Add files using upload-large-folder tool

.gitattributes

@@ -294,3 +294,16 @@ checkpoints/Llama-3.2-1B/babylm_reverse_partial_10M_seed0/artifacts/models--meta
 checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-900/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-100/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1700/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1200/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1100/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1600/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-600/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/GPT-2/babylm_reverse_full_10M_seed0/artifacts/models--gpt2/blobs/248dfc3911869ec493c76e65bf2fcf7f615828b0254c12b473182f0f81d3a707 filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-2700/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-400/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-2919/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1400/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-800/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1900/tokenizer.json filter=lfs diff=lfs merge=lfs -text

checkpoints/GPT-2/babylm_reverse_control_10M_seed0/runs/checkpoint-2100/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3621a4e42cc2c5cad279b3450fd9ef09ec5b4546548562775d917cab8173c52
+size 326089504

checkpoints/GPT-2/babylm_reverse_control_10M_seed0/runs/checkpoint-2200/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e1b7fbfd29153849db4684c989fd9536f3ec8949a70db40ec7a646b540d8320
+size 326089504

checkpoints/GPT-2/babylm_reverse_control_10M_seed0/runs/checkpoint-2300/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa77bada3cf328889026531cde3aced0ca7348eaf7dfe4d63fdb5a5453dae863
+size 326089504

checkpoints/GPT-2/babylm_reverse_control_10M_seed0/runs/checkpoint-2400/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:febd1939ba5ec58e6aee28d517d528577d1029323762745f64034e8a94ceda7e
+size 326089504

checkpoints/GPT-2/babylm_reverse_control_10M_seed0/runs/checkpoint-2500/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fc4646004037925fcefb11d3282b67c94a692c05be616969767b9b8b7f9732a
+size 326089504

checkpoints/GPT-2/babylm_reverse_control_10M_seed0/runs/checkpoint-2600/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4dbc180db5bbccea247d1aa54689010da96cf250136eb68add7c3eb2b5aebda6
+size 326089504

checkpoints/GPT-2/babylm_reverse_control_10M_seed0/runs/checkpoint-300/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3c9aa6b8aff52d50b47aee227a6dd86316334acc17165d36d1da41059471549
+size 326089504

checkpoints/GPT-2/babylm_reverse_full_10M_seed0/artifacts/models--gpt2/blobs/248dfc3911869ec493c76e65bf2fcf7f615828b0254c12b473182f0f81d3a707

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:248dfc3911869ec493c76e65bf2fcf7f615828b0254c12b473182f0f81d3a707
+size 548105171

checkpoints/GPT-2/babylm_reverse_full_10M_seed0/artifacts/models--gpt2/snapshots/607a30d783dfa663caf39e06633721c8d4cfcd7e/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:248dfc3911869ec493c76e65bf2fcf7f615828b0254c12b473182f0f81d3a707
+size 548105171

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-100/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbce2dc19c88327ccdd638378cf309339267b94959339b032dfd46918ad5f15e
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-1100/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e58a65c3d4eb283fe014174a890cfbbecc025f3a137b375dc09ea0372d0e7da
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-1200/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28b9d6469233b6f517cffe52eef1034c5aae0d6cafa930fb7e59f1feefdf4ac4
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-1300/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e47d5e2ac22b27408d78568cbf135ee5dcbccfddd4c98f85a0715f294ca8236
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-1400/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:777ec5eeb0591dd995361a0caf4686be57b66ceb13635213f7068a3e019e093c
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-1500/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7821603eb8e686f0804cc0300c8319d32845730696a41690e42094dfb96b2a3d
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-1600/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc275143ad3bf5c42fadd6fde0e0ed95ef73bab086244257694ee0f1f20a4b85
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-1700/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d32b436b597203c3bb9a72bc9db598262e90264454ae0c922146743e81a47bb
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-1800/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:719b6e0326e50aa26ad4f9e839be1d9c355596499cd156c99cd2ca72f12eea91
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-1900/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b93da9b1f889ff06893a0272c9789e61719115c2cd31c59b75fc45b7690b1ef4
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-200/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be3096f2cd2008ababe2056da753b61f62ce6711d69c3629a0a6b9ca6cc171ed
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-2000/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2eaab632ae2ef5e08af8b8e61edf8b8dd19262ae37a3a07ca3ee49085ecf2293
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-2700/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fda7ac84cde4f9f75fc3d58bd11610d90df061735584d346d8eb0a3c5e755cae
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-2919/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7593aafc3cf49f93f2f49c76c8652f670db002ec4244b7173e5776af928476ab
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-400/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70e695632c2b4ea1232d5e8ff06eef4707b8c1a92389243721f1bba7a878ef00
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-600/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e1b863cbbce509888f1fc064504a60412b84fa240b550e43ed394f28bd8d492
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-700/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4dca9ac06808d48d1c8224c697c45d4744b449b4e8032e6a12fdd7853f6a3c57
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-800/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6db5b63776b0a238d47e9a64800764dd780e4f3b85c1cf88a6c0c8a56940528d
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-900/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f741e772b467b67e4d2598373dc9efc8372a84b24a6ca00e4f31c85642d18e0
+size 326089504

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-900/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1adf48ea3a4dfba5404637be6f82c9a15ebd900e796e29b6ffeb0fdc3b1d475c
+size 16567

checkpoints/GPT-2/babylm_reverse_partial_10M_seed0/runs/checkpoint-900/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8eb0a6f24460a35a5849475791340c8d7dfb54225c5749b791b10e833071abfa
+size 6011

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1100/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:186b472c4073bfc086d5b4584aeaefc23f9c40f1a8199963bda08775617e1e12
+size 16567

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1100/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d8957d748b592f49a56723f1936535313c244cee56956dbc893c99d40f6ef84b
+size 16567

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1100/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be2cd13083af0074c65198c01c90c0a3c599fbe6e0ea705a8835ca51f393d0c0
+size 16567

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1100/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31b0601a70e7904c3660f5346767114aa9fa6d4a2cba625fac4b4b9ea3ada066
+size 563

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1100/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fda2ecf7b5daf3cb77324ea6c5b9cb8bf47119249a6d756ef2dd4d62540ac6f
+size 17210554

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1100/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4beb369b5d9b2e313f485f5348c5624fda1578b4b4b839a45918f8f15a4fbd4b
+size 6011

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1200/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f34c8972483cae48cde438f54c14179a5393e31298d634393f2e166c6a9bfb3e
+size 16567

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1200/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d0068d515416f5010564895d620c337c347a04bf8d5382015b5833708d52850
+size 16567

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1200/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1597fdb72beb896bfa337f48e1d75c0339960cd7266fb14e0c9cea974436401f
+size 563

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1200/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fda2ecf7b5daf3cb77324ea6c5b9cb8bf47119249a6d756ef2dd4d62540ac6f
+size 17210554

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1200/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4beb369b5d9b2e313f485f5348c5624fda1578b4b4b839a45918f8f15a4fbd4b
+size 6011

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7020c2ba12cf7959bf7033d7ad83dcf637e59ddbd3ff5b17fc4b9a1ebed3202
+size 16567

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f29ab0036e394c39267ad86bf247bf1558e2213add95383fce34a980f109b42e
+size 16567

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14413139ea5b777791c35297ed8a0a717bb3d6a3553268cc92b15fa61be1ae39
+size 16567

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3288dac9d8ea621655c7c1e03f2306d032714a31bcf52c99a781b6eb9a2538c3
+size 563

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fda2ecf7b5daf3cb77324ea6c5b9cb8bf47119249a6d756ef2dd4d62540ac6f
+size 17210554

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1300/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4beb369b5d9b2e313f485f5348c5624fda1578b4b4b839a45918f8f15a4fbd4b
+size 6011

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1400/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
+}

checkpoints/Llama-3.2-3B-FTP/babylm_reverse_full_10M_seed0/runs/checkpoint-1400/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"
+}