Upload full trainer checkpoint step=507 (profile=llama3_2_1b_base_grpo, run_type=grpo)
Browse files- meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/adapter_config.json +4 -4
- meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/adapter_model.safetensors +1 -1
- meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/optimizer.pt +1 -1
- meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/rng_state.pth +1 -1
- meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/scheduler.pt +1 -1
- meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/trainer_state.json +191 -191
- meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/training_args.bin +1 -1
- meta-llama__llama-3.2-1b/grpo/run_manifest.json +9 -171
meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/adapter_config.json
CHANGED
|
@@ -33,13 +33,13 @@
|
|
| 33 |
"rank_pattern": {},
|
| 34 |
"revision": null,
|
| 35 |
"target_modules": [
|
|
|
|
|
|
|
|
|
|
| 36 |
"down_proj",
|
| 37 |
"o_proj",
|
| 38 |
-
"q_proj",
|
| 39 |
"gate_proj",
|
| 40 |
-
"
|
| 41 |
-
"up_proj",
|
| 42 |
-
"v_proj"
|
| 43 |
],
|
| 44 |
"target_parameters": null,
|
| 45 |
"task_type": "CAUSAL_LM",
|
|
|
|
| 33 |
"rank_pattern": {},
|
| 34 |
"revision": null,
|
| 35 |
"target_modules": [
|
| 36 |
+
"up_proj",
|
| 37 |
+
"k_proj",
|
| 38 |
+
"v_proj",
|
| 39 |
"down_proj",
|
| 40 |
"o_proj",
|
|
|
|
| 41 |
"gate_proj",
|
| 42 |
+
"q_proj"
|
|
|
|
|
|
|
| 43 |
],
|
| 44 |
"target_parameters": null,
|
| 45 |
"task_type": "CAUSAL_LM",
|
meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 45118424
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4ffaf8361976d27fe67e70aff60a2b4f511b5d4dcf325ad2b661bf3dcf67788b
|
| 3 |
size 45118424
|
meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 23162187
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3c015b9ccd39081dc46df8e1131e8b46f973b8ba5d7930d46df13b6f10ebb5f5
|
| 3 |
size 23162187
|
meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e14d5d69b57745b112c7829210514ea094e9af0aa8ea5e30f32362a3d8e797f
|
| 3 |
size 14645
|
meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:45c99e06fc1c30a1b880d75b9b0017154f42da9a6b9f156c33d5915937be0b01
|
| 3 |
size 1465
|
meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/trainer_state.json
CHANGED
|
@@ -2,7 +2,7 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 500,
|
| 7 |
"global_step": 507,
|
| 8 |
"is_hyper_param_search": false,
|
|
@@ -15,25 +15,25 @@
|
|
| 15 |
"clip_ratio/low_mean": 0.0,
|
| 16 |
"clip_ratio/low_min": 0.0,
|
| 17 |
"clip_ratio/region_mean": 0.0,
|
| 18 |
-
"completion_length": 65.
|
| 19 |
"completions/clipped_ratio": 0.0,
|
| 20 |
-
"completions/max_length":
|
| 21 |
-
"completions/max_terminated_length":
|
| 22 |
-
"completions/mean_length":
|
| 23 |
-
"completions/mean_terminated_length":
|
| 24 |
-
"completions/min_length":
|
| 25 |
-
"completions/min_terminated_length":
|
| 26 |
-
"epoch": 8.
|
| 27 |
-
"frac_reward_zero_std": 0.
|
| 28 |
-
"grad_norm": 0.
|
| 29 |
"kl": 0.0,
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss": -0.
|
| 32 |
-
"num_tokens":
|
| 33 |
-
"reward":
|
| 34 |
-
"reward_std": 0.
|
| 35 |
-
"rewards/TranscriptCorrectionGrpoReward/mean":
|
| 36 |
-
"rewards/TranscriptCorrectionGrpoReward/std":
|
| 37 |
"step": 50
|
| 38 |
},
|
| 39 |
{
|
|
@@ -42,52 +42,52 @@
|
|
| 42 |
"clip_ratio/low_mean": 0.0,
|
| 43 |
"clip_ratio/low_min": 0.0,
|
| 44 |
"clip_ratio/region_mean": 0.0,
|
| 45 |
-
"completion_length":
|
| 46 |
"completions/clipped_ratio": 0.0,
|
| 47 |
-
"completions/max_length":
|
| 48 |
-
"completions/max_terminated_length":
|
| 49 |
-
"completions/mean_length":
|
| 50 |
-
"completions/mean_terminated_length":
|
| 51 |
-
"completions/min_length":
|
| 52 |
-
"completions/min_terminated_length":
|
| 53 |
-
"epoch": 0.
|
| 54 |
-
"frac_reward_zero_std": 0.
|
| 55 |
"grad_norm": 0.0,
|
| 56 |
"kl": 0.0,
|
| 57 |
-
"learning_rate":
|
| 58 |
-
"loss": 0.
|
| 59 |
-
"num_tokens":
|
| 60 |
-
"reward":
|
| 61 |
-
"reward_std": 0.
|
| 62 |
-
"rewards/TranscriptCorrectionGrpoReward/mean":
|
| 63 |
-
"rewards/TranscriptCorrectionGrpoReward/std":
|
| 64 |
"step": 100
|
| 65 |
},
|
| 66 |
{
|
| 67 |
"clip_ratio/high_max": 0.0,
|
| 68 |
"clip_ratio/high_mean": 0.0,
|
| 69 |
-
"clip_ratio/low_mean": 0.
|
| 70 |
-
"clip_ratio/low_min": 0.
|
| 71 |
-
"clip_ratio/region_mean": 0.
|
| 72 |
-
"completion_length":
|
| 73 |
"completions/clipped_ratio": 0.0,
|
| 74 |
-
"completions/max_length":
|
| 75 |
-
"completions/max_terminated_length":
|
| 76 |
-
"completions/mean_length":
|
| 77 |
-
"completions/mean_terminated_length":
|
| 78 |
-
"completions/min_length":
|
| 79 |
-
"completions/min_terminated_length":
|
| 80 |
-
"epoch": 0.
|
| 81 |
-
"frac_reward_zero_std": 0.
|
| 82 |
-
"grad_norm": 0.
|
| 83 |
"kl": 0.0,
|
| 84 |
-
"learning_rate":
|
| 85 |
-
"loss": 0.
|
| 86 |
-
"num_tokens":
|
| 87 |
-
"reward":
|
| 88 |
-
"reward_std": 0.
|
| 89 |
-
"rewards/TranscriptCorrectionGrpoReward/mean":
|
| 90 |
-
"rewards/TranscriptCorrectionGrpoReward/std":
|
| 91 |
"step": 150
|
| 92 |
},
|
| 93 |
{
|
|
@@ -96,52 +96,52 @@
|
|
| 96 |
"clip_ratio/low_mean": 0.0,
|
| 97 |
"clip_ratio/low_min": 0.0,
|
| 98 |
"clip_ratio/region_mean": 0.0,
|
| 99 |
-
"completion_length":
|
| 100 |
"completions/clipped_ratio": 0.0,
|
| 101 |
-
"completions/max_length":
|
| 102 |
-
"completions/max_terminated_length":
|
| 103 |
-
"completions/mean_length":
|
| 104 |
-
"completions/mean_terminated_length":
|
| 105 |
-
"completions/min_length":
|
| 106 |
-
"completions/min_terminated_length":
|
| 107 |
-
"epoch": 0.
|
| 108 |
-
"frac_reward_zero_std": 0.
|
| 109 |
-
"grad_norm": 0.
|
| 110 |
"kl": 0.0,
|
| 111 |
-
"learning_rate":
|
| 112 |
-
"loss": -0.
|
| 113 |
-
"num_tokens":
|
| 114 |
-
"reward":
|
| 115 |
-
"reward_std": 0.
|
| 116 |
-
"rewards/TranscriptCorrectionGrpoReward/mean":
|
| 117 |
-
"rewards/TranscriptCorrectionGrpoReward/std":
|
| 118 |
"step": 200
|
| 119 |
},
|
| 120 |
{
|
| 121 |
"clip_ratio/high_max": 0.0,
|
| 122 |
"clip_ratio/high_mean": 0.0,
|
| 123 |
-
"clip_ratio/low_mean": 0.
|
| 124 |
-
"clip_ratio/low_min": 0.
|
| 125 |
-
"clip_ratio/region_mean": 0.
|
| 126 |
-
"completion_length":
|
| 127 |
"completions/clipped_ratio": 0.0,
|
| 128 |
-
"completions/max_length":
|
| 129 |
-
"completions/max_terminated_length":
|
| 130 |
-
"completions/mean_length":
|
| 131 |
-
"completions/mean_terminated_length":
|
| 132 |
-
"completions/min_length":
|
| 133 |
-
"completions/min_terminated_length":
|
| 134 |
-
"epoch": 0.
|
| 135 |
-
"frac_reward_zero_std": 0.
|
| 136 |
-
"grad_norm": 0.
|
| 137 |
"kl": 0.0,
|
| 138 |
-
"learning_rate":
|
| 139 |
-
"loss": 0.
|
| 140 |
-
"num_tokens":
|
| 141 |
-
"reward":
|
| 142 |
-
"reward_std": 0.
|
| 143 |
-
"rewards/TranscriptCorrectionGrpoReward/mean":
|
| 144 |
-
"rewards/TranscriptCorrectionGrpoReward/std":
|
| 145 |
"step": 250
|
| 146 |
},
|
| 147 |
{
|
|
@@ -150,106 +150,106 @@
|
|
| 150 |
"clip_ratio/low_mean": 0.0,
|
| 151 |
"clip_ratio/low_min": 0.0,
|
| 152 |
"clip_ratio/region_mean": 0.0,
|
| 153 |
-
"completion_length":
|
| 154 |
"completions/clipped_ratio": 0.0,
|
| 155 |
-
"completions/max_length":
|
| 156 |
-
"completions/max_terminated_length":
|
| 157 |
-
"completions/mean_length":
|
| 158 |
-
"completions/mean_terminated_length":
|
| 159 |
-
"completions/min_length":
|
| 160 |
-
"completions/min_terminated_length":
|
| 161 |
-
"epoch": 0.
|
| 162 |
-
"frac_reward_zero_std": 0.
|
| 163 |
-
"grad_norm": 0.
|
| 164 |
"kl": 0.0,
|
| 165 |
-
"learning_rate":
|
| 166 |
-
"loss":
|
| 167 |
-
"num_tokens":
|
| 168 |
-
"reward":
|
| 169 |
-
"reward_std": 0.
|
| 170 |
-
"rewards/TranscriptCorrectionGrpoReward/mean":
|
| 171 |
-
"rewards/TranscriptCorrectionGrpoReward/std":
|
| 172 |
"step": 300
|
| 173 |
},
|
| 174 |
{
|
| 175 |
"clip_ratio/high_max": 0.0,
|
| 176 |
"clip_ratio/high_mean": 0.0,
|
| 177 |
-
"clip_ratio/low_mean":
|
| 178 |
-
"clip_ratio/low_min":
|
| 179 |
-
"clip_ratio/region_mean":
|
| 180 |
-
"completion_length":
|
| 181 |
"completions/clipped_ratio": 0.0,
|
| 182 |
-
"completions/max_length":
|
| 183 |
-
"completions/max_terminated_length":
|
| 184 |
-
"completions/mean_length": 61.
|
| 185 |
-
"completions/mean_terminated_length": 61.
|
| 186 |
-
"completions/min_length":
|
| 187 |
-
"completions/min_terminated_length":
|
| 188 |
-
"epoch": 0.
|
| 189 |
-
"frac_reward_zero_std": 0.
|
| 190 |
-
"grad_norm": 0.
|
| 191 |
"kl": 0.0,
|
| 192 |
-
"learning_rate":
|
| 193 |
-
"loss": -
|
| 194 |
-
"num_tokens":
|
| 195 |
-
"reward":
|
| 196 |
-
"reward_std": 0.
|
| 197 |
-
"rewards/TranscriptCorrectionGrpoReward/mean":
|
| 198 |
-
"rewards/TranscriptCorrectionGrpoReward/std":
|
| 199 |
"step": 350
|
| 200 |
},
|
| 201 |
{
|
| 202 |
"clip_ratio/high_max": 0.0,
|
| 203 |
"clip_ratio/high_mean": 0.0,
|
| 204 |
-
"clip_ratio/low_mean":
|
| 205 |
-
"clip_ratio/low_min":
|
| 206 |
-
"clip_ratio/region_mean":
|
| 207 |
-
"completion_length":
|
| 208 |
"completions/clipped_ratio": 0.0,
|
| 209 |
-
"completions/max_length":
|
| 210 |
-
"completions/max_terminated_length":
|
| 211 |
-
"completions/mean_length":
|
| 212 |
-
"completions/mean_terminated_length":
|
| 213 |
-
"completions/min_length":
|
| 214 |
-
"completions/min_terminated_length":
|
| 215 |
-
"epoch": 0.
|
| 216 |
-
"frac_reward_zero_std": 0.
|
| 217 |
-
"grad_norm": 0.
|
| 218 |
"kl": 0.0,
|
| 219 |
-
"learning_rate":
|
| 220 |
-
"loss": 0.
|
| 221 |
-
"num_tokens":
|
| 222 |
-
"reward":
|
| 223 |
-
"reward_std": 0.
|
| 224 |
-
"rewards/TranscriptCorrectionGrpoReward/mean":
|
| 225 |
-
"rewards/TranscriptCorrectionGrpoReward/std":
|
| 226 |
"step": 400
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"clip_ratio/high_max": 0.0,
|
| 230 |
"clip_ratio/high_mean": 0.0,
|
| 231 |
-
"clip_ratio/low_mean":
|
| 232 |
-
"clip_ratio/low_min":
|
| 233 |
-
"clip_ratio/region_mean":
|
| 234 |
-
"completion_length":
|
| 235 |
"completions/clipped_ratio": 0.0,
|
| 236 |
-
"completions/max_length":
|
| 237 |
-
"completions/max_terminated_length":
|
| 238 |
-
"completions/mean_length":
|
| 239 |
-
"completions/mean_terminated_length":
|
| 240 |
-
"completions/min_length":
|
| 241 |
-
"completions/min_terminated_length":
|
| 242 |
-
"epoch": 0.
|
| 243 |
-
"frac_reward_zero_std": 0.
|
| 244 |
-
"grad_norm": 0.
|
| 245 |
"kl": 0.0,
|
| 246 |
-
"learning_rate":
|
| 247 |
-
"loss": -0.
|
| 248 |
-
"num_tokens":
|
| 249 |
-
"reward": 1.
|
| 250 |
-
"reward_std": 0.
|
| 251 |
-
"rewards/TranscriptCorrectionGrpoReward/mean": 1.
|
| 252 |
-
"rewards/TranscriptCorrectionGrpoReward/std":
|
| 253 |
"step": 450
|
| 254 |
},
|
| 255 |
{
|
|
@@ -258,40 +258,40 @@
|
|
| 258 |
"clip_ratio/low_mean": 0.0,
|
| 259 |
"clip_ratio/low_min": 0.0,
|
| 260 |
"clip_ratio/region_mean": 0.0,
|
| 261 |
-
"completion_length":
|
| 262 |
"completions/clipped_ratio": 0.0,
|
| 263 |
-
"completions/max_length":
|
| 264 |
-
"completions/max_terminated_length":
|
| 265 |
-
"completions/mean_length":
|
| 266 |
-
"completions/mean_terminated_length":
|
| 267 |
-
"completions/min_length":
|
| 268 |
-
"completions/min_terminated_length":
|
| 269 |
-
"epoch": 0.
|
| 270 |
-
"frac_reward_zero_std": 0.
|
| 271 |
-
"grad_norm": 0.
|
| 272 |
"kl": 0.0,
|
| 273 |
-
"learning_rate":
|
| 274 |
-
"loss": 0.
|
| 275 |
-
"num_tokens":
|
| 276 |
-
"reward":
|
| 277 |
-
"reward_std": 0.
|
| 278 |
-
"rewards/TranscriptCorrectionGrpoReward/mean":
|
| 279 |
-
"rewards/TranscriptCorrectionGrpoReward/std": 0.
|
| 280 |
"step": 500
|
| 281 |
},
|
| 282 |
{
|
| 283 |
-
"eval_cer_subset": 0.
|
| 284 |
-
"eval_cer_subset_edit_distance":
|
| 285 |
"eval_cer_subset_groups": 250,
|
| 286 |
"eval_cer_subset_items": 250,
|
| 287 |
"eval_cer_subset_ref_chars": 61407,
|
| 288 |
-
"eval_loss": 0.
|
| 289 |
"step": 507
|
| 290 |
}
|
| 291 |
],
|
| 292 |
"logging_steps": 50,
|
| 293 |
-
"max_steps":
|
| 294 |
-
"num_input_tokens_seen":
|
| 295 |
"num_train_epochs": 4,
|
| 296 |
"save_steps": 507,
|
| 297 |
"stateful_callbacks": {
|
|
@@ -307,7 +307,7 @@
|
|
| 307 |
}
|
| 308 |
},
|
| 309 |
"total_flos": 0.0,
|
| 310 |
-
"train_batch_size":
|
| 311 |
"trial_name": null,
|
| 312 |
"trial_params": null
|
| 313 |
}
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.0009005168647093303,
|
| 6 |
"eval_steps": 500,
|
| 7 |
"global_step": 507,
|
| 8 |
"is_hyper_param_search": false,
|
|
|
|
| 15 |
"clip_ratio/low_mean": 0.0,
|
| 16 |
"clip_ratio/low_min": 0.0,
|
| 17 |
"clip_ratio/region_mean": 0.0,
|
| 18 |
+
"completion_length": 65.68,
|
| 19 |
"completions/clipped_ratio": 0.0,
|
| 20 |
+
"completions/max_length": 65.88,
|
| 21 |
+
"completions/max_terminated_length": 65.88,
|
| 22 |
+
"completions/mean_length": 65.68,
|
| 23 |
+
"completions/mean_terminated_length": 65.68,
|
| 24 |
+
"completions/min_length": 65.48,
|
| 25 |
+
"completions/min_terminated_length": 65.48,
|
| 26 |
+
"epoch": 8.88083693007229e-05,
|
| 27 |
+
"frac_reward_zero_std": 0.56,
|
| 28 |
+
"grad_norm": 0.5719226598739624,
|
| 29 |
"kl": 0.0,
|
| 30 |
+
"learning_rate": 5.439488466064254e-09,
|
| 31 |
+
"loss": -0.000577671229839325,
|
| 32 |
+
"num_tokens": 46552.0,
|
| 33 |
+
"reward": 2.9802322387695314e-09,
|
| 34 |
+
"reward_std": 0.47037978172302247,
|
| 35 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": 2.9802322387695314e-09,
|
| 36 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.4703797769546509,
|
| 37 |
"step": 50
|
| 38 |
},
|
| 39 |
{
|
|
|
|
| 42 |
"clip_ratio/low_mean": 0.0,
|
| 43 |
"clip_ratio/low_min": 0.0,
|
| 44 |
"clip_ratio/region_mean": 0.0,
|
| 45 |
+
"completion_length": 53.46,
|
| 46 |
"completions/clipped_ratio": 0.0,
|
| 47 |
+
"completions/max_length": 53.6,
|
| 48 |
+
"completions/max_terminated_length": 53.6,
|
| 49 |
+
"completions/mean_length": 53.46,
|
| 50 |
+
"completions/mean_terminated_length": 53.46,
|
| 51 |
+
"completions/min_length": 53.24,
|
| 52 |
+
"completions/min_terminated_length": 53.24,
|
| 53 |
+
"epoch": 0.0001776167386014458,
|
| 54 |
+
"frac_reward_zero_std": 0.64,
|
| 55 |
"grad_norm": 0.0,
|
| 56 |
"kl": 0.0,
|
| 57 |
+
"learning_rate": 1.0989986900823695e-08,
|
| 58 |
+
"loss": -0.0009160846471786499,
|
| 59 |
+
"num_tokens": 85556.0,
|
| 60 |
+
"reward": 3.874301910400391e-09,
|
| 61 |
+
"reward_std": 0.38485618114471437,
|
| 62 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": 3.874301910400391e-09,
|
| 63 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.38485618114471437,
|
| 64 |
"step": 100
|
| 65 |
},
|
| 66 |
{
|
| 67 |
"clip_ratio/high_max": 0.0,
|
| 68 |
"clip_ratio/high_mean": 0.0,
|
| 69 |
+
"clip_ratio/low_mean": 0.00011904762126505375,
|
| 70 |
+
"clip_ratio/low_min": 0.00011904762126505375,
|
| 71 |
+
"clip_ratio/region_mean": 0.00011904762126505375,
|
| 72 |
+
"completion_length": 62.48,
|
| 73 |
"completions/clipped_ratio": 0.0,
|
| 74 |
+
"completions/max_length": 62.6,
|
| 75 |
+
"completions/max_terminated_length": 62.6,
|
| 76 |
+
"completions/mean_length": 62.48,
|
| 77 |
+
"completions/mean_terminated_length": 62.48,
|
| 78 |
+
"completions/min_length": 62.24,
|
| 79 |
+
"completions/min_terminated_length": 62.24,
|
| 80 |
+
"epoch": 0.0002664251079021687,
|
| 81 |
+
"frac_reward_zero_std": 0.56,
|
| 82 |
+
"grad_norm": 0.2554231286048889,
|
| 83 |
"kl": 0.0,
|
| 84 |
+
"learning_rate": 1.6540485335583138e-08,
|
| 85 |
+
"loss": -0.0009111672639846802,
|
| 86 |
+
"num_tokens": 130380.0,
|
| 87 |
+
"reward": -1.1920928955078125e-09,
|
| 88 |
+
"reward_std": 0.4703797769546509,
|
| 89 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": -1.1920928955078125e-09,
|
| 90 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.47037978172302247,
|
| 91 |
"step": 150
|
| 92 |
},
|
| 93 |
{
|
|
|
|
| 96 |
"clip_ratio/low_mean": 0.0,
|
| 97 |
"clip_ratio/low_min": 0.0,
|
| 98 |
"clip_ratio/region_mean": 0.0,
|
| 99 |
+
"completion_length": 62.175,
|
| 100 |
"completions/clipped_ratio": 0.0,
|
| 101 |
+
"completions/max_length": 62.36,
|
| 102 |
+
"completions/max_terminated_length": 62.36,
|
| 103 |
+
"completions/mean_length": 62.175,
|
| 104 |
+
"completions/mean_terminated_length": 62.175,
|
| 105 |
+
"completions/min_length": 61.88,
|
| 106 |
+
"completions/min_terminated_length": 61.88,
|
| 107 |
+
"epoch": 0.0003552334772028916,
|
| 108 |
+
"frac_reward_zero_std": 0.56,
|
| 109 |
+
"grad_norm": 0.627755343914032,
|
| 110 |
"kl": 0.0,
|
| 111 |
+
"learning_rate": 2.2090983770342578e-08,
|
| 112 |
+
"loss": -0.0005262196063995361,
|
| 113 |
+
"num_tokens": 174959.0,
|
| 114 |
+
"reward": 1.1920928955078125e-09,
|
| 115 |
+
"reward_std": 0.47037978172302247,
|
| 116 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": 1.1920928955078125e-09,
|
| 117 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.4703797769546509,
|
| 118 |
"step": 200
|
| 119 |
},
|
| 120 |
{
|
| 121 |
"clip_ratio/high_max": 0.0,
|
| 122 |
"clip_ratio/high_mean": 0.0,
|
| 123 |
+
"clip_ratio/low_mean": 0.00016129031777381897,
|
| 124 |
+
"clip_ratio/low_min": 0.00016129031777381897,
|
| 125 |
+
"clip_ratio/region_mean": 0.00016129031777381897,
|
| 126 |
+
"completion_length": 65.835,
|
| 127 |
"completions/clipped_ratio": 0.0,
|
| 128 |
+
"completions/max_length": 66.08,
|
| 129 |
+
"completions/max_terminated_length": 66.08,
|
| 130 |
+
"completions/mean_length": 65.835,
|
| 131 |
+
"completions/mean_terminated_length": 65.835,
|
| 132 |
+
"completions/min_length": 65.6,
|
| 133 |
+
"completions/min_terminated_length": 65.6,
|
| 134 |
+
"epoch": 0.0004440418465036145,
|
| 135 |
+
"frac_reward_zero_std": 0.52,
|
| 136 |
+
"grad_norm": 0.0,
|
| 137 |
"kl": 0.0,
|
| 138 |
+
"learning_rate": 2.764148220510202e-08,
|
| 139 |
+
"loss": -0.00036669328808784483,
|
| 140 |
+
"num_tokens": 222270.0,
|
| 141 |
+
"reward": 3.874301910400391e-09,
|
| 142 |
+
"reward_std": 0.5131415796279907,
|
| 143 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": 3.874301910400391e-09,
|
| 144 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.5131415748596191,
|
| 145 |
"step": 250
|
| 146 |
},
|
| 147 |
{
|
|
|
|
| 150 |
"clip_ratio/low_mean": 0.0,
|
| 151 |
"clip_ratio/low_min": 0.0,
|
| 152 |
"clip_ratio/region_mean": 0.0,
|
| 153 |
+
"completion_length": 68.75,
|
| 154 |
"completions/clipped_ratio": 0.0,
|
| 155 |
+
"completions/max_length": 69.04,
|
| 156 |
+
"completions/max_terminated_length": 69.04,
|
| 157 |
+
"completions/mean_length": 68.75,
|
| 158 |
+
"completions/mean_terminated_length": 68.75,
|
| 159 |
+
"completions/min_length": 68.56,
|
| 160 |
+
"completions/min_terminated_length": 68.56,
|
| 161 |
+
"epoch": 0.0005328502158043374,
|
| 162 |
+
"frac_reward_zero_std": 0.6,
|
| 163 |
+
"grad_norm": 0.0,
|
| 164 |
"kl": 0.0,
|
| 165 |
+
"learning_rate": 3.3191980639861464e-08,
|
| 166 |
+
"loss": 0.0001394149661064148,
|
| 167 |
+
"num_tokens": 271740.0,
|
| 168 |
+
"reward": 6.258487701416015e-09,
|
| 169 |
+
"reward_std": 0.4276179838180542,
|
| 170 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": 6.258487701416015e-09,
|
| 171 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.42761797904968263,
|
| 172 |
"step": 300
|
| 173 |
},
|
| 174 |
{
|
| 175 |
"clip_ratio/high_max": 0.0,
|
| 176 |
"clip_ratio/high_mean": 0.0,
|
| 177 |
+
"clip_ratio/low_mean": 0.0,
|
| 178 |
+
"clip_ratio/low_min": 0.0,
|
| 179 |
+
"clip_ratio/region_mean": 0.0,
|
| 180 |
+
"completion_length": 61.07,
|
| 181 |
"completions/clipped_ratio": 0.0,
|
| 182 |
+
"completions/max_length": 61.2,
|
| 183 |
+
"completions/max_terminated_length": 61.2,
|
| 184 |
+
"completions/mean_length": 61.07,
|
| 185 |
+
"completions/mean_terminated_length": 61.07,
|
| 186 |
+
"completions/min_length": 60.96,
|
| 187 |
+
"completions/min_terminated_length": 60.96,
|
| 188 |
+
"epoch": 0.0006216585851050603,
|
| 189 |
+
"frac_reward_zero_std": 0.72,
|
| 190 |
+
"grad_norm": 0.0,
|
| 191 |
"kl": 0.0,
|
| 192 |
+
"learning_rate": 3.8742479074620904e-08,
|
| 193 |
+
"loss": -5.508854985237121e-05,
|
| 194 |
+
"num_tokens": 316354.0,
|
| 195 |
+
"reward": 2.384185791015625e-09,
|
| 196 |
+
"reward_std": 0.29933258533477786,
|
| 197 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": 2.384185791015625e-09,
|
| 198 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.2993325901031494,
|
| 199 |
"step": 350
|
| 200 |
},
|
| 201 |
{
|
| 202 |
"clip_ratio/high_max": 0.0,
|
| 203 |
"clip_ratio/high_mean": 0.0,
|
| 204 |
+
"clip_ratio/low_mean": 0.0,
|
| 205 |
+
"clip_ratio/low_min": 0.0,
|
| 206 |
+
"clip_ratio/region_mean": 0.0,
|
| 207 |
+
"completion_length": 64.22,
|
| 208 |
"completions/clipped_ratio": 0.0,
|
| 209 |
+
"completions/max_length": 64.6,
|
| 210 |
+
"completions/max_terminated_length": 64.6,
|
| 211 |
+
"completions/mean_length": 64.22,
|
| 212 |
+
"completions/mean_terminated_length": 64.22,
|
| 213 |
+
"completions/min_length": 63.92,
|
| 214 |
+
"completions/min_terminated_length": 63.92,
|
| 215 |
+
"epoch": 0.0007104669544057832,
|
| 216 |
+
"frac_reward_zero_std": 0.56,
|
| 217 |
+
"grad_norm": 0.0,
|
| 218 |
"kl": 0.0,
|
| 219 |
+
"learning_rate": 4.429297750938035e-08,
|
| 220 |
+
"loss": -0.0010024748742580413,
|
| 221 |
+
"num_tokens": 362550.0,
|
| 222 |
+
"reward": 5.960464477539063e-10,
|
| 223 |
+
"reward_std": 0.4703797769546509,
|
| 224 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": 5.960464477539063e-10,
|
| 225 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.4703797769546509,
|
| 226 |
"step": 400
|
| 227 |
},
|
| 228 |
{
|
| 229 |
"clip_ratio/high_max": 0.0,
|
| 230 |
"clip_ratio/high_mean": 0.0,
|
| 231 |
+
"clip_ratio/low_mean": 0.0,
|
| 232 |
+
"clip_ratio/low_min": 0.0,
|
| 233 |
+
"clip_ratio/region_mean": 0.0,
|
| 234 |
+
"completion_length": 65.25,
|
| 235 |
"completions/clipped_ratio": 0.0,
|
| 236 |
+
"completions/max_length": 65.36,
|
| 237 |
+
"completions/max_terminated_length": 65.36,
|
| 238 |
+
"completions/mean_length": 65.25,
|
| 239 |
+
"completions/mean_terminated_length": 65.25,
|
| 240 |
+
"completions/min_length": 65.04,
|
| 241 |
+
"completions/min_terminated_length": 65.04,
|
| 242 |
+
"epoch": 0.000799275323706506,
|
| 243 |
+
"frac_reward_zero_std": 0.6,
|
| 244 |
+
"grad_norm": 0.0,
|
| 245 |
"kl": 0.0,
|
| 246 |
+
"learning_rate": 4.984347594413978e-08,
|
| 247 |
+
"loss": -0.0002909022569656372,
|
| 248 |
+
"num_tokens": 409416.0,
|
| 249 |
+
"reward": 1.7881393432617187e-09,
|
| 250 |
+
"reward_std": 0.4276179838180542,
|
| 251 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": 1.7881393432617187e-09,
|
| 252 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.4276179838180542,
|
| 253 |
"step": 450
|
| 254 |
},
|
| 255 |
{
|
|
|
|
| 258 |
"clip_ratio/low_mean": 0.0,
|
| 259 |
"clip_ratio/low_min": 0.0,
|
| 260 |
"clip_ratio/region_mean": 0.0,
|
| 261 |
+
"completion_length": 65.76,
|
| 262 |
"completions/clipped_ratio": 0.0,
|
| 263 |
+
"completions/max_length": 65.92,
|
| 264 |
+
"completions/max_terminated_length": 65.92,
|
| 265 |
+
"completions/mean_length": 65.76,
|
| 266 |
+
"completions/mean_terminated_length": 65.76,
|
| 267 |
+
"completions/min_length": 65.68,
|
| 268 |
+
"completions/min_terminated_length": 65.68,
|
| 269 |
+
"epoch": 0.000888083693007229,
|
| 270 |
+
"frac_reward_zero_std": 0.84,
|
| 271 |
+
"grad_norm": 0.035873379558324814,
|
| 272 |
"kl": 0.0,
|
| 273 |
+
"learning_rate": 5.5393974378899236e-08,
|
| 274 |
+
"loss": -0.000893859937787056,
|
| 275 |
+
"num_tokens": 456024.0,
|
| 276 |
+
"reward": 0.0,
|
| 277 |
+
"reward_std": 0.17104718685150147,
|
| 278 |
+
"rewards/TranscriptCorrectionGrpoReward/mean": 0.0,
|
| 279 |
+
"rewards/TranscriptCorrectionGrpoReward/std": 0.17104719161987306,
|
| 280 |
"step": 500
|
| 281 |
},
|
| 282 |
{
|
| 283 |
+
"eval_cer_subset": 0.012913837184685785,
|
| 284 |
+
"eval_cer_subset_edit_distance": 793,
|
| 285 |
"eval_cer_subset_groups": 250,
|
| 286 |
"eval_cer_subset_items": 250,
|
| 287 |
"eval_cer_subset_ref_chars": 61407,
|
| 288 |
+
"eval_loss": 0.03959455178967374,
|
| 289 |
"step": 507
|
| 290 |
}
|
| 291 |
],
|
| 292 |
"logging_steps": 50,
|
| 293 |
+
"max_steps": 2252040,
|
| 294 |
+
"num_input_tokens_seen": 462420,
|
| 295 |
"num_train_epochs": 4,
|
| 296 |
"save_steps": 507,
|
| 297 |
"stateful_callbacks": {
|
|
|
|
| 307 |
}
|
| 308 |
},
|
| 309 |
"total_flos": 0.0,
|
| 310 |
+
"train_batch_size": 1,
|
| 311 |
"trial_name": null,
|
| 312 |
"trial_params": null
|
| 313 |
}
|
meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6737
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9bdad4bf60f02650b7130e20d3de1e81654197d408ff77954683d133ac0bce86
|
| 3 |
size 6737
|
meta-llama__llama-3.2-1b/grpo/run_manifest.json
CHANGED
|
@@ -1,187 +1,25 @@
|
|
| 1 |
{
|
| 2 |
"schema_version": 1,
|
| 3 |
-
"run_id": "
|
| 4 |
"run_type": "grpo",
|
| 5 |
"profile_name": "llama3_2_1b_base_grpo",
|
| 6 |
"model_name": "meta-llama/Llama-3.2-1B",
|
| 7 |
"status": "running",
|
| 8 |
"best_metric_name": "eval_cer_subset",
|
| 9 |
-
"best_metric_value": 0.
|
| 10 |
-
"best_step":
|
| 11 |
-
"best_checkpoint_step":
|
| 12 |
-
"latest_checkpoint_step":
|
| 13 |
"checkpoints": [
|
| 14 |
{
|
| 15 |
"step": 507,
|
| 16 |
-
"epoch": 0.
|
| 17 |
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507",
|
| 18 |
-
"timestamp": "2026-04-
|
| 19 |
-
},
|
| 20 |
-
{
|
| 21 |
-
"step": 1014,
|
| 22 |
-
"epoch": 0.0018010401273161306,
|
| 23 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-1014",
|
| 24 |
-
"timestamp": "2026-04-09T01:10:57.965692+00:00"
|
| 25 |
-
},
|
| 26 |
-
{
|
| 27 |
-
"step": 1521,
|
| 28 |
-
"epoch": 0.0027015601909741956,
|
| 29 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-1521",
|
| 30 |
-
"timestamp": "2026-04-09T01:17:04.646992+00:00"
|
| 31 |
-
},
|
| 32 |
-
{
|
| 33 |
-
"step": 2028,
|
| 34 |
-
"epoch": 0.003602080254632261,
|
| 35 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-2028",
|
| 36 |
-
"timestamp": "2026-04-09T01:23:03.985957+00:00"
|
| 37 |
-
},
|
| 38 |
-
{
|
| 39 |
-
"step": 2535,
|
| 40 |
-
"epoch": 0.004502600318290326,
|
| 41 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-2535",
|
| 42 |
-
"timestamp": "2026-04-09T01:29:18.820055+00:00"
|
| 43 |
-
},
|
| 44 |
-
{
|
| 45 |
-
"step": 3042,
|
| 46 |
-
"epoch": 0.005403120381948391,
|
| 47 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-3042",
|
| 48 |
-
"timestamp": "2026-04-09T01:34:32.562568+00:00"
|
| 49 |
-
},
|
| 50 |
-
{
|
| 51 |
-
"step": 3549,
|
| 52 |
-
"epoch": 0.006303640445606456,
|
| 53 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-3549",
|
| 54 |
-
"timestamp": "2026-04-09T01:39:39.758317+00:00"
|
| 55 |
-
},
|
| 56 |
-
{
|
| 57 |
-
"step": 4056,
|
| 58 |
-
"epoch": 0.007204160509264522,
|
| 59 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-4056",
|
| 60 |
-
"timestamp": "2026-04-09T01:44:53.215161+00:00"
|
| 61 |
-
},
|
| 62 |
-
{
|
| 63 |
-
"step": 4563,
|
| 64 |
-
"epoch": 0.008104680572922587,
|
| 65 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-4563",
|
| 66 |
-
"timestamp": "2026-04-09T01:50:12.247540+00:00"
|
| 67 |
-
},
|
| 68 |
-
{
|
| 69 |
-
"step": 5070,
|
| 70 |
-
"epoch": 0.009005200636580652,
|
| 71 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-5070",
|
| 72 |
-
"timestamp": "2026-04-09T01:55:16.037586+00:00"
|
| 73 |
-
},
|
| 74 |
-
{
|
| 75 |
-
"step": 5577,
|
| 76 |
-
"epoch": 0.009905720700238718,
|
| 77 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-5577",
|
| 78 |
-
"timestamp": "2026-04-09T02:00:26.191388+00:00"
|
| 79 |
-
},
|
| 80 |
-
{
|
| 81 |
-
"step": 6084,
|
| 82 |
-
"epoch": 0.010806240763896783,
|
| 83 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-6084",
|
| 84 |
-
"timestamp": "2026-04-09T02:05:34.159056+00:00"
|
| 85 |
-
},
|
| 86 |
-
{
|
| 87 |
-
"step": 6591,
|
| 88 |
-
"epoch": 0.011706760827554848,
|
| 89 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-6591",
|
| 90 |
-
"timestamp": "2026-04-09T02:10:28.495758+00:00"
|
| 91 |
-
},
|
| 92 |
-
{
|
| 93 |
-
"step": 7098,
|
| 94 |
-
"epoch": 0.012607280891212913,
|
| 95 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-7098",
|
| 96 |
-
"timestamp": "2026-04-09T02:15:29.948111+00:00"
|
| 97 |
-
},
|
| 98 |
-
{
|
| 99 |
-
"step": 7605,
|
| 100 |
-
"epoch": 0.01350780095487098,
|
| 101 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-7605",
|
| 102 |
-
"timestamp": "2026-04-09T02:20:27.241605+00:00"
|
| 103 |
-
},
|
| 104 |
-
{
|
| 105 |
-
"step": 8112,
|
| 106 |
-
"epoch": 0.014408321018529045,
|
| 107 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-8112",
|
| 108 |
-
"timestamp": "2026-04-09T02:26:44.879194+00:00"
|
| 109 |
-
},
|
| 110 |
-
{
|
| 111 |
-
"step": 8619,
|
| 112 |
-
"epoch": 0.01530884108218711,
|
| 113 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-8619",
|
| 114 |
-
"timestamp": "2026-04-09T02:32:09.353801+00:00"
|
| 115 |
-
},
|
| 116 |
-
{
|
| 117 |
-
"step": 9126,
|
| 118 |
-
"epoch": 0.016209361145845175,
|
| 119 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-9126",
|
| 120 |
-
"timestamp": "2026-04-09T02:37:18.094294+00:00"
|
| 121 |
-
},
|
| 122 |
-
{
|
| 123 |
-
"step": 9633,
|
| 124 |
-
"epoch": 0.017109881209503238,
|
| 125 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-9633",
|
| 126 |
-
"timestamp": "2026-04-09T02:42:25.354730+00:00"
|
| 127 |
-
},
|
| 128 |
-
{
|
| 129 |
-
"step": 10140,
|
| 130 |
-
"epoch": 0.018010401273161305,
|
| 131 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-10140",
|
| 132 |
-
"timestamp": "2026-04-09T02:47:29.896532+00:00"
|
| 133 |
-
},
|
| 134 |
-
{
|
| 135 |
-
"step": 10647,
|
| 136 |
-
"epoch": 0.01891092133681937,
|
| 137 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-10647",
|
| 138 |
-
"timestamp": "2026-04-09T02:52:33.759945+00:00"
|
| 139 |
-
},
|
| 140 |
-
{
|
| 141 |
-
"step": 11154,
|
| 142 |
-
"epoch": 0.019811441400477435,
|
| 143 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-11154",
|
| 144 |
-
"timestamp": "2026-04-09T02:57:47.308962+00:00"
|
| 145 |
-
},
|
| 146 |
-
{
|
| 147 |
-
"step": 11661,
|
| 148 |
-
"epoch": 0.020711961464135502,
|
| 149 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-11661",
|
| 150 |
-
"timestamp": "2026-04-09T03:03:03.600465+00:00"
|
| 151 |
-
},
|
| 152 |
-
{
|
| 153 |
-
"step": 12168,
|
| 154 |
-
"epoch": 0.021612481527793565,
|
| 155 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-12168",
|
| 156 |
-
"timestamp": "2026-04-09T03:08:25.241749+00:00"
|
| 157 |
-
},
|
| 158 |
-
{
|
| 159 |
-
"step": 12675,
|
| 160 |
-
"epoch": 0.022513001591451632,
|
| 161 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-12675",
|
| 162 |
-
"timestamp": "2026-04-09T03:13:36.439541+00:00"
|
| 163 |
-
},
|
| 164 |
-
{
|
| 165 |
-
"step": 13182,
|
| 166 |
-
"epoch": 0.023413521655109695,
|
| 167 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-13182",
|
| 168 |
-
"timestamp": "2026-04-09T03:18:46.544532+00:00"
|
| 169 |
-
},
|
| 170 |
-
{
|
| 171 |
-
"step": 13689,
|
| 172 |
-
"epoch": 0.024314041718767762,
|
| 173 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-13689",
|
| 174 |
-
"timestamp": "2026-04-09T03:24:16.394544+00:00"
|
| 175 |
-
},
|
| 176 |
-
{
|
| 177 |
-
"step": 14196,
|
| 178 |
-
"epoch": 0.025214561782425825,
|
| 179 |
-
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-14196",
|
| 180 |
-
"timestamp": "2026-04-09T03:29:35.818009+00:00"
|
| 181 |
}
|
| 182 |
],
|
| 183 |
"epoch_artifacts": [],
|
| 184 |
"resume_history": [],
|
| 185 |
-
"created_at": "2026-04-
|
| 186 |
-
"updated_at": "2026-04-09T03:
|
| 187 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"schema_version": 1,
|
| 3 |
+
"run_id": "09bed511-51e5-467b-b9b7-356e13c38e62",
|
| 4 |
"run_type": "grpo",
|
| 5 |
"profile_name": "llama3_2_1b_base_grpo",
|
| 6 |
"model_name": "meta-llama/Llama-3.2-1B",
|
| 7 |
"status": "running",
|
| 8 |
"best_metric_name": "eval_cer_subset",
|
| 9 |
+
"best_metric_value": 0.012913837184685785,
|
| 10 |
+
"best_step": 507,
|
| 11 |
+
"best_checkpoint_step": 507,
|
| 12 |
+
"latest_checkpoint_step": 507,
|
| 13 |
"checkpoints": [
|
| 14 |
{
|
| 15 |
"step": 507,
|
| 16 |
+
"epoch": 0.0009005168647093303,
|
| 17 |
"repo_path": "meta-llama__llama-3.2-1b/grpo/checkpoints/checkpoint-507",
|
| 18 |
+
"timestamp": "2026-04-09T03:42:39.674363+00:00"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
}
|
| 20 |
],
|
| 21 |
"epoch_artifacts": [],
|
| 22 |
"resume_history": [],
|
| 23 |
+
"created_at": "2026-04-09T03:35:11.674696+00:00",
|
| 24 |
+
"updated_at": "2026-04-09T03:42:39.674379+00:00"
|
| 25 |
}
|