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": 0.
|
| 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,
|
|
@@ -134,6 +134,48 @@
|
|
| 134 |
"eval_samples_per_second": 4.392,
|
| 135 |
"eval_steps_per_second": 1.098,
|
| 136 |
"step": 60
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 137 |
}
|
| 138 |
],
|
| 139 |
"logging_steps": 10,
|
|
@@ -153,7 +195,7 @@
|
|
| 153 |
"attributes": {}
|
| 154 |
}
|
| 155 |
},
|
| 156 |
-
"total_flos":
|
| 157 |
"train_batch_size": 4,
|
| 158 |
"trial_name": null,
|
| 159 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.903954802259887,
|
| 6 |
"eval_steps": 10,
|
| 7 |
+
"global_step": 80,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 134 |
"eval_samples_per_second": 4.392,
|
| 135 |
"eval_steps_per_second": 1.098,
|
| 136 |
"step": 60
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"entropy": 6.680863696336746,
|
| 140 |
+
"epoch": 0.7909604519774012,
|
| 141 |
+
"grad_norm": 1.34375,
|
| 142 |
+
"learning_rate": 0.00012674418604651164,
|
| 143 |
+
"loss": 6.45898666381836,
|
| 144 |
+
"mean_token_accuracy": 0.21339080817997455,
|
| 145 |
+
"num_tokens": 1181787.0,
|
| 146 |
+
"step": 70
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.7909604519774012,
|
| 150 |
+
"eval_entropy": 6.870167553424835,
|
| 151 |
+
"eval_loss": 6.5776286125183105,
|
| 152 |
+
"eval_mean_token_accuracy": 0.19155203737318516,
|
| 153 |
+
"eval_num_tokens": 1181787.0,
|
| 154 |
+
"eval_runtime": 43.9254,
|
| 155 |
+
"eval_samples_per_second": 4.371,
|
| 156 |
+
"eval_steps_per_second": 1.093,
|
| 157 |
+
"step": 70
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"entropy": 6.659492689371109,
|
| 161 |
+
"epoch": 0.903954802259887,
|
| 162 |
+
"grad_norm": 0.82421875,
|
| 163 |
+
"learning_rate": 0.00011511627906976746,
|
| 164 |
+
"loss": 6.499990081787109,
|
| 165 |
+
"mean_token_accuracy": 0.21302505303174257,
|
| 166 |
+
"num_tokens": 1360714.0,
|
| 167 |
+
"step": 80
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 0.903954802259887,
|
| 171 |
+
"eval_entropy": 6.760685175657272,
|
| 172 |
+
"eval_loss": 6.4918060302734375,
|
| 173 |
+
"eval_mean_token_accuracy": 0.20070527338733277,
|
| 174 |
+
"eval_num_tokens": 1360714.0,
|
| 175 |
+
"eval_runtime": 44.0105,
|
| 176 |
+
"eval_samples_per_second": 4.363,
|
| 177 |
+
"eval_steps_per_second": 1.091,
|
| 178 |
+
"step": 80
|
| 179 |
}
|
| 180 |
],
|
| 181 |
"logging_steps": 10,
|
|
|
|
| 195 |
"attributes": {}
|
| 196 |
}
|
| 197 |
},
|
| 198 |
+
"total_flos": 9.734138912925696e+16,
|
| 199 |
"train_batch_size": 4,
|
| 200 |
"trial_name": null,
|
| 201 |
"trial_params": null
|