Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +45 -3
trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 1.
|
| 6 |
"eval_steps": 10,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -260,6 +260,48 @@
|
|
| 260 |
"eval_samples_per_second": 4.362,
|
| 261 |
"eval_steps_per_second": 1.091,
|
| 262 |
"step": 120
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 263 |
}
|
| 264 |
],
|
| 265 |
"logging_steps": 10,
|
|
@@ -279,7 +321,7 @@
|
|
| 279 |
"attributes": {}
|
| 280 |
}
|
| 281 |
},
|
| 282 |
-
"total_flos": 1.
|
| 283 |
"train_batch_size": 4,
|
| 284 |
"trial_name": null,
|
| 285 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.576271186440678,
|
| 6 |
"eval_steps": 10,
|
| 7 |
+
"global_step": 140,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 260 |
"eval_samples_per_second": 4.362,
|
| 261 |
"eval_steps_per_second": 1.091,
|
| 262 |
"step": 120
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"entropy": 6.290887945890427,
|
| 266 |
+
"epoch": 1.463276836158192,
|
| 267 |
+
"grad_norm": 0.7265625,
|
| 268 |
+
"learning_rate": 5.697674418604652e-05,
|
| 269 |
+
"loss": 6.218046951293945,
|
| 270 |
+
"mean_token_accuracy": 0.24080509012565016,
|
| 271 |
+
"num_tokens": 2183142.0,
|
| 272 |
+
"step": 130
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.463276836158192,
|
| 276 |
+
"eval_entropy": 6.626457552115123,
|
| 277 |
+
"eval_loss": 6.388693332672119,
|
| 278 |
+
"eval_mean_token_accuracy": 0.20111992427458367,
|
| 279 |
+
"eval_num_tokens": 2183142.0,
|
| 280 |
+
"eval_runtime": 43.8362,
|
| 281 |
+
"eval_samples_per_second": 4.38,
|
| 282 |
+
"eval_steps_per_second": 1.095,
|
| 283 |
+
"step": 130
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"entropy": 6.49563906788826,
|
| 287 |
+
"epoch": 1.576271186440678,
|
| 288 |
+
"grad_norm": 0.6015625,
|
| 289 |
+
"learning_rate": 4.5348837209302326e-05,
|
| 290 |
+
"loss": 6.376762771606446,
|
| 291 |
+
"mean_token_accuracy": 0.21492539951577783,
|
| 292 |
+
"num_tokens": 2364034.0,
|
| 293 |
+
"step": 140
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.576271186440678,
|
| 297 |
+
"eval_entropy": 6.611844847599666,
|
| 298 |
+
"eval_loss": 6.380344390869141,
|
| 299 |
+
"eval_mean_token_accuracy": 0.20116183906793594,
|
| 300 |
+
"eval_num_tokens": 2364034.0,
|
| 301 |
+
"eval_runtime": 44.3142,
|
| 302 |
+
"eval_samples_per_second": 4.333,
|
| 303 |
+
"eval_steps_per_second": 1.083,
|
| 304 |
+
"step": 140
|
| 305 |
}
|
| 306 |
],
|
| 307 |
"logging_steps": 10,
|
|
|
|
| 321 |
"attributes": {}
|
| 322 |
}
|
| 323 |
},
|
| 324 |
+
"total_flos": 1.6869770537105203e+17,
|
| 325 |
"train_batch_size": 4,
|
| 326 |
"trial_name": null,
|
| 327 |
"trial_params": null
|