": 2.7867085933685303, + "learning_rate": 3.871336124857251e-06, + "loss": 1.4635, + "step": 900 + }, + { + "epoch": 0.9244992295839753, + "eval_loss": 1.8004201650619507, + "eval_runtime": 37.3928, + "eval_samples_per_second": 26.743, + "eval_steps_per_second": 1.123, + "step": 900 + }, + { + "epoch": 0.9255264509501797, + "grad_norm": 2.6903293132781982, + "learning_rate": 3.869432813094785e-06, + "loss": 1.5043, + "step": 901 + }, + { + "epoch": 0.9265536723163842, + "grad_norm": 2.6979103088378906, + "learning_rate": 3.867529501332318e-06, + "loss": 1.5168, + "step": 902 + }, + { + "epoch": 0.9275808936825886, + "grad_norm": 2.594578266143799, + "learning_rate": 3.8656261895698515e-06, + "loss": 1.5141, + "step": 903 + }, + { + "epoch": 0.928608115048793, + "grad_norm": 2.774486780166626, + "learning_rate": 3.863722877807386e-06, + "loss": 1.4982, + "step": 904 + }, + { + "epoch": 0.9296353364149974, + "grad_norm": 2.6801764965057373, + "learning_rate": 3.861819566044919e-06, + "loss": 1.5294, + "step": 905 + }, + { + "epoch": 0.9306625577812019, + "grad_norm": 2.681169033050537, + "learning_rate": 3.859916254282452e-06, + "loss": 1.5687, + "step": 906 + }, + { + "epoch": 0.9316897791474062, + "grad_norm": 2.6592564582824707, + "learning_rate": 3.858012942519985e-06, + "loss": 1.4967, + "step": 907 + }, + { + "epoch": 0.9327170005136107, + "grad_norm": 2.636106491088867, + "learning_rate": 3.856109630757518e-06, + "loss": 1.5108, + "step": 908 + }, + { + "epoch": 0.9337442218798151, + "grad_norm": 2.6918840408325195, + "learning_rate": 3.854206318995052e-06, + "loss": 1.4821, + "step": 909 + }, + { + "epoch": 0.9347714432460195, + "grad_norm": 2.7018930912017822, + "learning_rate": 3.852303007232585e-06, + "loss": 1.5336, + "step": 910 + }, + { + "epoch": 0.9347714432460195, + "eval_loss": 1.789487600326538, + "eval_runtime": 36.9792, + "eval_samples_per_second": 27.042, + "eval_steps_per_second": 1.136, + "step": 910 + }, + { + "epoch": 0.9357986646122239, + "grad_norm": 2.5881776809692383, + "learning_rate": 3.850399695470118e-06, + "loss": 1.4632, + "step": 911 + }, + { + "epoch": 0.9368258859784283, + "grad_norm": 2.55487322807312, + "learning_rate": 3.848496383707652e-06, + "loss": 1.4825, + "step": 912 + }, + { + "epoch": 0.9378531073446328, + "grad_norm": 2.5630884170532227, + "learning_rate": 3.846593071945185e-06, + "loss": 1.463, + "step": 913 + }, + { + "epoch": 0.9388803287108372, + "grad_norm": 2.6740293502807617, + "learning_rate": 3.8446897601827185e-06, + "loss": 1.4499, + "step": 914 + }, + { + "epoch": 0.9399075500770416, + "grad_norm": 2.641127109527588, + "learning_rate": 3.842786448420252e-06, + "loss": 1.5618, + "step": 915 + }, + { + "epoch": 0.940934771443246, + "grad_norm": 2.669304370880127, + "learning_rate": 3.840883136657785e-06, + "loss": 1.5725, + "step": 916 + }, + { + "epoch": 0.9419619928094505, + "grad_norm": 2.7020649909973145, + "learning_rate": 3.838979824895318e-06, + "loss": 1.5587, + "step": 917 + }, + { + "epoch": 0.9429892141756548, + "grad_norm": 2.6244914531707764, + "learning_rate": 3.837076513132851e-06, + "loss": 1.4363, + "step": 918 + }, + { + "epoch": 0.9440164355418593, + "grad_norm": 2.7254905700683594, + "learning_rate": 3.8351732013703844e-06, + "loss": 1.5205, + "step": 919 + }, + { + "epoch": 0.9450436569080637, + "grad_norm": 2.4565985202789307, + "learning_rate": 3.833269889607918e-06, + "loss": 1.4785, + "step": 920 + }, + { + "epoch": 0.9450436569080637, + "eval_loss": 1.8004498481750488, + "eval_runtime": 37.0945, + "eval_samples_per_second": 26.958, + "eval_steps_per_second": 1.132, + "step": 920 + }, + { + "epoch": 0.9460708782742681, + "grad_norm": 2.8681352138519287, + "learning_rate": 3.831366577845451e-06, + "loss": 1.4598, + "step": 921 + }, + { + "epoch": 0.9470980996404725, + "grad_norm": 2.6068029403686523, + "learning_rate": 3.8294632660829845e-06, + "loss": 1.4565, + "step": 922 + }, + { + "epoch": 0.9481253210066769, + "grad_norm": 2.764873504638672, + "learning_rate": 3.827559954320518e-06, + "loss": 1.4882, + "step": 923 + }, + { + "epoch": 0.9491525423728814, + "grad_norm": 2.655008316040039, + "learning_rate": 3.825656642558051e-06, + "loss": 1.5004, + "step": 924 + }, + { + "epoch": 0.9501797637390857, + "grad_norm": 2.7953972816467285, + "learning_rate": 3.823753330795585e-06, + "loss": 1.5172, + "step": 925 + }, + { + "epoch": 0.9512069851052902, + "grad_norm": 2.698564052581787, + "learning_rate": 3.821850019033118e-06, + "loss": 1.5022, + "step": 926 + }, + { + "epoch": 0.9522342064714946, + "grad_norm": 2.4997222423553467, + "learning_rate": 3.819946707270651e-06, + "loss": 1.4373, + "step": 927 + }, + { + "epoch": 0.9532614278376991, + "grad_norm": 2.6832621097564697, + "learning_rate": 3.818043395508185e-06, + "loss": 1.49, + "step": 928 + }, + { + "epoch": 0.9542886492039034, + "grad_norm": 2.734297513961792, + "learning_rate": 3.816140083745718e-06, + "loss": 1.499, + "step": 929 + }, + { + "epoch": 0.9553158705701078, + "grad_norm": 2.6073720455169678, + "learning_rate": 3.8142367719832514e-06, + "loss": 1.5701, + "step": 930 + }, + { + "epoch": 0.9553158705701078, + "eval_loss": 1.8036372661590576, + "eval_runtime": 37.141, + "eval_samples_per_second": 26.924, + "eval_steps_per_second": 1.131, + "step": 930 + }, + { + "epoch": 0.9563430919363123, + "grad_norm": 2.6502838134765625, + "learning_rate": 3.8123334602207844e-06, + "loss": 1.4666, + "step": 931 + }, + { + "epoch": 0.9573703133025167, + "grad_norm": 2.6244993209838867, + "learning_rate": 3.8104301484583177e-06, + "loss": 1.4659, + "step": 932 + }, + { + "epoch": 0.9583975346687211, + "grad_norm": 2.5014829635620117, + "learning_rate": 3.808526836695851e-06, + "loss": 1.4562, + "step": 933 + }, + { + "epoch": 0.9594247560349255, + "grad_norm": 2.6359620094299316, + "learning_rate": 3.8066235249333845e-06, + "loss": 1.5223, + "step": 934 + }, + { + "epoch": 0.96045197740113, + "grad_norm": 2.648331642150879, + "learning_rate": 3.804720213170918e-06, + "loss": 1.4924, + "step": 935 + }, + { + "epoch": 0.9614791987673343, + "grad_norm": 2.731069564819336, + "learning_rate": 3.8028169014084508e-06, + "loss": 1.4959, + "step": 936 + }, + { + "epoch": 0.9625064201335388, + "grad_norm": 2.5789313316345215, + "learning_rate": 3.800913589645984e-06, + "loss": 1.5125, + "step": 937 + }, + { + "epoch": 0.9635336414997432, + "grad_norm": 2.6203935146331787, + "learning_rate": 3.7990102778835175e-06, + "loss": 1.5064, + "step": 938 + }, + { + "epoch": 0.9645608628659477, + "grad_norm": 2.7040021419525146, + "learning_rate": 3.797106966121051e-06, + "loss": 1.5092, + "step": 939 + }, + { + "epoch": 0.965588084232152, + "grad_norm": 2.575338125228882, + "learning_rate": 3.7952036543585842e-06, + "loss": 1.5653, + "step": 940 + }, + { + "epoch": 0.965588084232152, + "eval_loss": 1.7945882081985474, + "eval_runtime": 37.1973, + "eval_samples_per_second": 26.884, + "eval_steps_per_second": 1.129, + "step": 940 + }, + { + "epoch": 0.9666153055983564, + "grad_norm": 2.597348690032959, + "learning_rate": 3.793300342596117e-06, + "loss": 1.5312, + "step": 941 + }, + { + "epoch": 0.9676425269645609, + "grad_norm": 2.590437650680542, + "learning_rate": 3.7913970308336505e-06, + "loss": 1.5347, + "step": 942 + }, + { + "epoch": 0.9686697483307652, + "grad_norm": 2.6253085136413574, + "learning_rate": 3.789493719071184e-06, + "loss": 1.5265, + "step": 943 + }, + { + "epoch": 0.9696969696969697, + "grad_norm": 2.5652995109558105, + "learning_rate": 3.7875904073087177e-06, + "loss": 1.4738, + "step": 944 + }, + { + "epoch": 0.9707241910631741, + "grad_norm": 2.552037239074707, + "learning_rate": 3.785687095546251e-06, + "loss": 1.4311, + "step": 945 + }, + { + "epoch": 0.9717514124293786, + "grad_norm": 2.806792974472046, + "learning_rate": 3.7837837837837844e-06, + "loss": 1.4857, + "step": 946 + }, + { + "epoch": 0.9727786337955829, + "grad_norm": 2.696113348007202, + "learning_rate": 3.7818804720213178e-06, + "loss": 1.4604, + "step": 947 + }, + { + "epoch": 0.9738058551617874, + "grad_norm": 2.972555637359619, + "learning_rate": 3.7799771602588507e-06, + "loss": 1.4328, + "step": 948 + }, + { + "epoch": 0.9748330765279918, + "grad_norm": 2.7260053157806396, + "learning_rate": 3.778073848496384e-06, + "loss": 1.4315, + "step": 949 + }, + { + "epoch": 0.9758602978941961, + "grad_norm": 2.501840353012085, + "learning_rate": 3.7761705367339174e-06, + "loss": 1.4913, + "step": 950 + }, + { + "epoch": 0.9758602978941961, + "eval_loss": 1.7844858169555664, + "eval_runtime": 37.1439, + "eval_samples_per_second": 26.922, + "eval_steps_per_second": 1.131, + "step": 950 + }, + { + "epoch": 0.9768875192604006, + "grad_norm": 2.704096794128418, + "learning_rate": 3.7742672249714508e-06, + "loss": 1.5138, + "step": 951 + }, + { + "epoch": 0.977914740626605, + "grad_norm": 2.7086358070373535, + "learning_rate": 3.772363913208984e-06, + "loss": 1.5228, + "step": 952 + }, + { + "epoch": 0.9789419619928095, + "grad_norm": 2.6690590381622314, + "learning_rate": 3.770460601446517e-06, + "loss": 1.4828, + "step": 953 + }, + { + "epoch": 0.9799691833590138, + "grad_norm": 2.6134684085845947, + "learning_rate": 3.7685572896840504e-06, + "loss": 1.509, + "step": 954 + }, + { + "epoch": 0.9809964047252183, + "grad_norm": 2.7675423622131348, + "learning_rate": 3.766653977921584e-06, + "loss": 1.4544, + "step": 955 + }, + { + "epoch": 0.9820236260914227, + "grad_norm": 2.8518829345703125, + "learning_rate": 3.764750666159117e-06, + "loss": 1.4042, + "step": 956 + }, + { + "epoch": 0.9830508474576272, + "grad_norm": 2.6791179180145264, + "learning_rate": 3.7628473543966505e-06, + "loss": 1.4961, + "step": 957 + }, + { + "epoch": 0.9840780688238315, + "grad_norm": 3.014099597930908, + "learning_rate": 3.7609440426341835e-06, + "loss": 1.4582, + "step": 958 + }, + { + "epoch": 0.9851052901900359, + "grad_norm": 2.7100412845611572, + "learning_rate": 3.759040730871717e-06, + "loss": 1.5491, + "step": 959 + }, + { + "epoch": 0.9861325115562404, + "grad_norm": 2.565263271331787, + "learning_rate": 3.75713741910925e-06, + "loss": 1.4437, + "step": 960 + }, + { + "epoch": 0.9861325115562404, + "eval_loss": 1.7909018993377686, + "eval_runtime": 37.1828, + "eval_samples_per_second": 26.894, + "eval_steps_per_second": 1.13, + "step": 960 + }, + { + "epoch": 0.9871597329224447, + "grad_norm": 2.8322439193725586, + "learning_rate": 3.7552341073467836e-06, + "loss": 1.4322, + "step": 961 + }, + { + "epoch": 0.9881869542886492, + "grad_norm": 2.738586902618408, + "learning_rate": 3.753330795584317e-06, + "loss": 1.4797, + "step": 962 + }, + { + "epoch": 0.9892141756548536, + "grad_norm": 2.711247205734253, + "learning_rate": 3.75142748382185e-06, + "loss": 1.5409, + "step": 963 + }, + { + "epoch": 0.9902413970210581, + "grad_norm": 2.5288846492767334, + "learning_rate": 3.749524172059384e-06, + "loss": 1.4408, + "step": 964 + }, + { + "epoch": 0.9912686183872624, + "grad_norm": 2.5907785892486572, + "learning_rate": 3.747620860296917e-06, + "loss": 1.4872, + "step": 965 + }, + { + "epoch": 0.9922958397534669, + "grad_norm": 2.4952504634857178, + "learning_rate": 3.7457175485344504e-06, + "loss": 1.5277, + "step": 966 + }, + { + "epoch": 0.9933230611196713, + "grad_norm": 2.633167266845703, + "learning_rate": 3.7438142367719837e-06, + "loss": 1.5258, + "step": 967 + }, + { + "epoch": 0.9943502824858758, + "grad_norm": 2.6295969486236572, + "learning_rate": 3.741910925009517e-06, + "loss": 1.5013, + "step": 968 + }, + { + "epoch": 0.9953775038520801, + "grad_norm": 2.4780771732330322, + "learning_rate": 3.7400076132470505e-06, + "loss": 1.4646, + "step": 969 + }, + { + "epoch": 0.9964047252182845, + "grad_norm": 2.585749626159668, + "learning_rate": 3.7381043014845834e-06, + "loss": 1.4814, + "step": 970 + }, + { + "epoch": 0.9964047252182845, + "eval_loss": 1.801538109779358, + "eval_runtime": 37.1396, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 970 + }, + { + "epoch": 0.997431946584489, + "grad_norm": 2.626293182373047, + "learning_rate": 3.7362009897221168e-06, + "loss": 1.5071, + "step": 971 + }, + { + "epoch": 0.9984591679506933, + "grad_norm": 2.7129111289978027, + "learning_rate": 3.73429767795965e-06, + "loss": 1.4911, + "step": 972 + }, + { + "epoch": 0.9994863893168978, + "grad_norm": 2.623265027999878, + "learning_rate": 3.7323943661971835e-06, + "loss": 1.506, + "step": 973 + }, + { + "epoch": 1.0005136106831023, + "grad_norm": 2.497767925262451, + "learning_rate": 3.730491054434717e-06, + "loss": 1.4119, + "step": 974 + }, + { + "epoch": 1.0015408320493067, + "grad_norm": 2.8463239669799805, + "learning_rate": 3.72858774267225e-06, + "loss": 1.3677, + "step": 975 + }, + { + "epoch": 1.002568053415511, + "grad_norm": 2.643601417541504, + "learning_rate": 3.726684430909783e-06, + "loss": 1.4573, + "step": 976 + }, + { + "epoch": 1.0035952747817154, + "grad_norm": 2.751330614089966, + "learning_rate": 3.7247811191473165e-06, + "loss": 1.4821, + "step": 977 + }, + { + "epoch": 1.00462249614792, + "grad_norm": 2.714946985244751, + "learning_rate": 3.72287780738485e-06, + "loss": 1.4898, + "step": 978 + }, + { + "epoch": 1.0056497175141244, + "grad_norm": 3.066772699356079, + "learning_rate": 3.7209744956223832e-06, + "loss": 1.464, + "step": 979 + }, + { + "epoch": 1.0066769388803287, + "grad_norm": 2.818673610687256, + "learning_rate": 3.719071183859916e-06, + "loss": 1.4743, + "step": 980 + }, + { + "epoch": 1.0066769388803287, + "eval_loss": 1.7895241975784302, + "eval_runtime": 37.1443, + "eval_samples_per_second": 26.922, + "eval_steps_per_second": 1.131, + "step": 980 + }, + { + "epoch": 1.007704160246533, + "grad_norm": 2.6721136569976807, + "learning_rate": 3.7171678720974495e-06, + "loss": 1.3996, + "step": 981 + }, + { + "epoch": 1.0087313816127375, + "grad_norm": 2.757756471633911, + "learning_rate": 3.715264560334983e-06, + "loss": 1.5009, + "step": 982 + }, + { + "epoch": 1.009758602978942, + "grad_norm": 2.8042495250701904, + "learning_rate": 3.7133612485725163e-06, + "loss": 1.4672, + "step": 983 + }, + { + "epoch": 1.0107858243451464, + "grad_norm": 2.662808418273926, + "learning_rate": 3.71145793681005e-06, + "loss": 1.4799, + "step": 984 + }, + { + "epoch": 1.0118130457113508, + "grad_norm": 2.6021053791046143, + "learning_rate": 3.7095546250475834e-06, + "loss": 1.4035, + "step": 985 + }, + { + "epoch": 1.0128402670775551, + "grad_norm": 2.7127013206481934, + "learning_rate": 3.7076513132851168e-06, + "loss": 1.5192, + "step": 986 + }, + { + "epoch": 1.0138674884437597, + "grad_norm": 2.6402854919433594, + "learning_rate": 3.7057480015226497e-06, + "loss": 1.4396, + "step": 987 + }, + { + "epoch": 1.014894709809964, + "grad_norm": 2.5758111476898193, + "learning_rate": 3.703844689760183e-06, + "loss": 1.4109, + "step": 988 + }, + { + "epoch": 1.0159219311761685, + "grad_norm": 2.7668240070343018, + "learning_rate": 3.7019413779977164e-06, + "loss": 1.4865, + "step": 989 + }, + { + "epoch": 1.0169491525423728, + "grad_norm": 2.840165615081787, + "learning_rate": 3.70003806623525e-06, + "loss": 1.4708, + "step": 990 + }, + { + "epoch": 1.0169491525423728, + "eval_loss": 1.7967512607574463, + "eval_runtime": 37.2322, + "eval_samples_per_second": 26.858, + "eval_steps_per_second": 1.128, + "step": 990 + }, + { + "epoch": 1.0179763739085772, + "grad_norm": 2.8218352794647217, + "learning_rate": 3.698134754472783e-06, + "loss": 1.4414, + "step": 991 + }, + { + "epoch": 1.0190035952747818, + "grad_norm": 2.742658853530884, + "learning_rate": 3.696231442710316e-06, + "loss": 1.5109, + "step": 992 + }, + { + "epoch": 1.0200308166409862, + "grad_norm": 2.6375293731689453, + "learning_rate": 3.6943281309478495e-06, + "loss": 1.4433, + "step": 993 + }, + { + "epoch": 1.0210580380071905, + "grad_norm": 2.914246082305908, + "learning_rate": 3.692424819185383e-06, + "loss": 1.4656, + "step": 994 + }, + { + "epoch": 1.022085259373395, + "grad_norm": 2.9181385040283203, + "learning_rate": 3.690521507422916e-06, + "loss": 1.4735, + "step": 995 + }, + { + "epoch": 1.0231124807395995, + "grad_norm": 2.755748748779297, + "learning_rate": 3.6886181956604496e-06, + "loss": 1.4196, + "step": 996 + }, + { + "epoch": 1.0241397021058039, + "grad_norm": 2.9245896339416504, + "learning_rate": 3.686714883897983e-06, + "loss": 1.4316, + "step": 997 + }, + { + "epoch": 1.0251669234720082, + "grad_norm": 2.839991569519043, + "learning_rate": 3.684811572135516e-06, + "loss": 1.4965, + "step": 998 + }, + { + "epoch": 1.0261941448382126, + "grad_norm": 2.904658079147339, + "learning_rate": 3.6829082603730492e-06, + "loss": 1.4962, + "step": 999 + }, + { + "epoch": 1.027221366204417, + "grad_norm": 2.811279535293579, + "learning_rate": 3.6810049486105826e-06, + "loss": 1.4514, + "step": 1000 + }, + { + "epoch": 1.027221366204417, + "eval_loss": 1.7917094230651855, + "eval_runtime": 37.315, + "eval_samples_per_second": 26.799, + "eval_steps_per_second": 1.126, + "step": 1000 + }, + { + "epoch": 1.0282485875706215, + "grad_norm": 2.730212926864624, + "learning_rate": 3.679101636848116e-06, + "loss": 1.4842, + "step": 1001 + }, + { + "epoch": 1.029275808936826, + "grad_norm": 2.8678042888641357, + "learning_rate": 3.6771983250856493e-06, + "loss": 1.4726, + "step": 1002 + }, + { + "epoch": 1.0303030303030303, + "grad_norm": 2.8228330612182617, + "learning_rate": 3.6752950133231822e-06, + "loss": 1.4501, + "step": 1003 + }, + { + "epoch": 1.0313302516692346, + "grad_norm": 2.8341903686523438, + "learning_rate": 3.6733917015607156e-06, + "loss": 1.4758, + "step": 1004 + }, + { + "epoch": 1.0323574730354392, + "grad_norm": 2.7462005615234375, + "learning_rate": 3.6714883897982494e-06, + "loss": 1.4783, + "step": 1005 + }, + { + "epoch": 1.0333846944016436, + "grad_norm": 2.920111894607544, + "learning_rate": 3.6695850780357828e-06, + "loss": 1.4393, + "step": 1006 + }, + { + "epoch": 1.034411915767848, + "grad_norm": 2.919542074203491, + "learning_rate": 3.667681766273316e-06, + "loss": 1.426, + "step": 1007 + }, + { + "epoch": 1.0354391371340523, + "grad_norm": 2.8268563747406006, + "learning_rate": 3.6657784545108495e-06, + "loss": 1.4391, + "step": 1008 + }, + { + "epoch": 1.0364663585002567, + "grad_norm": 2.711895227432251, + "learning_rate": 3.6638751427483824e-06, + "loss": 1.5107, + "step": 1009 + }, + { + "epoch": 1.0374935798664613, + "grad_norm": 2.7996091842651367, + "learning_rate": 3.6619718309859158e-06, + "loss": 1.4679, + "step": 1010 + }, + { + "epoch": 1.0374935798664613, + "eval_loss": 1.800447940826416, + "eval_runtime": 36.9218, + "eval_samples_per_second": 27.084, + "eval_steps_per_second": 1.138, + "step": 1010 + }, + { + "epoch": 1.0385208012326657, + "grad_norm": 2.8792128562927246, + "learning_rate": 3.660068519223449e-06, + "loss": 1.5156, + "step": 1011 + }, + { + "epoch": 1.03954802259887, + "grad_norm": 2.7948005199432373, + "learning_rate": 3.6581652074609825e-06, + "loss": 1.4709, + "step": 1012 + }, + { + "epoch": 1.0405752439650744, + "grad_norm": 2.7429840564727783, + "learning_rate": 3.656261895698516e-06, + "loss": 1.4414, + "step": 1013 + }, + { + "epoch": 1.041602465331279, + "grad_norm": 2.7512004375457764, + "learning_rate": 3.6543585839360492e-06, + "loss": 1.4326, + "step": 1014 + }, + { + "epoch": 1.0426296866974833, + "grad_norm": 2.7879600524902344, + "learning_rate": 3.652455272173582e-06, + "loss": 1.4292, + "step": 1015 + }, + { + "epoch": 1.0436569080636877, + "grad_norm": 2.568354368209839, + "learning_rate": 3.6505519604111155e-06, + "loss": 1.4383, + "step": 1016 + }, + { + "epoch": 1.044684129429892, + "grad_norm": 2.811404228210449, + "learning_rate": 3.648648648648649e-06, + "loss": 1.4573, + "step": 1017 + }, + { + "epoch": 1.0457113507960965, + "grad_norm": 2.909632921218872, + "learning_rate": 3.6467453368861823e-06, + "loss": 1.4934, + "step": 1018 + }, + { + "epoch": 1.046738572162301, + "grad_norm": 2.8014776706695557, + "learning_rate": 3.6448420251237156e-06, + "loss": 1.4433, + "step": 1019 + }, + { + "epoch": 1.0477657935285054, + "grad_norm": 2.8205742835998535, + "learning_rate": 3.6429387133612486e-06, + "loss": 1.4752, + "step": 1020 + }, + { + "epoch": 1.0477657935285054, + "eval_loss": 1.8070552349090576, + "eval_runtime": 37.1396, + "eval_samples_per_second": 26.925, + "eval_steps_per_second": 1.131, + "step": 1020 + }, + { + "epoch": 1.0487930148947098, + "grad_norm": 2.768143653869629, + "learning_rate": 3.641035401598782e-06, + "loss": 1.4928, + "step": 1021 + }, + { + "epoch": 1.0498202362609141, + "grad_norm": 2.8027846813201904, + "learning_rate": 3.6391320898363153e-06, + "loss": 1.4416, + "step": 1022 + }, + { + "epoch": 1.0508474576271187, + "grad_norm": 2.832526683807373, + "learning_rate": 3.6372287780738487e-06, + "loss": 1.4796, + "step": 1023 + }, + { + "epoch": 1.051874678993323, + "grad_norm": 2.8225502967834473, + "learning_rate": 3.635325466311382e-06, + "loss": 1.4553, + "step": 1024 + }, + { + "epoch": 1.0529019003595275, + "grad_norm": 2.7793405055999756, + "learning_rate": 3.633422154548916e-06, + "loss": 1.4756, + "step": 1025 + }, + { + "epoch": 1.0539291217257318, + "grad_norm": 2.872873544692993, + "learning_rate": 3.6315188427864487e-06, + "loss": 1.496, + "step": 1026 + }, + { + "epoch": 1.0549563430919364, + "grad_norm": 2.7574472427368164, + "learning_rate": 3.629615531023982e-06, + "loss": 1.464, + "step": 1027 + }, + { + "epoch": 1.0559835644581408, + "grad_norm": 2.7946009635925293, + "learning_rate": 3.6277122192615155e-06, + "loss": 1.4149, + "step": 1028 + }, + { + "epoch": 1.0570107858243452, + "grad_norm": 2.764500617980957, + "learning_rate": 3.625808907499049e-06, + "loss": 1.4756, + "step": 1029 + }, + { + "epoch": 1.0580380071905495, + "grad_norm": 2.659693479537964, + "learning_rate": 3.623905595736582e-06, + "loss": 1.4385, + "step": 1030 + }, + { + "epoch": 1.0580380071905495, + "eval_loss": 1.8091230392456055, + "eval_runtime": 37.109, + "eval_samples_per_second": 26.948, + "eval_steps_per_second": 1.132, + "step": 1030 + }, + { + "epoch": 1.059065228556754, + "grad_norm": 2.7500619888305664, + "learning_rate": 3.6220022839741156e-06, + "loss": 1.4334, + "step": 1031 + }, + { + "epoch": 1.0600924499229585, + "grad_norm": 2.7663776874542236, + "learning_rate": 3.6200989722116485e-06, + "loss": 1.4194, + "step": 1032 + }, + { + "epoch": 1.0611196712891628, + "grad_norm": 2.820913076400757, + "learning_rate": 3.618195660449182e-06, + "loss": 1.4453, + "step": 1033 + }, + { + "epoch": 1.0621468926553672, + "grad_norm": 2.8579916954040527, + "learning_rate": 3.6162923486867152e-06, + "loss": 1.4529, + "step": 1034 + }, + { + "epoch": 1.0631741140215716, + "grad_norm": 2.729501247406006, + "learning_rate": 3.6143890369242486e-06, + "loss": 1.3907, + "step": 1035 + }, + { + "epoch": 1.064201335387776, + "grad_norm": 2.7084193229675293, + "learning_rate": 3.612485725161782e-06, + "loss": 1.4252, + "step": 1036 + }, + { + "epoch": 1.0652285567539805, + "grad_norm": 2.7723402976989746, + "learning_rate": 3.610582413399315e-06, + "loss": 1.4339, + "step": 1037 + }, + { + "epoch": 1.066255778120185, + "grad_norm": 2.5925116539001465, + "learning_rate": 3.6086791016368482e-06, + "loss": 1.4252, + "step": 1038 + }, + { + "epoch": 1.0672829994863893, + "grad_norm": 2.7101070880889893, + "learning_rate": 3.6067757898743816e-06, + "loss": 1.4405, + "step": 1039 + }, + { + "epoch": 1.0683102208525936, + "grad_norm": 2.786576509475708, + "learning_rate": 3.604872478111915e-06, + "loss": 1.5049, + "step": 1040 + }, + { + "epoch": 1.0683102208525936, + "eval_loss": 1.8023844957351685, + "eval_runtime": 37.0823, + "eval_samples_per_second": 26.967, + "eval_steps_per_second": 1.133, + "step": 1040 + }, + { + "epoch": 1.0693374422187982, + "grad_norm": 2.8227438926696777, + "learning_rate": 3.6029691663494483e-06, + "loss": 1.5062, + "step": 1041 + }, + { + "epoch": 1.0703646635850026, + "grad_norm": 2.722092866897583, + "learning_rate": 3.6010658545869813e-06, + "loss": 1.4508, + "step": 1042 + }, + { + "epoch": 1.071391884951207, + "grad_norm": 2.719161033630371, + "learning_rate": 3.5991625428245146e-06, + "loss": 1.4791, + "step": 1043 + }, + { + "epoch": 1.0724191063174113, + "grad_norm": 2.8131134510040283, + "learning_rate": 3.597259231062048e-06, + "loss": 1.4811, + "step": 1044 + }, + { + "epoch": 1.073446327683616, + "grad_norm": 2.7374107837677, + "learning_rate": 3.5953559192995818e-06, + "loss": 1.4063, + "step": 1045 + }, + { + "epoch": 1.0744735490498203, + "grad_norm": 2.7651076316833496, + "learning_rate": 3.593452607537115e-06, + "loss": 1.4589, + "step": 1046 + }, + { + "epoch": 1.0755007704160247, + "grad_norm": 2.806370735168457, + "learning_rate": 3.5915492957746485e-06, + "loss": 1.5197, + "step": 1047 + }, + { + "epoch": 1.076527991782229, + "grad_norm": 2.8023886680603027, + "learning_rate": 3.589645984012182e-06, + "loss": 1.4427, + "step": 1048 + }, + { + "epoch": 1.0775552131484334, + "grad_norm": 2.7536637783050537, + "learning_rate": 3.587742672249715e-06, + "loss": 1.4753, + "step": 1049 + }, + { + "epoch": 1.078582434514638, + "grad_norm": 2.6758344173431396, + "learning_rate": 3.585839360487248e-06, + "loss": 1.4302, + "step": 1050 + }, + { + "epoch": 1.078582434514638, + "eval_loss": 1.7898778915405273, + "eval_runtime": 37.04, + "eval_samples_per_second": 26.998, + "eval_steps_per_second": 1.134, + "step": 1050 + }, + { + "epoch": 1.0796096558808423, + "grad_norm": 2.688101053237915, + "learning_rate": 3.5839360487247815e-06, + "loss": 1.4271, + "step": 1051 + }, + { + "epoch": 1.0806368772470467, + "grad_norm": 2.8667550086975098, + "learning_rate": 3.582032736962315e-06, + "loss": 1.4372, + "step": 1052 + }, + { + "epoch": 1.081664098613251, + "grad_norm": 2.9168481826782227, + "learning_rate": 3.5801294251998483e-06, + "loss": 1.5063, + "step": 1053 + }, + { + "epoch": 1.0826913199794557, + "grad_norm": 2.743194580078125, + "learning_rate": 3.578226113437381e-06, + "loss": 1.4041, + "step": 1054 + }, + { + "epoch": 1.08371854134566, + "grad_norm": 2.6846320629119873, + "learning_rate": 3.5763228016749146e-06, + "loss": 1.4427, + "step": 1055 + }, + { + "epoch": 1.0847457627118644, + "grad_norm": 3.1511945724487305, + "learning_rate": 3.574419489912448e-06, + "loss": 1.515, + "step": 1056 + }, + { + "epoch": 1.0857729840780688, + "grad_norm": 2.7326550483703613, + "learning_rate": 3.5725161781499813e-06, + "loss": 1.4529, + "step": 1057 + }, + { + "epoch": 1.0868002054442731, + "grad_norm": 2.7120213508605957, + "learning_rate": 3.5706128663875147e-06, + "loss": 1.4407, + "step": 1058 + }, + { + "epoch": 1.0878274268104777, + "grad_norm": 2.7596473693847656, + "learning_rate": 3.5687095546250476e-06, + "loss": 1.4606, + "step": 1059 + }, + { + "epoch": 1.088854648176682, + "grad_norm": 2.9972002506256104, + "learning_rate": 3.566806242862581e-06, + "loss": 1.4836, + "step": 1060 + }, + { + "epoch": 1.088854648176682, + "eval_loss": 1.7897655963897705, + "eval_runtime": 37.0403, + "eval_samples_per_second": 26.998, + "eval_steps_per_second": 1.134, + "step": 1060 + }, + { + "epoch": 1.0898818695428865, + "grad_norm": 2.8134756088256836, + "learning_rate": 3.5649029311001143e-06, + "loss": 1.4575, + "step": 1061 + }, + { + "epoch": 1.0909090909090908, + "grad_norm": 2.7311880588531494, + "learning_rate": 3.5629996193376477e-06, + "loss": 1.4326, + "step": 1062 + }, + { + "epoch": 1.0919363122752954, + "grad_norm": 2.7033157348632812, + "learning_rate": 3.561096307575181e-06, + "loss": 1.4457, + "step": 1063 + }, + { + "epoch": 1.0929635336414998, + "grad_norm": 2.8118138313293457, + "learning_rate": 3.559192995812714e-06, + "loss": 1.4869, + "step": 1064 + }, + { + "epoch": 1.0939907550077042, + "grad_norm": 2.8734021186828613, + "learning_rate": 3.557289684050248e-06, + "loss": 1.4964, + "step": 1065 + }, + { + "epoch": 1.0950179763739085, + "grad_norm": 2.8652546405792236, + "learning_rate": 3.555386372287781e-06, + "loss": 1.4856, + "step": 1066 + }, + { + "epoch": 1.0960451977401129, + "grad_norm": 2.6305630207061768, + "learning_rate": 3.5534830605253145e-06, + "loss": 1.49, + "step": 1067 + }, + { + "epoch": 1.0970724191063175, + "grad_norm": 2.9516141414642334, + "learning_rate": 3.551579748762848e-06, + "loss": 1.4617, + "step": 1068 + }, + { + "epoch": 1.0980996404725218, + "grad_norm": 2.862701416015625, + "learning_rate": 3.5496764370003812e-06, + "loss": 1.4481, + "step": 1069 + }, + { + "epoch": 1.0991268618387262, + "grad_norm": 2.829742193222046, + "learning_rate": 3.5477731252379146e-06, + "loss": 1.4309, + "step": 1070 + }, + { + "epoch": 1.0991268618387262, + "eval_loss": 1.7925201654434204, + "eval_runtime": 37.0708, + "eval_samples_per_second": 26.975, + "eval_steps_per_second": 1.133, + "step": 1070 + }, + { + "epoch": 1.1001540832049306, + "grad_norm": 2.786259174346924, + "learning_rate": 3.5458698134754475e-06, + "loss": 1.3945, + "step": 1071 + }, + { + "epoch": 1.1011813045711352, + "grad_norm": 2.766643762588501, + "learning_rate": 3.543966501712981e-06, + "loss": 1.419, + "step": 1072 + }, + { + "epoch": 1.1022085259373395, + "grad_norm": 2.830751419067383, + "learning_rate": 3.5420631899505142e-06, + "loss": 1.4222, + "step": 1073 + }, + { + "epoch": 1.103235747303544, + "grad_norm": 3.1409215927124023, + "learning_rate": 3.5401598781880476e-06, + "loss": 1.5262, + "step": 1074 + }, + { + "epoch": 1.1042629686697483, + "grad_norm": 2.678180694580078, + "learning_rate": 3.538256566425581e-06, + "loss": 1.3482, + "step": 1075 + }, + { + "epoch": 1.1052901900359529, + "grad_norm": 2.935002088546753, + "learning_rate": 3.536353254663114e-06, + "loss": 1.4864, + "step": 1076 + }, + { + "epoch": 1.1063174114021572, + "grad_norm": 2.91631817817688, + "learning_rate": 3.5344499429006473e-06, + "loss": 1.4334, + "step": 1077 + }, + { + "epoch": 1.1073446327683616, + "grad_norm": 2.6676135063171387, + "learning_rate": 3.5325466311381806e-06, + "loss": 1.4236, + "step": 1078 + }, + { + "epoch": 1.108371854134566, + "grad_norm": 2.779284715652466, + "learning_rate": 3.530643319375714e-06, + "loss": 1.463, + "step": 1079 + }, + { + "epoch": 1.1093990755007703, + "grad_norm": 2.9586293697357178, + "learning_rate": 3.5287400076132474e-06, + "loss": 1.3953, + "step": 1080 + }, + { + "epoch": 1.1093990755007703, + "eval_loss": 1.8054182529449463, + "eval_runtime": 37.0591, + "eval_samples_per_second": 26.984, + "eval_steps_per_second": 1.133, + "step": 1080 + }, + { + "epoch": 1.110426296866975, + "grad_norm": 2.990781545639038, + "learning_rate": 3.5268366958507803e-06, + "loss": 1.4249, + "step": 1081 + }, + { + "epoch": 1.1114535182331793, + "grad_norm": 2.8546760082244873, + "learning_rate": 3.5249333840883137e-06, + "loss": 1.4679, + "step": 1082 + }, + { + "epoch": 1.1124807395993837, + "grad_norm": 3.1174535751342773, + "learning_rate": 3.523030072325847e-06, + "loss": 1.4587, + "step": 1083 + }, + { + "epoch": 1.113507960965588, + "grad_norm": 3.0766773223876953, + "learning_rate": 3.5211267605633804e-06, + "loss": 1.422, + "step": 1084 + }, + { + "epoch": 1.1145351823317924, + "grad_norm": 2.842961311340332, + "learning_rate": 3.519223448800914e-06, + "loss": 1.4335, + "step": 1085 + }, + { + "epoch": 1.115562403697997, + "grad_norm": 2.682828426361084, + "learning_rate": 3.5173201370384475e-06, + "loss": 1.4401, + "step": 1086 + }, + { + "epoch": 1.1165896250642013, + "grad_norm": 2.9262278079986572, + "learning_rate": 3.515416825275981e-06, + "loss": 1.4399, + "step": 1087 + }, + { + "epoch": 1.1176168464304057, + "grad_norm": 3.0017175674438477, + "learning_rate": 3.513513513513514e-06, + "loss": 1.4466, + "step": 1088 + }, + { + "epoch": 1.11864406779661, + "grad_norm": 2.6254069805145264, + "learning_rate": 3.511610201751047e-06, + "loss": 1.4662, + "step": 1089 + }, + { + "epoch": 1.1196712891628147, + "grad_norm": 2.8206210136413574, + "learning_rate": 3.5097068899885806e-06, + "loss": 1.3802, + "step": 1090 + }, + { + "epoch": 1.1196712891628147, + "eval_loss": 1.7929506301879883, + "eval_runtime": 37.2033, + "eval_samples_per_second": 26.879, + "eval_steps_per_second": 1.129, + "step": 1090 + }, + { + "epoch": 1.120698510529019, + "grad_norm": 2.769279718399048, + "learning_rate": 3.507803578226114e-06, + "loss": 1.4483, + "step": 1091 + }, + { + "epoch": 1.1217257318952234, + "grad_norm": 2.872576951980591, + "learning_rate": 3.5059002664636473e-06, + "loss": 1.5144, + "step": 1092 + }, + { + "epoch": 1.1227529532614278, + "grad_norm": 2.690979480743408, + "learning_rate": 3.5039969547011802e-06, + "loss": 1.3938, + "step": 1093 + }, + { + "epoch": 1.1237801746276324, + "grad_norm": 2.7444441318511963, + "learning_rate": 3.5020936429387136e-06, + "loss": 1.4295, + "step": 1094 + }, + { + "epoch": 1.1248073959938367, + "grad_norm": 3.083296775817871, + "learning_rate": 3.500190331176247e-06, + "loss": 1.4406, + "step": 1095 + }, + { + "epoch": 1.125834617360041, + "grad_norm": 2.7987470626831055, + "learning_rate": 3.4982870194137803e-06, + "loss": 1.4937, + "step": 1096 + }, + { + "epoch": 1.1268618387262455, + "grad_norm": 2.9728713035583496, + "learning_rate": 3.4963837076513137e-06, + "loss": 1.4253, + "step": 1097 + }, + { + "epoch": 1.1278890600924498, + "grad_norm": 2.9592628479003906, + "learning_rate": 3.4944803958888466e-06, + "loss": 1.4339, + "step": 1098 + }, + { + "epoch": 1.1289162814586544, + "grad_norm": 2.6775944232940674, + "learning_rate": 3.49257708412638e-06, + "loss": 1.3673, + "step": 1099 + }, + { + "epoch": 1.1299435028248588, + "grad_norm": 2.9349329471588135, + "learning_rate": 3.4906737723639133e-06, + "loss": 1.4954, + "step": 1100 + }, + { + "epoch": 1.1299435028248588, + "eval_loss": 1.7869561910629272, + "eval_runtime": 37.3494, + "eval_samples_per_second": 26.774, + "eval_steps_per_second": 1.125, + "step": 1100 + }, + { + "epoch": 1.1309707241910631, + "grad_norm": 3.086017608642578, + "learning_rate": 3.4887704606014467e-06, + "loss": 1.5096, + "step": 1101 + }, + { + "epoch": 1.1319979455572675, + "grad_norm": 2.9253454208374023, + "learning_rate": 3.48686714883898e-06, + "loss": 1.4466, + "step": 1102 + }, + { + "epoch": 1.133025166923472, + "grad_norm": 2.7312660217285156, + "learning_rate": 3.484963837076513e-06, + "loss": 1.4934, + "step": 1103 + }, + { + "epoch": 1.1340523882896765, + "grad_norm": 2.9656598567962646, + "learning_rate": 3.4830605253140464e-06, + "loss": 1.4439, + "step": 1104 + }, + { + "epoch": 1.1350796096558808, + "grad_norm": 3.083383321762085, + "learning_rate": 3.4811572135515797e-06, + "loss": 1.3937, + "step": 1105 + }, + { + "epoch": 1.1361068310220852, + "grad_norm": 2.7466018199920654, + "learning_rate": 3.4792539017891135e-06, + "loss": 1.4244, + "step": 1106 + }, + { + "epoch": 1.1371340523882898, + "grad_norm": 2.755654811859131, + "learning_rate": 3.477350590026647e-06, + "loss": 1.3984, + "step": 1107 + }, + { + "epoch": 1.1381612737544942, + "grad_norm": 3.2099268436431885, + "learning_rate": 3.4754472782641802e-06, + "loss": 1.4775, + "step": 1108 + }, + { + "epoch": 1.1391884951206985, + "grad_norm": 2.843669891357422, + "learning_rate": 3.4735439665017136e-06, + "loss": 1.461, + "step": 1109 + }, + { + "epoch": 1.140215716486903, + "grad_norm": 2.816732883453369, + "learning_rate": 3.4716406547392465e-06, + "loss": 1.429, + "step": 1110 + }, + { + "epoch": 1.140215716486903, + "eval_loss": 1.7837378978729248, + "eval_runtime": 36.9198, + "eval_samples_per_second": 27.086, + "eval_steps_per_second": 1.138, + "step": 1110 + }, + { + "epoch": 1.1412429378531073, + "grad_norm": 2.885383367538452, + "learning_rate": 3.46973734297678e-06, + "loss": 1.4462, + "step": 1111 + }, + { + "epoch": 1.1422701592193119, + "grad_norm": 2.8761868476867676, + "learning_rate": 3.4678340312143133e-06, + "loss": 1.4468, + "step": 1112 + }, + { + "epoch": 1.1432973805855162, + "grad_norm": 2.9817779064178467, + "learning_rate": 3.4659307194518466e-06, + "loss": 1.4919, + "step": 1113 + }, + { + "epoch": 1.1443246019517206, + "grad_norm": 2.931018352508545, + "learning_rate": 3.46402740768938e-06, + "loss": 1.4836, + "step": 1114 + }, + { + "epoch": 1.145351823317925, + "grad_norm": 2.708481788635254, + "learning_rate": 3.462124095926913e-06, + "loss": 1.4092, + "step": 1115 + }, + { + "epoch": 1.1463790446841293, + "grad_norm": 2.8369429111480713, + "learning_rate": 3.4602207841644463e-06, + "loss": 1.4135, + "step": 1116 + }, + { + "epoch": 1.147406266050334, + "grad_norm": 2.8598222732543945, + "learning_rate": 3.4583174724019797e-06, + "loss": 1.4578, + "step": 1117 + }, + { + "epoch": 1.1484334874165383, + "grad_norm": 2.7808401584625244, + "learning_rate": 3.456414160639513e-06, + "loss": 1.5097, + "step": 1118 + }, + { + "epoch": 1.1494607087827426, + "grad_norm": 2.777153253555298, + "learning_rate": 3.4545108488770464e-06, + "loss": 1.4687, + "step": 1119 + }, + { + "epoch": 1.150487930148947, + "grad_norm": 2.750174045562744, + "learning_rate": 3.4526075371145793e-06, + "loss": 1.4513, + "step": 1120 + }, + { + "epoch": 1.150487930148947, + "eval_loss": 1.8012630939483643, + "eval_runtime": 37.0093, + "eval_samples_per_second": 27.02, + "eval_steps_per_second": 1.135, + "step": 1120 + }, + { + "epoch": 1.1515151515151516, + "grad_norm": 2.7932651042938232, + "learning_rate": 3.4507042253521127e-06, + "loss": 1.4574, + "step": 1121 + }, + { + "epoch": 1.152542372881356, + "grad_norm": 2.8366539478302, + "learning_rate": 3.448800913589646e-06, + "loss": 1.5257, + "step": 1122 + }, + { + "epoch": 1.1535695942475603, + "grad_norm": 2.7671217918395996, + "learning_rate": 3.4468976018271794e-06, + "loss": 1.4256, + "step": 1123 + }, + { + "epoch": 1.1545968156137647, + "grad_norm": 2.845287799835205, + "learning_rate": 3.4449942900647128e-06, + "loss": 1.5317, + "step": 1124 + }, + { + "epoch": 1.1556240369799693, + "grad_norm": 2.6672773361206055, + "learning_rate": 3.443090978302246e-06, + "loss": 1.4782, + "step": 1125 + }, + { + "epoch": 1.1566512583461737, + "grad_norm": 2.844306230545044, + "learning_rate": 3.44118766653978e-06, + "loss": 1.4771, + "step": 1126 + }, + { + "epoch": 1.157678479712378, + "grad_norm": 2.723320484161377, + "learning_rate": 3.439284354777313e-06, + "loss": 1.4568, + "step": 1127 + }, + { + "epoch": 1.1587057010785824, + "grad_norm": 2.725198984146118, + "learning_rate": 3.4373810430148462e-06, + "loss": 1.4682, + "step": 1128 + }, + { + "epoch": 1.1597329224447868, + "grad_norm": 2.8702471256256104, + "learning_rate": 3.4354777312523796e-06, + "loss": 1.4079, + "step": 1129 + }, + { + "epoch": 1.1607601438109914, + "grad_norm": 2.924783945083618, + "learning_rate": 3.433574419489913e-06, + "loss": 1.437, + "step": 1130 + }, + { + "epoch": 1.1607601438109914, + "eval_loss": 1.817325234413147, + "eval_runtime": 37.0048, + "eval_samples_per_second": 27.024, + "eval_steps_per_second": 1.135, + "step": 1130 + }, + { + "epoch": 1.1617873651771957, + "grad_norm": 2.8339531421661377, + "learning_rate": 3.4316711077274463e-06, + "loss": 1.469, + "step": 1131 + }, + { + "epoch": 1.1628145865434, + "grad_norm": 2.880077362060547, + "learning_rate": 3.4297677959649792e-06, + "loss": 1.4364, + "step": 1132 + }, + { + "epoch": 1.1638418079096045, + "grad_norm": 2.96932315826416, + "learning_rate": 3.4278644842025126e-06, + "loss": 1.4868, + "step": 1133 + }, + { + "epoch": 1.1648690292758088, + "grad_norm": 2.643693685531616, + "learning_rate": 3.425961172440046e-06, + "loss": 1.428, + "step": 1134 + }, + { + "epoch": 1.1658962506420134, + "grad_norm": 2.9450795650482178, + "learning_rate": 3.4240578606775793e-06, + "loss": 1.5155, + "step": 1135 + }, + { + "epoch": 1.1669234720082178, + "grad_norm": 2.7274043560028076, + "learning_rate": 3.4221545489151127e-06, + "loss": 1.4486, + "step": 1136 + }, + { + "epoch": 1.1679506933744221, + "grad_norm": 2.7501749992370605, + "learning_rate": 3.4202512371526456e-06, + "loss": 1.4254, + "step": 1137 + }, + { + "epoch": 1.1689779147406265, + "grad_norm": 3.0331339836120605, + "learning_rate": 3.418347925390179e-06, + "loss": 1.5338, + "step": 1138 + }, + { + "epoch": 1.170005136106831, + "grad_norm": 2.9705629348754883, + "learning_rate": 3.4164446136277124e-06, + "loss": 1.474, + "step": 1139 + }, + { + "epoch": 1.1710323574730355, + "grad_norm": 2.76196026802063, + "learning_rate": 3.4145413018652457e-06, + "loss": 1.4505, + "step": 1140 + }, + { + "epoch": 1.1710323574730355, + "eval_loss": 1.8043485879898071, + "eval_runtime": 37.0258, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 1140 + }, + { + "epoch": 1.1720595788392398, + "grad_norm": 2.6624460220336914, + "learning_rate": 3.412637990102779e-06, + "loss": 1.4739, + "step": 1141 + }, + { + "epoch": 1.1730868002054442, + "grad_norm": 2.7865853309631348, + "learning_rate": 3.4107346783403125e-06, + "loss": 1.4923, + "step": 1142 + }, + { + "epoch": 1.1741140215716488, + "grad_norm": 2.782196044921875, + "learning_rate": 3.4088313665778454e-06, + "loss": 1.4041, + "step": 1143 + }, + { + "epoch": 1.1751412429378532, + "grad_norm": 2.7796125411987305, + "learning_rate": 3.4069280548153788e-06, + "loss": 1.4446, + "step": 1144 + }, + { + "epoch": 1.1761684643040575, + "grad_norm": 2.8180930614471436, + "learning_rate": 3.405024743052912e-06, + "loss": 1.424, + "step": 1145 + }, + { + "epoch": 1.177195685670262, + "grad_norm": 2.8180930614471436, + "learning_rate": 3.405024743052912e-06, + "loss": 1.4457, + "step": 1146 + }, + { + "epoch": 1.1782229070364663, + "grad_norm": 2.922241449356079, + "learning_rate": 3.403121431290446e-06, + "loss": 1.4162, + "step": 1147 + }, + { + "epoch": 1.1792501284026708, + "grad_norm": 2.850069761276245, + "learning_rate": 3.4012181195279793e-06, + "loss": 1.432, + "step": 1148 + }, + { + "epoch": 1.1802773497688752, + "grad_norm": 2.863086462020874, + "learning_rate": 3.3993148077655126e-06, + "loss": 1.442, + "step": 1149 + }, + { + "epoch": 1.1813045711350796, + "grad_norm": 2.9545702934265137, + "learning_rate": 3.3974114960030456e-06, + "loss": 1.3931, + "step": 1150 + }, + { + "epoch": 1.1813045711350796, + "eval_loss": 1.812690258026123, + "eval_runtime": 37.0328, + "eval_samples_per_second": 27.003, + "eval_steps_per_second": 1.134, + "step": 1150 + }, + { + "epoch": 1.182331792501284, + "grad_norm": 2.9074196815490723, + "learning_rate": 3.395508184240579e-06, + "loss": 1.4199, + "step": 1151 + }, + { + "epoch": 1.1833590138674885, + "grad_norm": 3.1387181282043457, + "learning_rate": 3.3936048724781123e-06, + "loss": 1.509, + "step": 1152 + }, + { + "epoch": 1.184386235233693, + "grad_norm": 3.0677523612976074, + "learning_rate": 3.3917015607156457e-06, + "loss": 1.3952, + "step": 1153 + }, + { + "epoch": 1.1854134565998973, + "grad_norm": 2.8462986946105957, + "learning_rate": 3.389798248953179e-06, + "loss": 1.415, + "step": 1154 + }, + { + "epoch": 1.1864406779661016, + "grad_norm": 2.9588911533355713, + "learning_rate": 3.387894937190712e-06, + "loss": 1.464, + "step": 1155 + }, + { + "epoch": 1.1874678993323062, + "grad_norm": 2.7406837940216064, + "learning_rate": 3.3859916254282453e-06, + "loss": 1.4604, + "step": 1156 + }, + { + "epoch": 1.1884951206985106, + "grad_norm": 2.837615489959717, + "learning_rate": 3.3840883136657787e-06, + "loss": 1.5108, + "step": 1157 + }, + { + "epoch": 1.189522342064715, + "grad_norm": 2.6457152366638184, + "learning_rate": 3.382185001903312e-06, + "loss": 1.4247, + "step": 1158 + }, + { + "epoch": 1.1905495634309193, + "grad_norm": 2.887178659439087, + "learning_rate": 3.3802816901408454e-06, + "loss": 1.4609, + "step": 1159 + }, + { + "epoch": 1.1915767847971237, + "grad_norm": 2.6568970680236816, + "learning_rate": 3.3783783783783788e-06, + "loss": 1.4393, + "step": 1160 + }, + { + "epoch": 1.1915767847971237, + "eval_loss": 1.7956509590148926, + "eval_runtime": 37.0256, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 1160 + }, + { + "epoch": 1.1926040061633283, + "grad_norm": 2.731114625930786, + "learning_rate": 3.3764750666159117e-06, + "loss": 1.4399, + "step": 1161 + }, + { + "epoch": 1.1936312275295327, + "grad_norm": 2.8750929832458496, + "learning_rate": 3.374571754853445e-06, + "loss": 1.4318, + "step": 1162 + }, + { + "epoch": 1.194658448895737, + "grad_norm": 2.697359085083008, + "learning_rate": 3.3726684430909784e-06, + "loss": 1.3884, + "step": 1163 + }, + { + "epoch": 1.1956856702619414, + "grad_norm": 2.6850969791412354, + "learning_rate": 3.370765131328512e-06, + "loss": 1.4104, + "step": 1164 + }, + { + "epoch": 1.1967128916281458, + "grad_norm": 2.739121913909912, + "learning_rate": 3.368861819566045e-06, + "loss": 1.4227, + "step": 1165 + }, + { + "epoch": 1.1977401129943503, + "grad_norm": 2.7206380367279053, + "learning_rate": 3.366958507803578e-06, + "loss": 1.4587, + "step": 1166 + }, + { + "epoch": 1.1987673343605547, + "grad_norm": 2.651898145675659, + "learning_rate": 3.365055196041112e-06, + "loss": 1.4081, + "step": 1167 + }, + { + "epoch": 1.199794555726759, + "grad_norm": 2.7718563079833984, + "learning_rate": 3.3631518842786452e-06, + "loss": 1.4835, + "step": 1168 + }, + { + "epoch": 1.2008217770929634, + "grad_norm": 2.771209716796875, + "learning_rate": 3.3612485725161786e-06, + "loss": 1.4036, + "step": 1169 + }, + { + "epoch": 1.201848998459168, + "grad_norm": 2.6938540935516357, + "learning_rate": 3.359345260753712e-06, + "loss": 1.4987, + "step": 1170 + }, + { + "epoch": 1.201848998459168, + "eval_loss": 1.784244179725647, + "eval_runtime": 37.0121, + "eval_samples_per_second": 27.018, + "eval_steps_per_second": 1.135, + "step": 1170 + }, + { + "epoch": 1.2028762198253724, + "grad_norm": 2.884730339050293, + "learning_rate": 3.3574419489912453e-06, + "loss": 1.4778, + "step": 1171 + }, + { + "epoch": 1.2039034411915768, + "grad_norm": 3.020632266998291, + "learning_rate": 3.3555386372287783e-06, + "loss": 1.5133, + "step": 1172 + }, + { + "epoch": 1.2049306625577811, + "grad_norm": 2.8026037216186523, + "learning_rate": 3.3536353254663116e-06, + "loss": 1.405, + "step": 1173 + }, + { + "epoch": 1.2059578839239857, + "grad_norm": 2.763852596282959, + "learning_rate": 3.351732013703845e-06, + "loss": 1.4437, + "step": 1174 + }, + { + "epoch": 1.20698510529019, + "grad_norm": 2.897768259048462, + "learning_rate": 3.3498287019413784e-06, + "loss": 1.4548, + "step": 1175 + }, + { + "epoch": 1.2080123266563945, + "grad_norm": 2.8483922481536865, + "learning_rate": 3.3479253901789117e-06, + "loss": 1.4734, + "step": 1176 + }, + { + "epoch": 1.2090395480225988, + "grad_norm": 2.8507916927337646, + "learning_rate": 3.346022078416445e-06, + "loss": 1.378, + "step": 1177 + }, + { + "epoch": 1.2100667693888032, + "grad_norm": 2.74048113822937, + "learning_rate": 3.344118766653978e-06, + "loss": 1.4493, + "step": 1178 + }, + { + "epoch": 1.2110939907550078, + "grad_norm": 2.8064157962799072, + "learning_rate": 3.3422154548915114e-06, + "loss": 1.4221, + "step": 1179 + }, + { + "epoch": 1.2121212121212122, + "grad_norm": 2.721090793609619, + "learning_rate": 3.3403121431290448e-06, + "loss": 1.4603, + "step": 1180 + }, + { + "epoch": 1.2121212121212122, + "eval_loss": 1.7907730340957642, + "eval_runtime": 37.0168, + "eval_samples_per_second": 27.015, + "eval_steps_per_second": 1.135, + "step": 1180 + }, + { + "epoch": 1.2131484334874165, + "grad_norm": 2.706892967224121, + "learning_rate": 3.338408831366578e-06, + "loss": 1.3903, + "step": 1181 + }, + { + "epoch": 1.214175654853621, + "grad_norm": 3.0193893909454346, + "learning_rate": 3.3365055196041115e-06, + "loss": 1.4715, + "step": 1182 + }, + { + "epoch": 1.2152028762198253, + "grad_norm": 2.8713300228118896, + "learning_rate": 3.3346022078416444e-06, + "loss": 1.5017, + "step": 1183 + }, + { + "epoch": 1.2162300975860298, + "grad_norm": 2.7250728607177734, + "learning_rate": 3.3326988960791778e-06, + "loss": 1.4242, + "step": 1184 + }, + { + "epoch": 1.2172573189522342, + "grad_norm": 2.9367101192474365, + "learning_rate": 3.330795584316711e-06, + "loss": 1.3563, + "step": 1185 + }, + { + "epoch": 1.2182845403184386, + "grad_norm": 3.059776544570923, + "learning_rate": 3.3288922725542445e-06, + "loss": 1.4334, + "step": 1186 + }, + { + "epoch": 1.219311761684643, + "grad_norm": 2.996394157409668, + "learning_rate": 3.326988960791778e-06, + "loss": 1.4457, + "step": 1187 + }, + { + "epoch": 1.2203389830508475, + "grad_norm": 2.8958983421325684, + "learning_rate": 3.3250856490293117e-06, + "loss": 1.4904, + "step": 1188 + }, + { + "epoch": 1.221366204417052, + "grad_norm": 2.9497668743133545, + "learning_rate": 3.323182337266845e-06, + "loss": 1.4836, + "step": 1189 + }, + { + "epoch": 1.2223934257832563, + "grad_norm": 2.87447452545166, + "learning_rate": 3.321279025504378e-06, + "loss": 1.4664, + "step": 1190 + }, + { + "epoch": 1.2223934257832563, + "eval_loss": 1.7865233421325684, + "eval_runtime": 37.1144, + "eval_samples_per_second": 26.944, + "eval_steps_per_second": 1.132, + "step": 1190 + }, + { + "epoch": 1.2234206471494606, + "grad_norm": 2.7893576622009277, + "learning_rate": 3.3193757137419113e-06, + "loss": 1.4584, + "step": 1191 + }, + { + "epoch": 1.2244478685156652, + "grad_norm": 2.791938543319702, + "learning_rate": 3.3174724019794447e-06, + "loss": 1.4974, + "step": 1192 + }, + { + "epoch": 1.2254750898818696, + "grad_norm": 2.7816781997680664, + "learning_rate": 3.315569090216978e-06, + "loss": 1.4103, + "step": 1193 + }, + { + "epoch": 1.226502311248074, + "grad_norm": 2.938931703567505, + "learning_rate": 3.3136657784545114e-06, + "loss": 1.498, + "step": 1194 + }, + { + "epoch": 1.2275295326142783, + "grad_norm": 2.7500035762786865, + "learning_rate": 3.3117624666920443e-06, + "loss": 1.4044, + "step": 1195 + }, + { + "epoch": 1.2285567539804827, + "grad_norm": 2.758352756500244, + "learning_rate": 3.3098591549295777e-06, + "loss": 1.3968, + "step": 1196 + }, + { + "epoch": 1.2295839753466873, + "grad_norm": 3.144857406616211, + "learning_rate": 3.307955843167111e-06, + "loss": 1.459, + "step": 1197 + }, + { + "epoch": 1.2306111967128917, + "grad_norm": 2.777985095977783, + "learning_rate": 3.3060525314046444e-06, + "loss": 1.4419, + "step": 1198 + }, + { + "epoch": 1.231638418079096, + "grad_norm": 2.812570810317993, + "learning_rate": 3.304149219642178e-06, + "loss": 1.446, + "step": 1199 + }, + { + "epoch": 1.2326656394453004, + "grad_norm": 2.6672797203063965, + "learning_rate": 3.3022459078797107e-06, + "loss": 1.4615, + "step": 1200 + }, + { + "epoch": 1.2326656394453004, + "eval_loss": 1.7857756614685059, + "eval_runtime": 37.313, + "eval_samples_per_second": 26.8, + "eval_steps_per_second": 1.126, + "step": 1200 + }, + { + "epoch": 1.2336928608115048, + "grad_norm": 2.8981986045837402, + "learning_rate": 3.300342596117244e-06, + "loss": 1.4418, + "step": 1201 + }, + { + "epoch": 1.2347200821777093, + "grad_norm": 2.7738826274871826, + "learning_rate": 3.2984392843547775e-06, + "loss": 1.3926, + "step": 1202 + }, + { + "epoch": 1.2357473035439137, + "grad_norm": 2.735461950302124, + "learning_rate": 3.296535972592311e-06, + "loss": 1.3843, + "step": 1203 + }, + { + "epoch": 1.236774524910118, + "grad_norm": 2.965693950653076, + "learning_rate": 3.294632660829844e-06, + "loss": 1.4248, + "step": 1204 + }, + { + "epoch": 1.2378017462763227, + "grad_norm": 2.850008010864258, + "learning_rate": 3.292729349067377e-06, + "loss": 1.4277, + "step": 1205 + }, + { + "epoch": 1.238828967642527, + "grad_norm": 2.823054790496826, + "learning_rate": 3.2908260373049105e-06, + "loss": 1.4151, + "step": 1206 + }, + { + "epoch": 1.2398561890087314, + "grad_norm": 2.7468149662017822, + "learning_rate": 3.288922725542444e-06, + "loss": 1.4661, + "step": 1207 + }, + { + "epoch": 1.2408834103749358, + "grad_norm": 2.7379629611968994, + "learning_rate": 3.2870194137799776e-06, + "loss": 1.4219, + "step": 1208 + }, + { + "epoch": 1.2419106317411401, + "grad_norm": 2.7809221744537354, + "learning_rate": 3.285116102017511e-06, + "loss": 1.4002, + "step": 1209 + }, + { + "epoch": 1.2429378531073447, + "grad_norm": 2.9209654331207275, + "learning_rate": 3.2832127902550444e-06, + "loss": 1.4646, + "step": 1210 + }, + { + "epoch": 1.2429378531073447, + "eval_loss": 1.786431908607483, + "eval_runtime": 36.9218, + "eval_samples_per_second": 27.084, + "eval_steps_per_second": 1.138, + "step": 1210 + }, + { + "epoch": 1.243965074473549, + "grad_norm": 2.9019596576690674, + "learning_rate": 3.2813094784925777e-06, + "loss": 1.4526, + "step": 1211 + }, + { + "epoch": 1.2449922958397535, + "grad_norm": 2.7218263149261475, + "learning_rate": 3.2794061667301107e-06, + "loss": 1.3911, + "step": 1212 + }, + { + "epoch": 1.2460195172059578, + "grad_norm": 2.8646154403686523, + "learning_rate": 3.277502854967644e-06, + "loss": 1.4305, + "step": 1213 + }, + { + "epoch": 1.2470467385721622, + "grad_norm": 2.9065611362457275, + "learning_rate": 3.2755995432051774e-06, + "loss": 1.4239, + "step": 1214 + }, + { + "epoch": 1.2480739599383668, + "grad_norm": 2.7945196628570557, + "learning_rate": 3.2736962314427107e-06, + "loss": 1.4321, + "step": 1215 + }, + { + "epoch": 1.2491011813045712, + "grad_norm": 2.644244909286499, + "learning_rate": 3.271792919680244e-06, + "loss": 1.4411, + "step": 1216 + }, + { + "epoch": 1.2501284026707755, + "grad_norm": 2.9058117866516113, + "learning_rate": 3.269889607917777e-06, + "loss": 1.4942, + "step": 1217 + }, + { + "epoch": 1.25115562403698, + "grad_norm": 2.8242623805999756, + "learning_rate": 3.2679862961553104e-06, + "loss": 1.4033, + "step": 1218 + }, + { + "epoch": 1.2521828454031843, + "grad_norm": 2.859428644180298, + "learning_rate": 3.2660829843928438e-06, + "loss": 1.4298, + "step": 1219 + }, + { + "epoch": 1.2532100667693888, + "grad_norm": 2.7864537239074707, + "learning_rate": 3.264179672630377e-06, + "loss": 1.4827, + "step": 1220 + }, + { + "epoch": 1.2532100667693888, + "eval_loss": 1.7798999547958374, + "eval_runtime": 37.0394, + "eval_samples_per_second": 26.998, + "eval_steps_per_second": 1.134, + "step": 1220 + }, + { + "epoch": 1.2542372881355932, + "grad_norm": 2.7938263416290283, + "learning_rate": 3.2622763608679105e-06, + "loss": 1.4412, + "step": 1221 + }, + { + "epoch": 1.2552645095017976, + "grad_norm": 2.811281204223633, + "learning_rate": 3.2603730491054434e-06, + "loss": 1.4139, + "step": 1222 + }, + { + "epoch": 1.2562917308680022, + "grad_norm": 2.7984230518341064, + "learning_rate": 3.258469737342977e-06, + "loss": 1.4205, + "step": 1223 + }, + { + "epoch": 1.2573189522342065, + "grad_norm": 2.7457568645477295, + "learning_rate": 3.25656642558051e-06, + "loss": 1.396, + "step": 1224 + }, + { + "epoch": 1.258346173600411, + "grad_norm": 2.84529972076416, + "learning_rate": 3.2546631138180435e-06, + "loss": 1.4192, + "step": 1225 + }, + { + "epoch": 1.2593733949666153, + "grad_norm": 2.8183720111846924, + "learning_rate": 3.252759802055577e-06, + "loss": 1.4104, + "step": 1226 + }, + { + "epoch": 1.2604006163328196, + "grad_norm": 2.8872454166412354, + "learning_rate": 3.25085649029311e-06, + "loss": 1.4256, + "step": 1227 + }, + { + "epoch": 1.2614278376990242, + "grad_norm": 2.935094118118286, + "learning_rate": 3.248953178530644e-06, + "loss": 1.4558, + "step": 1228 + }, + { + "epoch": 1.2624550590652286, + "grad_norm": 2.69989013671875, + "learning_rate": 3.247049866768177e-06, + "loss": 1.4167, + "step": 1229 + }, + { + "epoch": 1.263482280431433, + "grad_norm": 2.781031847000122, + "learning_rate": 3.2451465550057103e-06, + "loss": 1.3745, + "step": 1230 + }, + { + "epoch": 1.263482280431433, + "eval_loss": 1.7843073606491089, + "eval_runtime": 37.0672, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 1230 + }, + { + "epoch": 1.2645095017976373, + "grad_norm": 2.974919080734253, + "learning_rate": 3.2432432432432437e-06, + "loss": 1.4221, + "step": 1231 + }, + { + "epoch": 1.2655367231638417, + "grad_norm": 3.095144271850586, + "learning_rate": 3.241339931480777e-06, + "loss": 1.4539, + "step": 1232 + }, + { + "epoch": 1.2665639445300463, + "grad_norm": 3.0127196311950684, + "learning_rate": 3.2394366197183104e-06, + "loss": 1.4559, + "step": 1233 + }, + { + "epoch": 1.2675911658962506, + "grad_norm": 2.7474801540374756, + "learning_rate": 3.2375333079558434e-06, + "loss": 1.4177, + "step": 1234 + }, + { + "epoch": 1.268618387262455, + "grad_norm": 2.743516445159912, + "learning_rate": 3.2356299961933767e-06, + "loss": 1.4471, + "step": 1235 + }, + { + "epoch": 1.2696456086286596, + "grad_norm": 2.7801740169525146, + "learning_rate": 3.23372668443091e-06, + "loss": 1.4588, + "step": 1236 + }, + { + "epoch": 1.270672829994864, + "grad_norm": 2.697547435760498, + "learning_rate": 3.2318233726684435e-06, + "loss": 1.4857, + "step": 1237 + }, + { + "epoch": 1.2717000513610683, + "grad_norm": 2.646890640258789, + "learning_rate": 3.229920060905977e-06, + "loss": 1.4068, + "step": 1238 + }, + { + "epoch": 1.2727272727272727, + "grad_norm": 2.694401979446411, + "learning_rate": 3.2280167491435098e-06, + "loss": 1.4148, + "step": 1239 + }, + { + "epoch": 1.273754494093477, + "grad_norm": 2.6139938831329346, + "learning_rate": 3.226113437381043e-06, + "loss": 1.4214, + "step": 1240 + }, + { + "epoch": 1.273754494093477, + "eval_loss": 1.7777915000915527, + "eval_runtime": 37.1013, + "eval_samples_per_second": 26.953, + "eval_steps_per_second": 1.132, + "step": 1240 + }, + { + "epoch": 1.2747817154596817, + "grad_norm": 2.7611944675445557, + "learning_rate": 3.2242101256185765e-06, + "loss": 1.4344, + "step": 1241 + }, + { + "epoch": 1.275808936825886, + "grad_norm": 2.7328102588653564, + "learning_rate": 3.22230681385611e-06, + "loss": 1.4891, + "step": 1242 + }, + { + "epoch": 1.2768361581920904, + "grad_norm": 2.822169065475464, + "learning_rate": 3.220403502093643e-06, + "loss": 1.4328, + "step": 1243 + }, + { + "epoch": 1.2778633795582948, + "grad_norm": 2.878005027770996, + "learning_rate": 3.218500190331176e-06, + "loss": 1.4779, + "step": 1244 + }, + { + "epoch": 1.2788906009244991, + "grad_norm": 2.725376844406128, + "learning_rate": 3.2165968785687095e-06, + "loss": 1.424, + "step": 1245 + }, + { + "epoch": 1.2799178222907037, + "grad_norm": 2.7598440647125244, + "learning_rate": 3.214693566806243e-06, + "loss": 1.4533, + "step": 1246 + }, + { + "epoch": 1.280945043656908, + "grad_norm": 2.8355820178985596, + "learning_rate": 3.2127902550437762e-06, + "loss": 1.4433, + "step": 1247 + }, + { + "epoch": 1.2819722650231125, + "grad_norm": 2.7520253658294678, + "learning_rate": 3.21088694328131e-06, + "loss": 1.4318, + "step": 1248 + }, + { + "epoch": 1.2829994863893168, + "grad_norm": 2.7606403827667236, + "learning_rate": 3.2089836315188434e-06, + "loss": 1.4242, + "step": 1249 + }, + { + "epoch": 1.2840267077555212, + "grad_norm": 2.784301280975342, + "learning_rate": 3.2070803197563767e-06, + "loss": 1.4253, + "step": 1250 + }, + { + "epoch": 1.2840267077555212, + "eval_loss": 1.7868103981018066, + "eval_runtime": 37.0911, + "eval_samples_per_second": 26.961, + "eval_steps_per_second": 1.132, + "step": 1250 + }, + { + "epoch": 1.2850539291217258, + "grad_norm": 2.8295185565948486, + "learning_rate": 3.2051770079939097e-06, + "loss": 1.4432, + "step": 1251 + }, + { + "epoch": 1.2860811504879301, + "grad_norm": 2.8488125801086426, + "learning_rate": 3.203273696231443e-06, + "loss": 1.4714, + "step": 1252 + }, + { + "epoch": 1.2871083718541345, + "grad_norm": 2.771249294281006, + "learning_rate": 3.2013703844689764e-06, + "loss": 1.4093, + "step": 1253 + }, + { + "epoch": 1.288135593220339, + "grad_norm": 2.782425880432129, + "learning_rate": 3.1994670727065098e-06, + "loss": 1.4597, + "step": 1254 + }, + { + "epoch": 1.2891628145865435, + "grad_norm": 2.88934588432312, + "learning_rate": 3.197563760944043e-06, + "loss": 1.4805, + "step": 1255 + }, + { + "epoch": 1.2901900359527478, + "grad_norm": 2.8694138526916504, + "learning_rate": 3.195660449181576e-06, + "loss": 1.4699, + "step": 1256 + }, + { + "epoch": 1.2912172573189522, + "grad_norm": 2.799614667892456, + "learning_rate": 3.1937571374191094e-06, + "loss": 1.4044, + "step": 1257 + }, + { + "epoch": 1.2922444786851566, + "grad_norm": 3.001746416091919, + "learning_rate": 3.191853825656643e-06, + "loss": 1.4305, + "step": 1258 + }, + { + "epoch": 1.2932717000513612, + "grad_norm": 2.8572680950164795, + "learning_rate": 3.189950513894176e-06, + "loss": 1.4252, + "step": 1259 + }, + { + "epoch": 1.2942989214175655, + "grad_norm": 2.942213535308838, + "learning_rate": 3.1880472021317095e-06, + "loss": 1.427, + "step": 1260 + }, + { + "epoch": 1.2942989214175655, + "eval_loss": 1.7929412126541138, + "eval_runtime": 37.087, + "eval_samples_per_second": 26.964, + "eval_steps_per_second": 1.132, + "step": 1260 + }, + { + "epoch": 1.29532614278377, + "grad_norm": 2.893172264099121, + "learning_rate": 3.1861438903692425e-06, + "loss": 1.379, + "step": 1261 + }, + { + "epoch": 1.2963533641499743, + "grad_norm": 2.9168691635131836, + "learning_rate": 3.184240578606776e-06, + "loss": 1.4124, + "step": 1262 + }, + { + "epoch": 1.2973805855161786, + "grad_norm": 2.7513487339019775, + "learning_rate": 3.182337266844309e-06, + "loss": 1.4343, + "step": 1263 + }, + { + "epoch": 1.2984078068823832, + "grad_norm": 2.8485500812530518, + "learning_rate": 3.1804339550818426e-06, + "loss": 1.467, + "step": 1264 + }, + { + "epoch": 1.2994350282485876, + "grad_norm": 2.6620397567749023, + "learning_rate": 3.178530643319376e-06, + "loss": 1.4191, + "step": 1265 + }, + { + "epoch": 1.300462249614792, + "grad_norm": 2.659393072128296, + "learning_rate": 3.176627331556909e-06, + "loss": 1.3934, + "step": 1266 + }, + { + "epoch": 1.3014894709809965, + "grad_norm": 2.8579442501068115, + "learning_rate": 3.1747240197944422e-06, + "loss": 1.4399, + "step": 1267 + }, + { + "epoch": 1.3025166923472007, + "grad_norm": 2.7417430877685547, + "learning_rate": 3.172820708031976e-06, + "loss": 1.3956, + "step": 1268 + }, + { + "epoch": 1.3035439137134053, + "grad_norm": 3.1082868576049805, + "learning_rate": 3.1709173962695094e-06, + "loss": 1.4768, + "step": 1269 + }, + { + "epoch": 1.3045711350796096, + "grad_norm": 3.099637985229492, + "learning_rate": 3.1690140845070427e-06, + "loss": 1.4679, + "step": 1270 + }, + { + "epoch": 1.3045711350796096, + "eval_loss": 1.7878395318984985, + "eval_runtime": 37.0795, + "eval_samples_per_second": 26.969, + "eval_steps_per_second": 1.133, + "step": 1270 + }, + { + "epoch": 1.305598356445814, + "grad_norm": 2.8106865882873535, + "learning_rate": 3.167110772744576e-06, + "loss": 1.4257, + "step": 1271 + }, + { + "epoch": 1.3066255778120186, + "grad_norm": 2.7419369220733643, + "learning_rate": 3.1652074609821095e-06, + "loss": 1.4197, + "step": 1272 + }, + { + "epoch": 1.307652799178223, + "grad_norm": 2.7077760696411133, + "learning_rate": 3.1633041492196424e-06, + "loss": 1.4188, + "step": 1273 + }, + { + "epoch": 1.3086800205444273, + "grad_norm": 2.6611413955688477, + "learning_rate": 3.1614008374571758e-06, + "loss": 1.4456, + "step": 1274 + }, + { + "epoch": 1.3097072419106317, + "grad_norm": 2.8461341857910156, + "learning_rate": 3.159497525694709e-06, + "loss": 1.4544, + "step": 1275 + }, + { + "epoch": 1.310734463276836, + "grad_norm": 2.7263684272766113, + "learning_rate": 3.1575942139322425e-06, + "loss": 1.4177, + "step": 1276 + }, + { + "epoch": 1.3117616846430407, + "grad_norm": 2.8571529388427734, + "learning_rate": 3.155690902169776e-06, + "loss": 1.4473, + "step": 1277 + }, + { + "epoch": 1.312788906009245, + "grad_norm": 2.914841651916504, + "learning_rate": 3.1537875904073088e-06, + "loss": 1.4117, + "step": 1278 + }, + { + "epoch": 1.3138161273754494, + "grad_norm": 2.7753429412841797, + "learning_rate": 3.151884278644842e-06, + "loss": 1.4359, + "step": 1279 + }, + { + "epoch": 1.3148433487416538, + "grad_norm": 2.624159097671509, + "learning_rate": 3.1499809668823755e-06, + "loss": 1.4072, + "step": 1280 + }, + { + "epoch": 1.3148433487416538, + "eval_loss": 1.7820065021514893, + "eval_runtime": 37.0679, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 1280 + }, + { + "epoch": 1.3158705701078581, + "grad_norm": 2.742490291595459, + "learning_rate": 3.148077655119909e-06, + "loss": 1.4681, + "step": 1281 + }, + { + "epoch": 1.3168977914740627, + "grad_norm": 2.6784660816192627, + "learning_rate": 3.1461743433574422e-06, + "loss": 1.406, + "step": 1282 + }, + { + "epoch": 1.317925012840267, + "grad_norm": 2.746730089187622, + "learning_rate": 3.144271031594975e-06, + "loss": 1.4538, + "step": 1283 + }, + { + "epoch": 1.3189522342064715, + "grad_norm": 2.746652603149414, + "learning_rate": 3.1423677198325085e-06, + "loss": 1.4574, + "step": 1284 + }, + { + "epoch": 1.319979455572676, + "grad_norm": 2.842672109603882, + "learning_rate": 3.140464408070042e-06, + "loss": 1.414, + "step": 1285 + }, + { + "epoch": 1.3210066769388804, + "grad_norm": 2.7332921028137207, + "learning_rate": 3.1385610963075753e-06, + "loss": 1.4115, + "step": 1286 + }, + { + "epoch": 1.3220338983050848, + "grad_norm": 2.6771159172058105, + "learning_rate": 3.1366577845451086e-06, + "loss": 1.4023, + "step": 1287 + }, + { + "epoch": 1.3230611196712891, + "grad_norm": 2.689605951309204, + "learning_rate": 3.134754472782642e-06, + "loss": 1.3897, + "step": 1288 + }, + { + "epoch": 1.3240883410374935, + "grad_norm": 2.7937700748443604, + "learning_rate": 3.1328511610201758e-06, + "loss": 1.3783, + "step": 1289 + }, + { + "epoch": 1.325115562403698, + "grad_norm": 2.643263816833496, + "learning_rate": 3.1309478492577087e-06, + "loss": 1.378, + "step": 1290 + }, + { + "epoch": 1.325115562403698, + "eval_loss": 1.7850252389907837, + "eval_runtime": 37.0816, + "eval_samples_per_second": 26.968, + "eval_steps_per_second": 1.133, + "step": 1290 + }, + { + "epoch": 1.3261427837699025, + "grad_norm": 2.7702224254608154, + "learning_rate": 3.129044537495242e-06, + "loss": 1.4236, + "step": 1291 + }, + { + "epoch": 1.3271700051361068, + "grad_norm": 2.783445358276367, + "learning_rate": 3.1271412257327754e-06, + "loss": 1.4454, + "step": 1292 + }, + { + "epoch": 1.3281972265023112, + "grad_norm": 2.556509494781494, + "learning_rate": 3.125237913970309e-06, + "loss": 1.3998, + "step": 1293 + }, + { + "epoch": 1.3292244478685156, + "grad_norm": 2.706526041030884, + "learning_rate": 3.123334602207842e-06, + "loss": 1.3781, + "step": 1294 + }, + { + "epoch": 1.3302516692347202, + "grad_norm": 2.9328322410583496, + "learning_rate": 3.121431290445375e-06, + "loss": 1.4747, + "step": 1295 + }, + { + "epoch": 1.3312788906009245, + "grad_norm": 3.0525310039520264, + "learning_rate": 3.1195279786829085e-06, + "loss": 1.5197, + "step": 1296 + }, + { + "epoch": 1.332306111967129, + "grad_norm": 2.715700149536133, + "learning_rate": 3.117624666920442e-06, + "loss": 1.4059, + "step": 1297 + }, + { + "epoch": 1.3333333333333333, + "grad_norm": 2.7344141006469727, + "learning_rate": 3.115721355157975e-06, + "loss": 1.4101, + "step": 1298 + }, + { + "epoch": 1.3343605546995376, + "grad_norm": 2.746990203857422, + "learning_rate": 3.1138180433955085e-06, + "loss": 1.3625, + "step": 1299 + }, + { + "epoch": 1.3353877760657422, + "grad_norm": 2.817866086959839, + "learning_rate": 3.1119147316330415e-06, + "loss": 1.4414, + "step": 1300 + }, + { + "epoch": 1.3353877760657422, + "eval_loss": 1.7872337102890015, + "eval_runtime": 37.2584, + "eval_samples_per_second": 26.84, + "eval_steps_per_second": 1.127, + "step": 1300 + }, + { + "epoch": 1.3364149974319466, + "grad_norm": 2.814664125442505, + "learning_rate": 3.110011419870575e-06, + "loss": 1.4768, + "step": 1301 + }, + { + "epoch": 1.337442218798151, + "grad_norm": 2.7719714641571045, + "learning_rate": 3.1081081081081082e-06, + "loss": 1.4128, + "step": 1302 + }, + { + "epoch": 1.3384694401643555, + "grad_norm": 2.8300364017486572, + "learning_rate": 3.1062047963456416e-06, + "loss": 1.4419, + "step": 1303 + }, + { + "epoch": 1.33949666153056, + "grad_norm": 2.943674087524414, + "learning_rate": 3.104301484583175e-06, + "loss": 1.4841, + "step": 1304 + }, + { + "epoch": 1.3405238828967643, + "grad_norm": 2.7583727836608887, + "learning_rate": 3.1023981728207083e-06, + "loss": 1.4674, + "step": 1305 + }, + { + "epoch": 1.3415511042629686, + "grad_norm": 2.716261386871338, + "learning_rate": 3.1004948610582412e-06, + "loss": 1.4292, + "step": 1306 + }, + { + "epoch": 1.342578325629173, + "grad_norm": 2.8233978748321533, + "learning_rate": 3.0985915492957746e-06, + "loss": 1.4208, + "step": 1307 + }, + { + "epoch": 1.3436055469953776, + "grad_norm": 2.829420804977417, + "learning_rate": 3.096688237533308e-06, + "loss": 1.3828, + "step": 1308 + }, + { + "epoch": 1.344632768361582, + "grad_norm": 2.791999101638794, + "learning_rate": 3.0947849257708418e-06, + "loss": 1.4751, + "step": 1309 + }, + { + "epoch": 1.3456599897277863, + "grad_norm": 2.691257953643799, + "learning_rate": 3.092881614008375e-06, + "loss": 1.423, + "step": 1310 + }, + { + "epoch": 1.3456599897277863, + "eval_loss": 1.7816330194473267, + "eval_runtime": 36.9065, + "eval_samples_per_second": 27.095, + "eval_steps_per_second": 1.138, + "step": 1310 + }, + { + "epoch": 1.3466872110939907, + "grad_norm": 2.8363473415374756, + "learning_rate": 3.0909783022459085e-06, + "loss": 1.4771, + "step": 1311 + }, + { + "epoch": 1.347714432460195, + "grad_norm": 2.720055103302002, + "learning_rate": 3.0890749904834414e-06, + "loss": 1.4126, + "step": 1312 + }, + { + "epoch": 1.3487416538263997, + "grad_norm": 2.8794236183166504, + "learning_rate": 3.0871716787209748e-06, + "loss": 1.4649, + "step": 1313 + }, + { + "epoch": 1.349768875192604, + "grad_norm": 2.760241746902466, + "learning_rate": 3.085268366958508e-06, + "loss": 1.4427, + "step": 1314 + }, + { + "epoch": 1.3507960965588084, + "grad_norm": 3.031611680984497, + "learning_rate": 3.0833650551960415e-06, + "loss": 1.4476, + "step": 1315 + }, + { + "epoch": 1.3518233179250128, + "grad_norm": 2.7548346519470215, + "learning_rate": 3.081461743433575e-06, + "loss": 1.4036, + "step": 1316 + }, + { + "epoch": 1.3528505392912171, + "grad_norm": 2.7737832069396973, + "learning_rate": 3.079558431671108e-06, + "loss": 1.4209, + "step": 1317 + }, + { + "epoch": 1.3538777606574217, + "grad_norm": 2.580578565597534, + "learning_rate": 3.077655119908641e-06, + "loss": 1.3899, + "step": 1318 + }, + { + "epoch": 1.354904982023626, + "grad_norm": 2.7918219566345215, + "learning_rate": 3.0757518081461745e-06, + "loss": 1.4205, + "step": 1319 + }, + { + "epoch": 1.3559322033898304, + "grad_norm": 2.807173490524292, + "learning_rate": 3.073848496383708e-06, + "loss": 1.4371, + "step": 1320 + }, + { + "epoch": 1.3559322033898304, + "eval_loss": 1.7834100723266602, + "eval_runtime": 37.0369, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 1320 + }, + { + "epoch": 1.356959424756035, + "grad_norm": 2.856569290161133, + "learning_rate": 3.0719451846212413e-06, + "loss": 1.4054, + "step": 1321 + }, + { + "epoch": 1.3579866461222394, + "grad_norm": 2.850403308868408, + "learning_rate": 3.0700418728587746e-06, + "loss": 1.4248, + "step": 1322 + }, + { + "epoch": 1.3590138674884438, + "grad_norm": 2.8618571758270264, + "learning_rate": 3.0681385610963076e-06, + "loss": 1.4066, + "step": 1323 + }, + { + "epoch": 1.3600410888546481, + "grad_norm": 3.0298640727996826, + "learning_rate": 3.066235249333841e-06, + "loss": 1.4479, + "step": 1324 + }, + { + "epoch": 1.3610683102208525, + "grad_norm": 2.9725687503814697, + "learning_rate": 3.0643319375713743e-06, + "loss": 1.4837, + "step": 1325 + }, + { + "epoch": 1.362095531587057, + "grad_norm": 2.8397116661071777, + "learning_rate": 3.0624286258089076e-06, + "loss": 1.4248, + "step": 1326 + }, + { + "epoch": 1.3631227529532615, + "grad_norm": 2.7257447242736816, + "learning_rate": 3.060525314046441e-06, + "loss": 1.4047, + "step": 1327 + }, + { + "epoch": 1.3641499743194658, + "grad_norm": 2.832534074783325, + "learning_rate": 3.058622002283974e-06, + "loss": 1.4491, + "step": 1328 + }, + { + "epoch": 1.3651771956856702, + "grad_norm": 2.6238863468170166, + "learning_rate": 3.0567186905215077e-06, + "loss": 1.457, + "step": 1329 + }, + { + "epoch": 1.3662044170518746, + "grad_norm": 2.7993323802948, + "learning_rate": 3.054815378759041e-06, + "loss": 1.458, + "step": 1330 + }, + { + "epoch": 1.3662044170518746, + "eval_loss": 1.7857885360717773, + "eval_runtime": 37.0668, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 1330 + }, + { + "epoch": 1.3672316384180792, + "grad_norm": 2.798299789428711, + "learning_rate": 3.0529120669965745e-06, + "loss": 1.4539, + "step": 1331 + }, + { + "epoch": 1.3682588597842835, + "grad_norm": 2.7995524406433105, + "learning_rate": 3.051008755234108e-06, + "loss": 1.3929, + "step": 1332 + }, + { + "epoch": 1.3692860811504879, + "grad_norm": 3.026519298553467, + "learning_rate": 3.049105443471641e-06, + "loss": 1.4, + "step": 1333 + }, + { + "epoch": 1.3703133025166925, + "grad_norm": 3.040426731109619, + "learning_rate": 3.0472021317091745e-06, + "loss": 1.5383, + "step": 1334 + }, + { + "epoch": 1.3713405238828968, + "grad_norm": 2.8738269805908203, + "learning_rate": 3.0452988199467075e-06, + "loss": 1.4581, + "step": 1335 + }, + { + "epoch": 1.3723677452491012, + "grad_norm": 2.895745038986206, + "learning_rate": 3.043395508184241e-06, + "loss": 1.4508, + "step": 1336 + }, + { + "epoch": 1.3733949666153056, + "grad_norm": 2.872739315032959, + "learning_rate": 3.041492196421774e-06, + "loss": 1.413, + "step": 1337 + }, + { + "epoch": 1.37442218798151, + "grad_norm": 2.717970132827759, + "learning_rate": 3.0395888846593076e-06, + "loss": 1.4433, + "step": 1338 + }, + { + "epoch": 1.3754494093477145, + "grad_norm": 2.799251079559326, + "learning_rate": 3.037685572896841e-06, + "loss": 1.4237, + "step": 1339 + }, + { + "epoch": 1.376476630713919, + "grad_norm": 2.8966126441955566, + "learning_rate": 3.035782261134374e-06, + "loss": 1.4386, + "step": 1340 + }, + { + "epoch": 1.376476630713919, + "eval_loss": 1.795745611190796, + "eval_runtime": 37.0421, + "eval_samples_per_second": 26.996, + "eval_steps_per_second": 1.134, + "step": 1340 + }, + { + "epoch": 1.3775038520801233, + "grad_norm": 3.0871410369873047, + "learning_rate": 3.0338789493719072e-06, + "loss": 1.4416, + "step": 1341 + }, + { + "epoch": 1.3785310734463276, + "grad_norm": 2.829460382461548, + "learning_rate": 3.0319756376094406e-06, + "loss": 1.4361, + "step": 1342 + }, + { + "epoch": 1.379558294812532, + "grad_norm": 3.0431859493255615, + "learning_rate": 3.030072325846974e-06, + "loss": 1.4431, + "step": 1343 + }, + { + "epoch": 1.3805855161787366, + "grad_norm": 2.966308116912842, + "learning_rate": 3.0281690140845073e-06, + "loss": 1.4022, + "step": 1344 + }, + { + "epoch": 1.381612737544941, + "grad_norm": 2.976332664489746, + "learning_rate": 3.0262657023220403e-06, + "loss": 1.4481, + "step": 1345 + }, + { + "epoch": 1.3826399589111453, + "grad_norm": 2.716144561767578, + "learning_rate": 3.0243623905595736e-06, + "loss": 1.4437, + "step": 1346 + }, + { + "epoch": 1.3836671802773497, + "grad_norm": 2.913194417953491, + "learning_rate": 3.022459078797107e-06, + "loss": 1.4414, + "step": 1347 + }, + { + "epoch": 1.384694401643554, + "grad_norm": 2.9889321327209473, + "learning_rate": 3.0205557670346404e-06, + "loss": 1.4333, + "step": 1348 + }, + { + "epoch": 1.3857216230097587, + "grad_norm": 2.657104730606079, + "learning_rate": 3.018652455272174e-06, + "loss": 1.3999, + "step": 1349 + }, + { + "epoch": 1.386748844375963, + "grad_norm": 2.988452196121216, + "learning_rate": 3.0167491435097075e-06, + "loss": 1.5192, + "step": 1350 + }, + { + "epoch": 1.386748844375963, + "eval_loss": 1.7963600158691406, + "eval_runtime": 37.062, + "eval_samples_per_second": 26.982, + "eval_steps_per_second": 1.133, + "step": 1350 + }, + { + "epoch": 1.3877760657421674, + "grad_norm": 2.9976272583007812, + "learning_rate": 3.014845831747241e-06, + "loss": 1.4288, + "step": 1351 + }, + { + "epoch": 1.388803287108372, + "grad_norm": 2.9255707263946533, + "learning_rate": 3.012942519984774e-06, + "loss": 1.4047, + "step": 1352 + }, + { + "epoch": 1.3898305084745763, + "grad_norm": 2.995906352996826, + "learning_rate": 3.011039208222307e-06, + "loss": 1.4151, + "step": 1353 + }, + { + "epoch": 1.3908577298407807, + "grad_norm": 2.745427370071411, + "learning_rate": 3.0091358964598405e-06, + "loss": 1.3784, + "step": 1354 + }, + { + "epoch": 1.391884951206985, + "grad_norm": 2.8206543922424316, + "learning_rate": 3.007232584697374e-06, + "loss": 1.4307, + "step": 1355 + }, + { + "epoch": 1.3929121725731894, + "grad_norm": 2.7975900173187256, + "learning_rate": 3.0053292729349073e-06, + "loss": 1.4363, + "step": 1356 + }, + { + "epoch": 1.393939393939394, + "grad_norm": 2.8240561485290527, + "learning_rate": 3.00342596117244e-06, + "loss": 1.4284, + "step": 1357 + }, + { + "epoch": 1.3949666153055984, + "grad_norm": 2.8556740283966064, + "learning_rate": 3.0015226494099736e-06, + "loss": 1.4252, + "step": 1358 + }, + { + "epoch": 1.3959938366718028, + "grad_norm": 2.828031539916992, + "learning_rate": 2.999619337647507e-06, + "loss": 1.4068, + "step": 1359 + }, + { + "epoch": 1.3970210580380071, + "grad_norm": 2.8133797645568848, + "learning_rate": 2.9977160258850403e-06, + "loss": 1.4114, + "step": 1360 + }, + { + "epoch": 1.3970210580380071, + "eval_loss": 1.7951425313949585, + "eval_runtime": 37.0796, + "eval_samples_per_second": 26.969, + "eval_steps_per_second": 1.133, + "step": 1360 + }, + { + "epoch": 1.3980482794042115, + "grad_norm": 2.8920187950134277, + "learning_rate": 2.9958127141225736e-06, + "loss": 1.3862, + "step": 1361 + }, + { + "epoch": 1.399075500770416, + "grad_norm": 2.936598777770996, + "learning_rate": 2.9939094023601066e-06, + "loss": 1.4565, + "step": 1362 + }, + { + "epoch": 1.4001027221366205, + "grad_norm": 2.8578219413757324, + "learning_rate": 2.99200609059764e-06, + "loss": 1.4291, + "step": 1363 + }, + { + "epoch": 1.4011299435028248, + "grad_norm": 2.8153412342071533, + "learning_rate": 2.9901027788351733e-06, + "loss": 1.3765, + "step": 1364 + }, + { + "epoch": 1.4021571648690292, + "grad_norm": 2.8321216106414795, + "learning_rate": 2.9881994670727067e-06, + "loss": 1.3914, + "step": 1365 + }, + { + "epoch": 1.4031843862352336, + "grad_norm": 2.7363576889038086, + "learning_rate": 2.98629615531024e-06, + "loss": 1.4091, + "step": 1366 + }, + { + "epoch": 1.4042116076014381, + "grad_norm": 2.8216500282287598, + "learning_rate": 2.984392843547773e-06, + "loss": 1.398, + "step": 1367 + }, + { + "epoch": 1.4052388289676425, + "grad_norm": 2.8419880867004395, + "learning_rate": 2.9824895317853063e-06, + "loss": 1.423, + "step": 1368 + }, + { + "epoch": 1.4062660503338469, + "grad_norm": 2.7820470333099365, + "learning_rate": 2.98058622002284e-06, + "loss": 1.4205, + "step": 1369 + }, + { + "epoch": 1.4072932717000515, + "grad_norm": 2.8953442573547363, + "learning_rate": 2.9786829082603735e-06, + "loss": 1.4667, + "step": 1370 + }, + { + "epoch": 1.4072932717000515, + "eval_loss": 1.7979086637496948, + "eval_runtime": 37.0854, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 1370 + }, + { + "epoch": 1.4083204930662558, + "grad_norm": 2.8208730220794678, + "learning_rate": 2.976779596497907e-06, + "loss": 1.4541, + "step": 1371 + }, + { + "epoch": 1.4093477144324602, + "grad_norm": 2.7347970008850098, + "learning_rate": 2.97487628473544e-06, + "loss": 1.4673, + "step": 1372 + }, + { + "epoch": 1.4103749357986646, + "grad_norm": 2.8528873920440674, + "learning_rate": 2.9729729729729736e-06, + "loss": 1.4267, + "step": 1373 + }, + { + "epoch": 1.411402157164869, + "grad_norm": 2.8298377990722656, + "learning_rate": 2.9710696612105065e-06, + "loss": 1.4713, + "step": 1374 + }, + { + "epoch": 1.4124293785310735, + "grad_norm": 2.847799301147461, + "learning_rate": 2.96916634944804e-06, + "loss": 1.4141, + "step": 1375 + }, + { + "epoch": 1.413456599897278, + "grad_norm": 2.8330819606781006, + "learning_rate": 2.9672630376855732e-06, + "loss": 1.4706, + "step": 1376 + }, + { + "epoch": 1.4144838212634823, + "grad_norm": 2.9275906085968018, + "learning_rate": 2.9653597259231066e-06, + "loss": 1.4371, + "step": 1377 + }, + { + "epoch": 1.4155110426296866, + "grad_norm": 2.7579028606414795, + "learning_rate": 2.96345641416064e-06, + "loss": 1.3921, + "step": 1378 + }, + { + "epoch": 1.416538263995891, + "grad_norm": 2.783045530319214, + "learning_rate": 2.961553102398173e-06, + "loss": 1.3295, + "step": 1379 + }, + { + "epoch": 1.4175654853620956, + "grad_norm": 2.946359157562256, + "learning_rate": 2.9596497906357063e-06, + "loss": 1.4655, + "step": 1380 + }, + { + "epoch": 1.4175654853620956, + "eval_loss": 1.7879260778427124, + "eval_runtime": 37.1105, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 1380 + }, + { + "epoch": 1.4185927067283, + "grad_norm": 3.1604318618774414, + "learning_rate": 2.9577464788732396e-06, + "loss": 1.4746, + "step": 1381 + }, + { + "epoch": 1.4196199280945043, + "grad_norm": 2.602548122406006, + "learning_rate": 2.955843167110773e-06, + "loss": 1.4024, + "step": 1382 + }, + { + "epoch": 1.420647149460709, + "grad_norm": 2.774137258529663, + "learning_rate": 2.9539398553483064e-06, + "loss": 1.4257, + "step": 1383 + }, + { + "epoch": 1.4216743708269133, + "grad_norm": 2.898256540298462, + "learning_rate": 2.9520365435858393e-06, + "loss": 1.4254, + "step": 1384 + }, + { + "epoch": 1.4227015921931176, + "grad_norm": 2.8015098571777344, + "learning_rate": 2.9501332318233727e-06, + "loss": 1.495, + "step": 1385 + }, + { + "epoch": 1.423728813559322, + "grad_norm": 2.7862651348114014, + "learning_rate": 2.948229920060906e-06, + "loss": 1.4355, + "step": 1386 + }, + { + "epoch": 1.4247560349255264, + "grad_norm": 2.7947680950164795, + "learning_rate": 2.9463266082984394e-06, + "loss": 1.4126, + "step": 1387 + }, + { + "epoch": 1.425783256291731, + "grad_norm": 2.8239989280700684, + "learning_rate": 2.9444232965359727e-06, + "loss": 1.417, + "step": 1388 + }, + { + "epoch": 1.4268104776579353, + "grad_norm": 2.845521926879883, + "learning_rate": 2.9425199847735057e-06, + "loss": 1.4406, + "step": 1389 + }, + { + "epoch": 1.4278376990241397, + "grad_norm": 2.9024460315704346, + "learning_rate": 2.94061667301104e-06, + "loss": 1.4083, + "step": 1390 + }, + { + "epoch": 1.4278376990241397, + "eval_loss": 1.7976808547973633, + "eval_runtime": 37.16, + "eval_samples_per_second": 26.911, + "eval_steps_per_second": 1.13, + "step": 1390 + }, + { + "epoch": 1.428864920390344, + "grad_norm": 2.7083208560943604, + "learning_rate": 2.938713361248573e-06, + "loss": 1.3428, + "step": 1391 + }, + { + "epoch": 1.4298921417565484, + "grad_norm": 2.7808196544647217, + "learning_rate": 2.936810049486106e-06, + "loss": 1.4262, + "step": 1392 + }, + { + "epoch": 1.430919363122753, + "grad_norm": 2.8307440280914307, + "learning_rate": 2.9349067377236396e-06, + "loss": 1.4328, + "step": 1393 + }, + { + "epoch": 1.4319465844889574, + "grad_norm": 2.8189167976379395, + "learning_rate": 2.933003425961173e-06, + "loss": 1.3878, + "step": 1394 + }, + { + "epoch": 1.4329738058551618, + "grad_norm": 2.7747530937194824, + "learning_rate": 2.9311001141987063e-06, + "loss": 1.404, + "step": 1395 + }, + { + "epoch": 1.4340010272213661, + "grad_norm": 2.747140884399414, + "learning_rate": 2.9291968024362392e-06, + "loss": 1.3954, + "step": 1396 + }, + { + "epoch": 1.4350282485875705, + "grad_norm": 2.665836811065674, + "learning_rate": 2.9272934906737726e-06, + "loss": 1.3868, + "step": 1397 + }, + { + "epoch": 1.436055469953775, + "grad_norm": 2.6918723583221436, + "learning_rate": 2.925390178911306e-06, + "loss": 1.3959, + "step": 1398 + }, + { + "epoch": 1.4370826913199795, + "grad_norm": 2.8271594047546387, + "learning_rate": 2.9234868671488393e-06, + "loss": 1.4114, + "step": 1399 + }, + { + "epoch": 1.4381099126861838, + "grad_norm": 3.037952184677124, + "learning_rate": 2.9215835553863727e-06, + "loss": 1.4267, + "step": 1400 + }, + { + "epoch": 1.4381099126861838, + "eval_loss": 1.7927908897399902, + "eval_runtime": 37.3671, + "eval_samples_per_second": 26.761, + "eval_steps_per_second": 1.124, + "step": 1400 + }, + { + "epoch": 1.4391371340523884, + "grad_norm": 3.0185632705688477, + "learning_rate": 2.9196802436239056e-06, + "loss": 1.4211, + "step": 1401 + }, + { + "epoch": 1.4401643554185928, + "grad_norm": 2.678468942642212, + "learning_rate": 2.917776931861439e-06, + "loss": 1.398, + "step": 1402 + }, + { + "epoch": 1.4411915767847971, + "grad_norm": 2.8228728771209717, + "learning_rate": 2.9158736200989723e-06, + "loss": 1.3946, + "step": 1403 + }, + { + "epoch": 1.4422187981510015, + "grad_norm": 2.6711578369140625, + "learning_rate": 2.9139703083365057e-06, + "loss": 1.3985, + "step": 1404 + }, + { + "epoch": 1.4432460195172059, + "grad_norm": 2.7891721725463867, + "learning_rate": 2.912066996574039e-06, + "loss": 1.4232, + "step": 1405 + }, + { + "epoch": 1.4442732408834105, + "grad_norm": 2.833613395690918, + "learning_rate": 2.910163684811572e-06, + "loss": 1.3629, + "step": 1406 + }, + { + "epoch": 1.4453004622496148, + "grad_norm": 3.0415570735931396, + "learning_rate": 2.9082603730491054e-06, + "loss": 1.4666, + "step": 1407 + }, + { + "epoch": 1.4463276836158192, + "grad_norm": 2.831063747406006, + "learning_rate": 2.9063570612866387e-06, + "loss": 1.4477, + "step": 1408 + }, + { + "epoch": 1.4473549049820236, + "grad_norm": 2.8253402709960938, + "learning_rate": 2.904453749524172e-06, + "loss": 1.3729, + "step": 1409 + }, + { + "epoch": 1.448382126348228, + "grad_norm": 2.9134843349456787, + "learning_rate": 2.902550437761706e-06, + "loss": 1.3958, + "step": 1410 + }, + { + "epoch": 1.448382126348228, + "eval_loss": 1.7833805084228516, + "eval_runtime": 36.9111, + "eval_samples_per_second": 27.092, + "eval_steps_per_second": 1.138, + "step": 1410 + }, + { + "epoch": 1.4494093477144325, + "grad_norm": 2.962355852127075, + "learning_rate": 2.9006471259992392e-06, + "loss": 1.4143, + "step": 1411 + }, + { + "epoch": 1.450436569080637, + "grad_norm": 2.738837242126465, + "learning_rate": 2.8987438142367726e-06, + "loss": 1.4051, + "step": 1412 + }, + { + "epoch": 1.4514637904468413, + "grad_norm": 2.766331195831299, + "learning_rate": 2.8968405024743055e-06, + "loss": 1.4628, + "step": 1413 + }, + { + "epoch": 1.4524910118130456, + "grad_norm": 2.6052863597869873, + "learning_rate": 2.894937190711839e-06, + "loss": 1.3662, + "step": 1414 + }, + { + "epoch": 1.45351823317925, + "grad_norm": 2.8454713821411133, + "learning_rate": 2.8930338789493723e-06, + "loss": 1.4395, + "step": 1415 + }, + { + "epoch": 1.4545454545454546, + "grad_norm": 2.683232545852661, + "learning_rate": 2.8911305671869056e-06, + "loss": 1.4243, + "step": 1416 + }, + { + "epoch": 1.455572675911659, + "grad_norm": 2.798797845840454, + "learning_rate": 2.889227255424439e-06, + "loss": 1.3627, + "step": 1417 + }, + { + "epoch": 1.4565998972778633, + "grad_norm": 2.821502923965454, + "learning_rate": 2.887323943661972e-06, + "loss": 1.4109, + "step": 1418 + }, + { + "epoch": 1.457627118644068, + "grad_norm": 2.8054518699645996, + "learning_rate": 2.8854206318995053e-06, + "loss": 1.3906, + "step": 1419 + }, + { + "epoch": 1.4586543400102723, + "grad_norm": 2.802000045776367, + "learning_rate": 2.8835173201370386e-06, + "loss": 1.4024, + "step": 1420 + }, + { + "epoch": 1.4586543400102723, + "eval_loss": 1.7884918451309204, + "eval_runtime": 37.0267, + "eval_samples_per_second": 27.008, + "eval_steps_per_second": 1.134, + "step": 1420 + }, + { + "epoch": 1.4596815613764766, + "grad_norm": 3.097747564315796, + "learning_rate": 2.881614008374572e-06, + "loss": 1.4695, + "step": 1421 + }, + { + "epoch": 1.460708782742681, + "grad_norm": 3.01665997505188, + "learning_rate": 2.8797106966121054e-06, + "loss": 1.4321, + "step": 1422 + }, + { + "epoch": 1.4617360041088854, + "grad_norm": 2.9308114051818848, + "learning_rate": 2.8778073848496383e-06, + "loss": 1.4084, + "step": 1423 + }, + { + "epoch": 1.46276322547509, + "grad_norm": 2.8440439701080322, + "learning_rate": 2.8759040730871717e-06, + "loss": 1.4473, + "step": 1424 + }, + { + "epoch": 1.4637904468412943, + "grad_norm": 2.6690547466278076, + "learning_rate": 2.874000761324705e-06, + "loss": 1.407, + "step": 1425 + }, + { + "epoch": 1.4648176682074987, + "grad_norm": 2.7961244583129883, + "learning_rate": 2.8720974495622384e-06, + "loss": 1.4414, + "step": 1426 + }, + { + "epoch": 1.465844889573703, + "grad_norm": 2.891948938369751, + "learning_rate": 2.8701941377997718e-06, + "loss": 1.4185, + "step": 1427 + }, + { + "epoch": 1.4668721109399074, + "grad_norm": 2.790811538696289, + "learning_rate": 2.8682908260373047e-06, + "loss": 1.4528, + "step": 1428 + }, + { + "epoch": 1.467899332306112, + "grad_norm": 2.9391868114471436, + "learning_rate": 2.866387514274838e-06, + "loss": 1.4291, + "step": 1429 + }, + { + "epoch": 1.4689265536723164, + "grad_norm": 2.9476356506347656, + "learning_rate": 2.864484202512372e-06, + "loss": 1.4373, + "step": 1430 + }, + { + "epoch": 1.4689265536723164, + "eval_loss": 1.7926347255706787, + "eval_runtime": 37.0308, + "eval_samples_per_second": 27.005, + "eval_steps_per_second": 1.134, + "step": 1430 + }, + { + "epoch": 1.4699537750385208, + "grad_norm": 2.907172203063965, + "learning_rate": 2.8625808907499052e-06, + "loss": 1.3983, + "step": 1431 + }, + { + "epoch": 1.4709809964047253, + "grad_norm": 2.8310680389404297, + "learning_rate": 2.8606775789874386e-06, + "loss": 1.4281, + "step": 1432 + }, + { + "epoch": 1.4720082177709297, + "grad_norm": 2.8614001274108887, + "learning_rate": 2.858774267224972e-06, + "loss": 1.3975, + "step": 1433 + }, + { + "epoch": 1.473035439137134, + "grad_norm": 3.1007702350616455, + "learning_rate": 2.8568709554625053e-06, + "loss": 1.4252, + "step": 1434 + }, + { + "epoch": 1.4740626605033385, + "grad_norm": 2.879520893096924, + "learning_rate": 2.8549676437000382e-06, + "loss": 1.468, + "step": 1435 + }, + { + "epoch": 1.4750898818695428, + "grad_norm": 2.9735469818115234, + "learning_rate": 2.8530643319375716e-06, + "loss": 1.4341, + "step": 1436 + }, + { + "epoch": 1.4761171032357474, + "grad_norm": 3.0768795013427734, + "learning_rate": 2.851161020175105e-06, + "loss": 1.4122, + "step": 1437 + }, + { + "epoch": 1.4771443246019518, + "grad_norm": 2.930915355682373, + "learning_rate": 2.8492577084126383e-06, + "loss": 1.3904, + "step": 1438 + }, + { + "epoch": 1.4781715459681561, + "grad_norm": 2.659090280532837, + "learning_rate": 2.8473543966501717e-06, + "loss": 1.4261, + "step": 1439 + }, + { + "epoch": 1.4791987673343605, + "grad_norm": 3.020371913909912, + "learning_rate": 2.8454510848877046e-06, + "loss": 1.4636, + "step": 1440 + }, + { + "epoch": 1.4791987673343605, + "eval_loss": 1.8039511442184448, + "eval_runtime": 37.0373, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 1440 + }, + { + "epoch": 1.4802259887005649, + "grad_norm": 3.131986141204834, + "learning_rate": 2.843547773125238e-06, + "loss": 1.4178, + "step": 1441 + }, + { + "epoch": 1.4812532100667695, + "grad_norm": 2.8436167240142822, + "learning_rate": 2.8416444613627714e-06, + "loss": 1.4149, + "step": 1442 + }, + { + "epoch": 1.4822804314329738, + "grad_norm": 2.8628804683685303, + "learning_rate": 2.8397411496003047e-06, + "loss": 1.3751, + "step": 1443 + }, + { + "epoch": 1.4833076527991782, + "grad_norm": 2.917836904525757, + "learning_rate": 2.837837837837838e-06, + "loss": 1.4191, + "step": 1444 + }, + { + "epoch": 1.4843348741653826, + "grad_norm": 2.951035737991333, + "learning_rate": 2.835934526075371e-06, + "loss": 1.4352, + "step": 1445 + }, + { + "epoch": 1.485362095531587, + "grad_norm": 2.7296388149261475, + "learning_rate": 2.8340312143129044e-06, + "loss": 1.3856, + "step": 1446 + }, + { + "epoch": 1.4863893168977915, + "grad_norm": 2.7064192295074463, + "learning_rate": 2.8321279025504377e-06, + "loss": 1.3782, + "step": 1447 + }, + { + "epoch": 1.487416538263996, + "grad_norm": 2.9402806758880615, + "learning_rate": 2.830224590787971e-06, + "loss": 1.4829, + "step": 1448 + }, + { + "epoch": 1.4884437596302003, + "grad_norm": 2.9216394424438477, + "learning_rate": 2.8283212790255045e-06, + "loss": 1.4173, + "step": 1449 + }, + { + "epoch": 1.4894709809964048, + "grad_norm": 2.7382750511169434, + "learning_rate": 2.8264179672630383e-06, + "loss": 1.4431, + "step": 1450 + }, + { + "epoch": 1.4894709809964048, + "eval_loss": 1.798187255859375, + "eval_runtime": 37.0401, + "eval_samples_per_second": 26.998, + "eval_steps_per_second": 1.134, + "step": 1450 + }, + { + "epoch": 1.4904982023626092, + "grad_norm": 2.781414270401001, + "learning_rate": 2.8245146555005716e-06, + "loss": 1.4096, + "step": 1451 + }, + { + "epoch": 1.4915254237288136, + "grad_norm": 2.8284265995025635, + "learning_rate": 2.8226113437381046e-06, + "loss": 1.3561, + "step": 1452 + }, + { + "epoch": 1.492552645095018, + "grad_norm": 2.6621487140655518, + "learning_rate": 2.820708031975638e-06, + "loss": 1.4232, + "step": 1453 + }, + { + "epoch": 1.4935798664612223, + "grad_norm": 2.9116251468658447, + "learning_rate": 2.8188047202131713e-06, + "loss": 1.4431, + "step": 1454 + }, + { + "epoch": 1.494607087827427, + "grad_norm": 2.749554395675659, + "learning_rate": 2.8169014084507046e-06, + "loss": 1.4603, + "step": 1455 + }, + { + "epoch": 1.4956343091936313, + "grad_norm": 2.699585199356079, + "learning_rate": 2.814998096688238e-06, + "loss": 1.414, + "step": 1456 + }, + { + "epoch": 1.4966615305598356, + "grad_norm": 2.6940858364105225, + "learning_rate": 2.813094784925771e-06, + "loss": 1.3673, + "step": 1457 + }, + { + "epoch": 1.49768875192604, + "grad_norm": 2.859602689743042, + "learning_rate": 2.8111914731633043e-06, + "loss": 1.4664, + "step": 1458 + }, + { + "epoch": 1.4987159732922444, + "grad_norm": 2.8793201446533203, + "learning_rate": 2.8092881614008377e-06, + "loss": 1.4168, + "step": 1459 + }, + { + "epoch": 1.499743194658449, + "grad_norm": 2.8155760765075684, + "learning_rate": 2.807384849638371e-06, + "loss": 1.3917, + "step": 1460 + }, + { + "epoch": 1.499743194658449, + "eval_loss": 1.780331015586853, + "eval_runtime": 37.0489, + "eval_samples_per_second": 26.991, + "eval_steps_per_second": 1.134, + "step": 1460 + }, + { + "epoch": 1.5007704160246533, + "grad_norm": 2.7490365505218506, + "learning_rate": 2.8054815378759044e-06, + "loss": 1.4326, + "step": 1461 + }, + { + "epoch": 1.5017976373908577, + "grad_norm": 2.8913350105285645, + "learning_rate": 2.8035782261134378e-06, + "loss": 1.403, + "step": 1462 + }, + { + "epoch": 1.5028248587570623, + "grad_norm": 2.83650279045105, + "learning_rate": 2.8016749143509707e-06, + "loss": 1.4107, + "step": 1463 + }, + { + "epoch": 1.5038520801232664, + "grad_norm": 2.838649272918701, + "learning_rate": 2.799771602588504e-06, + "loss": 1.4131, + "step": 1464 + }, + { + "epoch": 1.504879301489471, + "grad_norm": 2.8424947261810303, + "learning_rate": 2.7978682908260374e-06, + "loss": 1.4322, + "step": 1465 + }, + { + "epoch": 1.5059065228556754, + "grad_norm": 2.8207895755767822, + "learning_rate": 2.795964979063571e-06, + "loss": 1.4464, + "step": 1466 + }, + { + "epoch": 1.5069337442218798, + "grad_norm": 2.897221088409424, + "learning_rate": 2.794061667301104e-06, + "loss": 1.4237, + "step": 1467 + }, + { + "epoch": 1.5079609655880843, + "grad_norm": 2.766166925430298, + "learning_rate": 2.792158355538637e-06, + "loss": 1.4375, + "step": 1468 + }, + { + "epoch": 1.5089881869542885, + "grad_norm": 2.8115463256835938, + "learning_rate": 2.7902550437761705e-06, + "loss": 1.4265, + "step": 1469 + }, + { + "epoch": 1.510015408320493, + "grad_norm": 2.826871156692505, + "learning_rate": 2.7883517320137042e-06, + "loss": 1.3933, + "step": 1470 + }, + { + "epoch": 1.510015408320493, + "eval_loss": 1.780141830444336, + "eval_runtime": 37.0572, + "eval_samples_per_second": 26.985, + "eval_steps_per_second": 1.133, + "step": 1470 + }, + { + "epoch": 1.5110426296866974, + "grad_norm": 2.737398147583008, + "learning_rate": 2.7864484202512376e-06, + "loss": 1.3898, + "step": 1471 + }, + { + "epoch": 1.5120698510529018, + "grad_norm": 2.9325804710388184, + "learning_rate": 2.784545108488771e-06, + "loss": 1.4139, + "step": 1472 + }, + { + "epoch": 1.5130970724191064, + "grad_norm": 2.7400333881378174, + "learning_rate": 2.7826417967263043e-06, + "loss": 1.4122, + "step": 1473 + }, + { + "epoch": 1.5141242937853108, + "grad_norm": 2.835869550704956, + "learning_rate": 2.7807384849638373e-06, + "loss": 1.4634, + "step": 1474 + }, + { + "epoch": 1.5151515151515151, + "grad_norm": 2.7512247562408447, + "learning_rate": 2.7788351732013706e-06, + "loss": 1.409, + "step": 1475 + }, + { + "epoch": 1.5161787365177197, + "grad_norm": 2.745054244995117, + "learning_rate": 2.776931861438904e-06, + "loss": 1.4004, + "step": 1476 + }, + { + "epoch": 1.5172059578839239, + "grad_norm": 2.7766499519348145, + "learning_rate": 2.7750285496764374e-06, + "loss": 1.4197, + "step": 1477 + }, + { + "epoch": 1.5182331792501285, + "grad_norm": 2.793663263320923, + "learning_rate": 2.7731252379139707e-06, + "loss": 1.4205, + "step": 1478 + }, + { + "epoch": 1.5192604006163328, + "grad_norm": 2.674372434616089, + "learning_rate": 2.771221926151504e-06, + "loss": 1.3648, + "step": 1479 + }, + { + "epoch": 1.5202876219825372, + "grad_norm": 2.7583112716674805, + "learning_rate": 2.769318614389037e-06, + "loss": 1.4296, + "step": 1480 + }, + { + "epoch": 1.5202876219825372, + "eval_loss": 1.7793656587600708, + "eval_runtime": 37.05, + "eval_samples_per_second": 26.991, + "eval_steps_per_second": 1.134, + "step": 1480 + }, + { + "epoch": 1.5213148433487418, + "grad_norm": 2.776137351989746, + "learning_rate": 2.7674153026265704e-06, + "loss": 1.4267, + "step": 1481 + }, + { + "epoch": 1.522342064714946, + "grad_norm": 2.8750247955322266, + "learning_rate": 2.7655119908641037e-06, + "loss": 1.4311, + "step": 1482 + }, + { + "epoch": 1.5233692860811505, + "grad_norm": 2.740621566772461, + "learning_rate": 2.763608679101637e-06, + "loss": 1.4198, + "step": 1483 + }, + { + "epoch": 1.5243965074473549, + "grad_norm": 2.875997304916382, + "learning_rate": 2.7617053673391705e-06, + "loss": 1.4557, + "step": 1484 + }, + { + "epoch": 1.5254237288135593, + "grad_norm": 2.8406331539154053, + "learning_rate": 2.7598020555767034e-06, + "loss": 1.4504, + "step": 1485 + }, + { + "epoch": 1.5264509501797638, + "grad_norm": 2.6013126373291016, + "learning_rate": 2.7578987438142368e-06, + "loss": 1.3829, + "step": 1486 + }, + { + "epoch": 1.5274781715459682, + "grad_norm": 2.7469065189361572, + "learning_rate": 2.75599543205177e-06, + "loss": 1.4314, + "step": 1487 + }, + { + "epoch": 1.5285053929121726, + "grad_norm": 2.8129167556762695, + "learning_rate": 2.7540921202893035e-06, + "loss": 1.4491, + "step": 1488 + }, + { + "epoch": 1.529532614278377, + "grad_norm": 2.9677329063415527, + "learning_rate": 2.752188808526837e-06, + "loss": 1.4292, + "step": 1489 + }, + { + "epoch": 1.5305598356445813, + "grad_norm": 2.7910637855529785, + "learning_rate": 2.75028549676437e-06, + "loss": 1.4157, + "step": 1490 + }, + { + "epoch": 1.5305598356445813, + "eval_loss": 1.7833755016326904, + "eval_runtime": 37.1104, + "eval_samples_per_second": 26.947, + "eval_steps_per_second": 1.132, + "step": 1490 + }, + { + "epoch": 1.531587057010786, + "grad_norm": 2.7772984504699707, + "learning_rate": 2.7483821850019036e-06, + "loss": 1.4125, + "step": 1491 + }, + { + "epoch": 1.5326142783769903, + "grad_norm": 2.747514247894287, + "learning_rate": 2.746478873239437e-06, + "loss": 1.4057, + "step": 1492 + }, + { + "epoch": 1.5336414997431946, + "grad_norm": 2.7529642581939697, + "learning_rate": 2.7445755614769703e-06, + "loss": 1.4066, + "step": 1493 + }, + { + "epoch": 1.5346687211093992, + "grad_norm": 2.8375084400177, + "learning_rate": 2.7426722497145037e-06, + "loss": 1.3883, + "step": 1494 + }, + { + "epoch": 1.5356959424756034, + "grad_norm": 2.7483835220336914, + "learning_rate": 2.740768937952037e-06, + "loss": 1.3915, + "step": 1495 + }, + { + "epoch": 1.536723163841808, + "grad_norm": 2.837123394012451, + "learning_rate": 2.7388656261895704e-06, + "loss": 1.4082, + "step": 1496 + }, + { + "epoch": 1.5377503852080123, + "grad_norm": 2.909313440322876, + "learning_rate": 2.7369623144271033e-06, + "loss": 1.4815, + "step": 1497 + }, + { + "epoch": 1.5387776065742167, + "grad_norm": 2.931723117828369, + "learning_rate": 2.7350590026646367e-06, + "loss": 1.4099, + "step": 1498 + }, + { + "epoch": 1.5398048279404213, + "grad_norm": 2.9692134857177734, + "learning_rate": 2.73315569090217e-06, + "loss": 1.3637, + "step": 1499 + }, + { + "epoch": 1.5408320493066254, + "grad_norm": 2.974923610687256, + "learning_rate": 2.7312523791397034e-06, + "loss": 1.4505, + "step": 1500 + }, + { + "epoch": 1.5408320493066254, + "eval_loss": 1.7900456190109253, + "eval_runtime": 37.3025, + "eval_samples_per_second": 26.808, + "eval_steps_per_second": 1.126, + "step": 1500 + }, + { + "epoch": 1.54185927067283, + "grad_norm": 2.9976389408111572, + "learning_rate": 2.7293490673772368e-06, + "loss": 1.4726, + "step": 1501 + }, + { + "epoch": 1.5428864920390344, + "grad_norm": 2.8478479385375977, + "learning_rate": 2.7274457556147697e-06, + "loss": 1.4304, + "step": 1502 + }, + { + "epoch": 1.5439137134052388, + "grad_norm": 2.9100635051727295, + "learning_rate": 2.725542443852303e-06, + "loss": 1.4107, + "step": 1503 + }, + { + "epoch": 1.5449409347714433, + "grad_norm": 2.903132677078247, + "learning_rate": 2.7236391320898365e-06, + "loss": 1.4206, + "step": 1504 + }, + { + "epoch": 1.5459681561376477, + "grad_norm": 2.9177560806274414, + "learning_rate": 2.72173582032737e-06, + "loss": 1.3948, + "step": 1505 + }, + { + "epoch": 1.546995377503852, + "grad_norm": 2.7861175537109375, + "learning_rate": 2.719832508564903e-06, + "loss": 1.4123, + "step": 1506 + }, + { + "epoch": 1.5480225988700564, + "grad_norm": 2.787735939025879, + "learning_rate": 2.717929196802436e-06, + "loss": 1.4199, + "step": 1507 + }, + { + "epoch": 1.5490498202362608, + "grad_norm": 2.889784097671509, + "learning_rate": 2.7160258850399695e-06, + "loss": 1.3901, + "step": 1508 + }, + { + "epoch": 1.5500770416024654, + "grad_norm": 2.8383727073669434, + "learning_rate": 2.714122573277503e-06, + "loss": 1.4145, + "step": 1509 + }, + { + "epoch": 1.5511042629686698, + "grad_norm": 2.7701327800750732, + "learning_rate": 2.712219261515036e-06, + "loss": 1.4213, + "step": 1510 + }, + { + "epoch": 1.5511042629686698, + "eval_loss": 1.7863596677780151, + "eval_runtime": 36.8927, + "eval_samples_per_second": 27.106, + "eval_steps_per_second": 1.138, + "step": 1510 + }, + { + "epoch": 1.5521314843348741, + "grad_norm": 2.7612502574920654, + "learning_rate": 2.71031594975257e-06, + "loss": 1.4286, + "step": 1511 + }, + { + "epoch": 1.5531587057010787, + "grad_norm": 2.696126937866211, + "learning_rate": 2.7084126379901034e-06, + "loss": 1.4359, + "step": 1512 + }, + { + "epoch": 1.5541859270672829, + "grad_norm": 2.840223550796509, + "learning_rate": 2.7065093262276367e-06, + "loss": 1.4424, + "step": 1513 + }, + { + "epoch": 1.5552131484334875, + "grad_norm": 2.7309999465942383, + "learning_rate": 2.7046060144651697e-06, + "loss": 1.4567, + "step": 1514 + }, + { + "epoch": 1.5562403697996918, + "grad_norm": 2.8637242317199707, + "learning_rate": 2.702702702702703e-06, + "loss": 1.4699, + "step": 1515 + }, + { + "epoch": 1.5572675911658962, + "grad_norm": 2.653404474258423, + "learning_rate": 2.7007993909402364e-06, + "loss": 1.3445, + "step": 1516 + }, + { + "epoch": 1.5582948125321008, + "grad_norm": 2.9771206378936768, + "learning_rate": 2.6988960791777697e-06, + "loss": 1.4588, + "step": 1517 + }, + { + "epoch": 1.559322033898305, + "grad_norm": 2.888413667678833, + "learning_rate": 2.696992767415303e-06, + "loss": 1.425, + "step": 1518 + }, + { + "epoch": 1.5603492552645095, + "grad_norm": 2.8841593265533447, + "learning_rate": 2.695089455652836e-06, + "loss": 1.4101, + "step": 1519 + }, + { + "epoch": 1.5613764766307139, + "grad_norm": 2.990192413330078, + "learning_rate": 2.6931861438903694e-06, + "loss": 1.3897, + "step": 1520 + }, + { + "epoch": 1.5613764766307139, + "eval_loss": 1.7787587642669678, + "eval_runtime": 36.9999, + "eval_samples_per_second": 27.027, + "eval_steps_per_second": 1.135, + "step": 1520 + }, + { + "epoch": 1.5624036979969183, + "grad_norm": 2.8698887825012207, + "learning_rate": 2.6912828321279028e-06, + "loss": 1.4324, + "step": 1521 + }, + { + "epoch": 1.5634309193631228, + "grad_norm": 2.8924384117126465, + "learning_rate": 2.689379520365436e-06, + "loss": 1.4073, + "step": 1522 + }, + { + "epoch": 1.5644581407293272, + "grad_norm": 2.92336106300354, + "learning_rate": 2.6874762086029695e-06, + "loss": 1.4471, + "step": 1523 + }, + { + "epoch": 1.5654853620955316, + "grad_norm": 2.7212202548980713, + "learning_rate": 2.6855728968405024e-06, + "loss": 1.3902, + "step": 1524 + }, + { + "epoch": 1.5665125834617362, + "grad_norm": 2.711582660675049, + "learning_rate": 2.683669585078036e-06, + "loss": 1.4294, + "step": 1525 + }, + { + "epoch": 1.5675398048279403, + "grad_norm": 2.6806275844573975, + "learning_rate": 2.681766273315569e-06, + "loss": 1.3683, + "step": 1526 + }, + { + "epoch": 1.568567026194145, + "grad_norm": 2.920119524002075, + "learning_rate": 2.6798629615531025e-06, + "loss": 1.4194, + "step": 1527 + }, + { + "epoch": 1.5695942475603493, + "grad_norm": 2.7487800121307373, + "learning_rate": 2.677959649790636e-06, + "loss": 1.4091, + "step": 1528 + }, + { + "epoch": 1.5706214689265536, + "grad_norm": 2.7557766437530518, + "learning_rate": 2.676056338028169e-06, + "loss": 1.4453, + "step": 1529 + }, + { + "epoch": 1.5716486902927582, + "grad_norm": 2.7834057807922363, + "learning_rate": 2.674153026265702e-06, + "loss": 1.4381, + "step": 1530 + }, + { + "epoch": 1.5716486902927582, + "eval_loss": 1.7810354232788086, + "eval_runtime": 37.037, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 1530 + }, + { + "epoch": 1.5726759116589624, + "grad_norm": 3.107299327850342, + "learning_rate": 2.672249714503236e-06, + "loss": 1.4475, + "step": 1531 + }, + { + "epoch": 1.573703133025167, + "grad_norm": 2.813086748123169, + "learning_rate": 2.6703464027407693e-06, + "loss": 1.4062, + "step": 1532 + }, + { + "epoch": 1.5747303543913713, + "grad_norm": 2.781416416168213, + "learning_rate": 2.6684430909783027e-06, + "loss": 1.4268, + "step": 1533 + }, + { + "epoch": 1.5757575757575757, + "grad_norm": 2.681947708129883, + "learning_rate": 2.666539779215836e-06, + "loss": 1.4121, + "step": 1534 + }, + { + "epoch": 1.5767847971237803, + "grad_norm": 2.656013011932373, + "learning_rate": 2.6646364674533694e-06, + "loss": 1.4003, + "step": 1535 + }, + { + "epoch": 1.5778120184899846, + "grad_norm": 2.5161869525909424, + "learning_rate": 2.6627331556909024e-06, + "loss": 1.3127, + "step": 1536 + }, + { + "epoch": 1.578839239856189, + "grad_norm": 2.842507839202881, + "learning_rate": 2.6608298439284357e-06, + "loss": 1.4067, + "step": 1537 + }, + { + "epoch": 1.5798664612223934, + "grad_norm": 2.732778787612915, + "learning_rate": 2.658926532165969e-06, + "loss": 1.4071, + "step": 1538 + }, + { + "epoch": 1.5808936825885977, + "grad_norm": 2.802293300628662, + "learning_rate": 2.6570232204035024e-06, + "loss": 1.3637, + "step": 1539 + }, + { + "epoch": 1.5819209039548023, + "grad_norm": 2.927441358566284, + "learning_rate": 2.655119908641036e-06, + "loss": 1.405, + "step": 1540 + }, + { + "epoch": 1.5819209039548023, + "eval_loss": 1.7797114849090576, + "eval_runtime": 37.0514, + "eval_samples_per_second": 26.99, + "eval_steps_per_second": 1.134, + "step": 1540 + }, + { + "epoch": 1.5829481253210067, + "grad_norm": 2.959746837615967, + "learning_rate": 2.6532165968785687e-06, + "loss": 1.4123, + "step": 1541 + }, + { + "epoch": 1.583975346687211, + "grad_norm": 2.679551124572754, + "learning_rate": 2.651313285116102e-06, + "loss": 1.3742, + "step": 1542 + }, + { + "epoch": 1.5850025680534157, + "grad_norm": 2.7431442737579346, + "learning_rate": 2.6494099733536355e-06, + "loss": 1.3825, + "step": 1543 + }, + { + "epoch": 1.5860297894196198, + "grad_norm": 3.3630471229553223, + "learning_rate": 2.647506661591169e-06, + "loss": 1.5599, + "step": 1544 + }, + { + "epoch": 1.5870570107858244, + "grad_norm": 2.8524038791656494, + "learning_rate": 2.645603349828702e-06, + "loss": 1.4293, + "step": 1545 + }, + { + "epoch": 1.5880842321520288, + "grad_norm": 2.871917486190796, + "learning_rate": 2.643700038066235e-06, + "loss": 1.3716, + "step": 1546 + }, + { + "epoch": 1.5891114535182331, + "grad_norm": 2.7151851654052734, + "learning_rate": 2.6417967263037685e-06, + "loss": 1.43, + "step": 1547 + }, + { + "epoch": 1.5901386748844377, + "grad_norm": 2.7912635803222656, + "learning_rate": 2.639893414541302e-06, + "loss": 1.4401, + "step": 1548 + }, + { + "epoch": 1.5911658962506419, + "grad_norm": 2.9613499641418457, + "learning_rate": 2.6379901027788352e-06, + "loss": 1.4691, + "step": 1549 + }, + { + "epoch": 1.5921931176168465, + "grad_norm": 2.8831710815429688, + "learning_rate": 2.6360867910163686e-06, + "loss": 1.3975, + "step": 1550 + }, + { + "epoch": 1.5921931176168465, + "eval_loss": 1.7762643098831177, + "eval_runtime": 37.0485, + "eval_samples_per_second": 26.992, + "eval_steps_per_second": 1.134, + "step": 1550 + }, + { + "epoch": 1.5932203389830508, + "grad_norm": 2.6928372383117676, + "learning_rate": 2.6341834792539024e-06, + "loss": 1.4179, + "step": 1551 + }, + { + "epoch": 1.5942475603492552, + "grad_norm": 2.8088696002960205, + "learning_rate": 2.6322801674914357e-06, + "loss": 1.417, + "step": 1552 + }, + { + "epoch": 1.5952747817154598, + "grad_norm": 2.926135778427124, + "learning_rate": 2.6303768557289687e-06, + "loss": 1.4145, + "step": 1553 + }, + { + "epoch": 1.5963020030816641, + "grad_norm": 2.7283592224121094, + "learning_rate": 2.628473543966502e-06, + "loss": 1.3714, + "step": 1554 + }, + { + "epoch": 1.5973292244478685, + "grad_norm": 2.7970361709594727, + "learning_rate": 2.6265702322040354e-06, + "loss": 1.4393, + "step": 1555 + }, + { + "epoch": 1.5983564458140729, + "grad_norm": 2.717223644256592, + "learning_rate": 2.6246669204415688e-06, + "loss": 1.3857, + "step": 1556 + }, + { + "epoch": 1.5993836671802772, + "grad_norm": 2.817638635635376, + "learning_rate": 2.622763608679102e-06, + "loss": 1.3863, + "step": 1557 + }, + { + "epoch": 1.6004108885464818, + "grad_norm": 2.7856762409210205, + "learning_rate": 2.620860296916635e-06, + "loss": 1.4052, + "step": 1558 + }, + { + "epoch": 1.6014381099126862, + "grad_norm": 2.9281742572784424, + "learning_rate": 2.6189569851541684e-06, + "loss": 1.3417, + "step": 1559 + }, + { + "epoch": 1.6024653312788906, + "grad_norm": 2.8528225421905518, + "learning_rate": 2.617053673391702e-06, + "loss": 1.4486, + "step": 1560 + }, + { + "epoch": 1.6024653312788906, + "eval_loss": 1.771835207939148, + "eval_runtime": 37.0601, + "eval_samples_per_second": 26.983, + "eval_steps_per_second": 1.133, + "step": 1560 + }, + { + "epoch": 1.6034925526450952, + "grad_norm": 3.0976126194000244, + "learning_rate": 2.615150361629235e-06, + "loss": 1.4722, + "step": 1561 + }, + { + "epoch": 1.6045197740112993, + "grad_norm": 2.9993839263916016, + "learning_rate": 2.6132470498667685e-06, + "loss": 1.4265, + "step": 1562 + }, + { + "epoch": 1.605546995377504, + "grad_norm": 2.896975517272949, + "learning_rate": 2.6113437381043015e-06, + "loss": 1.4035, + "step": 1563 + }, + { + "epoch": 1.6065742167437083, + "grad_norm": 2.9382545948028564, + "learning_rate": 2.609440426341835e-06, + "loss": 1.3934, + "step": 1564 + }, + { + "epoch": 1.6076014381099126, + "grad_norm": 2.950702667236328, + "learning_rate": 2.607537114579368e-06, + "loss": 1.4272, + "step": 1565 + }, + { + "epoch": 1.6086286594761172, + "grad_norm": 2.888334274291992, + "learning_rate": 2.6056338028169015e-06, + "loss": 1.4088, + "step": 1566 + }, + { + "epoch": 1.6096558808423214, + "grad_norm": 2.6972687244415283, + "learning_rate": 2.603730491054435e-06, + "loss": 1.391, + "step": 1567 + }, + { + "epoch": 1.610683102208526, + "grad_norm": 2.834188461303711, + "learning_rate": 2.601827179291968e-06, + "loss": 1.4241, + "step": 1568 + }, + { + "epoch": 1.6117103235747303, + "grad_norm": 2.894808769226074, + "learning_rate": 2.599923867529501e-06, + "loss": 1.4297, + "step": 1569 + }, + { + "epoch": 1.6127375449409347, + "grad_norm": 2.9836509227752686, + "learning_rate": 2.5980205557670346e-06, + "loss": 1.4086, + "step": 1570 + }, + { + "epoch": 1.6127375449409347, + "eval_loss": 1.7795302867889404, + "eval_runtime": 37.0629, + "eval_samples_per_second": 26.981, + "eval_steps_per_second": 1.133, + "step": 1570 + }, + { + "epoch": 1.6137647663071393, + "grad_norm": 2.8716723918914795, + "learning_rate": 2.5961172440045684e-06, + "loss": 1.4425, + "step": 1571 + }, + { + "epoch": 1.6147919876733436, + "grad_norm": 2.8877105712890625, + "learning_rate": 2.5942139322421017e-06, + "loss": 1.4557, + "step": 1572 + }, + { + "epoch": 1.615819209039548, + "grad_norm": 2.8737246990203857, + "learning_rate": 2.592310620479635e-06, + "loss": 1.4015, + "step": 1573 + }, + { + "epoch": 1.6168464304057526, + "grad_norm": 2.7048048973083496, + "learning_rate": 2.5904073087171684e-06, + "loss": 1.361, + "step": 1574 + }, + { + "epoch": 1.6178736517719567, + "grad_norm": 2.79799222946167, + "learning_rate": 2.5885039969547014e-06, + "loss": 1.4325, + "step": 1575 + }, + { + "epoch": 1.6189008731381613, + "grad_norm": 2.788114309310913, + "learning_rate": 2.5866006851922347e-06, + "loss": 1.378, + "step": 1576 + }, + { + "epoch": 1.6199280945043657, + "grad_norm": 2.7423717975616455, + "learning_rate": 2.584697373429768e-06, + "loss": 1.3792, + "step": 1577 + }, + { + "epoch": 1.62095531587057, + "grad_norm": 2.777557849884033, + "learning_rate": 2.5827940616673015e-06, + "loss": 1.407, + "step": 1578 + }, + { + "epoch": 1.6219825372367747, + "grad_norm": 3.0097639560699463, + "learning_rate": 2.580890749904835e-06, + "loss": 1.4509, + "step": 1579 + }, + { + "epoch": 1.6230097586029788, + "grad_norm": 2.74116849899292, + "learning_rate": 2.5789874381423678e-06, + "loss": 1.3975, + "step": 1580 + }, + { + "epoch": 1.6230097586029788, + "eval_loss": 1.7757364511489868, + "eval_runtime": 37.0779, + "eval_samples_per_second": 26.97, + "eval_steps_per_second": 1.133, + "step": 1580 + }, + { + "epoch": 1.6240369799691834, + "grad_norm": 2.670633316040039, + "learning_rate": 2.577084126379901e-06, + "loss": 1.401, + "step": 1581 + }, + { + "epoch": 1.6250642013353878, + "grad_norm": 2.7129952907562256, + "learning_rate": 2.5751808146174345e-06, + "loss": 1.3663, + "step": 1582 + }, + { + "epoch": 1.6260914227015921, + "grad_norm": 2.839162826538086, + "learning_rate": 2.573277502854968e-06, + "loss": 1.4189, + "step": 1583 + }, + { + "epoch": 1.6271186440677967, + "grad_norm": 2.7310187816619873, + "learning_rate": 2.5713741910925012e-06, + "loss": 1.3899, + "step": 1584 + }, + { + "epoch": 1.6281458654340009, + "grad_norm": 2.6793200969696045, + "learning_rate": 2.569470879330034e-06, + "loss": 1.3926, + "step": 1585 + }, + { + "epoch": 1.6291730868002054, + "grad_norm": 2.946446418762207, + "learning_rate": 2.5675675675675675e-06, + "loss": 1.4539, + "step": 1586 + }, + { + "epoch": 1.6302003081664098, + "grad_norm": 2.683105945587158, + "learning_rate": 2.565664255805101e-06, + "loss": 1.399, + "step": 1587 + }, + { + "epoch": 1.6312275295326142, + "grad_norm": 2.8176918029785156, + "learning_rate": 2.5637609440426343e-06, + "loss": 1.427, + "step": 1588 + }, + { + "epoch": 1.6322547508988188, + "grad_norm": 2.646280288696289, + "learning_rate": 2.5618576322801676e-06, + "loss": 1.3738, + "step": 1589 + }, + { + "epoch": 1.6332819722650231, + "grad_norm": 2.9067888259887695, + "learning_rate": 2.5599543205177006e-06, + "loss": 1.4666, + "step": 1590 + }, + { + "epoch": 1.6332819722650231, + "eval_loss": 1.7719416618347168, + "eval_runtime": 37.1503, + "eval_samples_per_second": 26.918, + "eval_steps_per_second": 1.131, + "step": 1590 + }, + { + "epoch": 1.6343091936312275, + "grad_norm": 2.797149658203125, + "learning_rate": 2.558051008755234e-06, + "loss": 1.4162, + "step": 1591 + }, + { + "epoch": 1.635336414997432, + "grad_norm": 2.8294360637664795, + "learning_rate": 2.5561476969927677e-06, + "loss": 1.4658, + "step": 1592 + }, + { + "epoch": 1.6363636363636362, + "grad_norm": 2.738529682159424, + "learning_rate": 2.554244385230301e-06, + "loss": 1.3883, + "step": 1593 + }, + { + "epoch": 1.6373908577298408, + "grad_norm": 2.841029167175293, + "learning_rate": 2.5523410734678344e-06, + "loss": 1.4181, + "step": 1594 + }, + { + "epoch": 1.6384180790960452, + "grad_norm": 2.9257984161376953, + "learning_rate": 2.550437761705368e-06, + "loss": 1.3491, + "step": 1595 + }, + { + "epoch": 1.6394453004622496, + "grad_norm": 2.846874713897705, + "learning_rate": 2.548534449942901e-06, + "loss": 1.3832, + "step": 1596 + }, + { + "epoch": 1.6404725218284542, + "grad_norm": 2.9274935722351074, + "learning_rate": 2.546631138180434e-06, + "loss": 1.4215, + "step": 1597 + }, + { + "epoch": 1.6414997431946583, + "grad_norm": 2.9953150749206543, + "learning_rate": 2.5447278264179675e-06, + "loss": 1.4098, + "step": 1598 + }, + { + "epoch": 1.642526964560863, + "grad_norm": 2.7812561988830566, + "learning_rate": 2.542824514655501e-06, + "loss": 1.3806, + "step": 1599 + }, + { + "epoch": 1.6435541859270673, + "grad_norm": 2.9627180099487305, + "learning_rate": 2.540921202893034e-06, + "loss": 1.4623, + "step": 1600 + }, + { + "epoch": 1.6435541859270673, + "eval_loss": 1.7626967430114746, + "eval_runtime": 37.2795, + "eval_samples_per_second": 26.824, + "eval_steps_per_second": 1.127, + "step": 1600 + }, + { + "epoch": 1.6445814072932716, + "grad_norm": 2.8668668270111084, + "learning_rate": 2.5390178911305675e-06, + "loss": 1.4446, + "step": 1601 + }, + { + "epoch": 1.6456086286594762, + "grad_norm": 2.7962896823883057, + "learning_rate": 2.5371145793681005e-06, + "loss": 1.4155, + "step": 1602 + }, + { + "epoch": 1.6466358500256806, + "grad_norm": 2.76855731010437, + "learning_rate": 2.535211267605634e-06, + "loss": 1.4096, + "step": 1603 + }, + { + "epoch": 1.647663071391885, + "grad_norm": 2.7888379096984863, + "learning_rate": 2.533307955843167e-06, + "loss": 1.3544, + "step": 1604 + }, + { + "epoch": 1.6486902927580893, + "grad_norm": 2.9665539264678955, + "learning_rate": 2.5314046440807006e-06, + "loss": 1.4157, + "step": 1605 + }, + { + "epoch": 1.6497175141242937, + "grad_norm": 2.9601891040802, + "learning_rate": 2.529501332318234e-06, + "loss": 1.3963, + "step": 1606 + }, + { + "epoch": 1.6507447354904983, + "grad_norm": 2.897718667984009, + "learning_rate": 2.5275980205557673e-06, + "loss": 1.4498, + "step": 1607 + }, + { + "epoch": 1.6517719568567026, + "grad_norm": 2.9813809394836426, + "learning_rate": 2.5256947087933002e-06, + "loss": 1.4791, + "step": 1608 + }, + { + "epoch": 1.652799178222907, + "grad_norm": 3.025689125061035, + "learning_rate": 2.5237913970308336e-06, + "loss": 1.4616, + "step": 1609 + }, + { + "epoch": 1.6538263995891116, + "grad_norm": 2.6942813396453857, + "learning_rate": 2.521888085268367e-06, + "loss": 1.4112, + "step": 1610 + }, + { + "epoch": 1.6538263995891116, + "eval_loss": 1.775415301322937, + "eval_runtime": 36.8669, + "eval_samples_per_second": 27.125, + "eval_steps_per_second": 1.139, + "step": 1610 + }, + { + "epoch": 1.6548536209553157, + "grad_norm": 2.8190956115722656, + "learning_rate": 2.5199847735059003e-06, + "loss": 1.4559, + "step": 1611 + }, + { + "epoch": 1.6558808423215203, + "grad_norm": 2.7085635662078857, + "learning_rate": 2.518081461743434e-06, + "loss": 1.3905, + "step": 1612 + }, + { + "epoch": 1.6569080636877247, + "grad_norm": 2.856039524078369, + "learning_rate": 2.5161781499809675e-06, + "loss": 1.4223, + "step": 1613 + }, + { + "epoch": 1.657935285053929, + "grad_norm": 2.703293800354004, + "learning_rate": 2.5142748382185004e-06, + "loss": 1.3809, + "step": 1614 + }, + { + "epoch": 1.6589625064201337, + "grad_norm": 2.7698261737823486, + "learning_rate": 2.5123715264560338e-06, + "loss": 1.3732, + "step": 1615 + }, + { + "epoch": 1.6599897277863378, + "grad_norm": 2.8664398193359375, + "learning_rate": 2.510468214693567e-06, + "loss": 1.3994, + "step": 1616 + }, + { + "epoch": 1.6610169491525424, + "grad_norm": 2.843365430831909, + "learning_rate": 2.5085649029311005e-06, + "loss": 1.3466, + "step": 1617 + }, + { + "epoch": 1.6620441705187468, + "grad_norm": 2.7871944904327393, + "learning_rate": 2.506661591168634e-06, + "loss": 1.3096, + "step": 1618 + }, + { + "epoch": 1.6630713918849511, + "grad_norm": 2.9175424575805664, + "learning_rate": 2.504758279406167e-06, + "loss": 1.3931, + "step": 1619 + }, + { + "epoch": 1.6640986132511557, + "grad_norm": 2.926497220993042, + "learning_rate": 2.5028549676437e-06, + "loss": 1.4342, + "step": 1620 + }, + { + "epoch": 1.6640986132511557, + "eval_loss": 1.773324966430664, + "eval_runtime": 36.9745, + "eval_samples_per_second": 27.046, + "eval_steps_per_second": 1.136, + "step": 1620 + }, + { + "epoch": 1.66512583461736, + "grad_norm": 2.7389914989471436, + "learning_rate": 2.5009516558812335e-06, + "loss": 1.3808, + "step": 1621 + }, + { + "epoch": 1.6661530559835644, + "grad_norm": 2.7649383544921875, + "learning_rate": 2.499048344118767e-06, + "loss": 1.4307, + "step": 1622 + }, + { + "epoch": 1.667180277349769, + "grad_norm": 2.8508212566375732, + "learning_rate": 2.4971450323563002e-06, + "loss": 1.4699, + "step": 1623 + }, + { + "epoch": 1.6682074987159732, + "grad_norm": 2.8107385635375977, + "learning_rate": 2.4952417205938336e-06, + "loss": 1.3802, + "step": 1624 + }, + { + "epoch": 1.6692347200821778, + "grad_norm": 2.820322036743164, + "learning_rate": 2.4933384088313666e-06, + "loss": 1.3896, + "step": 1625 + }, + { + "epoch": 1.6702619414483821, + "grad_norm": 2.7912018299102783, + "learning_rate": 2.4914350970689e-06, + "loss": 1.3782, + "step": 1626 + }, + { + "epoch": 1.6712891628145865, + "grad_norm": 2.854904890060425, + "learning_rate": 2.4895317853064337e-06, + "loss": 1.386, + "step": 1627 + }, + { + "epoch": 1.672316384180791, + "grad_norm": 3.2605831623077393, + "learning_rate": 2.4876284735439666e-06, + "loss": 1.446, + "step": 1628 + }, + { + "epoch": 1.6733436055469952, + "grad_norm": 3.029327630996704, + "learning_rate": 2.4857251617815e-06, + "loss": 1.414, + "step": 1629 + }, + { + "epoch": 1.6743708269131998, + "grad_norm": 2.9044439792633057, + "learning_rate": 2.4838218500190334e-06, + "loss": 1.4219, + "step": 1630 + }, + { + "epoch": 1.6743708269131998, + "eval_loss": 1.778739094734192, + "eval_runtime": 36.9917, + "eval_samples_per_second": 27.033, + "eval_steps_per_second": 1.135, + "step": 1630 + }, + { + "epoch": 1.6753980482794042, + "grad_norm": 2.6776554584503174, + "learning_rate": 2.4819185382565667e-06, + "loss": 1.4001, + "step": 1631 + }, + { + "epoch": 1.6764252696456086, + "grad_norm": 2.841038227081299, + "learning_rate": 2.4800152264941e-06, + "loss": 1.406, + "step": 1632 + }, + { + "epoch": 1.6774524910118132, + "grad_norm": 2.848771095275879, + "learning_rate": 2.478111914731633e-06, + "loss": 1.3663, + "step": 1633 + }, + { + "epoch": 1.6784797123780173, + "grad_norm": 2.8403055667877197, + "learning_rate": 2.4762086029691664e-06, + "loss": 1.4065, + "step": 1634 + }, + { + "epoch": 1.6795069337442219, + "grad_norm": 3.0035014152526855, + "learning_rate": 2.4743052912066998e-06, + "loss": 1.5197, + "step": 1635 + }, + { + "epoch": 1.6805341551104263, + "grad_norm": 2.764261245727539, + "learning_rate": 2.472401979444233e-06, + "loss": 1.37, + "step": 1636 + }, + { + "epoch": 1.6815613764766306, + "grad_norm": 3.0496931076049805, + "learning_rate": 2.4704986676817665e-06, + "loss": 1.4657, + "step": 1637 + }, + { + "epoch": 1.6825885978428352, + "grad_norm": 2.8577051162719727, + "learning_rate": 2.4685953559193e-06, + "loss": 1.4019, + "step": 1638 + }, + { + "epoch": 1.6836158192090396, + "grad_norm": 2.8699984550476074, + "learning_rate": 2.466692044156833e-06, + "loss": 1.4508, + "step": 1639 + }, + { + "epoch": 1.684643040575244, + "grad_norm": 2.8038368225097656, + "learning_rate": 2.4647887323943666e-06, + "loss": 1.3859, + "step": 1640 + }, + { + "epoch": 1.684643040575244, + "eval_loss": 1.7762134075164795, + "eval_runtime": 37.0006, + "eval_samples_per_second": 27.027, + "eval_steps_per_second": 1.135, + "step": 1640 + }, + { + "epoch": 1.6856702619414485, + "grad_norm": 2.745168447494507, + "learning_rate": 2.4628854206319e-06, + "loss": 1.3581, + "step": 1641 + }, + { + "epoch": 1.6866974833076527, + "grad_norm": 2.8323488235473633, + "learning_rate": 2.460982108869433e-06, + "loss": 1.3661, + "step": 1642 + }, + { + "epoch": 1.6877247046738573, + "grad_norm": 2.8715322017669678, + "learning_rate": 2.4590787971069662e-06, + "loss": 1.4519, + "step": 1643 + }, + { + "epoch": 1.6887519260400616, + "grad_norm": 2.752286672592163, + "learning_rate": 2.4571754853444996e-06, + "loss": 1.3886, + "step": 1644 + }, + { + "epoch": 1.689779147406266, + "grad_norm": 2.7300102710723877, + "learning_rate": 2.455272173582033e-06, + "loss": 1.3767, + "step": 1645 + }, + { + "epoch": 1.6908063687724706, + "grad_norm": 2.803774356842041, + "learning_rate": 2.4533688618195663e-06, + "loss": 1.39, + "step": 1646 + }, + { + "epoch": 1.6918335901386747, + "grad_norm": 2.6127994060516357, + "learning_rate": 2.4514655500570997e-06, + "loss": 1.3186, + "step": 1647 + }, + { + "epoch": 1.6928608115048793, + "grad_norm": 2.718994140625, + "learning_rate": 2.449562238294633e-06, + "loss": 1.3693, + "step": 1648 + }, + { + "epoch": 1.6938880328710837, + "grad_norm": 2.635831117630005, + "learning_rate": 2.4476589265321664e-06, + "loss": 1.3342, + "step": 1649 + }, + { + "epoch": 1.694915254237288, + "grad_norm": 3.0353753566741943, + "learning_rate": 2.4457556147696993e-06, + "loss": 1.4433, + "step": 1650 + }, + { + "epoch": 1.694915254237288, + "eval_loss": 1.7774006128311157, + "eval_runtime": 37.0101, + "eval_samples_per_second": 27.02, + "eval_steps_per_second": 1.135, + "step": 1650 + }, + { + "epoch": 1.6959424756034926, + "grad_norm": 2.909377098083496, + "learning_rate": 2.4438523030072327e-06, + "loss": 1.4173, + "step": 1651 + }, + { + "epoch": 1.696969696969697, + "grad_norm": 2.9672727584838867, + "learning_rate": 2.441948991244766e-06, + "loss": 1.4274, + "step": 1652 + }, + { + "epoch": 1.6979969183359014, + "grad_norm": 2.822268486022949, + "learning_rate": 2.4400456794822994e-06, + "loss": 1.4012, + "step": 1653 + }, + { + "epoch": 1.6990241397021058, + "grad_norm": 2.8190255165100098, + "learning_rate": 2.438142367719833e-06, + "loss": 1.4251, + "step": 1654 + }, + { + "epoch": 1.7000513610683101, + "grad_norm": 2.8970980644226074, + "learning_rate": 2.4362390559573657e-06, + "loss": 1.3961, + "step": 1655 + }, + { + "epoch": 1.7010785824345147, + "grad_norm": 2.9281842708587646, + "learning_rate": 2.434335744194899e-06, + "loss": 1.3903, + "step": 1656 + }, + { + "epoch": 1.702105803800719, + "grad_norm": 3.0265564918518066, + "learning_rate": 2.432432432432433e-06, + "loss": 1.4319, + "step": 1657 + }, + { + "epoch": 1.7031330251669234, + "grad_norm": 2.785891056060791, + "learning_rate": 2.4305291206699662e-06, + "loss": 1.3561, + "step": 1658 + }, + { + "epoch": 1.704160246533128, + "grad_norm": 2.8406574726104736, + "learning_rate": 2.428625808907499e-06, + "loss": 1.4198, + "step": 1659 + }, + { + "epoch": 1.7051874678993322, + "grad_norm": 2.856170415878296, + "learning_rate": 2.4267224971450325e-06, + "loss": 1.446, + "step": 1660 + }, + { + "epoch": 1.7051874678993322, + "eval_loss": 1.768333077430725, + "eval_runtime": 37.0244, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 1660 + }, + { + "epoch": 1.7062146892655368, + "grad_norm": 2.7454917430877686, + "learning_rate": 2.424819185382566e-06, + "loss": 1.4046, + "step": 1661 + }, + { + "epoch": 1.7072419106317411, + "grad_norm": 2.9186973571777344, + "learning_rate": 2.4229158736200993e-06, + "loss": 1.3616, + "step": 1662 + }, + { + "epoch": 1.7082691319979455, + "grad_norm": 2.859410047531128, + "learning_rate": 2.4210125618576326e-06, + "loss": 1.3594, + "step": 1663 + }, + { + "epoch": 1.70929635336415, + "grad_norm": 2.8369641304016113, + "learning_rate": 2.4191092500951656e-06, + "loss": 1.3924, + "step": 1664 + }, + { + "epoch": 1.7103235747303542, + "grad_norm": 2.8959717750549316, + "learning_rate": 2.417205938332699e-06, + "loss": 1.4301, + "step": 1665 + }, + { + "epoch": 1.7113507960965588, + "grad_norm": 2.8268096446990967, + "learning_rate": 2.4153026265702323e-06, + "loss": 1.3397, + "step": 1666 + }, + { + "epoch": 1.7123780174627632, + "grad_norm": 2.717118740081787, + "learning_rate": 2.4133993148077657e-06, + "loss": 1.3706, + "step": 1667 + }, + { + "epoch": 1.7134052388289676, + "grad_norm": 2.865064859390259, + "learning_rate": 2.411496003045299e-06, + "loss": 1.4356, + "step": 1668 + }, + { + "epoch": 1.7144324601951721, + "grad_norm": 2.8731207847595215, + "learning_rate": 2.4095926912828324e-06, + "loss": 1.3786, + "step": 1669 + }, + { + "epoch": 1.7154596815613765, + "grad_norm": 2.7409722805023193, + "learning_rate": 2.4076893795203658e-06, + "loss": 1.3995, + "step": 1670 + }, + { + "epoch": 1.7154596815613765, + "eval_loss": 1.7796638011932373, + "eval_runtime": 37.0606, + "eval_samples_per_second": 26.983, + "eval_steps_per_second": 1.133, + "step": 1670 + }, + { + "epoch": 1.7164869029275809, + "grad_norm": 2.7279207706451416, + "learning_rate": 2.405786067757899e-06, + "loss": 1.3809, + "step": 1671 + }, + { + "epoch": 1.7175141242937855, + "grad_norm": 3.0839486122131348, + "learning_rate": 2.403882755995432e-06, + "loss": 1.472, + "step": 1672 + }, + { + "epoch": 1.7185413456599896, + "grad_norm": 2.810347557067871, + "learning_rate": 2.4019794442329654e-06, + "loss": 1.3999, + "step": 1673 + }, + { + "epoch": 1.7195685670261942, + "grad_norm": 2.843372106552124, + "learning_rate": 2.4000761324704988e-06, + "loss": 1.3667, + "step": 1674 + }, + { + "epoch": 1.7205957883923986, + "grad_norm": 2.8315467834472656, + "learning_rate": 2.398172820708032e-06, + "loss": 1.4193, + "step": 1675 + }, + { + "epoch": 1.721623009758603, + "grad_norm": 2.647193431854248, + "learning_rate": 2.3962695089455655e-06, + "loss": 1.3826, + "step": 1676 + }, + { + "epoch": 1.7226502311248075, + "grad_norm": 2.8349292278289795, + "learning_rate": 2.3943661971830984e-06, + "loss": 1.3776, + "step": 1677 + }, + { + "epoch": 1.7236774524910117, + "grad_norm": 2.723031520843506, + "learning_rate": 2.3924628854206322e-06, + "loss": 1.3851, + "step": 1678 + }, + { + "epoch": 1.7247046738572163, + "grad_norm": 2.7333242893218994, + "learning_rate": 2.3905595736581656e-06, + "loss": 1.3876, + "step": 1679 + }, + { + "epoch": 1.7257318952234206, + "grad_norm": 2.844578266143799, + "learning_rate": 2.388656261895699e-06, + "loss": 1.4353, + "step": 1680 + }, + { + "epoch": 1.7257318952234206, + "eval_loss": 1.7787370681762695, + "eval_runtime": 37.0536, + "eval_samples_per_second": 26.988, + "eval_steps_per_second": 1.133, + "step": 1680 + }, + { + "epoch": 1.726759116589625, + "grad_norm": 2.9545679092407227, + "learning_rate": 2.386752950133232e-06, + "loss": 1.4284, + "step": 1681 + }, + { + "epoch": 1.7277863379558296, + "grad_norm": 2.9255502223968506, + "learning_rate": 2.3848496383707653e-06, + "loss": 1.4102, + "step": 1682 + }, + { + "epoch": 1.7288135593220337, + "grad_norm": 3.032148838043213, + "learning_rate": 2.3829463266082986e-06, + "loss": 1.4044, + "step": 1683 + }, + { + "epoch": 1.7298407806882383, + "grad_norm": 2.960099935531616, + "learning_rate": 2.381043014845832e-06, + "loss": 1.4274, + "step": 1684 + }, + { + "epoch": 1.7308680020544427, + "grad_norm": 2.8444836139678955, + "learning_rate": 2.3791397030833653e-06, + "loss": 1.4059, + "step": 1685 + }, + { + "epoch": 1.731895223420647, + "grad_norm": 2.7982702255249023, + "learning_rate": 2.3772363913208983e-06, + "loss": 1.3865, + "step": 1686 + }, + { + "epoch": 1.7329224447868516, + "grad_norm": 2.762106418609619, + "learning_rate": 2.3753330795584316e-06, + "loss": 1.3728, + "step": 1687 + }, + { + "epoch": 1.733949666153056, + "grad_norm": 2.8059279918670654, + "learning_rate": 2.3734297677959654e-06, + "loss": 1.3924, + "step": 1688 + }, + { + "epoch": 1.7349768875192604, + "grad_norm": 2.913541316986084, + "learning_rate": 2.3715264560334984e-06, + "loss": 1.4347, + "step": 1689 + }, + { + "epoch": 1.736004108885465, + "grad_norm": 2.7966978549957275, + "learning_rate": 2.3696231442710317e-06, + "loss": 1.4246, + "step": 1690 + }, + { + "epoch": 1.736004108885465, + "eval_loss": 1.766000747680664, + "eval_runtime": 37.1692, + "eval_samples_per_second": 26.904, + "eval_steps_per_second": 1.13, + "step": 1690 + }, + { + "epoch": 1.7370313302516691, + "grad_norm": 2.795525550842285, + "learning_rate": 2.367719832508565e-06, + "loss": 1.4169, + "step": 1691 + }, + { + "epoch": 1.7380585516178737, + "grad_norm": 2.8152341842651367, + "learning_rate": 2.3658165207460985e-06, + "loss": 1.438, + "step": 1692 + }, + { + "epoch": 1.739085772984078, + "grad_norm": 2.6966261863708496, + "learning_rate": 2.363913208983632e-06, + "loss": 1.4164, + "step": 1693 + }, + { + "epoch": 1.7401129943502824, + "grad_norm": 2.875034809112549, + "learning_rate": 2.362009897221165e-06, + "loss": 1.4325, + "step": 1694 + }, + { + "epoch": 1.741140215716487, + "grad_norm": 2.9886977672576904, + "learning_rate": 2.360106585458698e-06, + "loss": 1.4277, + "step": 1695 + }, + { + "epoch": 1.7421674370826912, + "grad_norm": 2.735360860824585, + "learning_rate": 2.3582032736962315e-06, + "loss": 1.3688, + "step": 1696 + }, + { + "epoch": 1.7431946584488958, + "grad_norm": 2.823995351791382, + "learning_rate": 2.356299961933765e-06, + "loss": 1.4062, + "step": 1697 + }, + { + "epoch": 1.7442218798151001, + "grad_norm": 2.7787744998931885, + "learning_rate": 2.354396650171298e-06, + "loss": 1.3782, + "step": 1698 + }, + { + "epoch": 1.7452491011813045, + "grad_norm": 2.745983600616455, + "learning_rate": 2.3524933384088316e-06, + "loss": 1.3682, + "step": 1699 + }, + { + "epoch": 1.746276322547509, + "grad_norm": 2.853743553161621, + "learning_rate": 2.350590026646365e-06, + "loss": 1.4544, + "step": 1700 + }, + { + "epoch": 1.746276322547509, + "eval_loss": 1.7642710208892822, + "eval_runtime": 37.3227, + "eval_samples_per_second": 26.793, + "eval_steps_per_second": 1.125, + "step": 1700 + }, + { + "epoch": 1.7473035439137135, + "grad_norm": 2.8187296390533447, + "learning_rate": 2.3486867148838983e-06, + "loss": 1.3874, + "step": 1701 + }, + { + "epoch": 1.7483307652799178, + "grad_norm": 2.766657829284668, + "learning_rate": 2.3467834031214317e-06, + "loss": 1.37, + "step": 1702 + }, + { + "epoch": 1.7493579866461222, + "grad_norm": 2.796694755554199, + "learning_rate": 2.3448800913589646e-06, + "loss": 1.3909, + "step": 1703 + }, + { + "epoch": 1.7503852080123266, + "grad_norm": 2.8409624099731445, + "learning_rate": 2.342976779596498e-06, + "loss": 1.4091, + "step": 1704 + }, + { + "epoch": 1.7514124293785311, + "grad_norm": 2.7160699367523193, + "learning_rate": 2.3410734678340313e-06, + "loss": 1.3771, + "step": 1705 + }, + { + "epoch": 1.7524396507447355, + "grad_norm": 2.7986226081848145, + "learning_rate": 2.3391701560715647e-06, + "loss": 1.4323, + "step": 1706 + }, + { + "epoch": 1.7534668721109399, + "grad_norm": 2.814997911453247, + "learning_rate": 2.337266844309098e-06, + "loss": 1.3903, + "step": 1707 + }, + { + "epoch": 1.7544940934771445, + "grad_norm": 2.7355244159698486, + "learning_rate": 2.3353635325466314e-06, + "loss": 1.3955, + "step": 1708 + }, + { + "epoch": 1.7555213148433486, + "grad_norm": 2.8889238834381104, + "learning_rate": 2.3334602207841648e-06, + "loss": 1.4037, + "step": 1709 + }, + { + "epoch": 1.7565485362095532, + "grad_norm": 2.9253830909729004, + "learning_rate": 2.331556909021698e-06, + "loss": 1.3982, + "step": 1710 + }, + { + "epoch": 1.7565485362095532, + "eval_loss": 1.7730076313018799, + "eval_runtime": 36.931, + "eval_samples_per_second": 27.078, + "eval_steps_per_second": 1.137, + "step": 1710 + }, + { + "epoch": 1.7575757575757576, + "grad_norm": 2.8409945964813232, + "learning_rate": 2.3296535972592315e-06, + "loss": 1.4313, + "step": 1711 + }, + { + "epoch": 1.758602978941962, + "grad_norm": 2.7974541187286377, + "learning_rate": 2.3277502854967644e-06, + "loss": 1.375, + "step": 1712 + }, + { + "epoch": 1.7596302003081665, + "grad_norm": 2.828486204147339, + "learning_rate": 2.325846973734298e-06, + "loss": 1.4406, + "step": 1713 + }, + { + "epoch": 1.7606574216743707, + "grad_norm": 2.8244118690490723, + "learning_rate": 2.323943661971831e-06, + "loss": 1.3771, + "step": 1714 + }, + { + "epoch": 1.7616846430405753, + "grad_norm": 2.7830047607421875, + "learning_rate": 2.3220403502093645e-06, + "loss": 1.4408, + "step": 1715 + }, + { + "epoch": 1.7627118644067796, + "grad_norm": 2.889033079147339, + "learning_rate": 2.320137038446898e-06, + "loss": 1.438, + "step": 1716 + }, + { + "epoch": 1.763739085772984, + "grad_norm": 2.9460082054138184, + "learning_rate": 2.318233726684431e-06, + "loss": 1.4424, + "step": 1717 + }, + { + "epoch": 1.7647663071391886, + "grad_norm": 2.8492050170898438, + "learning_rate": 2.3163304149219646e-06, + "loss": 1.4354, + "step": 1718 + }, + { + "epoch": 1.765793528505393, + "grad_norm": 2.8115553855895996, + "learning_rate": 2.314427103159498e-06, + "loss": 1.4064, + "step": 1719 + }, + { + "epoch": 1.7668207498715973, + "grad_norm": 2.8483548164367676, + "learning_rate": 2.312523791397031e-06, + "loss": 1.3915, + "step": 1720 + }, + { + "epoch": 1.7668207498715973, + "eval_loss": 1.770641803741455, + "eval_runtime": 37.0317, + "eval_samples_per_second": 27.004, + "eval_steps_per_second": 1.134, + "step": 1720 + }, + { + "epoch": 1.767847971237802, + "grad_norm": 2.67576003074646, + "learning_rate": 2.3106204796345643e-06, + "loss": 1.3665, + "step": 1721 + }, + { + "epoch": 1.768875192604006, + "grad_norm": 2.815186023712158, + "learning_rate": 2.3087171678720976e-06, + "loss": 1.3975, + "step": 1722 + }, + { + "epoch": 1.7699024139702106, + "grad_norm": 2.9318761825561523, + "learning_rate": 2.306813856109631e-06, + "loss": 1.4506, + "step": 1723 + }, + { + "epoch": 1.770929635336415, + "grad_norm": 2.7512404918670654, + "learning_rate": 2.3049105443471644e-06, + "loss": 1.4047, + "step": 1724 + }, + { + "epoch": 1.7719568567026194, + "grad_norm": 2.8040406703948975, + "learning_rate": 2.3030072325846973e-06, + "loss": 1.3813, + "step": 1725 + }, + { + "epoch": 1.772984078068824, + "grad_norm": 2.7659668922424316, + "learning_rate": 2.3011039208222307e-06, + "loss": 1.373, + "step": 1726 + }, + { + "epoch": 1.774011299435028, + "grad_norm": 2.7986247539520264, + "learning_rate": 2.299200609059764e-06, + "loss": 1.3846, + "step": 1727 + }, + { + "epoch": 1.7750385208012327, + "grad_norm": 2.8309032917022705, + "learning_rate": 2.297297297297298e-06, + "loss": 1.3854, + "step": 1728 + }, + { + "epoch": 1.776065742167437, + "grad_norm": 2.7988357543945312, + "learning_rate": 2.2953939855348308e-06, + "loss": 1.3544, + "step": 1729 + }, + { + "epoch": 1.7770929635336414, + "grad_norm": 2.8682641983032227, + "learning_rate": 2.293490673772364e-06, + "loss": 1.4388, + "step": 1730 + }, + { + "epoch": 1.7770929635336414, + "eval_loss": 1.7706880569458008, + "eval_runtime": 37.0355, + "eval_samples_per_second": 27.001, + "eval_steps_per_second": 1.134, + "step": 1730 + }, + { + "epoch": 1.778120184899846, + "grad_norm": 2.8318183422088623, + "learning_rate": 2.2915873620098975e-06, + "loss": 1.4257, + "step": 1731 + }, + { + "epoch": 1.7791474062660502, + "grad_norm": 2.8488996028900146, + "learning_rate": 2.289684050247431e-06, + "loss": 1.4152, + "step": 1732 + }, + { + "epoch": 1.7801746276322548, + "grad_norm": 2.7117509841918945, + "learning_rate": 2.287780738484964e-06, + "loss": 1.3849, + "step": 1733 + }, + { + "epoch": 1.7812018489984591, + "grad_norm": 2.7878878116607666, + "learning_rate": 2.285877426722497e-06, + "loss": 1.4071, + "step": 1734 + }, + { + "epoch": 1.7822290703646635, + "grad_norm": 2.876696825027466, + "learning_rate": 2.2839741149600305e-06, + "loss": 1.4393, + "step": 1735 + }, + { + "epoch": 1.783256291730868, + "grad_norm": 2.8824517726898193, + "learning_rate": 2.282070803197564e-06, + "loss": 1.4225, + "step": 1736 + }, + { + "epoch": 1.7842835130970724, + "grad_norm": 2.8785459995269775, + "learning_rate": 2.2801674914350972e-06, + "loss": 1.4093, + "step": 1737 + }, + { + "epoch": 1.7853107344632768, + "grad_norm": 2.7482051849365234, + "learning_rate": 2.2782641796726306e-06, + "loss": 1.4004, + "step": 1738 + }, + { + "epoch": 1.7863379558294814, + "grad_norm": 2.812790632247925, + "learning_rate": 2.276360867910164e-06, + "loss": 1.4284, + "step": 1739 + }, + { + "epoch": 1.7873651771956856, + "grad_norm": 2.8116066455841064, + "learning_rate": 2.2744575561476973e-06, + "loss": 1.3601, + "step": 1740 + }, + { + "epoch": 1.7873651771956856, + "eval_loss": 1.780893325805664, + "eval_runtime": 37.0422, + "eval_samples_per_second": 26.996, + "eval_steps_per_second": 1.134, + "step": 1740 + }, + { + "epoch": 1.7883923985618901, + "grad_norm": 2.7929999828338623, + "learning_rate": 2.2725542443852307e-06, + "loss": 1.4034, + "step": 1741 + }, + { + "epoch": 1.7894196199280945, + "grad_norm": 2.8496506214141846, + "learning_rate": 2.2706509326227636e-06, + "loss": 1.4414, + "step": 1742 + }, + { + "epoch": 1.7904468412942989, + "grad_norm": 2.7631287574768066, + "learning_rate": 2.268747620860297e-06, + "loss": 1.4117, + "step": 1743 + }, + { + "epoch": 1.7914740626605035, + "grad_norm": 2.731952428817749, + "learning_rate": 2.2668443090978303e-06, + "loss": 1.4064, + "step": 1744 + }, + { + "epoch": 1.7925012840267076, + "grad_norm": 2.7945852279663086, + "learning_rate": 2.2649409973353637e-06, + "loss": 1.4329, + "step": 1745 + }, + { + "epoch": 1.7935285053929122, + "grad_norm": 2.7181437015533447, + "learning_rate": 2.263037685572897e-06, + "loss": 1.3821, + "step": 1746 + }, + { + "epoch": 1.7945557267591166, + "grad_norm": 2.792771816253662, + "learning_rate": 2.26113437381043e-06, + "loss": 1.3789, + "step": 1747 + }, + { + "epoch": 1.795582948125321, + "grad_norm": 2.9297096729278564, + "learning_rate": 2.259231062047964e-06, + "loss": 1.4049, + "step": 1748 + }, + { + "epoch": 1.7966101694915255, + "grad_norm": 2.9424288272857666, + "learning_rate": 2.257327750285497e-06, + "loss": 1.4275, + "step": 1749 + }, + { + "epoch": 1.7976373908577299, + "grad_norm": 2.880979537963867, + "learning_rate": 2.2554244385230305e-06, + "loss": 1.3911, + "step": 1750 + }, + { + "epoch": 1.7976373908577299, + "eval_loss": 1.7886379957199097, + "eval_runtime": 37.0413, + "eval_samples_per_second": 26.997, + "eval_steps_per_second": 1.134, + "step": 1750 + }, + { + "epoch": 1.7986646122239343, + "grad_norm": 2.9970686435699463, + "learning_rate": 2.2535211267605635e-06, + "loss": 1.3976, + "step": 1751 + }, + { + "epoch": 1.7996918335901386, + "grad_norm": 3.0055971145629883, + "learning_rate": 2.251617814998097e-06, + "loss": 1.4029, + "step": 1752 + }, + { + "epoch": 1.800719054956343, + "grad_norm": 2.8585352897644043, + "learning_rate": 2.24971450323563e-06, + "loss": 1.3279, + "step": 1753 + }, + { + "epoch": 1.8017462763225476, + "grad_norm": 2.8452632427215576, + "learning_rate": 2.2478111914731636e-06, + "loss": 1.4141, + "step": 1754 + }, + { + "epoch": 1.802773497688752, + "grad_norm": 2.642672538757324, + "learning_rate": 2.245907879710697e-06, + "loss": 1.2864, + "step": 1755 + }, + { + "epoch": 1.8038007190549563, + "grad_norm": 2.8887314796447754, + "learning_rate": 2.24400456794823e-06, + "loss": 1.4301, + "step": 1756 + }, + { + "epoch": 1.804827940421161, + "grad_norm": 2.735727071762085, + "learning_rate": 2.2421012561857632e-06, + "loss": 1.3908, + "step": 1757 + }, + { + "epoch": 1.805855161787365, + "grad_norm": 2.861366033554077, + "learning_rate": 2.240197944423297e-06, + "loss": 1.3757, + "step": 1758 + }, + { + "epoch": 1.8068823831535696, + "grad_norm": 2.7908987998962402, + "learning_rate": 2.23829463266083e-06, + "loss": 1.3758, + "step": 1759 + }, + { + "epoch": 1.807909604519774, + "grad_norm": 2.602125644683838, + "learning_rate": 2.2363913208983633e-06, + "loss": 1.369, + "step": 1760 + }, + { + "epoch": 1.807909604519774, + "eval_loss": 1.7980055809020996, + "eval_runtime": 37.0615, + "eval_samples_per_second": 26.982, + "eval_steps_per_second": 1.133, + "step": 1760 + }, + { + "epoch": 1.8089368258859784, + "grad_norm": 2.693953275680542, + "learning_rate": 2.2344880091358967e-06, + "loss": 1.4095, + "step": 1761 + }, + { + "epoch": 1.809964047252183, + "grad_norm": 2.772446632385254, + "learning_rate": 2.23258469737343e-06, + "loss": 1.386, + "step": 1762 + }, + { + "epoch": 1.810991268618387, + "grad_norm": 2.7845568656921387, + "learning_rate": 2.2306813856109634e-06, + "loss": 1.4293, + "step": 1763 + }, + { + "epoch": 1.8120184899845917, + "grad_norm": 2.786252498626709, + "learning_rate": 2.2287780738484963e-06, + "loss": 1.3805, + "step": 1764 + }, + { + "epoch": 1.813045711350796, + "grad_norm": 2.859053611755371, + "learning_rate": 2.2268747620860297e-06, + "loss": 1.4585, + "step": 1765 + }, + { + "epoch": 1.8140729327170004, + "grad_norm": 2.8749146461486816, + "learning_rate": 2.224971450323563e-06, + "loss": 1.393, + "step": 1766 + }, + { + "epoch": 1.815100154083205, + "grad_norm": 2.740787982940674, + "learning_rate": 2.2230681385610964e-06, + "loss": 1.3309, + "step": 1767 + }, + { + "epoch": 1.8161273754494094, + "grad_norm": 2.834662914276123, + "learning_rate": 2.2211648267986298e-06, + "loss": 1.3961, + "step": 1768 + }, + { + "epoch": 1.8171545968156138, + "grad_norm": 2.8033571243286133, + "learning_rate": 2.219261515036163e-06, + "loss": 1.3747, + "step": 1769 + }, + { + "epoch": 1.8181818181818183, + "grad_norm": 2.9068171977996826, + "learning_rate": 2.2173582032736965e-06, + "loss": 1.4091, + "step": 1770 + }, + { + "epoch": 1.8181818181818183, + "eval_loss": 1.7923928499221802, + "eval_runtime": 37.067, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 1770 + }, + { + "epoch": 1.8192090395480225, + "grad_norm": 2.8439955711364746, + "learning_rate": 2.21545489151123e-06, + "loss": 1.4281, + "step": 1771 + }, + { + "epoch": 1.820236260914227, + "grad_norm": 2.861231565475464, + "learning_rate": 2.2135515797487632e-06, + "loss": 1.3887, + "step": 1772 + }, + { + "epoch": 1.8212634822804314, + "grad_norm": 2.986788034439087, + "learning_rate": 2.211648267986296e-06, + "loss": 1.395, + "step": 1773 + }, + { + "epoch": 1.8222907036466358, + "grad_norm": 2.755415678024292, + "learning_rate": 2.2097449562238295e-06, + "loss": 1.3885, + "step": 1774 + }, + { + "epoch": 1.8233179250128404, + "grad_norm": 2.7685859203338623, + "learning_rate": 2.207841644461363e-06, + "loss": 1.3861, + "step": 1775 + }, + { + "epoch": 1.8243451463790445, + "grad_norm": 2.970322608947754, + "learning_rate": 2.2059383326988963e-06, + "loss": 1.4436, + "step": 1776 + }, + { + "epoch": 1.8253723677452491, + "grad_norm": 2.997922658920288, + "learning_rate": 2.2040350209364296e-06, + "loss": 1.3978, + "step": 1777 + }, + { + "epoch": 1.8263995891114535, + "grad_norm": 2.734039545059204, + "learning_rate": 2.2021317091739626e-06, + "loss": 1.3623, + "step": 1778 + }, + { + "epoch": 1.8274268104776579, + "grad_norm": 2.8555188179016113, + "learning_rate": 2.2002283974114963e-06, + "loss": 1.3998, + "step": 1779 + }, + { + "epoch": 1.8284540318438625, + "grad_norm": 2.8464975357055664, + "learning_rate": 2.1983250856490297e-06, + "loss": 1.3346, + "step": 1780 + }, + { + "epoch": 1.8284540318438625, + "eval_loss": 1.7876062393188477, + "eval_runtime": 37.0771, + "eval_samples_per_second": 26.971, + "eval_steps_per_second": 1.133, + "step": 1780 + }, + { + "epoch": 1.8294812532100666, + "grad_norm": 3.012516975402832, + "learning_rate": 2.1964217738865626e-06, + "loss": 1.3803, + "step": 1781 + }, + { + "epoch": 1.8305084745762712, + "grad_norm": 2.941331148147583, + "learning_rate": 2.194518462124096e-06, + "loss": 1.4033, + "step": 1782 + }, + { + "epoch": 1.8315356959424756, + "grad_norm": 3.090533494949341, + "learning_rate": 2.1926151503616294e-06, + "loss": 1.443, + "step": 1783 + }, + { + "epoch": 1.83256291730868, + "grad_norm": 2.890427827835083, + "learning_rate": 2.1907118385991627e-06, + "loss": 1.4089, + "step": 1784 + }, + { + "epoch": 1.8335901386748845, + "grad_norm": 2.8417627811431885, + "learning_rate": 2.188808526836696e-06, + "loss": 1.3722, + "step": 1785 + }, + { + "epoch": 1.8346173600410889, + "grad_norm": 2.7787647247314453, + "learning_rate": 2.1869052150742295e-06, + "loss": 1.4011, + "step": 1786 + }, + { + "epoch": 1.8356445814072933, + "grad_norm": 2.7441141605377197, + "learning_rate": 2.1850019033117624e-06, + "loss": 1.3262, + "step": 1787 + }, + { + "epoch": 1.8366718027734978, + "grad_norm": 2.873112678527832, + "learning_rate": 2.1830985915492958e-06, + "loss": 1.3896, + "step": 1788 + }, + { + "epoch": 1.837699024139702, + "grad_norm": 2.8022804260253906, + "learning_rate": 2.1811952797868295e-06, + "loss": 1.4035, + "step": 1789 + }, + { + "epoch": 1.8387262455059066, + "grad_norm": 2.8220114707946777, + "learning_rate": 2.1792919680243625e-06, + "loss": 1.3748, + "step": 1790 + }, + { + "epoch": 1.8387262455059066, + "eval_loss": 1.7839077711105347, + "eval_runtime": 37.0886, + "eval_samples_per_second": 26.962, + "eval_steps_per_second": 1.132, + "step": 1790 + }, + { + "epoch": 1.839753466872111, + "grad_norm": 2.769643783569336, + "learning_rate": 2.177388656261896e-06, + "loss": 1.4084, + "step": 1791 + }, + { + "epoch": 1.8407806882383153, + "grad_norm": 2.8555970191955566, + "learning_rate": 2.1754853444994292e-06, + "loss": 1.4772, + "step": 1792 + }, + { + "epoch": 1.84180790960452, + "grad_norm": 2.777965545654297, + "learning_rate": 2.1735820327369626e-06, + "loss": 1.3024, + "step": 1793 + }, + { + "epoch": 1.842835130970724, + "grad_norm": 2.8747634887695312, + "learning_rate": 2.171678720974496e-06, + "loss": 1.4195, + "step": 1794 + }, + { + "epoch": 1.8438623523369286, + "grad_norm": 2.8822247982025146, + "learning_rate": 2.169775409212029e-06, + "loss": 1.4151, + "step": 1795 + }, + { + "epoch": 1.844889573703133, + "grad_norm": 2.8203372955322266, + "learning_rate": 2.1678720974495622e-06, + "loss": 1.4223, + "step": 1796 + }, + { + "epoch": 1.8459167950693374, + "grad_norm": 2.632070779800415, + "learning_rate": 2.1659687856870956e-06, + "loss": 1.4141, + "step": 1797 + }, + { + "epoch": 1.846944016435542, + "grad_norm": 2.746530771255493, + "learning_rate": 2.164065473924629e-06, + "loss": 1.4351, + "step": 1798 + }, + { + "epoch": 1.8479712378017463, + "grad_norm": 2.7693254947662354, + "learning_rate": 2.1621621621621623e-06, + "loss": 1.3588, + "step": 1799 + }, + { + "epoch": 1.8489984591679507, + "grad_norm": 2.7944774627685547, + "learning_rate": 2.1602588503996957e-06, + "loss": 1.4181, + "step": 1800 + }, + { + "epoch": 1.8489984591679507, + "eval_loss": 1.780879259109497, + "eval_runtime": 37.2375, + "eval_samples_per_second": 26.855, + "eval_steps_per_second": 1.128, + "step": 1800 + }, + { + "epoch": 1.850025680534155, + "grad_norm": 2.827951669692993, + "learning_rate": 2.158355538637229e-06, + "loss": 1.3857, + "step": 1801 + }, + { + "epoch": 1.8510529019003594, + "grad_norm": 2.7634339332580566, + "learning_rate": 2.1564522268747624e-06, + "loss": 1.387, + "step": 1802 + }, + { + "epoch": 1.852080123266564, + "grad_norm": 3.014620542526245, + "learning_rate": 2.1545489151122958e-06, + "loss": 1.4595, + "step": 1803 + }, + { + "epoch": 1.8531073446327684, + "grad_norm": 2.7546815872192383, + "learning_rate": 2.1526456033498287e-06, + "loss": 1.4098, + "step": 1804 + }, + { + "epoch": 1.8541345659989728, + "grad_norm": 2.7237050533294678, + "learning_rate": 2.150742291587362e-06, + "loss": 1.4099, + "step": 1805 + }, + { + "epoch": 1.8551617873651773, + "grad_norm": 2.8354947566986084, + "learning_rate": 2.1488389798248954e-06, + "loss": 1.395, + "step": 1806 + }, + { + "epoch": 1.8561890087313815, + "grad_norm": 2.863032102584839, + "learning_rate": 2.146935668062429e-06, + "loss": 1.4275, + "step": 1807 + }, + { + "epoch": 1.857216230097586, + "grad_norm": 2.6963112354278564, + "learning_rate": 2.145032356299962e-06, + "loss": 1.3567, + "step": 1808 + }, + { + "epoch": 1.8582434514637904, + "grad_norm": 2.7628366947174072, + "learning_rate": 2.1431290445374955e-06, + "loss": 1.4036, + "step": 1809 + }, + { + "epoch": 1.8592706728299948, + "grad_norm": 2.790919065475464, + "learning_rate": 2.141225732775029e-06, + "loss": 1.3679, + "step": 1810 + }, + { + "epoch": 1.8592706728299948, + "eval_loss": 1.7804644107818604, + "eval_runtime": 36.9549, + "eval_samples_per_second": 27.06, + "eval_steps_per_second": 1.137, + "step": 1810 + }, + { + "epoch": 1.8602978941961994, + "grad_norm": 2.913525104522705, + "learning_rate": 2.1393224210125623e-06, + "loss": 1.4015, + "step": 1811 + }, + { + "epoch": 1.8613251155624035, + "grad_norm": 2.8357040882110596, + "learning_rate": 2.137419109250095e-06, + "loss": 1.3912, + "step": 1812 + }, + { + "epoch": 1.8623523369286081, + "grad_norm": 2.989529848098755, + "learning_rate": 2.1355157974876286e-06, + "loss": 1.4397, + "step": 1813 + }, + { + "epoch": 1.8633795582948125, + "grad_norm": 2.8843417167663574, + "learning_rate": 2.133612485725162e-06, + "loss": 1.4343, + "step": 1814 + }, + { + "epoch": 1.8644067796610169, + "grad_norm": 2.882571220397949, + "learning_rate": 2.1317091739626953e-06, + "loss": 1.342, + "step": 1815 + }, + { + "epoch": 1.8654340010272215, + "grad_norm": 2.8986611366271973, + "learning_rate": 2.1298058622002286e-06, + "loss": 1.3687, + "step": 1816 + }, + { + "epoch": 1.8664612223934258, + "grad_norm": 2.8779523372650146, + "learning_rate": 2.1279025504377616e-06, + "loss": 1.4005, + "step": 1817 + }, + { + "epoch": 1.8674884437596302, + "grad_norm": 2.7829904556274414, + "learning_rate": 2.125999238675295e-06, + "loss": 1.3843, + "step": 1818 + }, + { + "epoch": 1.8685156651258348, + "grad_norm": 3.0068447589874268, + "learning_rate": 2.1240959269128287e-06, + "loss": 1.4347, + "step": 1819 + }, + { + "epoch": 1.869542886492039, + "grad_norm": 2.945225715637207, + "learning_rate": 2.122192615150362e-06, + "loss": 1.3934, + "step": 1820 + }, + { + "epoch": 1.869542886492039, + "eval_loss": 1.7796295881271362, + "eval_runtime": 37.0423, + "eval_samples_per_second": 26.996, + "eval_steps_per_second": 1.134, + "step": 1820 + }, + { + "epoch": 1.8705701078582435, + "grad_norm": 2.79140305519104, + "learning_rate": 2.120289303387895e-06, + "loss": 1.3387, + "step": 1821 + }, + { + "epoch": 1.8715973292244479, + "grad_norm": 2.897294282913208, + "learning_rate": 2.1183859916254284e-06, + "loss": 1.3905, + "step": 1822 + }, + { + "epoch": 1.8726245505906522, + "grad_norm": 2.8181400299072266, + "learning_rate": 2.1164826798629618e-06, + "loss": 1.4219, + "step": 1823 + }, + { + "epoch": 1.8736517719568568, + "grad_norm": 2.695333480834961, + "learning_rate": 2.114579368100495e-06, + "loss": 1.38, + "step": 1824 + }, + { + "epoch": 1.874678993323061, + "grad_norm": 2.7668285369873047, + "learning_rate": 2.1126760563380285e-06, + "loss": 1.3613, + "step": 1825 + }, + { + "epoch": 1.8757062146892656, + "grad_norm": 2.723405599594116, + "learning_rate": 2.1107727445755614e-06, + "loss": 1.3344, + "step": 1826 + }, + { + "epoch": 1.87673343605547, + "grad_norm": 3.01149320602417, + "learning_rate": 2.1088694328130948e-06, + "loss": 1.4672, + "step": 1827 + }, + { + "epoch": 1.8777606574216743, + "grad_norm": 2.7828991413116455, + "learning_rate": 2.106966121050628e-06, + "loss": 1.3671, + "step": 1828 + }, + { + "epoch": 1.878787878787879, + "grad_norm": 2.792518138885498, + "learning_rate": 2.1050628092881615e-06, + "loss": 1.4182, + "step": 1829 + }, + { + "epoch": 1.879815100154083, + "grad_norm": 2.935760736465454, + "learning_rate": 2.103159497525695e-06, + "loss": 1.4623, + "step": 1830 + }, + { + "epoch": 1.879815100154083, + "eval_loss": 1.7818807363510132, + "eval_runtime": 37.0809, + "eval_samples_per_second": 26.968, + "eval_steps_per_second": 1.133, + "step": 1830 + }, + { + "epoch": 1.8808423215202876, + "grad_norm": 2.849071741104126, + "learning_rate": 2.1012561857632282e-06, + "loss": 1.3777, + "step": 1831 + }, + { + "epoch": 1.881869542886492, + "grad_norm": 2.978045701980591, + "learning_rate": 2.0993528740007616e-06, + "loss": 1.4325, + "step": 1832 + }, + { + "epoch": 1.8828967642526964, + "grad_norm": 2.8763439655303955, + "learning_rate": 2.097449562238295e-06, + "loss": 1.4417, + "step": 1833 + }, + { + "epoch": 1.883923985618901, + "grad_norm": 2.891221523284912, + "learning_rate": 2.095546250475828e-06, + "loss": 1.3881, + "step": 1834 + }, + { + "epoch": 1.8849512069851053, + "grad_norm": 2.9743025302886963, + "learning_rate": 2.0936429387133613e-06, + "loss": 1.3956, + "step": 1835 + }, + { + "epoch": 1.8859784283513097, + "grad_norm": 2.7557995319366455, + "learning_rate": 2.0917396269508946e-06, + "loss": 1.4511, + "step": 1836 + }, + { + "epoch": 1.8870056497175143, + "grad_norm": 2.878021240234375, + "learning_rate": 2.089836315188428e-06, + "loss": 1.4206, + "step": 1837 + }, + { + "epoch": 1.8880328710837184, + "grad_norm": 2.9022324085235596, + "learning_rate": 2.0879330034259614e-06, + "loss": 1.4256, + "step": 1838 + }, + { + "epoch": 1.889060092449923, + "grad_norm": 2.8805174827575684, + "learning_rate": 2.0860296916634947e-06, + "loss": 1.3752, + "step": 1839 + }, + { + "epoch": 1.8900873138161274, + "grad_norm": 2.887437105178833, + "learning_rate": 2.084126379901028e-06, + "loss": 1.4636, + "step": 1840 + }, + { + "epoch": 1.8900873138161274, + "eval_loss": 1.775463581085205, + "eval_runtime": 37.0928, + "eval_samples_per_second": 26.959, + "eval_steps_per_second": 1.132, + "step": 1840 + }, + { + "epoch": 1.8911145351823317, + "grad_norm": 2.8477346897125244, + "learning_rate": 2.0822230681385614e-06, + "loss": 1.3653, + "step": 1841 + }, + { + "epoch": 1.8921417565485363, + "grad_norm": 2.939516067504883, + "learning_rate": 2.080319756376095e-06, + "loss": 1.4216, + "step": 1842 + }, + { + "epoch": 1.8931689779147405, + "grad_norm": 2.869184970855713, + "learning_rate": 2.0784164446136277e-06, + "loss": 1.4138, + "step": 1843 + }, + { + "epoch": 1.894196199280945, + "grad_norm": 2.7639729976654053, + "learning_rate": 2.076513132851161e-06, + "loss": 1.3837, + "step": 1844 + }, + { + "epoch": 1.8952234206471494, + "grad_norm": 2.9656383991241455, + "learning_rate": 2.0746098210886945e-06, + "loss": 1.4168, + "step": 1845 + }, + { + "epoch": 1.8962506420133538, + "grad_norm": 2.80155873298645, + "learning_rate": 2.072706509326228e-06, + "loss": 1.3879, + "step": 1846 + }, + { + "epoch": 1.8972778633795584, + "grad_norm": 2.742859363555908, + "learning_rate": 2.070803197563761e-06, + "loss": 1.397, + "step": 1847 + }, + { + "epoch": 1.8983050847457628, + "grad_norm": 2.790229082107544, + "learning_rate": 2.068899885801294e-06, + "loss": 1.3894, + "step": 1848 + }, + { + "epoch": 1.8993323061119671, + "grad_norm": 2.87921142578125, + "learning_rate": 2.066996574038828e-06, + "loss": 1.4122, + "step": 1849 + }, + { + "epoch": 1.9003595274781715, + "grad_norm": 2.937082052230835, + "learning_rate": 2.0650932622763613e-06, + "loss": 1.435, + "step": 1850 + }, + { + "epoch": 1.9003595274781715, + "eval_loss": 1.7634403705596924, + "eval_runtime": 37.0377, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 1850 + }, + { + "epoch": 1.9013867488443759, + "grad_norm": 2.830686569213867, + "learning_rate": 2.0631899505138942e-06, + "loss": 1.3248, + "step": 1851 + }, + { + "epoch": 1.9024139702105805, + "grad_norm": 2.878131151199341, + "learning_rate": 2.0612866387514276e-06, + "loss": 1.3778, + "step": 1852 + }, + { + "epoch": 1.9034411915767848, + "grad_norm": 2.894076108932495, + "learning_rate": 2.059383326988961e-06, + "loss": 1.4073, + "step": 1853 + }, + { + "epoch": 1.9044684129429892, + "grad_norm": 2.816612958908081, + "learning_rate": 2.0574800152264943e-06, + "loss": 1.3873, + "step": 1854 + }, + { + "epoch": 1.9054956343091938, + "grad_norm": 2.905581474304199, + "learning_rate": 2.0555767034640277e-06, + "loss": 1.3644, + "step": 1855 + }, + { + "epoch": 1.906522855675398, + "grad_norm": 2.7126173973083496, + "learning_rate": 2.053673391701561e-06, + "loss": 1.3402, + "step": 1856 + }, + { + "epoch": 1.9075500770416025, + "grad_norm": 2.8092000484466553, + "learning_rate": 2.051770079939094e-06, + "loss": 1.4095, + "step": 1857 + }, + { + "epoch": 1.9085772984078069, + "grad_norm": 2.7571353912353516, + "learning_rate": 2.0498667681766273e-06, + "loss": 1.3657, + "step": 1858 + }, + { + "epoch": 1.9096045197740112, + "grad_norm": 2.8373515605926514, + "learning_rate": 2.047963456414161e-06, + "loss": 1.3643, + "step": 1859 + }, + { + "epoch": 1.9106317411402158, + "grad_norm": 2.825209379196167, + "learning_rate": 2.046060144651694e-06, + "loss": 1.3859, + "step": 1860 + }, + { + "epoch": 1.9106317411402158, + "eval_loss": 1.7664082050323486, + "eval_runtime": 37.1245, + "eval_samples_per_second": 26.936, + "eval_steps_per_second": 1.131, + "step": 1860 + }, + { + "epoch": 1.91165896250642, + "grad_norm": 2.7947168350219727, + "learning_rate": 2.0441568328892274e-06, + "loss": 1.4002, + "step": 1861 + }, + { + "epoch": 1.9126861838726246, + "grad_norm": 2.8314290046691895, + "learning_rate": 2.0422535211267608e-06, + "loss": 1.4017, + "step": 1862 + }, + { + "epoch": 1.913713405238829, + "grad_norm": 2.7938215732574463, + "learning_rate": 2.040350209364294e-06, + "loss": 1.379, + "step": 1863 + }, + { + "epoch": 1.9147406266050333, + "grad_norm": 2.7426047325134277, + "learning_rate": 2.0384468976018275e-06, + "loss": 1.3615, + "step": 1864 + }, + { + "epoch": 1.915767847971238, + "grad_norm": 2.916954278945923, + "learning_rate": 2.0365435858393604e-06, + "loss": 1.3756, + "step": 1865 + }, + { + "epoch": 1.9167950693374423, + "grad_norm": 2.7564570903778076, + "learning_rate": 2.034640274076894e-06, + "loss": 1.3716, + "step": 1866 + }, + { + "epoch": 1.9178222907036466, + "grad_norm": 2.911621332168579, + "learning_rate": 2.032736962314427e-06, + "loss": 1.4378, + "step": 1867 + }, + { + "epoch": 1.9188495120698512, + "grad_norm": 2.805112361907959, + "learning_rate": 2.0308336505519605e-06, + "loss": 1.3823, + "step": 1868 + }, + { + "epoch": 1.9198767334360554, + "grad_norm": 3.079866647720337, + "learning_rate": 2.028930338789494e-06, + "loss": 1.4388, + "step": 1869 + }, + { + "epoch": 1.92090395480226, + "grad_norm": 2.9153823852539062, + "learning_rate": 2.0270270270270273e-06, + "loss": 1.359, + "step": 1870 + }, + { + "epoch": 1.92090395480226, + "eval_loss": 1.7608962059020996, + "eval_runtime": 37.1265, + "eval_samples_per_second": 26.935, + "eval_steps_per_second": 1.131, + "step": 1870 + }, + { + "epoch": 1.9219311761684643, + "grad_norm": 2.8992526531219482, + "learning_rate": 2.0251237152645606e-06, + "loss": 1.3814, + "step": 1871 + }, + { + "epoch": 1.9229583975346687, + "grad_norm": 2.8838071823120117, + "learning_rate": 2.023220403502094e-06, + "loss": 1.3642, + "step": 1872 + }, + { + "epoch": 1.9239856189008733, + "grad_norm": 2.8734290599823, + "learning_rate": 2.0213170917396273e-06, + "loss": 1.3366, + "step": 1873 + }, + { + "epoch": 1.9250128402670774, + "grad_norm": 2.929142951965332, + "learning_rate": 2.0194137799771603e-06, + "loss": 1.3725, + "step": 1874 + }, + { + "epoch": 1.926040061633282, + "grad_norm": 2.984711170196533, + "learning_rate": 2.0175104682146937e-06, + "loss": 1.3825, + "step": 1875 + }, + { + "epoch": 1.9270672829994864, + "grad_norm": 2.916740894317627, + "learning_rate": 2.015607156452227e-06, + "loss": 1.3261, + "step": 1876 + }, + { + "epoch": 1.9280945043656907, + "grad_norm": 2.8688554763793945, + "learning_rate": 2.0137038446897604e-06, + "loss": 1.365, + "step": 1877 + }, + { + "epoch": 1.9291217257318953, + "grad_norm": 2.932525396347046, + "learning_rate": 2.0118005329272937e-06, + "loss": 1.3654, + "step": 1878 + }, + { + "epoch": 1.9301489470980995, + "grad_norm": 2.8874945640563965, + "learning_rate": 2.0098972211648267e-06, + "loss": 1.3745, + "step": 1879 + }, + { + "epoch": 1.931176168464304, + "grad_norm": 2.7608230113983154, + "learning_rate": 2.0079939094023605e-06, + "loss": 1.3718, + "step": 1880 + }, + { + "epoch": 1.931176168464304, + "eval_loss": 1.7662792205810547, + "eval_runtime": 37.1426, + "eval_samples_per_second": 26.923, + "eval_steps_per_second": 1.131, + "step": 1880 + }, + { + "epoch": 1.9322033898305084, + "grad_norm": 2.7790932655334473, + "learning_rate": 2.006090597639894e-06, + "loss": 1.4227, + "step": 1881 + }, + { + "epoch": 1.9332306111967128, + "grad_norm": 2.9143898487091064, + "learning_rate": 2.0041872858774268e-06, + "loss": 1.4396, + "step": 1882 + }, + { + "epoch": 1.9342578325629174, + "grad_norm": 2.744253158569336, + "learning_rate": 2.00228397411496e-06, + "loss": 1.4123, + "step": 1883 + }, + { + "epoch": 1.9352850539291218, + "grad_norm": 2.7352776527404785, + "learning_rate": 2.0003806623524935e-06, + "loss": 1.3936, + "step": 1884 + }, + { + "epoch": 1.9363122752953261, + "grad_norm": 2.7074668407440186, + "learning_rate": 1.998477350590027e-06, + "loss": 1.3201, + "step": 1885 + }, + { + "epoch": 1.9373394966615307, + "grad_norm": 2.848219871520996, + "learning_rate": 1.9965740388275602e-06, + "loss": 1.4048, + "step": 1886 + }, + { + "epoch": 1.9383667180277349, + "grad_norm": 2.7833354473114014, + "learning_rate": 1.994670727065093e-06, + "loss": 1.4007, + "step": 1887 + }, + { + "epoch": 1.9393939393939394, + "grad_norm": 2.83203387260437, + "learning_rate": 1.9927674153026265e-06, + "loss": 1.3471, + "step": 1888 + }, + { + "epoch": 1.9404211607601438, + "grad_norm": 2.99426007270813, + "learning_rate": 1.99086410354016e-06, + "loss": 1.407, + "step": 1889 + }, + { + "epoch": 1.9414483821263482, + "grad_norm": 2.8964831829071045, + "learning_rate": 1.9889607917776937e-06, + "loss": 1.3784, + "step": 1890 + }, + { + "epoch": 1.9414483821263482, + "eval_loss": 1.7859200239181519, + "eval_runtime": 37.16, + "eval_samples_per_second": 26.911, + "eval_steps_per_second": 1.13, + "step": 1890 + }, + { + "epoch": 1.9424756034925528, + "grad_norm": 2.938148021697998, + "learning_rate": 1.9870574800152266e-06, + "loss": 1.3932, + "step": 1891 + }, + { + "epoch": 1.943502824858757, + "grad_norm": 2.9096012115478516, + "learning_rate": 1.98515416825276e-06, + "loss": 1.4051, + "step": 1892 + }, + { + "epoch": 1.9445300462249615, + "grad_norm": 2.8964526653289795, + "learning_rate": 1.9832508564902933e-06, + "loss": 1.3961, + "step": 1893 + }, + { + "epoch": 1.9455572675911659, + "grad_norm": 2.681792974472046, + "learning_rate": 1.9813475447278267e-06, + "loss": 1.3175, + "step": 1894 + }, + { + "epoch": 1.9465844889573702, + "grad_norm": 2.8340206146240234, + "learning_rate": 1.97944423296536e-06, + "loss": 1.3955, + "step": 1895 + }, + { + "epoch": 1.9476117103235748, + "grad_norm": 2.9196391105651855, + "learning_rate": 1.977540921202893e-06, + "loss": 1.3833, + "step": 1896 + }, + { + "epoch": 1.9486389316897792, + "grad_norm": 2.7774484157562256, + "learning_rate": 1.9756376094404264e-06, + "loss": 1.3645, + "step": 1897 + }, + { + "epoch": 1.9496661530559836, + "grad_norm": 2.7222559452056885, + "learning_rate": 1.9737342976779597e-06, + "loss": 1.342, + "step": 1898 + }, + { + "epoch": 1.950693374422188, + "grad_norm": 2.816084623336792, + "learning_rate": 1.971830985915493e-06, + "loss": 1.4255, + "step": 1899 + }, + { + "epoch": 1.9517205957883923, + "grad_norm": 2.7530157566070557, + "learning_rate": 1.9699276741530264e-06, + "loss": 1.3792, + "step": 1900 + }, + { + "epoch": 1.9517205957883923, + "eval_loss": 1.7724789381027222, + "eval_runtime": 37.3499, + "eval_samples_per_second": 26.774, + "eval_steps_per_second": 1.125, + "step": 1900 + }, + { + "epoch": 1.9527478171545969, + "grad_norm": 2.7674543857574463, + "learning_rate": 1.96802436239056e-06, + "loss": 1.3888, + "step": 1901 + }, + { + "epoch": 1.9537750385208013, + "grad_norm": 2.9137115478515625, + "learning_rate": 1.966121050628093e-06, + "loss": 1.3656, + "step": 1902 + }, + { + "epoch": 1.9548022598870056, + "grad_norm": 2.8127288818359375, + "learning_rate": 1.9642177388656265e-06, + "loss": 1.4109, + "step": 1903 + }, + { + "epoch": 1.9558294812532102, + "grad_norm": 2.906191110610962, + "learning_rate": 1.9623144271031595e-06, + "loss": 1.4258, + "step": 1904 + }, + { + "epoch": 1.9568567026194144, + "grad_norm": 2.9624578952789307, + "learning_rate": 1.960411115340693e-06, + "loss": 1.4328, + "step": 1905 + }, + { + "epoch": 1.957883923985619, + "grad_norm": 2.8785276412963867, + "learning_rate": 1.958507803578226e-06, + "loss": 1.3912, + "step": 1906 + }, + { + "epoch": 1.9589111453518233, + "grad_norm": 2.848287343978882, + "learning_rate": 1.9566044918157596e-06, + "loss": 1.4052, + "step": 1907 + }, + { + "epoch": 1.9599383667180277, + "grad_norm": 2.766186237335205, + "learning_rate": 1.954701180053293e-06, + "loss": 1.4064, + "step": 1908 + }, + { + "epoch": 1.9609655880842323, + "grad_norm": 2.8182034492492676, + "learning_rate": 1.952797868290826e-06, + "loss": 1.3724, + "step": 1909 + }, + { + "epoch": 1.9619928094504364, + "grad_norm": 2.826488494873047, + "learning_rate": 1.9508945565283596e-06, + "loss": 1.421, + "step": 1910 + }, + { + "epoch": 1.9619928094504364, + "eval_loss": 1.780763864517212, + "eval_runtime": 36.9249, + "eval_samples_per_second": 27.082, + "eval_steps_per_second": 1.137, + "step": 1910 + }, + { + "epoch": 1.963020030816641, + "grad_norm": 2.876213312149048, + "learning_rate": 1.948991244765893e-06, + "loss": 1.3885, + "step": 1911 + }, + { + "epoch": 1.9640472521828454, + "grad_norm": 2.897611141204834, + "learning_rate": 1.9470879330034264e-06, + "loss": 1.4126, + "step": 1912 + }, + { + "epoch": 1.9650744735490497, + "grad_norm": 2.806605100631714, + "learning_rate": 1.9451846212409593e-06, + "loss": 1.3506, + "step": 1913 + }, + { + "epoch": 1.9661016949152543, + "grad_norm": 2.879645824432373, + "learning_rate": 1.9432813094784927e-06, + "loss": 1.409, + "step": 1914 + }, + { + "epoch": 1.9671289162814587, + "grad_norm": 2.871046304702759, + "learning_rate": 1.941377997716026e-06, + "loss": 1.3417, + "step": 1915 + }, + { + "epoch": 1.968156137647663, + "grad_norm": 2.8778395652770996, + "learning_rate": 1.9394746859535594e-06, + "loss": 1.3779, + "step": 1916 + }, + { + "epoch": 1.9691833590138677, + "grad_norm": 2.8736093044281006, + "learning_rate": 1.9375713741910928e-06, + "loss": 1.3355, + "step": 1917 + }, + { + "epoch": 1.9702105803800718, + "grad_norm": 2.822600841522217, + "learning_rate": 1.9356680624286257e-06, + "loss": 1.3779, + "step": 1918 + }, + { + "epoch": 1.9712378017462764, + "grad_norm": 2.779498338699341, + "learning_rate": 1.933764750666159e-06, + "loss": 1.3328, + "step": 1919 + }, + { + "epoch": 1.9722650231124808, + "grad_norm": 2.9568161964416504, + "learning_rate": 1.931861438903693e-06, + "loss": 1.3902, + "step": 1920 + }, + { + "epoch": 1.9722650231124808, + "eval_loss": 1.769534707069397, + "eval_runtime": 37.0243, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 1920 + }, + { + "epoch": 1.9732922444786851, + "grad_norm": 2.7917582988739014, + "learning_rate": 1.929958127141226e-06, + "loss": 1.3517, + "step": 1921 + }, + { + "epoch": 1.9743194658448897, + "grad_norm": 2.8624138832092285, + "learning_rate": 1.928054815378759e-06, + "loss": 1.3765, + "step": 1922 + }, + { + "epoch": 1.9753466872110939, + "grad_norm": 2.874619960784912, + "learning_rate": 1.9261515036162925e-06, + "loss": 1.4548, + "step": 1923 + }, + { + "epoch": 1.9763739085772984, + "grad_norm": 2.921558141708374, + "learning_rate": 1.924248191853826e-06, + "loss": 1.4122, + "step": 1924 + }, + { + "epoch": 1.9774011299435028, + "grad_norm": 2.853748321533203, + "learning_rate": 1.9223448800913592e-06, + "loss": 1.3909, + "step": 1925 + }, + { + "epoch": 1.9784283513097072, + "grad_norm": 2.828735828399658, + "learning_rate": 1.9204415683288926e-06, + "loss": 1.3748, + "step": 1926 + }, + { + "epoch": 1.9794555726759118, + "grad_norm": 2.8109824657440186, + "learning_rate": 1.9185382565664255e-06, + "loss": 1.4051, + "step": 1927 + }, + { + "epoch": 1.980482794042116, + "grad_norm": 3.035722494125366, + "learning_rate": 1.916634944803959e-06, + "loss": 1.42, + "step": 1928 + }, + { + "epoch": 1.9815100154083205, + "grad_norm": 2.9289391040802, + "learning_rate": 1.9147316330414923e-06, + "loss": 1.3965, + "step": 1929 + }, + { + "epoch": 1.9825372367745249, + "grad_norm": 2.853398561477661, + "learning_rate": 1.9128283212790256e-06, + "loss": 1.3869, + "step": 1930 + }, + { + "epoch": 1.9825372367745249, + "eval_loss": 1.7659740447998047, + "eval_runtime": 37.0479, + "eval_samples_per_second": 26.992, + "eval_steps_per_second": 1.134, + "step": 1930 + }, + { + "epoch": 1.9835644581407292, + "grad_norm": 2.9305245876312256, + "learning_rate": 1.910925009516559e-06, + "loss": 1.4265, + "step": 1931 + }, + { + "epoch": 1.9845916795069338, + "grad_norm": 2.8339171409606934, + "learning_rate": 1.9090216977540924e-06, + "loss": 1.3934, + "step": 1932 + }, + { + "epoch": 1.9856189008731382, + "grad_norm": 2.929807424545288, + "learning_rate": 1.9071183859916257e-06, + "loss": 1.4339, + "step": 1933 + }, + { + "epoch": 1.9866461222393426, + "grad_norm": 2.9070687294006348, + "learning_rate": 1.9052150742291589e-06, + "loss": 1.4004, + "step": 1934 + }, + { + "epoch": 1.9876733436055471, + "grad_norm": 2.9756429195404053, + "learning_rate": 1.9033117624666922e-06, + "loss": 1.404, + "step": 1935 + }, + { + "epoch": 1.9887005649717513, + "grad_norm": 2.831841468811035, + "learning_rate": 1.9014084507042254e-06, + "loss": 1.3668, + "step": 1936 + }, + { + "epoch": 1.9897277863379559, + "grad_norm": 2.8649723529815674, + "learning_rate": 1.8995051389417587e-06, + "loss": 1.3625, + "step": 1937 + }, + { + "epoch": 1.9907550077041603, + "grad_norm": 2.7570254802703857, + "learning_rate": 1.8976018271792921e-06, + "loss": 1.3427, + "step": 1938 + }, + { + "epoch": 1.9917822290703646, + "grad_norm": 2.9655392169952393, + "learning_rate": 1.8956985154168253e-06, + "loss": 1.4039, + "step": 1939 + }, + { + "epoch": 1.9928094504365692, + "grad_norm": 2.840308427810669, + "learning_rate": 1.8937952036543588e-06, + "loss": 1.3678, + "step": 1940 + }, + { + "epoch": 1.9928094504365692, + "eval_loss": 1.7668676376342773, + "eval_runtime": 37.0279, + "eval_samples_per_second": 27.007, + "eval_steps_per_second": 1.134, + "step": 1940 + }, + { + "epoch": 1.9938366718027734, + "grad_norm": 2.7608296871185303, + "learning_rate": 1.8918918918918922e-06, + "loss": 1.3353, + "step": 1941 + }, + { + "epoch": 1.994863893168978, + "grad_norm": 2.861088275909424, + "learning_rate": 1.8899885801294253e-06, + "loss": 1.3866, + "step": 1942 + }, + { + "epoch": 1.9958911145351823, + "grad_norm": 2.8978993892669678, + "learning_rate": 1.8880852683669587e-06, + "loss": 1.4054, + "step": 1943 + }, + { + "epoch": 1.9969183359013867, + "grad_norm": 2.850844144821167, + "learning_rate": 1.886181956604492e-06, + "loss": 1.3692, + "step": 1944 + }, + { + "epoch": 1.9979455572675913, + "grad_norm": 2.8984055519104004, + "learning_rate": 1.8842786448420252e-06, + "loss": 1.3681, + "step": 1945 + }, + { + "epoch": 1.9989727786337956, + "grad_norm": 2.8746683597564697, + "learning_rate": 1.8823753330795586e-06, + "loss": 1.4047, + "step": 1946 + }, + { + "epoch": 2.0, + "grad_norm": 2.7244315147399902, + "learning_rate": 1.8804720213170917e-06, + "loss": 1.3003, + "step": 1947 + }, + { + "epoch": 2.0010272213662046, + "grad_norm": 3.10217022895813, + "learning_rate": 1.878568709554625e-06, + "loss": 1.3375, + "step": 1948 + }, + { + "epoch": 2.0020544427324087, + "grad_norm": 3.1777870655059814, + "learning_rate": 1.8766653977921585e-06, + "loss": 1.3775, + "step": 1949 + }, + { + "epoch": 2.0030816640986133, + "grad_norm": 3.008085250854492, + "learning_rate": 1.874762086029692e-06, + "loss": 1.3435, + "step": 1950 + }, + { + "epoch": 2.0030816640986133, + "eval_loss": 1.7830486297607422, + "eval_runtime": 37.0192, + "eval_samples_per_second": 27.013, + "eval_steps_per_second": 1.135, + "step": 1950 + }, + { + "epoch": 2.0041088854648175, + "grad_norm": 3.228102445602417, + "learning_rate": 1.8728587742672252e-06, + "loss": 1.3697, + "step": 1951 + }, + { + "epoch": 2.005136106831022, + "grad_norm": 3.392749071121216, + "learning_rate": 1.8709554625047585e-06, + "loss": 1.3813, + "step": 1952 + }, + { + "epoch": 2.0061633281972266, + "grad_norm": 3.264998197555542, + "learning_rate": 1.8690521507422917e-06, + "loss": 1.3203, + "step": 1953 + }, + { + "epoch": 2.007190549563431, + "grad_norm": 3.344672679901123, + "learning_rate": 1.867148838979825e-06, + "loss": 1.3413, + "step": 1954 + }, + { + "epoch": 2.0082177709296354, + "grad_norm": 3.138148307800293, + "learning_rate": 1.8652455272173584e-06, + "loss": 1.3606, + "step": 1955 + }, + { + "epoch": 2.00924499229584, + "grad_norm": 3.1569597721099854, + "learning_rate": 1.8633422154548916e-06, + "loss": 1.3846, + "step": 1956 + }, + { + "epoch": 2.010272213662044, + "grad_norm": 3.0843050479888916, + "learning_rate": 1.861438903692425e-06, + "loss": 1.282, + "step": 1957 + }, + { + "epoch": 2.0112994350282487, + "grad_norm": 3.3988428115844727, + "learning_rate": 1.859535591929958e-06, + "loss": 1.3183, + "step": 1958 + }, + { + "epoch": 2.012326656394453, + "grad_norm": 3.3206543922424316, + "learning_rate": 1.8576322801674915e-06, + "loss": 1.3874, + "step": 1959 + }, + { + "epoch": 2.0133538777606574, + "grad_norm": 3.1254804134368896, + "learning_rate": 1.855728968405025e-06, + "loss": 1.3335, + "step": 1960 + }, + { + "epoch": 2.0133538777606574, + "eval_loss": 1.780193567276001, + "eval_runtime": 37.0457, + "eval_samples_per_second": 26.994, + "eval_steps_per_second": 1.134, + "step": 1960 + }, + { + "epoch": 2.014381099126862, + "grad_norm": 3.2736592292785645, + "learning_rate": 1.8538256566425584e-06, + "loss": 1.3489, + "step": 1961 + }, + { + "epoch": 2.015408320493066, + "grad_norm": 3.1620168685913086, + "learning_rate": 1.8519223448800915e-06, + "loss": 1.3898, + "step": 1962 + }, + { + "epoch": 2.0164355418592708, + "grad_norm": 3.3050506114959717, + "learning_rate": 1.850019033117625e-06, + "loss": 1.3292, + "step": 1963 + }, + { + "epoch": 2.017462763225475, + "grad_norm": 3.192668914794922, + "learning_rate": 1.848115721355158e-06, + "loss": 1.3319, + "step": 1964 + }, + { + "epoch": 2.0184899845916795, + "grad_norm": 3.1518166065216064, + "learning_rate": 1.8462124095926914e-06, + "loss": 1.3433, + "step": 1965 + }, + { + "epoch": 2.019517205957884, + "grad_norm": 3.365323305130005, + "learning_rate": 1.8443090978302248e-06, + "loss": 1.362, + "step": 1966 + }, + { + "epoch": 2.0205444273240882, + "grad_norm": 3.4271507263183594, + "learning_rate": 1.842405786067758e-06, + "loss": 1.3867, + "step": 1967 + }, + { + "epoch": 2.021571648690293, + "grad_norm": 3.167879581451416, + "learning_rate": 1.8405024743052913e-06, + "loss": 1.3341, + "step": 1968 + }, + { + "epoch": 2.022598870056497, + "grad_norm": 3.1220595836639404, + "learning_rate": 1.8385991625428247e-06, + "loss": 1.379, + "step": 1969 + }, + { + "epoch": 2.0236260914227016, + "grad_norm": 3.03648042678833, + "learning_rate": 1.8366958507803578e-06, + "loss": 1.3095, + "step": 1970 + }, + { + "epoch": 2.0236260914227016, + "eval_loss": 1.7782403230667114, + "eval_runtime": 37.0535, + "eval_samples_per_second": 26.988, + "eval_steps_per_second": 1.133, + "step": 1970 + }, + { + "epoch": 2.024653312788906, + "grad_norm": 3.254167318344116, + "learning_rate": 1.8347925390178914e-06, + "loss": 1.3625, + "step": 1971 + }, + { + "epoch": 2.0256805341551103, + "grad_norm": 3.0616867542266846, + "learning_rate": 1.8328892272554247e-06, + "loss": 1.3059, + "step": 1972 + }, + { + "epoch": 2.026707755521315, + "grad_norm": 3.236483097076416, + "learning_rate": 1.8309859154929579e-06, + "loss": 1.3737, + "step": 1973 + }, + { + "epoch": 2.0277349768875195, + "grad_norm": 3.1941466331481934, + "learning_rate": 1.8290826037304913e-06, + "loss": 1.35, + "step": 1974 + }, + { + "epoch": 2.0287621982537236, + "grad_norm": 3.0594732761383057, + "learning_rate": 1.8271792919680246e-06, + "loss": 1.3261, + "step": 1975 + }, + { + "epoch": 2.029789419619928, + "grad_norm": 3.157635450363159, + "learning_rate": 1.8252759802055578e-06, + "loss": 1.3224, + "step": 1976 + }, + { + "epoch": 2.0308166409861323, + "grad_norm": 3.380322217941284, + "learning_rate": 1.8233726684430911e-06, + "loss": 1.3742, + "step": 1977 + }, + { + "epoch": 2.031843862352337, + "grad_norm": 3.397022247314453, + "learning_rate": 1.8214693566806243e-06, + "loss": 1.324, + "step": 1978 + }, + { + "epoch": 2.0328710837185415, + "grad_norm": 3.187364101409912, + "learning_rate": 1.8195660449181576e-06, + "loss": 1.3435, + "step": 1979 + }, + { + "epoch": 2.0338983050847457, + "grad_norm": 3.1330549716949463, + "learning_rate": 1.817662733155691e-06, + "loss": 1.3472, + "step": 1980 + }, + { + "epoch": 2.0338983050847457, + "eval_loss": 1.7825566530227661, + "eval_runtime": 37.0315, + "eval_samples_per_second": 27.004, + "eval_steps_per_second": 1.134, + "step": 1980 + }, + { + "epoch": 2.0349255264509503, + "grad_norm": 3.1697194576263428, + "learning_rate": 1.8157594213932244e-06, + "loss": 1.347, + "step": 1981 + }, + { + "epoch": 2.0359527478171544, + "grad_norm": 3.1682229042053223, + "learning_rate": 1.8138561096307577e-06, + "loss": 1.3588, + "step": 1982 + }, + { + "epoch": 2.036979969183359, + "grad_norm": 3.1456754207611084, + "learning_rate": 1.811952797868291e-06, + "loss": 1.2981, + "step": 1983 + }, + { + "epoch": 2.0380071905495636, + "grad_norm": 3.1854794025421143, + "learning_rate": 1.8100494861058242e-06, + "loss": 1.3429, + "step": 1984 + }, + { + "epoch": 2.0390344119157677, + "grad_norm": 3.2500100135803223, + "learning_rate": 1.8081461743433576e-06, + "loss": 1.3572, + "step": 1985 + }, + { + "epoch": 2.0400616332819723, + "grad_norm": 3.0852582454681396, + "learning_rate": 1.806242862580891e-06, + "loss": 1.3081, + "step": 1986 + }, + { + "epoch": 2.0410888546481765, + "grad_norm": 3.3680732250213623, + "learning_rate": 1.8043395508184241e-06, + "loss": 1.3584, + "step": 1987 + }, + { + "epoch": 2.042116076014381, + "grad_norm": 3.1346354484558105, + "learning_rate": 1.8024362390559575e-06, + "loss": 1.3437, + "step": 1988 + }, + { + "epoch": 2.0431432973805856, + "grad_norm": 3.1831793785095215, + "learning_rate": 1.8005329272934906e-06, + "loss": 1.3206, + "step": 1989 + }, + { + "epoch": 2.04417051874679, + "grad_norm": 3.3129804134368896, + "learning_rate": 1.798629615531024e-06, + "loss": 1.3992, + "step": 1990 + }, + { + "epoch": 2.04417051874679, + "eval_loss": 1.791827917098999, + "eval_runtime": 37.1302, + "eval_samples_per_second": 26.932, + "eval_steps_per_second": 1.131, + "step": 1990 + }, + { + "epoch": 2.0451977401129944, + "grad_norm": 3.2966983318328857, + "learning_rate": 1.7967263037685576e-06, + "loss": 1.3374, + "step": 1991 + }, + { + "epoch": 2.046224961479199, + "grad_norm": 3.2134904861450195, + "learning_rate": 1.794822992006091e-06, + "loss": 1.3304, + "step": 1992 + }, + { + "epoch": 2.047252182845403, + "grad_norm": 3.3718373775482178, + "learning_rate": 1.792919680243624e-06, + "loss": 1.3898, + "step": 1993 + }, + { + "epoch": 2.0482794042116077, + "grad_norm": 3.1467645168304443, + "learning_rate": 1.7910163684811574e-06, + "loss": 1.3089, + "step": 1994 + }, + { + "epoch": 2.049306625577812, + "grad_norm": 3.1504383087158203, + "learning_rate": 1.7891130567186906e-06, + "loss": 1.3398, + "step": 1995 + }, + { + "epoch": 2.0503338469440164, + "grad_norm": 3.2479350566864014, + "learning_rate": 1.787209744956224e-06, + "loss": 1.3099, + "step": 1996 + }, + { + "epoch": 2.051361068310221, + "grad_norm": 3.311795234680176, + "learning_rate": 1.7853064331937573e-06, + "loss": 1.3369, + "step": 1997 + }, + { + "epoch": 2.052388289676425, + "grad_norm": 3.1674633026123047, + "learning_rate": 1.7834031214312905e-06, + "loss": 1.3732, + "step": 1998 + }, + { + "epoch": 2.0534155110426298, + "grad_norm": 3.1175575256347656, + "learning_rate": 1.7814998096688238e-06, + "loss": 1.3159, + "step": 1999 + }, + { + "epoch": 2.054442732408834, + "grad_norm": 3.216402530670166, + "learning_rate": 1.779596497906357e-06, + "loss": 1.3479, + "step": 2000 + }, + { + "epoch": 2.054442732408834, + "eval_loss": 1.7845674753189087, + "eval_runtime": 37.2755, + "eval_samples_per_second": 26.827, + "eval_steps_per_second": 1.127, + "step": 2000 + }, + { + "epoch": 2.0554699537750385, + "grad_norm": 3.2530016899108887, + "learning_rate": 1.7776931861438906e-06, + "loss": 1.3924, + "step": 2001 + }, + { + "epoch": 2.056497175141243, + "grad_norm": 3.222146511077881, + "learning_rate": 1.775789874381424e-06, + "loss": 1.3826, + "step": 2002 + }, + { + "epoch": 2.0575243965074472, + "grad_norm": 3.3821890354156494, + "learning_rate": 1.7738865626189573e-06, + "loss": 1.3664, + "step": 2003 + }, + { + "epoch": 2.058551617873652, + "grad_norm": 3.2899715900421143, + "learning_rate": 1.7719832508564904e-06, + "loss": 1.3, + "step": 2004 + }, + { + "epoch": 2.0595788392398564, + "grad_norm": 3.2912356853485107, + "learning_rate": 1.7700799390940238e-06, + "loss": 1.3297, + "step": 2005 + }, + { + "epoch": 2.0606060606060606, + "grad_norm": 3.397883892059326, + "learning_rate": 1.768176627331557e-06, + "loss": 1.3622, + "step": 2006 + }, + { + "epoch": 2.061633281972265, + "grad_norm": 3.230801820755005, + "learning_rate": 1.7662733155690903e-06, + "loss": 1.353, + "step": 2007 + }, + { + "epoch": 2.0626605033384693, + "grad_norm": 3.2354795932769775, + "learning_rate": 1.7643700038066237e-06, + "loss": 1.3777, + "step": 2008 + }, + { + "epoch": 2.063687724704674, + "grad_norm": 3.086163282394409, + "learning_rate": 1.7624666920441568e-06, + "loss": 1.3087, + "step": 2009 + }, + { + "epoch": 2.0647149460708785, + "grad_norm": 3.1875600814819336, + "learning_rate": 1.7605633802816902e-06, + "loss": 1.3045, + "step": 2010 + }, + { + "epoch": 2.0647149460708785, + "eval_loss": 1.7808178663253784, + "eval_runtime": 36.9437, + "eval_samples_per_second": 27.068, + "eval_steps_per_second": 1.137, + "step": 2010 + }, + { + "epoch": 2.0657421674370826, + "grad_norm": 3.33947491645813, + "learning_rate": 1.7586600685192238e-06, + "loss": 1.3992, + "step": 2011 + }, + { + "epoch": 2.066769388803287, + "grad_norm": 3.314852476119995, + "learning_rate": 1.756756756756757e-06, + "loss": 1.4015, + "step": 2012 + }, + { + "epoch": 2.0677966101694913, + "grad_norm": 3.1022720336914062, + "learning_rate": 1.7548534449942903e-06, + "loss": 1.3664, + "step": 2013 + }, + { + "epoch": 2.068823831535696, + "grad_norm": 3.2397265434265137, + "learning_rate": 1.7529501332318236e-06, + "loss": 1.3429, + "step": 2014 + }, + { + "epoch": 2.0698510529019005, + "grad_norm": 3.1528069972991943, + "learning_rate": 1.7510468214693568e-06, + "loss": 1.3176, + "step": 2015 + }, + { + "epoch": 2.0708782742681047, + "grad_norm": 3.157923460006714, + "learning_rate": 1.7491435097068902e-06, + "loss": 1.3156, + "step": 2016 + }, + { + "epoch": 2.0719054956343093, + "grad_norm": 3.094817876815796, + "learning_rate": 1.7472401979444233e-06, + "loss": 1.3552, + "step": 2017 + }, + { + "epoch": 2.0729327170005134, + "grad_norm": 3.0657732486724854, + "learning_rate": 1.7453368861819567e-06, + "loss": 1.2914, + "step": 2018 + }, + { + "epoch": 2.073959938366718, + "grad_norm": 3.0663399696350098, + "learning_rate": 1.74343357441949e-06, + "loss": 1.344, + "step": 2019 + }, + { + "epoch": 2.0749871597329226, + "grad_norm": 3.1718454360961914, + "learning_rate": 1.7415302626570232e-06, + "loss": 1.3987, + "step": 2020 + }, + { + "epoch": 2.0749871597329226, + "eval_loss": 1.7751562595367432, + "eval_runtime": 37.0572, + "eval_samples_per_second": 26.985, + "eval_steps_per_second": 1.133, + "step": 2020 + }, + { + "epoch": 2.0760143810991267, + "grad_norm": 3.0701513290405273, + "learning_rate": 1.7396269508945568e-06, + "loss": 1.3327, + "step": 2021 + }, + { + "epoch": 2.0770416024653313, + "grad_norm": 3.1184256076812744, + "learning_rate": 1.7377236391320901e-06, + "loss": 1.3111, + "step": 2022 + }, + { + "epoch": 2.078068823831536, + "grad_norm": 3.2951152324676514, + "learning_rate": 1.7358203273696233e-06, + "loss": 1.3774, + "step": 2023 + }, + { + "epoch": 2.07909604519774, + "grad_norm": 3.2582433223724365, + "learning_rate": 1.7339170156071566e-06, + "loss": 1.3332, + "step": 2024 + }, + { + "epoch": 2.0801232665639446, + "grad_norm": 3.1642305850982666, + "learning_rate": 1.73201370384469e-06, + "loss": 1.3294, + "step": 2025 + }, + { + "epoch": 2.081150487930149, + "grad_norm": 3.2165608406066895, + "learning_rate": 1.7301103920822231e-06, + "loss": 1.344, + "step": 2026 + }, + { + "epoch": 2.0821777092963534, + "grad_norm": 3.247485637664795, + "learning_rate": 1.7282070803197565e-06, + "loss": 1.3496, + "step": 2027 + }, + { + "epoch": 2.083204930662558, + "grad_norm": 3.192835569381714, + "learning_rate": 1.7263037685572897e-06, + "loss": 1.3345, + "step": 2028 + }, + { + "epoch": 2.084232152028762, + "grad_norm": 3.2298715114593506, + "learning_rate": 1.724400456794823e-06, + "loss": 1.3666, + "step": 2029 + }, + { + "epoch": 2.0852593733949667, + "grad_norm": 3.091574192047119, + "learning_rate": 1.7224971450323564e-06, + "loss": 1.3321, + "step": 2030 + }, + { + "epoch": 2.0852593733949667, + "eval_loss": 1.7775636911392212, + "eval_runtime": 37.0648, + "eval_samples_per_second": 26.98, + "eval_steps_per_second": 1.133, + "step": 2030 + }, + { + "epoch": 2.086286594761171, + "grad_norm": 3.235821008682251, + "learning_rate": 1.72059383326989e-06, + "loss": 1.3372, + "step": 2031 + }, + { + "epoch": 2.0873138161273754, + "grad_norm": 3.333073377609253, + "learning_rate": 1.7186905215074231e-06, + "loss": 1.3926, + "step": 2032 + }, + { + "epoch": 2.08834103749358, + "grad_norm": 3.235518217086792, + "learning_rate": 1.7167872097449565e-06, + "loss": 1.3179, + "step": 2033 + }, + { + "epoch": 2.089368258859784, + "grad_norm": 3.1183924674987793, + "learning_rate": 1.7148838979824896e-06, + "loss": 1.3402, + "step": 2034 + }, + { + "epoch": 2.0903954802259888, + "grad_norm": 3.245620012283325, + "learning_rate": 1.712980586220023e-06, + "loss": 1.3351, + "step": 2035 + }, + { + "epoch": 2.091422701592193, + "grad_norm": 3.413208246231079, + "learning_rate": 1.7110772744575563e-06, + "loss": 1.3199, + "step": 2036 + }, + { + "epoch": 2.0924499229583975, + "grad_norm": 3.183093786239624, + "learning_rate": 1.7091739626950895e-06, + "loss": 1.3295, + "step": 2037 + }, + { + "epoch": 2.093477144324602, + "grad_norm": 3.2691562175750732, + "learning_rate": 1.7072706509326229e-06, + "loss": 1.3623, + "step": 2038 + }, + { + "epoch": 2.0945043656908062, + "grad_norm": 3.2356081008911133, + "learning_rate": 1.7053673391701562e-06, + "loss": 1.3547, + "step": 2039 + }, + { + "epoch": 2.095531587057011, + "grad_norm": 3.138723134994507, + "learning_rate": 1.7034640274076894e-06, + "loss": 1.3189, + "step": 2040 + }, + { + "epoch": 2.095531587057011, + "eval_loss": 1.7775871753692627, + "eval_runtime": 37.0654, + "eval_samples_per_second": 26.979, + "eval_steps_per_second": 1.133, + "step": 2040 + }, + { + "epoch": 2.0965588084232154, + "grad_norm": 3.2420573234558105, + "learning_rate": 1.701560715645223e-06, + "loss": 1.3862, + "step": 2041 + }, + { + "epoch": 2.0975860297894195, + "grad_norm": 3.137073278427124, + "learning_rate": 1.6996574038827563e-06, + "loss": 1.3076, + "step": 2042 + }, + { + "epoch": 2.098613251155624, + "grad_norm": 3.122952461242676, + "learning_rate": 1.6977540921202895e-06, + "loss": 1.3493, + "step": 2043 + }, + { + "epoch": 2.0996404725218283, + "grad_norm": 3.1556475162506104, + "learning_rate": 1.6958507803578228e-06, + "loss": 1.3337, + "step": 2044 + }, + { + "epoch": 2.100667693888033, + "grad_norm": 3.2341697216033936, + "learning_rate": 1.693947468595356e-06, + "loss": 1.3575, + "step": 2045 + }, + { + "epoch": 2.1016949152542375, + "grad_norm": 3.411771774291992, + "learning_rate": 1.6920441568328893e-06, + "loss": 1.3754, + "step": 2046 + }, + { + "epoch": 2.1027221366204416, + "grad_norm": 3.0538063049316406, + "learning_rate": 1.6901408450704227e-06, + "loss": 1.2938, + "step": 2047 + }, + { + "epoch": 2.103749357986646, + "grad_norm": 3.2879412174224854, + "learning_rate": 1.6882375333079559e-06, + "loss": 1.3313, + "step": 2048 + }, + { + "epoch": 2.1047765793528503, + "grad_norm": 3.1848442554473877, + "learning_rate": 1.6863342215454892e-06, + "loss": 1.3835, + "step": 2049 + }, + { + "epoch": 2.105803800719055, + "grad_norm": 3.1774306297302246, + "learning_rate": 1.6844309097830226e-06, + "loss": 1.3741, + "step": 2050 + }, + { + "epoch": 2.105803800719055, + "eval_loss": 1.77629554271698, + "eval_runtime": 37.0728, + "eval_samples_per_second": 26.974, + "eval_steps_per_second": 1.133, + "step": 2050 + }, + { + "epoch": 2.1068310220852595, + "grad_norm": 3.218514919281006, + "learning_rate": 1.682527598020556e-06, + "loss": 1.3592, + "step": 2051 + }, + { + "epoch": 2.1078582434514637, + "grad_norm": 3.2372989654541016, + "learning_rate": 1.6806242862580893e-06, + "loss": 1.376, + "step": 2052 + }, + { + "epoch": 2.1088854648176683, + "grad_norm": 3.115527629852295, + "learning_rate": 1.6787209744956227e-06, + "loss": 1.3192, + "step": 2053 + }, + { + "epoch": 2.109912686183873, + "grad_norm": 3.06538987159729, + "learning_rate": 1.6768176627331558e-06, + "loss": 1.303, + "step": 2054 + }, + { + "epoch": 2.110939907550077, + "grad_norm": 3.075857162475586, + "learning_rate": 1.6749143509706892e-06, + "loss": 1.3504, + "step": 2055 + }, + { + "epoch": 2.1119671289162816, + "grad_norm": 3.2284908294677734, + "learning_rate": 1.6730110392082225e-06, + "loss": 1.3284, + "step": 2056 + }, + { + "epoch": 2.1129943502824857, + "grad_norm": 3.3034918308258057, + "learning_rate": 1.6711077274457557e-06, + "loss": 1.349, + "step": 2057 + }, + { + "epoch": 2.1140215716486903, + "grad_norm": 3.1239676475524902, + "learning_rate": 1.669204415683289e-06, + "loss": 1.2957, + "step": 2058 + }, + { + "epoch": 2.115048793014895, + "grad_norm": 3.4618728160858154, + "learning_rate": 1.6673011039208222e-06, + "loss": 1.3626, + "step": 2059 + }, + { + "epoch": 2.116076014381099, + "grad_norm": 3.4596455097198486, + "learning_rate": 1.6653977921583556e-06, + "loss": 1.3436, + "step": 2060 + }, + { + "epoch": 2.116076014381099, + "eval_loss": 1.7774609327316284, + "eval_runtime": 37.12, + "eval_samples_per_second": 26.94, + "eval_steps_per_second": 1.131, + "step": 2060 + }, + { + "epoch": 2.1171032357473036, + "grad_norm": 3.6342878341674805, + "learning_rate": 1.663494480395889e-06, + "loss": 1.3618, + "step": 2061 + }, + { + "epoch": 2.118130457113508, + "grad_norm": 3.329286575317383, + "learning_rate": 1.6615911686334225e-06, + "loss": 1.3187, + "step": 2062 + }, + { + "epoch": 2.1191576784797124, + "grad_norm": 3.2518863677978516, + "learning_rate": 1.6596878568709557e-06, + "loss": 1.3412, + "step": 2063 + }, + { + "epoch": 2.120184899845917, + "grad_norm": 3.1159286499023438, + "learning_rate": 1.657784545108489e-06, + "loss": 1.331, + "step": 2064 + }, + { + "epoch": 2.121212121212121, + "grad_norm": 3.352390766143799, + "learning_rate": 1.6558812333460222e-06, + "loss": 1.3668, + "step": 2065 + }, + { + "epoch": 2.1222393425783257, + "grad_norm": 3.1647236347198486, + "learning_rate": 1.6539779215835555e-06, + "loss": 1.3574, + "step": 2066 + }, + { + "epoch": 2.12326656394453, + "grad_norm": 3.268143653869629, + "learning_rate": 1.652074609821089e-06, + "loss": 1.3247, + "step": 2067 + }, + { + "epoch": 2.1242937853107344, + "grad_norm": 3.081690788269043, + "learning_rate": 1.650171298058622e-06, + "loss": 1.3269, + "step": 2068 + }, + { + "epoch": 2.125321006676939, + "grad_norm": 3.1442677974700928, + "learning_rate": 1.6482679862961554e-06, + "loss": 1.329, + "step": 2069 + }, + { + "epoch": 2.126348228043143, + "grad_norm": 3.434091091156006, + "learning_rate": 1.6463646745336886e-06, + "loss": 1.3249, + "step": 2070 + }, + { + "epoch": 2.126348228043143, + "eval_loss": 1.7745627164840698, + "eval_runtime": 37.115, + "eval_samples_per_second": 26.943, + "eval_steps_per_second": 1.132, + "step": 2070 + }, + { + "epoch": 2.1273754494093478, + "grad_norm": 3.312300443649292, + "learning_rate": 1.644461362771222e-06, + "loss": 1.3091, + "step": 2071 + }, + { + "epoch": 2.128402670775552, + "grad_norm": 3.319504976272583, + "learning_rate": 1.6425580510087555e-06, + "loss": 1.3439, + "step": 2072 + }, + { + "epoch": 2.1294298921417565, + "grad_norm": 3.2242252826690674, + "learning_rate": 1.6406547392462889e-06, + "loss": 1.3068, + "step": 2073 + }, + { + "epoch": 2.130457113507961, + "grad_norm": 3.1730830669403076, + "learning_rate": 1.638751427483822e-06, + "loss": 1.3079, + "step": 2074 + }, + { + "epoch": 2.131484334874165, + "grad_norm": 3.2944724559783936, + "learning_rate": 1.6368481157213554e-06, + "loss": 1.3502, + "step": 2075 + }, + { + "epoch": 2.13251155624037, + "grad_norm": 3.1474711894989014, + "learning_rate": 1.6349448039588885e-06, + "loss": 1.3308, + "step": 2076 + }, + { + "epoch": 2.1335387776065744, + "grad_norm": 3.291567802429199, + "learning_rate": 1.6330414921964219e-06, + "loss": 1.3834, + "step": 2077 + }, + { + "epoch": 2.1345659989727785, + "grad_norm": 3.2176096439361572, + "learning_rate": 1.6311381804339553e-06, + "loss": 1.3134, + "step": 2078 + }, + { + "epoch": 2.135593220338983, + "grad_norm": 3.2517168521881104, + "learning_rate": 1.6292348686714884e-06, + "loss": 1.357, + "step": 2079 + }, + { + "epoch": 2.1366204417051873, + "grad_norm": 3.1429731845855713, + "learning_rate": 1.6273315569090218e-06, + "loss": 1.3327, + "step": 2080 + }, + { + "epoch": 2.1366204417051873, + "eval_loss": 1.7769756317138672, + "eval_runtime": 37.1282, + "eval_samples_per_second": 26.934, + "eval_steps_per_second": 1.131, + "step": 2080 + }, + { + "epoch": 2.137647663071392, + "grad_norm": 3.3201053142547607, + "learning_rate": 1.625428245146555e-06, + "loss": 1.4052, + "step": 2081 + }, + { + "epoch": 2.1386748844375965, + "grad_norm": 3.2188704013824463, + "learning_rate": 1.6235249333840885e-06, + "loss": 1.3829, + "step": 2082 + }, + { + "epoch": 2.1397021058038006, + "grad_norm": 3.2554514408111572, + "learning_rate": 1.6216216216216219e-06, + "loss": 1.3356, + "step": 2083 + }, + { + "epoch": 2.140729327170005, + "grad_norm": 3.286280870437622, + "learning_rate": 1.6197183098591552e-06, + "loss": 1.3627, + "step": 2084 + }, + { + "epoch": 2.1417565485362093, + "grad_norm": 3.4556727409362793, + "learning_rate": 1.6178149980966884e-06, + "loss": 1.35, + "step": 2085 + }, + { + "epoch": 2.142783769902414, + "grad_norm": 3.186457395553589, + "learning_rate": 1.6159116863342217e-06, + "loss": 1.3218, + "step": 2086 + }, + { + "epoch": 2.1438109912686185, + "grad_norm": 3.360995054244995, + "learning_rate": 1.6140083745717549e-06, + "loss": 1.3895, + "step": 2087 + }, + { + "epoch": 2.1448382126348227, + "grad_norm": 3.1155898571014404, + "learning_rate": 1.6121050628092882e-06, + "loss": 1.3113, + "step": 2088 + }, + { + "epoch": 2.1458654340010272, + "grad_norm": 3.300079584121704, + "learning_rate": 1.6102017510468216e-06, + "loss": 1.379, + "step": 2089 + }, + { + "epoch": 2.146892655367232, + "grad_norm": 3.1357364654541016, + "learning_rate": 1.6082984392843548e-06, + "loss": 1.2912, + "step": 2090 + }, + { + "epoch": 2.146892655367232, + "eval_loss": 1.7777626514434814, + "eval_runtime": 37.1337, + "eval_samples_per_second": 26.93, + "eval_steps_per_second": 1.131, + "step": 2090 + }, + { + "epoch": 2.147919876733436, + "grad_norm": 3.267508029937744, + "learning_rate": 1.6063951275218881e-06, + "loss": 1.3489, + "step": 2091 + }, + { + "epoch": 2.1489470980996406, + "grad_norm": 3.3493881225585938, + "learning_rate": 1.6044918157594217e-06, + "loss": 1.395, + "step": 2092 + }, + { + "epoch": 2.1499743194658447, + "grad_norm": 3.3308827877044678, + "learning_rate": 1.6025885039969548e-06, + "loss": 1.3225, + "step": 2093 + }, + { + "epoch": 2.1510015408320493, + "grad_norm": 3.253615617752075, + "learning_rate": 1.6006851922344882e-06, + "loss": 1.3537, + "step": 2094 + }, + { + "epoch": 2.152028762198254, + "grad_norm": 3.2310163974761963, + "learning_rate": 1.5987818804720216e-06, + "loss": 1.3581, + "step": 2095 + }, + { + "epoch": 2.153055983564458, + "grad_norm": 3.228498935699463, + "learning_rate": 1.5968785687095547e-06, + "loss": 1.3591, + "step": 2096 + }, + { + "epoch": 2.1540832049306626, + "grad_norm": 3.145155191421509, + "learning_rate": 1.594975256947088e-06, + "loss": 1.3068, + "step": 2097 + }, + { + "epoch": 2.1551104262968668, + "grad_norm": 3.19036865234375, + "learning_rate": 1.5930719451846212e-06, + "loss": 1.3645, + "step": 2098 + }, + { + "epoch": 2.1561376476630714, + "grad_norm": 3.1367108821868896, + "learning_rate": 1.5911686334221546e-06, + "loss": 1.3138, + "step": 2099 + }, + { + "epoch": 2.157164869029276, + "grad_norm": 3.268315076828003, + "learning_rate": 1.589265321659688e-06, + "loss": 1.3677, + "step": 2100 + }, + { + "epoch": 2.157164869029276, + "eval_loss": 1.7752312421798706, + "eval_runtime": 37.3358, + "eval_samples_per_second": 26.784, + "eval_steps_per_second": 1.125, + "step": 2100 + }, + { + "epoch": 2.15819209039548, + "grad_norm": 3.0902626514434814, + "learning_rate": 1.5873620098972211e-06, + "loss": 1.2737, + "step": 2101 + }, + { + "epoch": 2.1592193117616847, + "grad_norm": 3.183992385864258, + "learning_rate": 1.5854586981347547e-06, + "loss": 1.318, + "step": 2102 + }, + { + "epoch": 2.1602465331278893, + "grad_norm": 3.348939895629883, + "learning_rate": 1.583555386372288e-06, + "loss": 1.3618, + "step": 2103 + }, + { + "epoch": 2.1612737544940934, + "grad_norm": 3.395472526550293, + "learning_rate": 1.5816520746098212e-06, + "loss": 1.378, + "step": 2104 + }, + { + "epoch": 2.162300975860298, + "grad_norm": 3.367633819580078, + "learning_rate": 1.5797487628473546e-06, + "loss": 1.3318, + "step": 2105 + }, + { + "epoch": 2.163328197226502, + "grad_norm": 3.1082911491394043, + "learning_rate": 1.577845451084888e-06, + "loss": 1.3049, + "step": 2106 + }, + { + "epoch": 2.1643554185927067, + "grad_norm": 3.3265581130981445, + "learning_rate": 1.575942139322421e-06, + "loss": 1.3368, + "step": 2107 + }, + { + "epoch": 2.1653826399589113, + "grad_norm": 3.167185068130493, + "learning_rate": 1.5740388275599544e-06, + "loss": 1.2669, + "step": 2108 + }, + { + "epoch": 2.1664098613251155, + "grad_norm": 3.3141398429870605, + "learning_rate": 1.5721355157974876e-06, + "loss": 1.3532, + "step": 2109 + }, + { + "epoch": 2.16743708269132, + "grad_norm": 3.2805464267730713, + "learning_rate": 1.570232204035021e-06, + "loss": 1.3291, + "step": 2110 + }, + { + "epoch": 2.16743708269132, + "eval_loss": 1.7720998525619507, + "eval_runtime": 36.915, + "eval_samples_per_second": 27.089, + "eval_steps_per_second": 1.138, + "step": 2110 + }, + { + "epoch": 2.168464304057524, + "grad_norm": 3.2257862091064453, + "learning_rate": 1.5683288922725543e-06, + "loss": 1.3417, + "step": 2111 + }, + { + "epoch": 2.169491525423729, + "grad_norm": 3.1542367935180664, + "learning_rate": 1.5664255805100879e-06, + "loss": 1.3475, + "step": 2112 + }, + { + "epoch": 2.1705187467899334, + "grad_norm": 3.1991381645202637, + "learning_rate": 1.564522268747621e-06, + "loss": 1.3259, + "step": 2113 + }, + { + "epoch": 2.1715459681561375, + "grad_norm": 3.0904557704925537, + "learning_rate": 1.5626189569851544e-06, + "loss": 1.3489, + "step": 2114 + }, + { + "epoch": 2.172573189522342, + "grad_norm": 3.1544106006622314, + "learning_rate": 1.5607156452226875e-06, + "loss": 1.2626, + "step": 2115 + }, + { + "epoch": 2.1736004108885463, + "grad_norm": 3.2527761459350586, + "learning_rate": 1.558812333460221e-06, + "loss": 1.3497, + "step": 2116 + }, + { + "epoch": 2.174627632254751, + "grad_norm": 3.3398337364196777, + "learning_rate": 1.5569090216977543e-06, + "loss": 1.3492, + "step": 2117 + }, + { + "epoch": 2.1756548536209555, + "grad_norm": 3.246671199798584, + "learning_rate": 1.5550057099352874e-06, + "loss": 1.3495, + "step": 2118 + }, + { + "epoch": 2.1766820749871596, + "grad_norm": 3.4486782550811768, + "learning_rate": 1.5531023981728208e-06, + "loss": 1.3541, + "step": 2119 + }, + { + "epoch": 2.177709296353364, + "grad_norm": 3.209132432937622, + "learning_rate": 1.5511990864103542e-06, + "loss": 1.3134, + "step": 2120 + }, + { + "epoch": 2.177709296353364, + "eval_loss": 1.7801611423492432, + "eval_runtime": 36.9888, + "eval_samples_per_second": 27.035, + "eval_steps_per_second": 1.135, + "step": 2120 + }, + { + "epoch": 2.1787365177195683, + "grad_norm": 3.225480794906616, + "learning_rate": 1.5492957746478873e-06, + "loss": 1.3486, + "step": 2121 + }, + { + "epoch": 2.179763739085773, + "grad_norm": 3.2712321281433105, + "learning_rate": 1.5473924628854209e-06, + "loss": 1.3659, + "step": 2122 + }, + { + "epoch": 2.1807909604519775, + "grad_norm": 3.233598470687866, + "learning_rate": 1.5454891511229542e-06, + "loss": 1.3295, + "step": 2123 + }, + { + "epoch": 2.1818181818181817, + "grad_norm": 3.167868137359619, + "learning_rate": 1.5435858393604874e-06, + "loss": 1.3372, + "step": 2124 + }, + { + "epoch": 2.1828454031843862, + "grad_norm": 3.364309310913086, + "learning_rate": 1.5416825275980208e-06, + "loss": 1.356, + "step": 2125 + }, + { + "epoch": 2.183872624550591, + "grad_norm": 3.125401020050049, + "learning_rate": 1.539779215835554e-06, + "loss": 1.3306, + "step": 2126 + }, + { + "epoch": 2.184899845916795, + "grad_norm": 3.256223201751709, + "learning_rate": 1.5378759040730873e-06, + "loss": 1.3256, + "step": 2127 + }, + { + "epoch": 2.1859270672829996, + "grad_norm": 3.309673547744751, + "learning_rate": 1.5359725923106206e-06, + "loss": 1.3753, + "step": 2128 + }, + { + "epoch": 2.1869542886492037, + "grad_norm": 3.238278388977051, + "learning_rate": 1.5340692805481538e-06, + "loss": 1.3459, + "step": 2129 + }, + { + "epoch": 2.1879815100154083, + "grad_norm": 3.16030216217041, + "learning_rate": 1.5321659687856871e-06, + "loss": 1.3108, + "step": 2130 + }, + { + "epoch": 2.1879815100154083, + "eval_loss": 1.7832409143447876, + "eval_runtime": 37.0244, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 2130 + }, + { + "epoch": 2.189008731381613, + "grad_norm": 3.124526262283325, + "learning_rate": 1.5302626570232205e-06, + "loss": 1.3097, + "step": 2131 + }, + { + "epoch": 2.190035952747817, + "grad_norm": 3.112478256225586, + "learning_rate": 1.5283593452607539e-06, + "loss": 1.278, + "step": 2132 + }, + { + "epoch": 2.1910631741140216, + "grad_norm": 3.4559900760650635, + "learning_rate": 1.5264560334982872e-06, + "loss": 1.385, + "step": 2133 + }, + { + "epoch": 2.1920903954802258, + "grad_norm": 3.1703872680664062, + "learning_rate": 1.5245527217358206e-06, + "loss": 1.3299, + "step": 2134 + }, + { + "epoch": 2.1931176168464304, + "grad_norm": 3.314497947692871, + "learning_rate": 1.5226494099733537e-06, + "loss": 1.3474, + "step": 2135 + }, + { + "epoch": 2.194144838212635, + "grad_norm": 3.0944998264312744, + "learning_rate": 1.520746098210887e-06, + "loss": 1.3214, + "step": 2136 + }, + { + "epoch": 2.195172059578839, + "grad_norm": 3.2073676586151123, + "learning_rate": 1.5188427864484205e-06, + "loss": 1.3766, + "step": 2137 + }, + { + "epoch": 2.1961992809450437, + "grad_norm": 3.0689923763275146, + "learning_rate": 1.5169394746859536e-06, + "loss": 1.319, + "step": 2138 + }, + { + "epoch": 2.1972265023112483, + "grad_norm": 3.1764934062957764, + "learning_rate": 1.515036162923487e-06, + "loss": 1.3342, + "step": 2139 + }, + { + "epoch": 2.1982537236774524, + "grad_norm": 3.1994433403015137, + "learning_rate": 1.5131328511610201e-06, + "loss": 1.3462, + "step": 2140 + }, + { + "epoch": 2.1982537236774524, + "eval_loss": 1.7849186658859253, + "eval_runtime": 37.0183, + "eval_samples_per_second": 27.014, + "eval_steps_per_second": 1.135, + "step": 2140 + }, + { + "epoch": 2.199280945043657, + "grad_norm": 3.37168025970459, + "learning_rate": 1.5112295393985535e-06, + "loss": 1.3805, + "step": 2141 + }, + { + "epoch": 2.200308166409861, + "grad_norm": 3.1596593856811523, + "learning_rate": 1.509326227636087e-06, + "loss": 1.3543, + "step": 2142 + }, + { + "epoch": 2.2013353877760657, + "grad_norm": 3.086030960083008, + "learning_rate": 1.5074229158736204e-06, + "loss": 1.2834, + "step": 2143 + }, + { + "epoch": 2.2023626091422703, + "grad_norm": 3.2379868030548096, + "learning_rate": 1.5055196041111536e-06, + "loss": 1.342, + "step": 2144 + }, + { + "epoch": 2.2033898305084745, + "grad_norm": 3.233773708343506, + "learning_rate": 1.503616292348687e-06, + "loss": 1.3455, + "step": 2145 + }, + { + "epoch": 2.204417051874679, + "grad_norm": 3.4700264930725098, + "learning_rate": 1.50171298058622e-06, + "loss": 1.3741, + "step": 2146 + }, + { + "epoch": 2.205444273240883, + "grad_norm": 3.1174497604370117, + "learning_rate": 1.4998096688237535e-06, + "loss": 1.2824, + "step": 2147 + }, + { + "epoch": 2.206471494607088, + "grad_norm": 3.1174497604370117, + "learning_rate": 1.4998096688237535e-06, + "loss": 1.356, + "step": 2148 + }, + { + "epoch": 2.2074987159732924, + "grad_norm": 3.3671200275421143, + "learning_rate": 1.4979063570612868e-06, + "loss": 1.3421, + "step": 2149 + }, + { + "epoch": 2.2085259373394965, + "grad_norm": 3.2470571994781494, + "learning_rate": 1.49600304529882e-06, + "loss": 1.3181, + "step": 2150 + }, + { + "epoch": 2.2085259373394965, + "eval_loss": 1.7855594158172607, + "eval_runtime": 37.0613, + "eval_samples_per_second": 26.982, + "eval_steps_per_second": 1.133, + "step": 2150 + }, + { + "epoch": 2.209553158705701, + "grad_norm": 3.3434650897979736, + "learning_rate": 1.4940997335363533e-06, + "loss": 1.3418, + "step": 2151 + }, + { + "epoch": 2.2105803800719057, + "grad_norm": 3.169682502746582, + "learning_rate": 1.4921964217738865e-06, + "loss": 1.3279, + "step": 2152 + }, + { + "epoch": 2.21160760143811, + "grad_norm": 3.3032784461975098, + "learning_rate": 1.49029311001142e-06, + "loss": 1.3582, + "step": 2153 + }, + { + "epoch": 2.2126348228043144, + "grad_norm": 3.2480883598327637, + "learning_rate": 1.4883897982489534e-06, + "loss": 1.362, + "step": 2154 + }, + { + "epoch": 2.2136620441705186, + "grad_norm": 3.2822015285491943, + "learning_rate": 1.4864864864864868e-06, + "loss": 1.3299, + "step": 2155 + }, + { + "epoch": 2.214689265536723, + "grad_norm": 3.2832584381103516, + "learning_rate": 1.48458317472402e-06, + "loss": 1.3374, + "step": 2156 + }, + { + "epoch": 2.2157164869029278, + "grad_norm": 3.057344913482666, + "learning_rate": 1.4826798629615533e-06, + "loss": 1.2969, + "step": 2157 + }, + { + "epoch": 2.216743708269132, + "grad_norm": 3.279812812805176, + "learning_rate": 1.4807765511990864e-06, + "loss": 1.3083, + "step": 2158 + }, + { + "epoch": 2.2177709296353365, + "grad_norm": 3.273972272872925, + "learning_rate": 1.4788732394366198e-06, + "loss": 1.3285, + "step": 2159 + }, + { + "epoch": 2.2187981510015407, + "grad_norm": 3.231541633605957, + "learning_rate": 1.4769699276741532e-06, + "loss": 1.3178, + "step": 2160 + }, + { + "epoch": 2.2187981510015407, + "eval_loss": 1.7819184064865112, + "eval_runtime": 37.0481, + "eval_samples_per_second": 26.992, + "eval_steps_per_second": 1.134, + "step": 2160 + }, + { + "epoch": 2.2198253723677452, + "grad_norm": 3.125030755996704, + "learning_rate": 1.4750666159116863e-06, + "loss": 1.2916, + "step": 2161 + }, + { + "epoch": 2.22085259373395, + "grad_norm": 3.172628164291382, + "learning_rate": 1.4731633041492197e-06, + "loss": 1.3554, + "step": 2162 + }, + { + "epoch": 2.221879815100154, + "grad_norm": 3.4434096813201904, + "learning_rate": 1.4712599923867528e-06, + "loss": 1.3769, + "step": 2163 + }, + { + "epoch": 2.2229070364663586, + "grad_norm": 3.1751928329467773, + "learning_rate": 1.4693566806242864e-06, + "loss": 1.2994, + "step": 2164 + }, + { + "epoch": 2.2239342578325627, + "grad_norm": 3.153481960296631, + "learning_rate": 1.4674533688618198e-06, + "loss": 1.3245, + "step": 2165 + }, + { + "epoch": 2.2249614791987673, + "grad_norm": 3.2735719680786133, + "learning_rate": 1.4655500570993531e-06, + "loss": 1.3312, + "step": 2166 + }, + { + "epoch": 2.225988700564972, + "grad_norm": 3.34511137008667, + "learning_rate": 1.4636467453368863e-06, + "loss": 1.3805, + "step": 2167 + }, + { + "epoch": 2.227015921931176, + "grad_norm": 3.2341556549072266, + "learning_rate": 1.4617434335744197e-06, + "loss": 1.3694, + "step": 2168 + }, + { + "epoch": 2.2280431432973806, + "grad_norm": 3.2977921962738037, + "learning_rate": 1.4598401218119528e-06, + "loss": 1.3546, + "step": 2169 + }, + { + "epoch": 2.2290703646635848, + "grad_norm": 3.2472712993621826, + "learning_rate": 1.4579368100494862e-06, + "loss": 1.3382, + "step": 2170 + }, + { + "epoch": 2.2290703646635848, + "eval_loss": 1.7886496782302856, + "eval_runtime": 37.0389, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 2170 + }, + { + "epoch": 2.2300975860297894, + "grad_norm": 3.219923257827759, + "learning_rate": 1.4560334982870195e-06, + "loss": 1.3479, + "step": 2171 + }, + { + "epoch": 2.231124807395994, + "grad_norm": 3.297858476638794, + "learning_rate": 1.4541301865245527e-06, + "loss": 1.3507, + "step": 2172 + }, + { + "epoch": 2.232152028762198, + "grad_norm": 3.2059547901153564, + "learning_rate": 1.452226874762086e-06, + "loss": 1.3111, + "step": 2173 + }, + { + "epoch": 2.2331792501284027, + "grad_norm": 3.320009231567383, + "learning_rate": 1.4503235629996196e-06, + "loss": 1.3853, + "step": 2174 + }, + { + "epoch": 2.2342064714946073, + "grad_norm": 3.116455554962158, + "learning_rate": 1.4484202512371528e-06, + "loss": 1.2811, + "step": 2175 + }, + { + "epoch": 2.2352336928608114, + "grad_norm": 3.2926032543182373, + "learning_rate": 1.4465169394746861e-06, + "loss": 1.3899, + "step": 2176 + }, + { + "epoch": 2.236260914227016, + "grad_norm": 3.2553927898406982, + "learning_rate": 1.4446136277122195e-06, + "loss": 1.289, + "step": 2177 + }, + { + "epoch": 2.23728813559322, + "grad_norm": 3.3934881687164307, + "learning_rate": 1.4427103159497526e-06, + "loss": 1.314, + "step": 2178 + }, + { + "epoch": 2.2383153569594247, + "grad_norm": 3.380460262298584, + "learning_rate": 1.440807004187286e-06, + "loss": 1.3433, + "step": 2179 + }, + { + "epoch": 2.2393425783256293, + "grad_norm": 3.6436474323272705, + "learning_rate": 1.4389036924248192e-06, + "loss": 1.366, + "step": 2180 + }, + { + "epoch": 2.2393425783256293, + "eval_loss": 1.7869333028793335, + "eval_runtime": 37.0766, + "eval_samples_per_second": 26.971, + "eval_steps_per_second": 1.133, + "step": 2180 + }, + { + "epoch": 2.2403697996918335, + "grad_norm": 3.26050066947937, + "learning_rate": 1.4370003806623525e-06, + "loss": 1.3365, + "step": 2181 + }, + { + "epoch": 2.241397021058038, + "grad_norm": 3.31622576713562, + "learning_rate": 1.4350970688998859e-06, + "loss": 1.3857, + "step": 2182 + }, + { + "epoch": 2.242424242424242, + "grad_norm": 3.2586722373962402, + "learning_rate": 1.433193757137419e-06, + "loss": 1.3216, + "step": 2183 + }, + { + "epoch": 2.243451463790447, + "grad_norm": 3.00472092628479, + "learning_rate": 1.4312904453749526e-06, + "loss": 1.2854, + "step": 2184 + }, + { + "epoch": 2.2444786851566514, + "grad_norm": 3.261831760406494, + "learning_rate": 1.429387133612486e-06, + "loss": 1.3284, + "step": 2185 + }, + { + "epoch": 2.2455059065228555, + "grad_norm": 3.1558244228363037, + "learning_rate": 1.4274838218500191e-06, + "loss": 1.3478, + "step": 2186 + }, + { + "epoch": 2.24653312788906, + "grad_norm": 3.3475098609924316, + "learning_rate": 1.4255805100875525e-06, + "loss": 1.3973, + "step": 2187 + }, + { + "epoch": 2.2475603492552647, + "grad_norm": 3.2336196899414062, + "learning_rate": 1.4236771983250858e-06, + "loss": 1.3334, + "step": 2188 + }, + { + "epoch": 2.248587570621469, + "grad_norm": 3.107243537902832, + "learning_rate": 1.421773886562619e-06, + "loss": 1.3104, + "step": 2189 + }, + { + "epoch": 2.2496147919876734, + "grad_norm": 3.0853683948516846, + "learning_rate": 1.4198705748001524e-06, + "loss": 1.326, + "step": 2190 + }, + { + "epoch": 2.2496147919876734, + "eval_loss": 1.7847614288330078, + "eval_runtime": 37.1233, + "eval_samples_per_second": 26.937, + "eval_steps_per_second": 1.131, + "step": 2190 + }, + { + "epoch": 2.2506420133538776, + "grad_norm": 3.3024446964263916, + "learning_rate": 1.4179672630376855e-06, + "loss": 1.3083, + "step": 2191 + }, + { + "epoch": 2.251669234720082, + "grad_norm": 3.2713446617126465, + "learning_rate": 1.4160639512752189e-06, + "loss": 1.3272, + "step": 2192 + }, + { + "epoch": 2.2526964560862868, + "grad_norm": 3.2788100242614746, + "learning_rate": 1.4141606395127522e-06, + "loss": 1.2954, + "step": 2193 + }, + { + "epoch": 2.253723677452491, + "grad_norm": 3.1132969856262207, + "learning_rate": 1.4122573277502858e-06, + "loss": 1.3539, + "step": 2194 + }, + { + "epoch": 2.2547508988186955, + "grad_norm": 3.1661081314086914, + "learning_rate": 1.410354015987819e-06, + "loss": 1.3275, + "step": 2195 + }, + { + "epoch": 2.2557781201848996, + "grad_norm": 3.156219959259033, + "learning_rate": 1.4084507042253523e-06, + "loss": 1.3608, + "step": 2196 + }, + { + "epoch": 2.2568053415511042, + "grad_norm": 3.0919189453125, + "learning_rate": 1.4065473924628855e-06, + "loss": 1.3308, + "step": 2197 + }, + { + "epoch": 2.257832562917309, + "grad_norm": 3.244119882583618, + "learning_rate": 1.4046440807004188e-06, + "loss": 1.3188, + "step": 2198 + }, + { + "epoch": 2.258859784283513, + "grad_norm": 3.2616937160491943, + "learning_rate": 1.4027407689379522e-06, + "loss": 1.3105, + "step": 2199 + }, + { + "epoch": 2.2598870056497176, + "grad_norm": 3.2117950916290283, + "learning_rate": 1.4008374571754854e-06, + "loss": 1.3396, + "step": 2200 + }, + { + "epoch": 2.2598870056497176, + "eval_loss": 1.7757095098495483, + "eval_runtime": 37.2922, + "eval_samples_per_second": 26.815, + "eval_steps_per_second": 1.126, + "step": 2200 + }, + { + "epoch": 2.260914227015922, + "grad_norm": 3.098792791366577, + "learning_rate": 1.3989341454130187e-06, + "loss": 1.3305, + "step": 2201 + }, + { + "epoch": 2.2619414483821263, + "grad_norm": 3.301877737045288, + "learning_rate": 1.397030833650552e-06, + "loss": 1.3512, + "step": 2202 + }, + { + "epoch": 2.262968669748331, + "grad_norm": 3.3404908180236816, + "learning_rate": 1.3951275218880852e-06, + "loss": 1.4076, + "step": 2203 + }, + { + "epoch": 2.263995891114535, + "grad_norm": 3.272155284881592, + "learning_rate": 1.3932242101256188e-06, + "loss": 1.303, + "step": 2204 + }, + { + "epoch": 2.2650231124807396, + "grad_norm": 3.3755571842193604, + "learning_rate": 1.3913208983631522e-06, + "loss": 1.3385, + "step": 2205 + }, + { + "epoch": 2.266050333846944, + "grad_norm": 3.267378568649292, + "learning_rate": 1.3894175866006853e-06, + "loss": 1.2971, + "step": 2206 + }, + { + "epoch": 2.2670775552131484, + "grad_norm": 3.285639762878418, + "learning_rate": 1.3875142748382187e-06, + "loss": 1.275, + "step": 2207 + }, + { + "epoch": 2.268104776579353, + "grad_norm": 3.270244598388672, + "learning_rate": 1.385610963075752e-06, + "loss": 1.3392, + "step": 2208 + }, + { + "epoch": 2.269131997945557, + "grad_norm": 3.3014471530914307, + "learning_rate": 1.3837076513132852e-06, + "loss": 1.3665, + "step": 2209 + }, + { + "epoch": 2.2701592193117617, + "grad_norm": 3.4669065475463867, + "learning_rate": 1.3818043395508186e-06, + "loss": 1.3543, + "step": 2210 + }, + { + "epoch": 2.2701592193117617, + "eval_loss": 1.7823984622955322, + "eval_runtime": 36.8945, + "eval_samples_per_second": 27.104, + "eval_steps_per_second": 1.138, + "step": 2210 + }, + { + "epoch": 2.2711864406779663, + "grad_norm": 3.119457483291626, + "learning_rate": 1.3799010277883517e-06, + "loss": 1.3349, + "step": 2211 + }, + { + "epoch": 2.2722136620441704, + "grad_norm": 3.258059024810791, + "learning_rate": 1.377997716025885e-06, + "loss": 1.3373, + "step": 2212 + }, + { + "epoch": 2.273240883410375, + "grad_norm": 3.176013946533203, + "learning_rate": 1.3760944042634184e-06, + "loss": 1.3661, + "step": 2213 + }, + { + "epoch": 2.2742681047765796, + "grad_norm": 3.2498974800109863, + "learning_rate": 1.3741910925009518e-06, + "loss": 1.3518, + "step": 2214 + }, + { + "epoch": 2.2752953261427837, + "grad_norm": 3.280646562576294, + "learning_rate": 1.3722877807384852e-06, + "loss": 1.3542, + "step": 2215 + }, + { + "epoch": 2.2763225475089883, + "grad_norm": 3.2371726036071777, + "learning_rate": 1.3703844689760185e-06, + "loss": 1.3276, + "step": 2216 + }, + { + "epoch": 2.2773497688751925, + "grad_norm": 3.15509033203125, + "learning_rate": 1.3684811572135517e-06, + "loss": 1.2877, + "step": 2217 + }, + { + "epoch": 2.278376990241397, + "grad_norm": 3.234192132949829, + "learning_rate": 1.366577845451085e-06, + "loss": 1.3538, + "step": 2218 + }, + { + "epoch": 2.279404211607601, + "grad_norm": 3.3198421001434326, + "learning_rate": 1.3646745336886184e-06, + "loss": 1.3544, + "step": 2219 + }, + { + "epoch": 2.280431432973806, + "grad_norm": 3.238799810409546, + "learning_rate": 1.3627712219261515e-06, + "loss": 1.315, + "step": 2220 + }, + { + "epoch": 2.280431432973806, + "eval_loss": 1.7820954322814941, + "eval_runtime": 37.0114, + "eval_samples_per_second": 27.019, + "eval_steps_per_second": 1.135, + "step": 2220 + }, + { + "epoch": 2.2814586543400104, + "grad_norm": 3.232816457748413, + "learning_rate": 1.360867910163685e-06, + "loss": 1.3295, + "step": 2221 + }, + { + "epoch": 2.2824858757062145, + "grad_norm": 3.228085994720459, + "learning_rate": 1.358964598401218e-06, + "loss": 1.3379, + "step": 2222 + }, + { + "epoch": 2.283513097072419, + "grad_norm": 3.1206276416778564, + "learning_rate": 1.3570612866387514e-06, + "loss": 1.3182, + "step": 2223 + }, + { + "epoch": 2.2845403184386237, + "grad_norm": 3.3304741382598877, + "learning_rate": 1.355157974876285e-06, + "loss": 1.3784, + "step": 2224 + }, + { + "epoch": 2.285567539804828, + "grad_norm": 3.2442893981933594, + "learning_rate": 1.3532546631138184e-06, + "loss": 1.3434, + "step": 2225 + }, + { + "epoch": 2.2865947611710324, + "grad_norm": 3.1287336349487305, + "learning_rate": 1.3513513513513515e-06, + "loss": 1.327, + "step": 2226 + }, + { + "epoch": 2.2876219825372366, + "grad_norm": 3.189202308654785, + "learning_rate": 1.3494480395888849e-06, + "loss": 1.326, + "step": 2227 + }, + { + "epoch": 2.288649203903441, + "grad_norm": 3.1488254070281982, + "learning_rate": 1.347544727826418e-06, + "loss": 1.3087, + "step": 2228 + }, + { + "epoch": 2.2896764252696458, + "grad_norm": 3.2201285362243652, + "learning_rate": 1.3456414160639514e-06, + "loss": 1.3484, + "step": 2229 + }, + { + "epoch": 2.29070364663585, + "grad_norm": 3.3324484825134277, + "learning_rate": 1.3437381043014847e-06, + "loss": 1.3355, + "step": 2230 + }, + { + "epoch": 2.29070364663585, + "eval_loss": 1.7727630138397217, + "eval_runtime": 37.0251, + "eval_samples_per_second": 27.009, + "eval_steps_per_second": 1.134, + "step": 2230 + }, + { + "epoch": 2.2917308680020545, + "grad_norm": 3.1624255180358887, + "learning_rate": 1.341834792539018e-06, + "loss": 1.3145, + "step": 2231 + }, + { + "epoch": 2.2927580893682586, + "grad_norm": 3.161055326461792, + "learning_rate": 1.3399314807765513e-06, + "loss": 1.3412, + "step": 2232 + }, + { + "epoch": 2.2937853107344632, + "grad_norm": 3.23097562789917, + "learning_rate": 1.3380281690140844e-06, + "loss": 1.3384, + "step": 2233 + }, + { + "epoch": 2.294812532100668, + "grad_norm": 3.3151564598083496, + "learning_rate": 1.336124857251618e-06, + "loss": 1.3043, + "step": 2234 + }, + { + "epoch": 2.295839753466872, + "grad_norm": 3.2156801223754883, + "learning_rate": 1.3342215454891513e-06, + "loss": 1.311, + "step": 2235 + }, + { + "epoch": 2.2968669748330766, + "grad_norm": 3.254260540008545, + "learning_rate": 1.3323182337266847e-06, + "loss": 1.3254, + "step": 2236 + }, + { + "epoch": 2.297894196199281, + "grad_norm": 3.37776517868042, + "learning_rate": 1.3304149219642179e-06, + "loss": 1.382, + "step": 2237 + }, + { + "epoch": 2.2989214175654853, + "grad_norm": 3.2818682193756104, + "learning_rate": 1.3285116102017512e-06, + "loss": 1.3286, + "step": 2238 + }, + { + "epoch": 2.29994863893169, + "grad_norm": 3.2190120220184326, + "learning_rate": 1.3266082984392844e-06, + "loss": 1.3348, + "step": 2239 + }, + { + "epoch": 2.300975860297894, + "grad_norm": 3.1184210777282715, + "learning_rate": 1.3247049866768177e-06, + "loss": 1.3166, + "step": 2240 + }, + { + "epoch": 2.300975860297894, + "eval_loss": 1.7773677110671997, + "eval_runtime": 37.0239, + "eval_samples_per_second": 27.01, + "eval_steps_per_second": 1.134, + "step": 2240 + }, + { + "epoch": 2.3020030816640986, + "grad_norm": 3.2418594360351562, + "learning_rate": 1.322801674914351e-06, + "loss": 1.3055, + "step": 2241 + }, + { + "epoch": 2.303030303030303, + "grad_norm": 3.1460630893707275, + "learning_rate": 1.3208983631518843e-06, + "loss": 1.3482, + "step": 2242 + }, + { + "epoch": 2.3040575243965074, + "grad_norm": 3.3080341815948486, + "learning_rate": 1.3189950513894176e-06, + "loss": 1.3316, + "step": 2243 + }, + { + "epoch": 2.305084745762712, + "grad_norm": 3.311420440673828, + "learning_rate": 1.3170917396269512e-06, + "loss": 1.3382, + "step": 2244 + }, + { + "epoch": 2.306111967128916, + "grad_norm": 3.226576328277588, + "learning_rate": 1.3151884278644843e-06, + "loss": 1.3567, + "step": 2245 + }, + { + "epoch": 2.3071391884951207, + "grad_norm": 3.1826727390289307, + "learning_rate": 1.3132851161020177e-06, + "loss": 1.304, + "step": 2246 + }, + { + "epoch": 2.3081664098613253, + "grad_norm": 3.3982961177825928, + "learning_rate": 1.311381804339551e-06, + "loss": 1.3933, + "step": 2247 + }, + { + "epoch": 2.3091936312275294, + "grad_norm": 3.2186927795410156, + "learning_rate": 1.3094784925770842e-06, + "loss": 1.3115, + "step": 2248 + }, + { + "epoch": 2.310220852593734, + "grad_norm": 3.240696430206299, + "learning_rate": 1.3075751808146176e-06, + "loss": 1.3409, + "step": 2249 + }, + { + "epoch": 2.3112480739599386, + "grad_norm": 3.1434874534606934, + "learning_rate": 1.3056718690521507e-06, + "loss": 1.311, + "step": 2250 + }, + { + "epoch": 2.3112480739599386, + "eval_loss": 1.778112530708313, + "eval_runtime": 37.0348, + "eval_samples_per_second": 27.002, + "eval_steps_per_second": 1.134, + "step": 2250 + }, + { + "epoch": 2.3122752953261427, + "grad_norm": 3.2286510467529297, + "learning_rate": 1.303768557289684e-06, + "loss": 1.3151, + "step": 2251 + }, + { + "epoch": 2.3133025166923473, + "grad_norm": 3.27567195892334, + "learning_rate": 1.3018652455272175e-06, + "loss": 1.323, + "step": 2252 + }, + { + "epoch": 2.3143297380585515, + "grad_norm": 3.267167568206787, + "learning_rate": 1.2999619337647506e-06, + "loss": 1.3192, + "step": 2253 + }, + { + "epoch": 2.315356959424756, + "grad_norm": 3.297957420349121, + "learning_rate": 1.2980586220022842e-06, + "loss": 1.3502, + "step": 2254 + }, + { + "epoch": 2.3163841807909606, + "grad_norm": 3.302241802215576, + "learning_rate": 1.2961553102398175e-06, + "loss": 1.3725, + "step": 2255 + }, + { + "epoch": 2.317411402157165, + "grad_norm": 3.3519294261932373, + "learning_rate": 1.2942519984773507e-06, + "loss": 1.3419, + "step": 2256 + }, + { + "epoch": 2.3184386235233694, + "grad_norm": 3.213895559310913, + "learning_rate": 1.292348686714884e-06, + "loss": 1.3151, + "step": 2257 + }, + { + "epoch": 2.3194658448895735, + "grad_norm": 3.2269561290740967, + "learning_rate": 1.2904453749524174e-06, + "loss": 1.3596, + "step": 2258 + }, + { + "epoch": 2.320493066255778, + "grad_norm": 3.199706792831421, + "learning_rate": 1.2885420631899506e-06, + "loss": 1.2856, + "step": 2259 + }, + { + "epoch": 2.3215202876219827, + "grad_norm": 3.2250821590423584, + "learning_rate": 1.286638751427484e-06, + "loss": 1.3246, + "step": 2260 + }, + { + "epoch": 2.3215202876219827, + "eval_loss": 1.7836729288101196, + "eval_runtime": 37.0349, + "eval_samples_per_second": 27.002, + "eval_steps_per_second": 1.134, + "step": 2260 + }, + { + "epoch": 2.322547508988187, + "grad_norm": 3.313812494277954, + "learning_rate": 1.284735439665017e-06, + "loss": 1.3475, + "step": 2261 + }, + { + "epoch": 2.3235747303543914, + "grad_norm": 3.4544384479522705, + "learning_rate": 1.2828321279025504e-06, + "loss": 1.3582, + "step": 2262 + }, + { + "epoch": 2.324601951720596, + "grad_norm": 3.2306017875671387, + "learning_rate": 1.2809288161400838e-06, + "loss": 1.3306, + "step": 2263 + }, + { + "epoch": 2.3256291730868, + "grad_norm": 3.2713663578033447, + "learning_rate": 1.279025504377617e-06, + "loss": 1.3128, + "step": 2264 + }, + { + "epoch": 2.3266563944530048, + "grad_norm": 3.4028589725494385, + "learning_rate": 1.2771221926151505e-06, + "loss": 1.3843, + "step": 2265 + }, + { + "epoch": 2.327683615819209, + "grad_norm": 3.26240611076355, + "learning_rate": 1.275218880852684e-06, + "loss": 1.3215, + "step": 2266 + }, + { + "epoch": 2.3287108371854135, + "grad_norm": 3.265568971633911, + "learning_rate": 1.273315569090217e-06, + "loss": 1.3157, + "step": 2267 + }, + { + "epoch": 2.3297380585516176, + "grad_norm": 3.3437466621398926, + "learning_rate": 1.2714122573277504e-06, + "loss": 1.3891, + "step": 2268 + }, + { + "epoch": 2.3307652799178222, + "grad_norm": 3.258962392807007, + "learning_rate": 1.2695089455652838e-06, + "loss": 1.3849, + "step": 2269 + }, + { + "epoch": 2.331792501284027, + "grad_norm": 3.243682622909546, + "learning_rate": 1.267605633802817e-06, + "loss": 1.3493, + "step": 2270 + }, + { + "epoch": 2.331792501284027, + "eval_loss": 1.7861374616622925, + "eval_runtime": 37.0589, + "eval_samples_per_second": 26.984, + "eval_steps_per_second": 1.133, + "step": 2270 + }, + { + "epoch": 2.332819722650231, + "grad_norm": 3.2758984565734863, + "learning_rate": 1.2657023220403503e-06, + "loss": 1.3433, + "step": 2271 + }, + { + "epoch": 2.3338469440164356, + "grad_norm": 3.3338420391082764, + "learning_rate": 1.2637990102778836e-06, + "loss": 1.3269, + "step": 2272 + }, + { + "epoch": 2.33487416538264, + "grad_norm": 3.283289670944214, + "learning_rate": 1.2618956985154168e-06, + "loss": 1.325, + "step": 2273 + }, + { + "epoch": 2.3359013867488443, + "grad_norm": 3.3016412258148193, + "learning_rate": 1.2599923867529502e-06, + "loss": 1.3744, + "step": 2274 + }, + { + "epoch": 2.336928608115049, + "grad_norm": 3.1569316387176514, + "learning_rate": 1.2580890749904837e-06, + "loss": 1.3356, + "step": 2275 + }, + { + "epoch": 2.337955829481253, + "grad_norm": 3.427867889404297, + "learning_rate": 1.2561857632280169e-06, + "loss": 1.3608, + "step": 2276 + }, + { + "epoch": 2.3389830508474576, + "grad_norm": 3.4001643657684326, + "learning_rate": 1.2542824514655502e-06, + "loss": 1.3933, + "step": 2277 + }, + { + "epoch": 2.340010272213662, + "grad_norm": 3.304152488708496, + "learning_rate": 1.2523791397030834e-06, + "loss": 1.3507, + "step": 2278 + }, + { + "epoch": 2.3410374935798663, + "grad_norm": 3.2333521842956543, + "learning_rate": 1.2504758279406168e-06, + "loss": 1.3282, + "step": 2279 + }, + { + "epoch": 2.342064714946071, + "grad_norm": 3.3327064514160156, + "learning_rate": 1.2485725161781501e-06, + "loss": 1.3499, + "step": 2280 + }, + { + "epoch": 2.342064714946071, + "eval_loss": 1.778182864189148, + "eval_runtime": 37.0637, + "eval_samples_per_second": 26.981, + "eval_steps_per_second": 1.133, + "step": 2280 + }, + { + "epoch": 2.343091936312275, + "grad_norm": 3.3308699131011963, + "learning_rate": 1.2466692044156833e-06, + "loss": 1.4328, + "step": 2281 + }, + { + "epoch": 2.3441191576784797, + "grad_norm": 3.1699163913726807, + "learning_rate": 1.2447658926532168e-06, + "loss": 1.298, + "step": 2282 + }, + { + "epoch": 2.3451463790446843, + "grad_norm": 3.2453134059906006, + "learning_rate": 1.24286258089075e-06, + "loss": 1.3354, + "step": 2283 + }, + { + "epoch": 2.3461736004108884, + "grad_norm": 3.4059667587280273, + "learning_rate": 1.2409592691282834e-06, + "loss": 1.3068, + "step": 2284 + }, + { + "epoch": 2.347200821777093, + "grad_norm": 3.3648319244384766, + "learning_rate": 1.2390559573658165e-06, + "loss": 1.3485, + "step": 2285 + }, + { + "epoch": 2.3482280431432976, + "grad_norm": 3.3277859687805176, + "learning_rate": 1.2371526456033499e-06, + "loss": 1.3486, + "step": 2286 + }, + { + "epoch": 2.3492552645095017, + "grad_norm": 3.3068926334381104, + "learning_rate": 1.2352493338408832e-06, + "loss": 1.303, + "step": 2287 + }, + { + "epoch": 2.3502824858757063, + "grad_norm": 3.399491310119629, + "learning_rate": 1.2333460220784166e-06, + "loss": 1.411, + "step": 2288 + }, + { + "epoch": 2.3513097072419105, + "grad_norm": 3.3704006671905518, + "learning_rate": 1.23144271031595e-06, + "loss": 1.3676, + "step": 2289 + }, + { + "epoch": 2.352336928608115, + "grad_norm": 3.285747528076172, + "learning_rate": 1.2295393985534831e-06, + "loss": 1.3763, + "step": 2290 + }, + { + "epoch": 2.352336928608115, + "eval_loss": 1.7770358324050903, + "eval_runtime": 37.1157, + "eval_samples_per_second": 26.943, + "eval_steps_per_second": 1.132, + "step": 2290 + }, + { + "epoch": 2.3533641499743196, + "grad_norm": 3.123500347137451, + "learning_rate": 1.2276360867910165e-06, + "loss": 1.2622, + "step": 2291 + }, + { + "epoch": 2.354391371340524, + "grad_norm": 3.286098003387451, + "learning_rate": 1.2257327750285498e-06, + "loss": 1.3803, + "step": 2292 + }, + { + "epoch": 2.3554185927067284, + "grad_norm": 3.24936842918396, + "learning_rate": 1.2238294632660832e-06, + "loss": 1.3044, + "step": 2293 + }, + { + "epoch": 2.3564458140729325, + "grad_norm": 3.107374668121338, + "learning_rate": 1.2219261515036164e-06, + "loss": 1.3125, + "step": 2294 + }, + { + "epoch": 2.357473035439137, + "grad_norm": 3.3370137214660645, + "learning_rate": 1.2200228397411497e-06, + "loss": 1.3631, + "step": 2295 + }, + { + "epoch": 2.3585002568053417, + "grad_norm": 3.2840030193328857, + "learning_rate": 1.2181195279786829e-06, + "loss": 1.3832, + "step": 2296 + }, + { + "epoch": 2.359527478171546, + "grad_norm": 3.2381551265716553, + "learning_rate": 1.2162162162162164e-06, + "loss": 1.328, + "step": 2297 + }, + { + "epoch": 2.3605546995377504, + "grad_norm": 3.285858631134033, + "learning_rate": 1.2143129044537496e-06, + "loss": 1.3198, + "step": 2298 + }, + { + "epoch": 2.361581920903955, + "grad_norm": 3.2457311153411865, + "learning_rate": 1.212409592691283e-06, + "loss": 1.3269, + "step": 2299 + }, + { + "epoch": 2.362609142270159, + "grad_norm": 3.2986185550689697, + "learning_rate": 1.2105062809288163e-06, + "loss": 1.3354, + "step": 2300 + }, + { + "epoch": 2.362609142270159, + "eval_loss": 1.7782814502716064, + "eval_runtime": 37.3426, + "eval_samples_per_second": 26.779, + "eval_steps_per_second": 1.125, + "step": 2300 + }, + { + "epoch": 2.3636363636363638, + "grad_norm": 3.4224026203155518, + "learning_rate": 1.2086029691663495e-06, + "loss": 1.3492, + "step": 2301 + }, + { + "epoch": 2.364663585002568, + "grad_norm": 3.1536035537719727, + "learning_rate": 1.2066996574038828e-06, + "loss": 1.2877, + "step": 2302 + }, + { + "epoch": 2.3656908063687725, + "grad_norm": 3.4149744510650635, + "learning_rate": 1.2047963456414162e-06, + "loss": 1.3481, + "step": 2303 + }, + { + "epoch": 2.366718027734977, + "grad_norm": 3.3409876823425293, + "learning_rate": 1.2028930338789496e-06, + "loss": 1.3415, + "step": 2304 + }, + { + "epoch": 2.3677452491011812, + "grad_norm": 3.2326085567474365, + "learning_rate": 1.2009897221164827e-06, + "loss": 1.2796, + "step": 2305 + }, + { + "epoch": 2.368772470467386, + "grad_norm": 3.2900753021240234, + "learning_rate": 1.199086410354016e-06, + "loss": 1.3451, + "step": 2306 + }, + { + "epoch": 2.36979969183359, + "grad_norm": 3.2022976875305176, + "learning_rate": 1.1971830985915492e-06, + "loss": 1.3125, + "step": 2307 + }, + { + "epoch": 2.3708269131997946, + "grad_norm": 3.2597901821136475, + "learning_rate": 1.1952797868290828e-06, + "loss": 1.3201, + "step": 2308 + }, + { + "epoch": 2.371854134565999, + "grad_norm": 3.2959396839141846, + "learning_rate": 1.193376475066616e-06, + "loss": 1.3424, + "step": 2309 + }, + { + "epoch": 2.3728813559322033, + "grad_norm": 3.3182787895202637, + "learning_rate": 1.1914731633041493e-06, + "loss": 1.3494, + "step": 2310 + }, + { + "epoch": 2.3728813559322033, + "eval_loss": 1.7703481912612915, + "eval_runtime": 36.9381, + "eval_samples_per_second": 27.072, + "eval_steps_per_second": 1.137, + "step": 2310 + }, + { + "epoch": 2.373908577298408, + "grad_norm": 3.2580761909484863, + "learning_rate": 1.1895698515416827e-06, + "loss": 1.3175, + "step": 2311 + }, + { + "epoch": 2.3749357986646125, + "grad_norm": 3.2173686027526855, + "learning_rate": 1.1876665397792158e-06, + "loss": 1.3174, + "step": 2312 + }, + { + "epoch": 2.3759630200308166, + "grad_norm": 3.355337619781494, + "learning_rate": 1.1857632280167492e-06, + "loss": 1.3483, + "step": 2313 + }, + { + "epoch": 2.376990241397021, + "grad_norm": 3.0853991508483887, + "learning_rate": 1.1838599162542825e-06, + "loss": 1.2876, + "step": 2314 + }, + { + "epoch": 2.3780174627632253, + "grad_norm": 3.384265661239624, + "learning_rate": 1.181956604491816e-06, + "loss": 1.3543, + "step": 2315 + }, + { + "epoch": 2.37904468412943, + "grad_norm": 3.1869325637817383, + "learning_rate": 1.180053292729349e-06, + "loss": 1.336, + "step": 2316 + }, + { + "epoch": 2.380071905495634, + "grad_norm": 3.4381766319274902, + "learning_rate": 1.1781499809668824e-06, + "loss": 1.3976, + "step": 2317 + }, + { + "epoch": 2.3810991268618387, + "grad_norm": 3.508650064468384, + "learning_rate": 1.1762466692044158e-06, + "loss": 1.3316, + "step": 2318 + }, + { + "epoch": 2.3821263482280433, + "grad_norm": 3.449836015701294, + "learning_rate": 1.1743433574419491e-06, + "loss": 1.3663, + "step": 2319 + }, + { + "epoch": 2.3831535695942474, + "grad_norm": 3.234224796295166, + "learning_rate": 1.1724400456794823e-06, + "loss": 1.3148, + "step": 2320 + }, + { + "epoch": 2.3831535695942474, + "eval_loss": 1.7687087059020996, + "eval_runtime": 37.0375, + "eval_samples_per_second": 27.0, + "eval_steps_per_second": 1.134, + "step": 2320 + }, + { + "epoch": 2.384180790960452, + "grad_norm": 3.3927574157714844, + "learning_rate": 1.1705367339170157e-06, + "loss": 1.359, + "step": 2321 + }, + { + "epoch": 2.3852080123266566, + "grad_norm": 3.2600550651550293, + "learning_rate": 1.168633422154549e-06, + "loss": 1.3357, + "step": 2322 + }, + { + "epoch": 2.3862352336928607, + "grad_norm": 3.1594438552856445, + "learning_rate": 1.1667301103920824e-06, + "loss": 1.3015, + "step": 2323 + }, + { + "epoch": 2.3872624550590653, + "grad_norm": 3.1696481704711914, + "learning_rate": 1.1648267986296157e-06, + "loss": 1.266, + "step": 2324 + }, + { + "epoch": 2.3882896764252695, + "grad_norm": 3.3665976524353027, + "learning_rate": 1.162923486867149e-06, + "loss": 1.321, + "step": 2325 + }, + { + "epoch": 2.389316897791474, + "grad_norm": 3.3364579677581787, + "learning_rate": 1.1610201751046823e-06, + "loss": 1.3363, + "step": 2326 + }, + { + "epoch": 2.3903441191576786, + "grad_norm": 3.304212808609009, + "learning_rate": 1.1591168633422154e-06, + "loss": 1.3945, + "step": 2327 + }, + { + "epoch": 2.391371340523883, + "grad_norm": 3.4038922786712646, + "learning_rate": 1.157213551579749e-06, + "loss": 1.3503, + "step": 2328 + }, + { + "epoch": 2.3923985618900874, + "grad_norm": 3.3038649559020996, + "learning_rate": 1.1553102398172821e-06, + "loss": 1.3332, + "step": 2329 + }, + { + "epoch": 2.3934257832562915, + "grad_norm": 3.179196834564209, + "learning_rate": 1.1534069280548155e-06, + "loss": 1.3223, + "step": 2330 + }, + { + "epoch": 2.3934257832562915, + "eval_loss": 1.7737306356430054, + "eval_runtime": 37.0672, + "eval_samples_per_second": 26.978, + "eval_steps_per_second": 1.133, + "step": 2330 + }, + { + "epoch": 2.394453004622496, + "grad_norm": 3.4174447059631348, + "learning_rate": 1.1515036162923487e-06, + "loss": 1.3388, + "step": 2331 + }, + { + "epoch": 2.3954802259887007, + "grad_norm": 3.310110092163086, + "learning_rate": 1.149600304529882e-06, + "loss": 1.3083, + "step": 2332 + }, + { + "epoch": 2.396507447354905, + "grad_norm": 3.1463305950164795, + "learning_rate": 1.1476969927674154e-06, + "loss": 1.3179, + "step": 2333 + }, + { + "epoch": 2.3975346687211094, + "grad_norm": 3.266430616378784, + "learning_rate": 1.1457936810049487e-06, + "loss": 1.3333, + "step": 2334 + }, + { + "epoch": 2.398561890087314, + "grad_norm": 3.307996988296509, + "learning_rate": 1.143890369242482e-06, + "loss": 1.3619, + "step": 2335 + }, + { + "epoch": 2.399589111453518, + "grad_norm": 3.1686081886291504, + "learning_rate": 1.1419870574800153e-06, + "loss": 1.2897, + "step": 2336 + }, + { + "epoch": 2.4006163328197228, + "grad_norm": 3.222095489501953, + "learning_rate": 1.1400837457175486e-06, + "loss": 1.3231, + "step": 2337 + }, + { + "epoch": 2.401643554185927, + "grad_norm": 3.1941967010498047, + "learning_rate": 1.138180433955082e-06, + "loss": 1.2919, + "step": 2338 + }, + { + "epoch": 2.4026707755521315, + "grad_norm": 3.2475600242614746, + "learning_rate": 1.1362771221926153e-06, + "loss": 1.3664, + "step": 2339 + }, + { + "epoch": 2.403697996918336, + "grad_norm": 3.280923366546631, + "learning_rate": 1.1343738104301485e-06, + "loss": 1.3338, + "step": 2340 + }, + { + "epoch": 2.403697996918336, + "eval_loss": 1.7804691791534424, + "eval_runtime": 37.0607, + "eval_samples_per_second": 26.983, + "eval_steps_per_second": 1.133, + "step": 2340 + }, + { + "epoch": 2.4047252182845402, + "grad_norm": 3.2299935817718506, + "learning_rate": 1.1324704986676819e-06, + "loss": 1.3092, + "step": 2341 + }, + { + "epoch": 2.405752439650745, + "grad_norm": 3.5989112854003906, + "learning_rate": 1.130567186905215e-06, + "loss": 1.3777, + "step": 2342 + }, + { + "epoch": 2.406779661016949, + "grad_norm": 3.2064805030822754, + "learning_rate": 1.1286638751427486e-06, + "loss": 1.319, + "step": 2343 + }, + { + "epoch": 2.4078068823831535, + "grad_norm": 3.3006231784820557, + "learning_rate": 1.1267605633802817e-06, + "loss": 1.3258, + "step": 2344 + }, + { + "epoch": 2.408834103749358, + "grad_norm": 3.4369914531707764, + "learning_rate": 1.124857251617815e-06, + "loss": 1.3569, + "step": 2345 + }, + { + "epoch": 2.4098613251155623, + "grad_norm": 3.346527576446533, + "learning_rate": 1.1229539398553485e-06, + "loss": 1.3831, + "step": 2346 + }, + { + "epoch": 2.410888546481767, + "grad_norm": 3.2306876182556152, + "learning_rate": 1.1210506280928816e-06, + "loss": 1.3246, + "step": 2347 + }, + { + "epoch": 2.4119157678479715, + "grad_norm": 3.29376482963562, + "learning_rate": 1.119147316330415e-06, + "loss": 1.3601, + "step": 2348 + }, + { + "epoch": 2.4129429892141756, + "grad_norm": 3.4046168327331543, + "learning_rate": 1.1172440045679483e-06, + "loss": 1.3327, + "step": 2349 + }, + { + "epoch": 2.41397021058038, + "grad_norm": 3.2365424633026123, + "learning_rate": 1.1153406928054817e-06, + "loss": 1.3292, + "step": 2350 + }, + { + "epoch": 2.41397021058038, + "eval_loss": 1.7834311723709106, + "eval_runtime": 37.0449, + "eval_samples_per_second": 26.994, + "eval_steps_per_second": 1.134, + "step": 2350 + }, + { + "epoch": 2.4149974319465843, + "grad_norm": 3.2570419311523438, + "learning_rate": 1.1134373810430148e-06, + "loss": 1.3581, + "step": 2351 + }, + { + "epoch": 2.416024653312789, + "grad_norm": 3.294447183609009, + "learning_rate": 1.1115340692805482e-06, + "loss": 1.3619, + "step": 2352 + }, + { + "epoch": 2.4170518746789935, + "grad_norm": 3.2618157863616943, + "learning_rate": 1.1096307575180816e-06, + "loss": 1.3325, + "step": 2353 + }, + { + "epoch": 2.4180790960451977, + "grad_norm": 3.111448049545288, + "learning_rate": 1.107727445755615e-06, + "loss": 1.2961, + "step": 2354 + }, + { + "epoch": 2.4191063174114023, + "grad_norm": 3.2396726608276367, + "learning_rate": 1.105824133993148e-06, + "loss": 1.3093, + "step": 2355 + }, + { + "epoch": 2.4201335387776064, + "grad_norm": 3.3896162509918213, + "learning_rate": 1.1039208222306814e-06, + "loss": 1.3349, + "step": 2356 + }, + { + "epoch": 2.421160760143811, + "grad_norm": 3.399331569671631, + "learning_rate": 1.1020175104682148e-06, + "loss": 1.3324, + "step": 2357 + }, + { + "epoch": 2.4221879815100156, + "grad_norm": 3.366804599761963, + "learning_rate": 1.1001141987057482e-06, + "loss": 1.3233, + "step": 2358 + }, + { + "epoch": 2.4232152028762197, + "grad_norm": 3.3795650005340576, + "learning_rate": 1.0982108869432813e-06, + "loss": 1.3315, + "step": 2359 + }, + { + "epoch": 2.4242424242424243, + "grad_norm": 3.236072540283203, + "learning_rate": 1.0963075751808147e-06, + "loss": 1.3601, + "step": 2360 + }, + { + "epoch": 2.4242424242424243, + "eval_loss": 1.7841907739639282, + "eval_runtime": 37.0389, + "eval_samples_per_second": 26.999, + "eval_steps_per_second": 1.134, + "step": 2360 + }, + { + "epoch": 2.425269645608629, + "grad_norm": 3.1747965812683105, + "learning_rate": 1.094404263418348e-06, + "loss": 1.333, + "step": 2361 + }, + { + "epoch": 2.426296866974833, + "grad_norm": 3.4078290462493896, + "learning_rate": 1.0925009516558812e-06, + "loss": 1.3287, + "step": 2362 + }, + { + "epoch": 2.4273240883410376, + "grad_norm": 3.251595973968506, + "learning_rate": 1.0905976398934148e-06, + "loss": 1.3169, + "step": 2363 + }, + { + "epoch": 2.428351309707242, + "grad_norm": 3.375819683074951, + "learning_rate": 1.088694328130948e-06, + "loss": 1.3433, + "step": 2364 + }, + { + "epoch": 2.4293785310734464, + "grad_norm": 3.214198350906372, + "learning_rate": 1.0867910163684813e-06, + "loss": 1.294, + "step": 2365 + }, + { + "epoch": 2.4304057524396505, + "grad_norm": 3.227752208709717, + "learning_rate": 1.0848877046060144e-06, + "loss": 1.3269, + "step": 2366 + }, + { + "epoch": 2.431432973805855, + "grad_norm": 3.2604594230651855, + "learning_rate": 1.0829843928435478e-06, + "loss": 1.3227, + "step": 2367 + }, + { + "epoch": 2.4324601951720597, + "grad_norm": 3.2288079261779785, + "learning_rate": 1.0810810810810812e-06, + "loss": 1.3294, + "step": 2368 + }, + { + "epoch": 2.433487416538264, + "grad_norm": 3.29543137550354, + "learning_rate": 1.0791777693186145e-06, + "loss": 1.3422, + "step": 2369 + }, + { + "epoch": 2.4345146379044684, + "grad_norm": 3.2168960571289062, + "learning_rate": 1.0772744575561479e-06, + "loss": 1.3242, + "step": 2370 + }, + { + "epoch": 2.4345146379044684, + "eval_loss": 1.7844047546386719, + "eval_runtime": 37.0449, + "eval_samples_per_second": 26.994, + "eval_steps_per_second": 1.134, + "step": 2370 + }, + { + "epoch": 2.435541859270673, + "grad_norm": 3.308483600616455, + "learning_rate": 1.075371145793681e-06, + "loss": 1.3175, + "step": 2371 + }, + { + "epoch": 2.436569080636877, + "grad_norm": 3.209135055541992, + "learning_rate": 1.0734678340312144e-06, + "loss": 1.3281, + "step": 2372 + }, + { + "epoch": 2.4375963020030817, + "grad_norm": 3.09364652633667, + "learning_rate": 1.0715645222687478e-06, + "loss": 1.3185, + "step": 2373 + }, + { + "epoch": 2.438623523369286, + "grad_norm": 3.0643625259399414, + "learning_rate": 1.0696612105062811e-06, + "loss": 1.3146, + "step": 2374 + }, + { + "epoch": 2.4396507447354905, + "grad_norm": 3.3155925273895264, + "learning_rate": 1.0677578987438143e-06, + "loss": 1.3418, + "step": 2375 + }, + { + "epoch": 2.440677966101695, + "grad_norm": 3.1828720569610596, + "learning_rate": 1.0658545869813476e-06, + "loss": 1.2959, + "step": 2376 + }, + { + "epoch": 2.441705187467899, + "grad_norm": 3.2486345767974854, + "learning_rate": 1.0639512752188808e-06, + "loss": 1.3191, + "step": 2377 + }, + { + "epoch": 2.442732408834104, + "grad_norm": 3.186289072036743, + "learning_rate": 1.0620479634564144e-06, + "loss": 1.3268, + "step": 2378 + }, + { + "epoch": 2.443759630200308, + "grad_norm": 3.1231343746185303, + "learning_rate": 1.0601446516939475e-06, + "loss": 1.2935, + "step": 2379 + }, + { + "epoch": 2.4447868515665125, + "grad_norm": 3.161511182785034, + "learning_rate": 1.0582413399314809e-06, + "loss": 1.3001, + "step": 2380 + }, + { + "epoch": 2.4447868515665125, + "eval_loss": 1.7855908870697021, + "eval_runtime": 37.0424, + "eval_samples_per_second": 26.996, + "eval_steps_per_second": 1.134, + "step": 2380 + }, + { + "epoch": 2.445814072932717, + "grad_norm": 3.1266775131225586, + "learning_rate": 1.0563380281690142e-06, + "loss": 1.3053, + "step": 2381 + }, + { + "epoch": 2.4468412942989213, + "grad_norm": 3.149704933166504, + "learning_rate": 1.0544347164065474e-06, + "loss": 1.3062, + "step": 2382 + }, + { + "epoch": 2.447868515665126, + "grad_norm": 3.2709426879882812, + "learning_rate": 1.0525314046440808e-06, + "loss": 1.3383, + "step": 2383 + }, + { + "epoch": 2.4488957370313305, + "grad_norm": 3.200308322906494, + "learning_rate": 1.0506280928816141e-06, + "loss": 1.3066, + "step": 2384 + }, + { + "epoch": 2.4499229583975346, + "grad_norm": 3.2325589656829834, + "learning_rate": 1.0487247811191475e-06, + "loss": 1.3194, + "step": 2385 + }, + { + "epoch": 2.450950179763739, + "grad_norm": 3.2231247425079346, + "learning_rate": 1.0468214693566806e-06, + "loss": 1.3276, + "step": 2386 + }, + { + "epoch": 2.4519774011299433, + "grad_norm": 3.4579389095306396, + "learning_rate": 1.044918157594214e-06, + "loss": 1.3599, + "step": 2387 + }, + { + "epoch": 2.453004622496148, + "grad_norm": 3.1717240810394287, + "learning_rate": 1.0430148458317474e-06, + "loss": 1.3218, + "step": 2388 + }, + { + "epoch": 2.4540318438623525, + "grad_norm": 3.1802005767822266, + "learning_rate": 1.0411115340692807e-06, + "loss": 1.2873, + "step": 2389 + }, + { + "epoch": 2.4550590652285567, + "grad_norm": 3.3396799564361572, + "learning_rate": 1.0392082223068139e-06, + "loss": 1.3695, + "step": 2390 + }, + { + "epoch": 2.4550590652285567, + "eval_loss": 1.782988429069519, + "eval_runtime": 37.0862, + "eval_samples_per_second": 26.964, + "eval_steps_per_second": 1.132, + "step": 2390 + }, + { + "epoch": 2.4560862865947612, + "grad_norm": 3.324970006942749, + "learning_rate": 1.0373049105443472e-06, + "loss": 1.3275, + "step": 2391 + }, + { + "epoch": 2.4571135079609654, + "grad_norm": 3.3089895248413086, + "learning_rate": 1.0354015987818806e-06, + "loss": 1.3371, + "step": 2392 + }, + { + "epoch": 2.45814072932717, + "grad_norm": 3.1750428676605225, + "learning_rate": 1.033498287019414e-06, + "loss": 1.3021, + "step": 2393 + }, + { + "epoch": 2.4591679506933746, + "grad_norm": 3.3353590965270996, + "learning_rate": 1.0315949752569471e-06, + "loss": 1.3501, + "step": 2394 + }, + { + "epoch": 2.4601951720595787, + "grad_norm": 3.318697929382324, + "learning_rate": 1.0296916634944805e-06, + "loss": 1.2996, + "step": 2395 + }, + { + "epoch": 2.4612223934257833, + "grad_norm": 3.438951253890991, + "learning_rate": 1.0277883517320138e-06, + "loss": 1.334, + "step": 2396 + }, + { + "epoch": 2.462249614791988, + "grad_norm": 3.406104803085327, + "learning_rate": 1.025885039969547e-06, + "loss": 1.3665, + "step": 2397 + }, + { + "epoch": 2.463276836158192, + "grad_norm": 3.1286909580230713, + "learning_rate": 1.0239817282070806e-06, + "loss": 1.2641, + "step": 2398 + }, + { + "epoch": 2.4643040575243966, + "grad_norm": 3.285829544067383, + "learning_rate": 1.0220784164446137e-06, + "loss": 1.2917, + "step": 2399 + }, + { + "epoch": 2.4653312788906008, + "grad_norm": 3.352050542831421, + "learning_rate": 1.020175104682147e-06, + "loss": 1.3239, + "step": 2400 + }, + { + "epoch": 2.4653312788906008, + "eval_loss": 1.7855092287063599, + "eval_runtime": 37.2406, + "eval_samples_per_second": 26.852, + "eval_steps_per_second": 1.128, + "step": 2400 + }, + { + "epoch": 2.4663585002568054, + "grad_norm": 3.3660738468170166, + "learning_rate": 1.0182717929196802e-06, + "loss": 1.3201, + "step": 2401 + }, + { + "epoch": 2.4673857216230095, + "grad_norm": 3.3929147720336914, + "learning_rate": 1.0163684811572136e-06, + "loss": 1.3413, + "step": 2402 + }, + { + "epoch": 2.468412942989214, + "grad_norm": 3.0892271995544434, + "learning_rate": 1.014465169394747e-06, + "loss": 1.2744, + "step": 2403 + }, + { + "epoch": 2.4694401643554187, + "grad_norm": 3.190429449081421, + "learning_rate": 1.0125618576322803e-06, + "loss": 1.3146, + "step": 2404 + }, + { + "epoch": 2.470467385721623, + "grad_norm": 3.2355215549468994, + "learning_rate": 1.0106585458698137e-06, + "loss": 1.2959, + "step": 2405 + }, + { + "epoch": 2.4714946070878274, + "grad_norm": 3.207818031311035, + "learning_rate": 1.0087552341073468e-06, + "loss": 1.3046, + "step": 2406 + }, + { + "epoch": 2.472521828454032, + "grad_norm": 3.3168411254882812, + "learning_rate": 1.0068519223448802e-06, + "loss": 1.3838, + "step": 2407 + }, + { + "epoch": 2.473549049820236, + "grad_norm": 3.3504858016967773, + "learning_rate": 1.0049486105824133e-06, + "loss": 1.3596, + "step": 2408 + }, + { + "epoch": 2.4745762711864407, + "grad_norm": 3.269547939300537, + "learning_rate": 1.003045298819947e-06, + "loss": 1.3057, + "step": 2409 + }, + { + "epoch": 2.4756034925526453, + "grad_norm": 3.31899094581604, + "learning_rate": 1.00114198705748e-06, + "loss": 1.2969, + "step": 2410 + }, + { + "epoch": 2.4756034925526453, + "eval_loss": 1.7794221639633179, + "eval_runtime": 36.9306, + "eval_samples_per_second": 27.078, + "eval_steps_per_second": 1.137, + "step": 2410 + }, + { + "epoch": 2.4766307139188495, + "grad_norm": 3.4301860332489014, + "learning_rate": 9.992386752950134e-07, + "loss": 1.2956, + "step": 2411 + }, + { + "epoch": 2.477657935285054, + "grad_norm": 3.3026981353759766, + "learning_rate": 9.973353635325466e-07, + "loss": 1.2898, + "step": 2412 + }, + { + "epoch": 2.478685156651258, + "grad_norm": 3.268594980239868, + "learning_rate": 9.9543205177008e-07, + "loss": 1.3039, + "step": 2413 + }, + { + "epoch": 2.479712378017463, + "grad_norm": 3.2551498413085938, + "learning_rate": 9.935287400076133e-07, + "loss": 1.307, + "step": 2414 + }, + { + "epoch": 2.480739599383667, + "grad_norm": 3.2159249782562256, + "learning_rate": 9.916254282451467e-07, + "loss": 1.2874, + "step": 2415 + }, + { + "epoch": 2.4817668207498715, + "grad_norm": 3.4053354263305664, + "learning_rate": 9.8972211648268e-07, + "loss": 1.3227, + "step": 2416 + }, + { + "epoch": 2.482794042116076, + "grad_norm": 3.2459702491760254, + "learning_rate": 9.878188047202132e-07, + "loss": 1.3401, + "step": 2417 + }, + { + "epoch": 2.4838212634822803, + "grad_norm": 3.245961904525757, + "learning_rate": 9.859154929577465e-07, + "loss": 1.3137, + "step": 2418 + }, + { + "epoch": 2.484848484848485, + "grad_norm": 3.1001832485198975, + "learning_rate": 9.8401218119528e-07, + "loss": 1.2932, + "step": 2419 + }, + { + "epoch": 2.4858757062146895, + "grad_norm": 3.3570878505706787, + "learning_rate": 9.821088694328133e-07, + "loss": 1.3085, + "step": 2420 + }, + { + "epoch": 2.4858757062146895, + "eval_loss": 1.775665044784546, + "eval_runtime": 37.0412, + "eval_samples_per_second": 26.997, + "eval_steps_per_second": 1.134, + "step": 2420 + }, + { + "epoch": 2.4869029275808936, + "grad_norm": 3.271428346633911, + "learning_rate": 9.802055576703464e-07, + "loss": 1.3534, + "step": 2421 + }, + { + "epoch": 2.487930148947098, + "grad_norm": 3.348170518875122, + "learning_rate": 9.783022459078798e-07, + "loss": 1.3249, + "step": 2422 + }, + { + "epoch": 2.4889573703133023, + "grad_norm": 3.30704665184021, + "learning_rate": 9.76398934145413e-07, + "loss": 1.3222, + "step": 2423 + }, + { + "epoch": 2.489984591679507, + "grad_norm": 3.2533040046691895, + "learning_rate": 9.744956223829465e-07, + "loss": 1.3297, + "step": 2424 + }, + { + "epoch": 2.4910118130457115, + "grad_norm": 3.3263676166534424, + "learning_rate": 9.725923106204797e-07, + "loss": 1.3449, + "step": 2425 + }, + { + "epoch": 2.4920390344119157, + "grad_norm": 3.4388599395751953, + "learning_rate": 9.70688998858013e-07, + "loss": 1.3675, + "step": 2426 + }, + { + "epoch": 2.4930662557781202, + "grad_norm": 3.3554348945617676, + "learning_rate": 9.687856870955464e-07, + "loss": 1.3566, + "step": 2427 + }, + { + "epoch": 2.4940934771443244, + "grad_norm": 3.2762863636016846, + "learning_rate": 9.668823753330795e-07, + "loss": 1.3436, + "step": 2428 + }, + { + "epoch": 2.495120698510529, + "grad_norm": 3.267756700515747, + "learning_rate": 9.64979063570613e-07, + "loss": 1.2814, + "step": 2429 + }, + { + "epoch": 2.4961479198767336, + "grad_norm": 3.1882898807525635, + "learning_rate": 9.630757518081463e-07, + "loss": 1.32, + "step": 2430 + }, + { + "epoch": 2.4961479198767336, + "eval_loss": 1.773284673690796, + "eval_runtime": 37.0787, + "eval_samples_per_second": 26.97, + "eval_steps_per_second": 1.133, + "step": 2430 + }, + { + "epoch": 2.4971751412429377, + "grad_norm": 3.3851265907287598, + "learning_rate": 9.611724400456796e-07, + "loss": 1.3142, + "step": 2431 + }, + { + "epoch": 2.4982023626091423, + "grad_norm": 3.1993184089660645, + "learning_rate": 9.592691282832128e-07, + "loss": 1.3096, + "step": 2432 + }, + { + "epoch": 2.499229583975347, + "grad_norm": 3.3335726261138916, + "learning_rate": 9.573658165207461e-07, + "loss": 1.3204, + "step": 2433 + }, + { + "epoch": 2.500256805341551, + "grad_norm": 3.4124698638916016, + "learning_rate": 9.554625047582795e-07, + "loss": 1.3443, + "step": 2434 + }, + { + "epoch": 2.5012840267077556, + "grad_norm": 3.179321050643921, + "learning_rate": 9.535591929958129e-07, + "loss": 1.2961, + "step": 2435 + }, + { + "epoch": 2.50231124807396, + "grad_norm": 3.289297342300415, + "learning_rate": 9.516558812333461e-07, + "loss": 1.3301, + "step": 2436 + }, + { + "epoch": 2.5033384694401644, + "grad_norm": 3.3984432220458984, + "learning_rate": 9.497525694708794e-07, + "loss": 1.3457, + "step": 2437 + }, + { + "epoch": 2.5043656908063685, + "grad_norm": 3.187577247619629, + "learning_rate": 9.478492577084126e-07, + "loss": 1.3035, + "step": 2438 + }, + { + "epoch": 2.505392912172573, + "grad_norm": 3.1600699424743652, + "learning_rate": 9.459459459459461e-07, + "loss": 1.3141, + "step": 2439 + }, + { + "epoch": 2.5064201335387777, + "grad_norm": 3.2465484142303467, + "learning_rate": 9.440426341834794e-07, + "loss": 1.3051, + "step": 2440 + }, + { + "epoch": 2.5064201335387777, + "eval_loss": 1.774383783340454, + "eval_runtime": 37.0854, + "eval_samples_per_second": 26.965, + "eval_steps_per_second": 1.133, + "step": 2440 + }, + { + "epoch": 2.507447354904982, + "grad_norm": 3.156062602996826, + "learning_rate": 9.421393224210126e-07, + "loss": 1.3282, + "step": 2441 + }, + { + "epoch": 2.5084745762711864, + "grad_norm": 3.276942729949951, + "learning_rate": 9.402360106585459e-07, + "loss": 1.3267, + "step": 2442 + }, + { + "epoch": 2.509501797637391, + "grad_norm": 3.232865571975708, + "learning_rate": 9.383326988960792e-07, + "loss": 1.2864, + "step": 2443 + }, + { + "epoch": 2.510529019003595, + "grad_norm": 3.2006256580352783, + "learning_rate": 9.364293871336126e-07, + "loss": 1.2825, + "step": 2444 + }, + { + "epoch": 2.5115562403697997, + "grad_norm": 3.208831548690796, + "learning_rate": 9.345260753711458e-07, + "loss": 1.3212, + "step": 2445 + }, + { + "epoch": 2.5125834617360043, + "grad_norm": 3.470818042755127, + "learning_rate": 9.326227636086792e-07, + "loss": 1.3391, + "step": 2446 + }, + { + "epoch": 2.5136106831022085, + "grad_norm": 3.102159261703491, + "learning_rate": 9.307194518462125e-07, + "loss": 1.281, + "step": 2447 + }, + { + "epoch": 2.514637904468413, + "grad_norm": 3.358527183532715, + "learning_rate": 9.288161400837457e-07, + "loss": 1.3998, + "step": 2448 + }, + { + "epoch": 2.515665125834617, + "grad_norm": 3.2137224674224854, + "learning_rate": 9.269128283212792e-07, + "loss": 1.2889, + "step": 2449 + }, + { + "epoch": 2.516692347200822, + "grad_norm": 3.3090152740478516, + "learning_rate": 9.250095165588125e-07, + "loss": 1.3237, + "step": 2450 + }, + { + "epoch": 2.516692347200822, + "eval_loss": 1.7762006521224976, + "eval_runtime": 37.0787, + "eval_samples_per_second": 26.97, + "eval_steps_per_second": 1.133, + "step": 2450 + }, + { + "epoch": 2.517719568567026, + "grad_norm": 3.2170157432556152, + "learning_rate": 9.231062047963457e-07, + "loss": 1.3697, + "step": 2451 + }, + { + "epoch": 2.5187467899332305, + "grad_norm": 3.3608391284942627, + "learning_rate": 9.21202893033879e-07, + "loss": 1.3788, + "step": 2452 + }, + { + "epoch": 2.519774011299435, + "grad_norm": 3.219870090484619, + "learning_rate": 9.192995812714123e-07, + "loss": 1.3388, + "step": 2453 + }, + { + "epoch": 2.5208012326656393, + "grad_norm": 3.2731239795684814, + "learning_rate": 9.173962695089457e-07, + "loss": 1.3316, + "step": 2454 + }, + { + "epoch": 2.521828454031844, + "grad_norm": 3.22361421585083, + "learning_rate": 9.154929577464789e-07, + "loss": 1.3276, + "step": 2455 + }, + { + "epoch": 2.5228556753980484, + "grad_norm": 3.3190758228302, + "learning_rate": 9.135896459840123e-07, + "loss": 1.3341, + "step": 2456 + }, + { + "epoch": 2.5238828967642526, + "grad_norm": 3.310340642929077, + "learning_rate": 9.116863342215456e-07, + "loss": 1.3302, + "step": 2457 + }, + { + "epoch": 2.524910118130457, + "grad_norm": 3.40218448638916, + "learning_rate": 9.097830224590788e-07, + "loss": 1.3637, + "step": 2458 + }, + { + "epoch": 2.5259373394966618, + "grad_norm": 3.247948408126831, + "learning_rate": 9.078797106966122e-07, + "loss": 1.2817, + "step": 2459 + }, + { + "epoch": 2.526964560862866, + "grad_norm": 3.2093722820281982, + "learning_rate": 9.059763989341455e-07, + "loss": 1.3595, + "step": 2460 + }, + { + "epoch": 2.526964560862866, + "eval_loss": 1.7752339839935303, + "eval_runtime": 37.0975, + "eval_samples_per_second": 26.956, + "eval_steps_per_second": 1.132, + "step": 2460 + }, + { + "epoch": 2.5279917822290705, + "grad_norm": 3.136906385421753, + "learning_rate": 9.040730871716788e-07, + "loss": 1.2999, + "step": 2461 + }, + { + "epoch": 2.5290190035952747, + "grad_norm": 3.21834397315979, + "learning_rate": 9.021697754092121e-07, + "loss": 1.3255, + "step": 2462 + }, + { + "epoch": 2.5300462249614792, + "grad_norm": 3.209012508392334, + "learning_rate": 9.002664636467453e-07, + "loss": 1.3303, + "step": 2463 + }, + { + "epoch": 2.5310734463276834, + "grad_norm": 3.41849684715271, + "learning_rate": 8.983631518842788e-07, + "loss": 1.3389, + "step": 2464 + }, + { + "epoch": 2.532100667693888, + "grad_norm": 3.4514036178588867, + "learning_rate": 8.96459840121812e-07, + "loss": 1.3609, + "step": 2465 + }, + { + "epoch": 2.5331278890600926, + "grad_norm": 3.216822385787964, + "learning_rate": 8.945565283593453e-07, + "loss": 1.3249, + "step": 2466 + }, + { + "epoch": 2.5341551104262967, + "grad_norm": 3.2723228931427, + "learning_rate": 8.926532165968787e-07, + "loss": 1.3289, + "step": 2467 + }, + { + "epoch": 2.5351823317925013, + "grad_norm": 3.3893849849700928, + "learning_rate": 8.907499048344119e-07, + "loss": 1.3416, + "step": 2468 + }, + { + "epoch": 2.536209553158706, + "grad_norm": 3.221137285232544, + "learning_rate": 8.888465930719453e-07, + "loss": 1.2838, + "step": 2469 + }, + { + "epoch": 2.53723677452491, + "grad_norm": 3.2613704204559326, + "learning_rate": 8.869432813094786e-07, + "loss": 1.2877, + "step": 2470 + }, + { + "epoch": 2.53723677452491, + "eval_loss": 1.7816756963729858, + "eval_runtime": 37.1286, + "eval_samples_per_second": 26.933, + "eval_steps_per_second": 1.131, + "step": 2470 + }, + { + "epoch": 2.5382639958911146, + "grad_norm": 3.375920057296753, + "learning_rate": 8.850399695470119e-07, + "loss": 1.3325, + "step": 2471 + }, + { + "epoch": 2.539291217257319, + "grad_norm": 3.352329730987549, + "learning_rate": 8.831366577845452e-07, + "loss": 1.3432, + "step": 2472 + }, + { + "epoch": 2.5403184386235234, + "grad_norm": 3.358915090560913, + "learning_rate": 8.812333460220784e-07, + "loss": 1.3001, + "step": 2473 + }, + { + "epoch": 2.541345659989728, + "grad_norm": 3.193372964859009, + "learning_rate": 8.793300342596119e-07, + "loss": 1.2725, + "step": 2474 + }, + { + "epoch": 2.542372881355932, + "grad_norm": 3.3925843238830566, + "learning_rate": 8.774267224971451e-07, + "loss": 1.3274, + "step": 2475 + }, + { + "epoch": 2.5434001027221367, + "grad_norm": 3.3297488689422607, + "learning_rate": 8.755234107346784e-07, + "loss": 1.3639, + "step": 2476 + }, + { + "epoch": 2.544427324088341, + "grad_norm": 3.317533493041992, + "learning_rate": 8.736200989722117e-07, + "loss": 1.2929, + "step": 2477 + }, + { + "epoch": 2.5454545454545454, + "grad_norm": 3.2257425785064697, + "learning_rate": 8.71716787209745e-07, + "loss": 1.3219, + "step": 2478 + }, + { + "epoch": 2.54648176682075, + "grad_norm": 3.264587879180908, + "learning_rate": 8.698134754472784e-07, + "loss": 1.3602, + "step": 2479 + }, + { + "epoch": 2.547508988186954, + "grad_norm": 3.2435789108276367, + "learning_rate": 8.679101636848116e-07, + "loss": 1.262, + "step": 2480 + }, + { + "epoch": 2.547508988186954, + "eval_loss": 1.7793283462524414, + "eval_runtime": 37.1342, + "eval_samples_per_second": 26.929, + "eval_steps_per_second": 1.131, + "step": 2480 + }, + { + "epoch": 2.5485362095531587, + "grad_norm": 3.3055450916290283, + "learning_rate": 8.66006851922345e-07, + "loss": 1.3343, + "step": 2481 + }, + { + "epoch": 2.5495634309193633, + "grad_norm": 3.144941806793213, + "learning_rate": 8.641035401598783e-07, + "loss": 1.3228, + "step": 2482 + }, + { + "epoch": 2.5505906522855675, + "grad_norm": 3.2669522762298584, + "learning_rate": 8.622002283974115e-07, + "loss": 1.3119, + "step": 2483 + }, + { + "epoch": 2.551617873651772, + "grad_norm": 3.20893931388855, + "learning_rate": 8.60296916634945e-07, + "loss": 1.3551, + "step": 2484 + }, + { + "epoch": 2.5526450950179767, + "grad_norm": 3.2025182247161865, + "learning_rate": 8.583936048724782e-07, + "loss": 1.3146, + "step": 2485 + }, + { + "epoch": 2.553672316384181, + "grad_norm": 3.3333630561828613, + "learning_rate": 8.564902931100115e-07, + "loss": 1.3045, + "step": 2486 + }, + { + "epoch": 2.554699537750385, + "grad_norm": 3.3790230751037598, + "learning_rate": 8.545869813475448e-07, + "loss": 1.3243, + "step": 2487 + }, + { + "epoch": 2.5557267591165895, + "grad_norm": 3.2351009845733643, + "learning_rate": 8.526836695850781e-07, + "loss": 1.3264, + "step": 2488 + }, + { + "epoch": 2.556753980482794, + "grad_norm": 3.356888771057129, + "learning_rate": 8.507803578226115e-07, + "loss": 1.3461, + "step": 2489 + }, + { + "epoch": 2.5577812018489983, + "grad_norm": 3.36618971824646, + "learning_rate": 8.488770460601447e-07, + "loss": 1.3315, + "step": 2490 + }, + { + "epoch": 2.5577812018489983, + "eval_loss": 1.7821745872497559, + "eval_runtime": 37.2092, + "eval_samples_per_second": 26.875, + "eval_steps_per_second": 1.129, + "step": 2490 + }, + { + "epoch": 2.558808423215203, + "grad_norm": 3.404245376586914, + "learning_rate": 8.46973734297678e-07, + "loss": 1.297, + "step": 2491 + }, + { + "epoch": 2.5598356445814074, + "grad_norm": 3.181821823120117, + "learning_rate": 8.450704225352114e-07, + "loss": 1.3111, + "step": 2492 + }, + { + "epoch": 2.5608628659476116, + "grad_norm": 3.281442165374756, + "learning_rate": 8.431671107727446e-07, + "loss": 1.301, + "step": 2493 + }, + { + "epoch": 2.561890087313816, + "grad_norm": 3.3224937915802, + "learning_rate": 8.41263799010278e-07, + "loss": 1.3311, + "step": 2494 + }, + { + "epoch": 2.5629173086800208, + "grad_norm": 3.2567851543426514, + "learning_rate": 8.393604872478113e-07, + "loss": 1.3212, + "step": 2495 + }, + { + "epoch": 2.563944530046225, + "grad_norm": 3.241591691970825, + "learning_rate": 8.374571754853446e-07, + "loss": 1.3058, + "step": 2496 + }, + { + "epoch": 2.5649717514124295, + "grad_norm": 3.2794666290283203, + "learning_rate": 8.355538637228778e-07, + "loss": 1.3086, + "step": 2497 + }, + { + "epoch": 2.5659989727786336, + "grad_norm": 3.243818759918213, + "learning_rate": 8.336505519604111e-07, + "loss": 1.3095, + "step": 2498 + }, + { + "epoch": 2.5670261941448382, + "grad_norm": 3.2905161380767822, + "learning_rate": 8.317472401979445e-07, + "loss": 1.3393, + "step": 2499 + }, + { + "epoch": 2.5680534155110424, + "grad_norm": 3.3779470920562744, + "learning_rate": 8.298439284354778e-07, + "loss": 1.3732, + "step": 2500 + }, + { + "epoch": 2.5680534155110424, + "eval_loss": 1.78448486328125, + "eval_runtime": 37.3763, + "eval_samples_per_second": 26.755, + "eval_steps_per_second": 1.124, + "step": 2500 + } + ], + "logging_steps": 1, + "max_steps": 2919, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 100, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 7.7932735234048e+17, + "train_batch_size": 3, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8 --- /dev/null +++ b/checkpoints/Llama-3.2-3B/babylm_reverse_full_10M_seed0/runs/checkpoint-2500/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters)