Import experiment archive from bart (batch 4)
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_004500_rank1.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_004500_rank2.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_004500_rank3.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005000_rank0.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005000_rank1.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005000_rank2.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005000_rank3.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005500_rank0.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005500_rank1.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005500_rank2.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005500_rank3.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006000_rank0.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006000_rank1.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006000_rank2.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006000_rank3.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006500_rank0.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006500_rank1.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006500_rank2.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006500_rank3.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007000_rank0.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007000_rank1.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007000_rank2.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007000_rank3.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007500_rank0.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007500_rank1.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007500_rank2.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007500_rank3.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008000_rank0.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008000_rank1.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008000_rank2.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008000_rank3.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008352_rank0.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008352_rank1.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008352_rank2.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008352_rank3.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/config.json +63 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/evals/core.json +56 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/evals/samples.json +48 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/evals/val_bpb.json +12 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/run.json +10 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-nanochat-default-v1/config.json +36 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-nanochat-default-v1/run.json +11 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/checkpoints/meta_000007.json +103 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/checkpoints/model_000007.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/checkpoints/optim_000007_rank0.pt +3 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/config.json +32 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/run.json +10 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0/checkpoints/eval_metrics.json +30 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0/checkpoints/meta_000052.json +161 -0
- experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0/checkpoints/model_000052.pt +3 -0
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_004500_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eecc5243a2ed93ac65bea8de8373ff27e5208b62b6b17d31b14c9a68eb37394b
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_004500_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:29b34641c29769169a1ed6b4afd58e9e307067cada1e9e88c72966e0144f438d
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_004500_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c13ffbd7e162bf3c9a022393fd3901c55e8db1cf896cb08c0bdee97c2b6dc83c
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005000_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ff0cc947e18af1a04987dc2f64fa0a29d22418d34efdff24027a1798c7cea94
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005000_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:085b891a2d75abe7ff9687ec43f89c444604eb83bbf0d5f903b21675eb51160f
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005000_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7f20218827152f212517747857afab2b7349aa5a043d96c23f296a5907d717ba
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005000_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b94746b3fa690d69a9ba3e5cbde7c7da3e1f0aabb50b73c9f3b1326f4fc38743
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005500_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:225af4d6d4d822abbce34e63d621eac376be5b8acf857ba8997e36582762573a
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005500_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bda69aaa71c4ac52b0c4658c58b7d07f8671608af2f71a512e983da688149ad6
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005500_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c39f5feb695f84f82218defbe6c535b506173b550fd3ae6fe0caa20278817f3f
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_005500_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b806a61d71740cbb0e2b088ea4c42165334ecd9863494c27a343b548b0d84326
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006000_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7e168383bc13bbc56df91c21f3d46e835001ed592f90f849af53993f9c965a3
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006000_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a068ed26f38b48c5c7ef70a1771da690ab6cdd1749d03f902277b5b2a553f94e
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006000_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f63a2df3dbbe97e7bbf952c0fc0eb732423d49fe7e994c9b632e1342d9ac8868
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006000_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:60613cd93ddac30dd28676e47ae16588c2ad85aa09b9bd4f63f17a5ad73dfc07
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006500_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7a5fd4763414c513d80406c35ecdf9f26ec5d1d9d27a676e39dfd01c8e6992b7
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006500_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b21033cf7c8929fe01be78ca45b157751d65c7dff5cb57940226319a9d76d9c
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006500_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99ac47912efb2feda2492d9f772370071b090843ee13e55a08e1e31d7068926a
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_006500_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:10a5f118f4f3537eaf2230fbeef6fe6e9298958787e807116fadf09f5fc0e04e
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007000_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d27d602d30b1a4d7aa9605fa175739fb161b32e02909ec30c231e3fde3be8264
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007000_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e638779e6afd341c5453a5de23e7c834f305837bd70cc5a0fc12b277cd7ee9fc
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007000_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a1afe22caa4f2819cfc397846b74b8c26272f6266a16746d31f3a87bac3a5c37
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007000_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:acbcb96c068e6d855d7619f22c370140f31a894c4d23f105c7522214d8a51ae4
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007500_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c8beda301552c587b21e718b3f0eb383006dbb232a26d5951b9132153346c4d
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007500_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7db4c7619d2c479f1b6a41805b4fc74865c7f4563cf71d8105af4ecc6f9ffd27
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007500_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cdc736fbeef52ca1bfe7e975b93ad550bfc35129ba2e25eddad9ecf59d3fc3bf
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_007500_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:30523b0f3a84b56579b1843151587d357fbae481de2e0885f92b484ed6356e2e
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008000_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:11ff33d778dc9ba30dc3fda9174bfe1b73fd251370a901f1e396ab7de520028e
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008000_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fa40a10a4058f06d01d3393321e236973fa78f2d137d1e02d7ca4c119b7305e6
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008000_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:22d887c0fc65e71df1f3d74b5ea3656ca860658d8249aa8cac96421e89a7a835
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008000_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fc131b1f33fa81372bcc1ecb759f2b7ddf2acf55652b07e1a14b26b96a909e73
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008352_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ef67bbc1c6e6f73cf3ccfe50d3df3e916b2cf913a2ef1fb6a14123e784e4bc13
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008352_rank1.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:839d78c2d9e329f91167b3ca2a706ba931d4fc946101826a6bbac664b3c366b1
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008352_rank2.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7987c59cd7ef08458fcdf58f4e737501e3adf2e5a32cd0e44015ba70ee68d29b
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints/optim_008352_rank3.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4b5c08a5e6fd49981ac6887b43d4d835742ef20e8eefa060670635dedc0c69de
|
| 3 |
+
size 1434913073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/config.json
ADDED
|
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"schema_version": 1,
|
| 3 |
+
"stage": "base",
|
| 4 |
+
"experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1",
|
| 5 |
+
"dataset": {
|
| 6 |
+
"adapter": "parquet_shards",
|
| 7 |
+
"repo": "jbduran/think-dataset-clean-1930s",
|
| 8 |
+
"revision": "main",
|
| 9 |
+
"validation_shard": 472,
|
| 10 |
+
"num_train_shards": 180,
|
| 11 |
+
"download_workers": 4
|
| 12 |
+
},
|
| 13 |
+
"tokenizer": {
|
| 14 |
+
"mode": "reuse",
|
| 15 |
+
"source_experiment_id": "clean1930s-d24-r12-ctx4096-fulltok-v1",
|
| 16 |
+
"vocab_size": 32768
|
| 17 |
+
},
|
| 18 |
+
"pretokenize": {
|
| 19 |
+
"enabled": true,
|
| 20 |
+
"slack": 1.03,
|
| 21 |
+
"require_no_wrap": true,
|
| 22 |
+
"val_tokens": 20971520,
|
| 23 |
+
"shard_tokens": 100000000,
|
| 24 |
+
"tokenizer_threads": 8
|
| 25 |
+
},
|
| 26 |
+
"training": {
|
| 27 |
+
"depth": 24,
|
| 28 |
+
"scaling_params": 729810624,
|
| 29 |
+
"target_param_data_ratio": 12,
|
| 30 |
+
"max_seq_len": 4096,
|
| 31 |
+
"window_pattern": "SSSL",
|
| 32 |
+
"device_batch_size": 8,
|
| 33 |
+
"total_batch_size": 1048576,
|
| 34 |
+
"fp8": true,
|
| 35 |
+
"fp8_recipe": "tensorwise",
|
| 36 |
+
"seed": 42,
|
| 37 |
+
"save_every": 500,
|
| 38 |
+
"eval_every": 250,
|
| 39 |
+
"eval_tokens": 2097152,
|
| 40 |
+
"core_metric_every": -1,
|
| 41 |
+
"sample_every": -1
|
| 42 |
+
},
|
| 43 |
+
"artifacts": {
|
| 44 |
+
"repo": "jbduran/think.nano",
|
| 45 |
+
"keep_local_checkpoints": 1
|
| 46 |
+
},
|
| 47 |
+
"wandb": {
|
| 48 |
+
"entity": "jbduran-thinkingmachinesncsu",
|
| 49 |
+
"project": "think.nano",
|
| 50 |
+
"name": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1",
|
| 51 |
+
"group": "clean1930s-d24",
|
| 52 |
+
"tags": [
|
| 53 |
+
"think-dataset-clean-1930s",
|
| 54 |
+
"d24",
|
| 55 |
+
"ratio12",
|
| 56 |
+
"ctx4096",
|
| 57 |
+
"sssl",
|
| 58 |
+
"fp8"
|
| 59 |
+
]
|
| 60 |
+
},
|
| 61 |
+
"config_fingerprint": "166e7e695c1b536a",
|
| 62 |
+
"artifact_path": "experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1"
|
| 63 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/evals/core.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": "base_model (step 8352)",
|
| 3 |
+
"step": 8352,
|
| 4 |
+
"bpb": {},
|
| 5 |
+
"core_metric": 0.13542864491271994,
|
| 6 |
+
"core_results": {
|
| 7 |
+
"hellaswag_zeroshot": 0.32364070415496826,
|
| 8 |
+
"jeopardy": 0.0037789323832839727,
|
| 9 |
+
"bigbench_qa_wikidata": 0.2516116201877594,
|
| 10 |
+
"arc_easy": 0.41750839352607727,
|
| 11 |
+
"arc_challenge": 0.239761084318161,
|
| 12 |
+
"copa": 0.6100000143051147,
|
| 13 |
+
"commonsense_qa": 0.3071253001689911,
|
| 14 |
+
"piqa": 0.5805222988128662,
|
| 15 |
+
"openbook_qa": 0.2580000162124634,
|
| 16 |
+
"lambada_openai": 0.35338637232780457,
|
| 17 |
+
"hellaswag": 0.32911768555641174,
|
| 18 |
+
"winograd": 0.6703296899795532,
|
| 19 |
+
"winogrande": 0.49408048391342163,
|
| 20 |
+
"bigbench_dyck_languages": 0.12700000405311584,
|
| 21 |
+
"agi_eval_lsat_ar": 0.260869562625885,
|
| 22 |
+
"bigbench_cs_algorithms": 0.38333332538604736,
|
| 23 |
+
"bigbench_operators": 0.11428572237491608,
|
| 24 |
+
"bigbench_repeat_copy_logic": 0.0,
|
| 25 |
+
"squad": 0.114096499979496,
|
| 26 |
+
"coqa": 0.16610296070575714,
|
| 27 |
+
"boolq": 0.5969418883323669,
|
| 28 |
+
"bigbench_language_identification": 0.25699999928474426
|
| 29 |
+
},
|
| 30 |
+
"centered_results": {
|
| 31 |
+
"hellaswag_zeroshot": 0.09818760553995769,
|
| 32 |
+
"jeopardy": 0.0037789323832839727,
|
| 33 |
+
"bigbench_qa_wikidata": 0.2516116201877594,
|
| 34 |
+
"arc_easy": 0.22334452470143637,
|
| 35 |
+
"arc_challenge": -0.013651887575785318,
|
| 36 |
+
"copa": 0.2200000286102295,
|
| 37 |
+
"commonsense_qa": 0.13390662521123883,
|
| 38 |
+
"piqa": 0.16104459762573242,
|
| 39 |
+
"openbook_qa": 0.010666688283284506,
|
| 40 |
+
"lambada_openai": 0.35338637232780457,
|
| 41 |
+
"hellaswag": 0.105490247408549,
|
| 42 |
+
"winograd": 0.34065937995910645,
|
| 43 |
+
"winogrande": -0.011839032173156738,
|
| 44 |
+
"bigbench_dyck_languages": 0.12700000405311584,
|
| 45 |
+
"agi_eval_lsat_ar": 0.07608695328235625,
|
| 46 |
+
"bigbench_cs_algorithms": 0.38333332538604736,
|
| 47 |
+
"bigbench_operators": 0.11428572237491608,
|
| 48 |
+
"bigbench_repeat_copy_logic": 0.0,
|
| 49 |
+
"squad": 0.114096499979496,
|
| 50 |
+
"coqa": 0.16610296070575714,
|
| 51 |
+
"boolq": -0.060679241230613294,
|
| 52 |
+
"bigbench_language_identification": 0.1826182610393226
|
| 53 |
+
},
|
| 54 |
+
"conditioned_samples": [],
|
| 55 |
+
"unconditioned_samples": []
|
| 56 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/evals/samples.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": "base_model (step 8352)",
|
| 3 |
+
"step": 8352,
|
| 4 |
+
"bpb": {},
|
| 5 |
+
"core_metric": null,
|
| 6 |
+
"core_results": null,
|
| 7 |
+
"centered_results": null,
|
| 8 |
+
"conditioned_samples": [
|
| 9 |
+
{
|
| 10 |
+
"prompt": "The capital of France is",
|
| 11 |
+
"text": "<|bos|>The capital of France is the city of Paris, which is the capital of France. The capital of France"
|
| 12 |
+
},
|
| 13 |
+
{
|
| 14 |
+
"prompt": "The chemical symbol of gold is",
|
| 15 |
+
"text": "<|bos|>The chemical symbol of gold is Au, and the symbol of silver is Ag. The symbol of gold is Au"
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"prompt": "If yesterday was Friday, then tomorrow will be",
|
| 19 |
+
"text": "<|bos|>If yesterday was Friday, then tomorrow will be Saturday. If yesterday was Friday, then tomorrow will be Saturday. If yesterday was"
|
| 20 |
+
},
|
| 21 |
+
{
|
| 22 |
+
"prompt": "The opposite of hot is",
|
| 23 |
+
"text": "<|bos|>The opposite of hot is cold.\n\nThe opposite of cold is hot.\n\nThe opposite of hot is cold.\n\n"
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"prompt": "The planets of the solar system are:",
|
| 27 |
+
"text": "<|bos|>The planets of the solar system are: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus,"
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"prompt": "My favorite color is",
|
| 31 |
+
"text": "<|bos|>My favorite color is a dark brown, with a tinge of red. It is a very pretty color"
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"prompt": "If 5*x + 3 = 13, then x is",
|
| 35 |
+
"text": "<|bos|>If 5*x + 3 = 13, then x is 13, and the equation is 13x + 3 = 13"
|
| 36 |
+
}
|
| 37 |
+
],
|
| 38 |
+
"unconditioned_samples": [
|
| 39 |
+
"<|bos|>PREFACE.\n\nTHIS essay was contributed to the \"Psychological Bulletin,\" of the University of Chicago, during the past summer; the temporary arrangements for its printing prevented its meeting the demand for it. Fair play in the matter of results in the revision of Shakespeare's plays has not been given in this field in this company, Dunning, Uzziel, and others engaging to contribute whatever they have to say on the revision and construction of even a very few plays, and the magazine enterprise to furnish more plays of a classic nature, although comprising many dramas of merit and of the utmost possible variety, is by no means of sufficient scope and interest",
|
| 40 |
+
"<|bos|>\n\nARTON LIBRARY OF THE DEZA SUPERINTENDENTS AND BOOKKEEPERS\n\nThe following Outline of Business\n\nthe State or Territory in which\n\nFoss work is to be carried on.\n\nPLAN OF BUSINESS Department 1. Traveling\n\n2. Records, coupons, checks, mercantile accounts, vouchers, &c.\n\n3. General office\n\n4. Subsidiary agencies. Persons dealing with department men.\n\nPlanwork.\n\nPlan.\n\n1. Plans for the method of review of machinery, etc. Uncertainty.\n\n2. Plans for securing uniformity of practice over a large field",
|
| 41 |
+
"<|bos|>!\n\nIt's the coss is mighty wonderous weak Thefe minutes yhere hain marry teef The tryfector\n\nOf my fandd in Life who ought to hast The pist o' a President\n\nApril 10. J,\"\n\nAmong the Arms in cison\n\nThe Arms of the President *The Arms of Washington\n\nOn October 5, (?) in another and more noticeable storm-swept plagiarism, the President bore the baptismal emblem of the United States; on the way to Washington, January 1, 1799, occurred the cancel that has made possible the interpolations of the following",
|
| 42 |
+
"<|bos|> ministered unto Him with great joy.\"\n\nXX. 1. CHIRIT OF THE BLESSED.\n\nThe full contrast to this joy is the hunger that followed when He left His throne in heaven, ascended, and took the place of the \"Name\" lifted up on Calvary.\n\n2. The Fulfilment of the Tabernacle: In our imagination let 50 Enter: first, the Church universal: prayer, glorification, rejoicing, nothing but an eternal army and the eternal joy which forms one side of this-and well! for consequently this is that infant Church which a voice descended, and the Redeemer Sunday broke through the wall that",
|
| 43 |
+
"<|bos|>eed\n\nNew York paid\n\nEach separate $1.25 and i piece by akely ried in Watts averaged 75 cents.\n\nRetail price 50 cents-No longer find it worth while to order second-hand books. Already considerable order great publishing interests are sending agents to England or Scotland to our mmission depot expressly to collect the $1.25 per set, on which income is made over to us. The influens can gain fifteen per cent, and this can be taken care of at once as it is here almost never debited. Its great advantages that has the most of A. L. Burlingame's",
|
| 44 |
+
"<|bos|>PREFACE\n\nWHEN PERSHAD was about to issue and was being attacked by the fanatical cuttlefish of Northern India, some weeks before at least the Calcutta of 1862 stood face to face with a panic-crazed gale, the doctors began one day on their platform to advise the world how they were going to appease the spirit of the Fish-god. The Elders and the Lady-Seconds criticized emphatically the pros and cons, and the whole city stared.\n\nOne man said: \"I am Oriental; the people of this English country do not believe in all this agnostic nonsense, and the skies when they are",
|
| 45 |
+
"<|bos|>iman have published,\n\nTo be completed in about 100 Monthly Parts, each illustrated by a Steel Engraving, a Work consisting of a Library of Original 180 Engravings, 3 vols. 8vo. \u00a35. 5s. strongly bound in cloth, as a Work for convenient\n\nSale by STEEL Engraved by W. L. STEENERSON, each neatly illustrated by a Steel Engraving; the Plates are in a very high state of\n\nPerfection, and warranted to give a correct Image of each Novel.\n\nThe Work is published in Monthly Parts to contain Historical and Biographical Sketches of the principal",
|
| 46 |
+
"<|bos|>MYSTER LIBRARY.\n\n3 2044 009 603 166 fidh\n\nQUE\n\n7808, W"
|
| 47 |
+
]
|
| 48 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/evals/val_bpb.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": "base_model (step 8352)",
|
| 3 |
+
"step": 8352,
|
| 4 |
+
"bpb": {
|
| 5 |
+
"val": 0.8435203902195944
|
| 6 |
+
},
|
| 7 |
+
"core_metric": null,
|
| 8 |
+
"core_results": null,
|
| 9 |
+
"centered_results": null,
|
| 10 |
+
"conditioned_samples": [],
|
| 11 |
+
"unconditioned_samples": []
|
| 12 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/run.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1",
|
| 3 |
+
"stage": "base",
|
| 4 |
+
"base_experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1",
|
| 5 |
+
"parent_experiment_id": null,
|
| 6 |
+
"parent_checkpoint_step": null,
|
| 7 |
+
"config_fingerprint": "166e7e695c1b536a",
|
| 8 |
+
"wandb_run_id": "e69c1e59",
|
| 9 |
+
"created_at": 1784557543
|
| 10 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-nanochat-default-v1/config.json
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"schema_version": 1,
|
| 3 |
+
"stage": "sft",
|
| 4 |
+
"experiment_suffix": "nanochat-default-v1",
|
| 5 |
+
"data": {
|
| 6 |
+
"recipe": "nanochat-default",
|
| 7 |
+
"mmlu_epochs": 3,
|
| 8 |
+
"gsm8k_epochs": 4
|
| 9 |
+
},
|
| 10 |
+
"training": {
|
| 11 |
+
"num_iterations": -1,
|
| 12 |
+
"load_optimizer": 0,
|
| 13 |
+
"device_batch_size": 8,
|
| 14 |
+
"eval_every": -1,
|
| 15 |
+
"chatcore_every": 100000,
|
| 16 |
+
"chatcore_max_sample": 32,
|
| 17 |
+
"save_every": 200
|
| 18 |
+
},
|
| 19 |
+
"artifacts": {
|
| 20 |
+
"repo": "jbduran/think.nano"
|
| 21 |
+
},
|
| 22 |
+
"wandb": {
|
| 23 |
+
"enabled": false,
|
| 24 |
+
"entity": "jbduran-thinkingmachinesncsu",
|
| 25 |
+
"project": "think.nano",
|
| 26 |
+
"group": "clean1930s-d24-r12",
|
| 27 |
+
"tags": [
|
| 28 |
+
"sft",
|
| 29 |
+
"smoltalk",
|
| 30 |
+
"mmlu3",
|
| 31 |
+
"gsm8k4"
|
| 32 |
+
]
|
| 33 |
+
},
|
| 34 |
+
"config_fingerprint": "103d7c3526aa35c0",
|
| 35 |
+
"artifact_path": "experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-nanochat-default-v1"
|
| 36 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-nanochat-default-v1/run.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-nanochat-default-v1",
|
| 3 |
+
"stage": "sft",
|
| 4 |
+
"base_experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1",
|
| 5 |
+
"parent_experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1",
|
| 6 |
+
"parent_checkpoint_step": 8352,
|
| 7 |
+
"branch_parent_step": null,
|
| 8 |
+
"config_fingerprint": "103d7c3526aa35c0",
|
| 9 |
+
"wandb_run_id": null,
|
| 10 |
+
"created_at": 1786512279
|
| 11 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/checkpoints/meta_000007.json
ADDED
|
@@ -0,0 +1,103 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"step": 7,
|
| 3 |
+
"training_complete": true,
|
| 4 |
+
"val_bpb": 0.7205382880492125,
|
| 5 |
+
"model_config": {
|
| 6 |
+
"sequence_len": 4096,
|
| 7 |
+
"vocab_size": 32768,
|
| 8 |
+
"n_layer": 24,
|
| 9 |
+
"n_head": 12,
|
| 10 |
+
"n_kv_head": 12,
|
| 11 |
+
"n_embd": 1536,
|
| 12 |
+
"window_pattern": "SSSL"
|
| 13 |
+
},
|
| 14 |
+
"user_config": {
|
| 15 |
+
"run": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic",
|
| 16 |
+
"wandb_run_id": "0837926f",
|
| 17 |
+
"wandb_group": "think-d12",
|
| 18 |
+
"wandb_tags": "sft,pre1930,ratio20",
|
| 19 |
+
"device_type": "",
|
| 20 |
+
"model_tag": null,
|
| 21 |
+
"model_step": null,
|
| 22 |
+
"base_checkpoint_dir": "/content/nanochat_cache/experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints",
|
| 23 |
+
"base_step": 8352,
|
| 24 |
+
"checkpoint_dir": "/content/nanochat_cache/experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/checkpoints",
|
| 25 |
+
"tokenizer_dir": "/content/nanochat_cache/experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/tokenizer",
|
| 26 |
+
"resume_from_step": null,
|
| 27 |
+
"experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic",
|
| 28 |
+
"experiment_config": "/content/nanochat_cache/experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/config.json",
|
| 29 |
+
"parent_cumulative_flops": 4.5291244139996774e+19,
|
| 30 |
+
"tokenizer_fingerprint": "21e99d5cdeeaa660",
|
| 31 |
+
"git_commit_sha": "60aa7005265901858554bb7ca98cd788de721e4a",
|
| 32 |
+
"load_optimizer": 1,
|
| 33 |
+
"num_iterations": -1,
|
| 34 |
+
"max_seq_len": null,
|
| 35 |
+
"device_batch_size": 8,
|
| 36 |
+
"total_batch_size": null,
|
| 37 |
+
"embedding_lr": null,
|
| 38 |
+
"unembedding_lr": null,
|
| 39 |
+
"matrix_lr": null,
|
| 40 |
+
"init_lr_frac": 0.8,
|
| 41 |
+
"warmup_ratio": 0.0,
|
| 42 |
+
"warmdown_ratio": 0.5,
|
| 43 |
+
"final_lr_frac": 0.0,
|
| 44 |
+
"eval_every": -1,
|
| 45 |
+
"eval_tokens": 20971520,
|
| 46 |
+
"chatcore_every": -1,
|
| 47 |
+
"chatcore_max_cat": -1,
|
| 48 |
+
"chatcore_max_sample": 24,
|
| 49 |
+
"save_every": -1,
|
| 50 |
+
"recipe": "pre1930",
|
| 51 |
+
"pre1930_epochs": 5,
|
| 52 |
+
"mmlu_epochs": 3,
|
| 53 |
+
"gsm8k_epochs": 4,
|
| 54 |
+
"resolved_experiment_config": {
|
| 55 |
+
"schema_version": 1,
|
| 56 |
+
"stage": "sft",
|
| 57 |
+
"experiment_suffix": "pre1930-authentic",
|
| 58 |
+
"data": {
|
| 59 |
+
"recipe": "pre1930",
|
| 60 |
+
"pre1930_epochs": 5
|
| 61 |
+
},
|
| 62 |
+
"training": {
|
| 63 |
+
"num_iterations": -1,
|
| 64 |
+
"device_batch_size": 8,
|
| 65 |
+
"eval_every": -1,
|
| 66 |
+
"chatcore_every": -1,
|
| 67 |
+
"save_every": -1
|
| 68 |
+
},
|
| 69 |
+
"artifacts": {
|
| 70 |
+
"repo": "jbduran/think.nano"
|
| 71 |
+
},
|
| 72 |
+
"wandb": {
|
| 73 |
+
"enabled": true,
|
| 74 |
+
"entity": "jbduran-thinkingmachinesncsu",
|
| 75 |
+
"project": "think.nano",
|
| 76 |
+
"group": "think-d12",
|
| 77 |
+
"tags": [
|
| 78 |
+
"sft",
|
| 79 |
+
"pre1930",
|
| 80 |
+
"ratio20"
|
| 81 |
+
]
|
| 82 |
+
},
|
| 83 |
+
"config_fingerprint": "d3378357cef17359",
|
| 84 |
+
"artifact_path": "experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic"
|
| 85 |
+
},
|
| 86 |
+
"stage": "sft",
|
| 87 |
+
"base_experiment_id": null,
|
| 88 |
+
"parent_experiment_id": null,
|
| 89 |
+
"parent_checkpoint_step": null,
|
| 90 |
+
"config_fingerprint": "d3378357cef17359"
|
| 91 |
+
},
|
| 92 |
+
"loop_state": {
|
| 93 |
+
"step": 7,
|
| 94 |
+
"total_training_time": 0.0,
|
| 95 |
+
"min_val_bpb": 0.7205382880492125,
|
| 96 |
+
"smooth_train_loss": 1.187201474390888,
|
| 97 |
+
"mfu": 35.33209107765201,
|
| 98 |
+
"tok_per_sec": 21315,
|
| 99 |
+
"stage_training_flops": 3.79596155387904e+16,
|
| 100 |
+
"inherited_parent_flops": 4.5291244139996774e+19,
|
| 101 |
+
"cumulative_pipeline_training_flops": 4.5329203755535565e+19
|
| 102 |
+
}
|
| 103 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/checkpoints/model_000007.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3c16eb054f43f54f05624b42a4ec3fbcab666b80d99c3c87ac6ee8b34261a131
|
| 3 |
+
size 4227936073
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/checkpoints/optim_000007_rank0.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c53d48431c7459eae0e562672d69b2dc15a15f91e1243ce120d54ab6b64ba458
|
| 3 |
+
size 5739601781
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"schema_version": 1,
|
| 3 |
+
"stage": "sft",
|
| 4 |
+
"experiment_suffix": "pre1930-authentic",
|
| 5 |
+
"data": {
|
| 6 |
+
"recipe": "pre1930",
|
| 7 |
+
"pre1930_epochs": 5
|
| 8 |
+
},
|
| 9 |
+
"training": {
|
| 10 |
+
"num_iterations": -1,
|
| 11 |
+
"device_batch_size": 8,
|
| 12 |
+
"eval_every": -1,
|
| 13 |
+
"chatcore_every": -1,
|
| 14 |
+
"save_every": -1
|
| 15 |
+
},
|
| 16 |
+
"artifacts": {
|
| 17 |
+
"repo": "jbduran/think.nano"
|
| 18 |
+
},
|
| 19 |
+
"wandb": {
|
| 20 |
+
"enabled": true,
|
| 21 |
+
"entity": "jbduran-thinkingmachinesncsu",
|
| 22 |
+
"project": "think.nano",
|
| 23 |
+
"group": "think-d12",
|
| 24 |
+
"tags": [
|
| 25 |
+
"sft",
|
| 26 |
+
"pre1930",
|
| 27 |
+
"ratio20"
|
| 28 |
+
]
|
| 29 |
+
},
|
| 30 |
+
"config_fingerprint": "d3378357cef17359",
|
| 31 |
+
"artifact_path": "experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic"
|
| 32 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic/run.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-authentic",
|
| 3 |
+
"stage": "sft",
|
| 4 |
+
"base_experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1",
|
| 5 |
+
"parent_experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1",
|
| 6 |
+
"parent_checkpoint_step": null,
|
| 7 |
+
"config_fingerprint": "d3378357cef17359",
|
| 8 |
+
"wandb_run_id": "0837926f",
|
| 9 |
+
"created_at": 1784577842
|
| 10 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0/checkpoints/eval_metrics.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0",
|
| 3 |
+
"recipe": "curriculum",
|
| 4 |
+
"step": 52,
|
| 5 |
+
"val_bpb": 0.8869868096604953,
|
| 6 |
+
"min_val_bpb": 0.8464286315951827,
|
| 7 |
+
"per_route_bpb": {},
|
| 8 |
+
"per_domain_bpb": {},
|
| 9 |
+
"curriculum_summary": null,
|
| 10 |
+
"chatcore": {
|
| 11 |
+
"chatcore_metric": -0.006895505470088888,
|
| 12 |
+
"chatcore_cat": -0.011492509116814813,
|
| 13 |
+
"suite": {
|
| 14 |
+
"tasks": [
|
| 15 |
+
"ARC-Easy",
|
| 16 |
+
"ARC-Challenge",
|
| 17 |
+
"MMLU",
|
| 18 |
+
"GSM8K",
|
| 19 |
+
"SpellingBee"
|
| 20 |
+
],
|
| 21 |
+
"max_generative_problems": 32,
|
| 22 |
+
"generative_answer_format": "End your response with #### followed by the final numeric answer (for example: #### 42)."
|
| 23 |
+
},
|
| 24 |
+
"ARC-Easy": 0.24494949494949494,
|
| 25 |
+
"ARC-Challenge": 0.2354948805460751,
|
| 26 |
+
"MMLU": 0.24369747899159663,
|
| 27 |
+
"GSM8K": 0.0,
|
| 28 |
+
"SpellingBee": 0.0
|
| 29 |
+
}
|
| 30 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0/checkpoints/meta_000052.json
ADDED
|
@@ -0,0 +1,161 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"step": 52,
|
| 3 |
+
"training_complete": true,
|
| 4 |
+
"val_bpb": 0.8869868096604953,
|
| 5 |
+
"model_config": {
|
| 6 |
+
"sequence_len": 4096,
|
| 7 |
+
"vocab_size": 32768,
|
| 8 |
+
"n_layer": 24,
|
| 9 |
+
"n_head": 12,
|
| 10 |
+
"n_kv_head": 12,
|
| 11 |
+
"n_embd": 1536,
|
| 12 |
+
"window_pattern": "SSSL"
|
| 13 |
+
},
|
| 14 |
+
"user_config": {
|
| 15 |
+
"run": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0",
|
| 16 |
+
"wandb_run_id": "fb8277f3",
|
| 17 |
+
"wandb_group": "think-d12",
|
| 18 |
+
"wandb_tags": "sft,curriculum,c0,baseline",
|
| 19 |
+
"device_type": "",
|
| 20 |
+
"model_tag": null,
|
| 21 |
+
"model_step": null,
|
| 22 |
+
"base_checkpoint_dir": "/workspace/nanochat-state/experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/base_checkpoints",
|
| 23 |
+
"base_step": 8352,
|
| 24 |
+
"checkpoint_dir": "/workspace/nanochat-state/experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0/checkpoints",
|
| 25 |
+
"tokenizer_dir": "/workspace/nanochat-state/experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/tokenizer",
|
| 26 |
+
"resume_from_step": null,
|
| 27 |
+
"experiment_id": "clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0",
|
| 28 |
+
"experiment_config": "/workspace/nanochat-state/experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0/config.json",
|
| 29 |
+
"parent_cumulative_flops": 4.5291244139996774e+19,
|
| 30 |
+
"tokenizer_fingerprint": "21e99d5cdeeaa660",
|
| 31 |
+
"git_commit_sha": "",
|
| 32 |
+
"load_optimizer": 0,
|
| 33 |
+
"num_iterations": -1,
|
| 34 |
+
"max_seq_len": null,
|
| 35 |
+
"device_batch_size": 8,
|
| 36 |
+
"total_batch_size": null,
|
| 37 |
+
"embedding_lr": null,
|
| 38 |
+
"unembedding_lr": null,
|
| 39 |
+
"matrix_lr": null,
|
| 40 |
+
"init_lr_frac": 0.8,
|
| 41 |
+
"warmup_ratio": 0.03,
|
| 42 |
+
"warmdown_ratio": 0.5,
|
| 43 |
+
"final_lr_frac": 0.0,
|
| 44 |
+
"eval_every": 200,
|
| 45 |
+
"eval_tokens": 20971520,
|
| 46 |
+
"chatcore_every": -1,
|
| 47 |
+
"chatcore_max_cat": -1,
|
| 48 |
+
"chatcore_max_sample": 24,
|
| 49 |
+
"save_every": -1,
|
| 50 |
+
"recipe": "curriculum",
|
| 51 |
+
"curriculum_config": "",
|
| 52 |
+
"pre1930_epochs": 5,
|
| 53 |
+
"mmlu_epochs": 3,
|
| 54 |
+
"gsm8k_epochs": 4,
|
| 55 |
+
"authentic_epochs": 0,
|
| 56 |
+
"knowledge_qa_epochs": 0,
|
| 57 |
+
"multiturn_qa_epochs": 0,
|
| 58 |
+
"reasoning_qa_epochs": 0,
|
| 59 |
+
"stem_reasoning_epochs": 0,
|
| 60 |
+
"narrative_grounded_epochs": 0,
|
| 61 |
+
"narrative_fiction_epochs": 0,
|
| 62 |
+
"opinion_qa_epochs": 0,
|
| 63 |
+
"how_to_qa_epochs": 0,
|
| 64 |
+
"verse_qa_epochs": 0,
|
| 65 |
+
"composition_qa_epochs": 0,
|
| 66 |
+
"resolved_experiment_config": {
|
| 67 |
+
"schema_version": 1,
|
| 68 |
+
"stage": "sft",
|
| 69 |
+
"experiment_suffix": "pre1930-curriculum-c0",
|
| 70 |
+
"data": {
|
| 71 |
+
"recipe": "curriculum",
|
| 72 |
+
"curriculum": {
|
| 73 |
+
"name": "C0",
|
| 74 |
+
"mode": "flat",
|
| 75 |
+
"epochs": 3,
|
| 76 |
+
"threshold_default": 90,
|
| 77 |
+
"routes": {
|
| 78 |
+
"knowledge_qa": {
|
| 79 |
+
"count": 18000
|
| 80 |
+
},
|
| 81 |
+
"multiturn_qa": {
|
| 82 |
+
"count": 18000
|
| 83 |
+
},
|
| 84 |
+
"reasoning_qa": {
|
| 85 |
+
"count": 12000
|
| 86 |
+
},
|
| 87 |
+
"narrative_grounded": {
|
| 88 |
+
"count": 12000
|
| 89 |
+
},
|
| 90 |
+
"opinion_qa": {
|
| 91 |
+
"count": 10000
|
| 92 |
+
},
|
| 93 |
+
"composition_qa": {
|
| 94 |
+
"count": 10000
|
| 95 |
+
},
|
| 96 |
+
"how_to_qa": {
|
| 97 |
+
"count": 7000
|
| 98 |
+
},
|
| 99 |
+
"verse_qa": {
|
| 100 |
+
"count": 6000
|
| 101 |
+
},
|
| 102 |
+
"narrative_fiction": {
|
| 103 |
+
"count": 6000
|
| 104 |
+
},
|
| 105 |
+
"stem_reasoning": {
|
| 106 |
+
"count": 4000
|
| 107 |
+
}
|
| 108 |
+
},
|
| 109 |
+
"calibration_qa": {
|
| 110 |
+
"count": null
|
| 111 |
+
},
|
| 112 |
+
"authentic": {
|
| 113 |
+
"count": 12257
|
| 114 |
+
}
|
| 115 |
+
}
|
| 116 |
+
},
|
| 117 |
+
"training": {
|
| 118 |
+
"num_iterations": -1,
|
| 119 |
+
"load_optimizer": 0,
|
| 120 |
+
"device_batch_size": 8,
|
| 121 |
+
"warmup_ratio": 0.03,
|
| 122 |
+
"eval_every": 200,
|
| 123 |
+
"chatcore_every": -1,
|
| 124 |
+
"save_every": -1
|
| 125 |
+
},
|
| 126 |
+
"artifacts": {
|
| 127 |
+
"repo": "jbduran/think.nano"
|
| 128 |
+
},
|
| 129 |
+
"wandb": {
|
| 130 |
+
"enabled": true,
|
| 131 |
+
"entity": "jbduran-thinkingmachinesncsu",
|
| 132 |
+
"project": "think.nano",
|
| 133 |
+
"group": "think-d12",
|
| 134 |
+
"tags": [
|
| 135 |
+
"sft",
|
| 136 |
+
"curriculum",
|
| 137 |
+
"c0",
|
| 138 |
+
"baseline"
|
| 139 |
+
]
|
| 140 |
+
},
|
| 141 |
+
"config_fingerprint": "4ce60ef1ccb8a8c8",
|
| 142 |
+
"artifact_path": "experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0"
|
| 143 |
+
},
|
| 144 |
+
"stage": "sft",
|
| 145 |
+
"base_experiment_id": null,
|
| 146 |
+
"parent_experiment_id": null,
|
| 147 |
+
"parent_checkpoint_step": null,
|
| 148 |
+
"config_fingerprint": "4ce60ef1ccb8a8c8"
|
| 149 |
+
},
|
| 150 |
+
"loop_state": {
|
| 151 |
+
"step": 52,
|
| 152 |
+
"total_training_time": 2071.3676829338074,
|
| 153 |
+
"min_val_bpb": 0.8464286315951827,
|
| 154 |
+
"smooth_train_loss": 1.608637469780549,
|
| 155 |
+
"mfu": 35.219659061221186,
|
| 156 |
+
"tok_per_sec": 21247,
|
| 157 |
+
"stage_training_flops": 2.819857154310144e+17,
|
| 158 |
+
"inherited_parent_flops": 4.5291244139996774e+19,
|
| 159 |
+
"cumulative_pipeline_training_flops": 4.557322985542779e+19
|
| 160 |
+
}
|
| 161 |
+
}
|
experiments/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1/sft/clean1930s-d24-r12-ctx4096-sssl-fulltok-v1-pre1930-curriculum-c0/checkpoints/model_000052.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:70c91228c289d6b4d6928d866710be6bda73f21a0e8c1f2a772cd0cd593e2806
|
| 3 |
+
size 4227936073
|