Delete checkpoint-1000/trainer_state.json
Browse files
checkpoint-1000/trainer_state.json
DELETED
|
@@ -1,78 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_global_step": null,
|
| 3 |
-
"best_metric": null,
|
| 4 |
-
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.10828370330265295,
|
| 6 |
-
"eval_steps": 500,
|
| 7 |
-
"global_step": 1000,
|
| 8 |
-
"is_hyper_param_search": false,
|
| 9 |
-
"is_local_process_zero": true,
|
| 10 |
-
"is_world_process_zero": true,
|
| 11 |
-
"log_history": [
|
| 12 |
-
{
|
| 13 |
-
"epoch": 0.02707092582566324,
|
| 14 |
-
"grad_norm": 0.9713733196258545,
|
| 15 |
-
"learning_rate": 9.910124526258799e-05,
|
| 16 |
-
"loss": 3.4945,
|
| 17 |
-
"step": 250
|
| 18 |
-
},
|
| 19 |
-
{
|
| 20 |
-
"epoch": 0.05414185165132648,
|
| 21 |
-
"grad_norm": 0.9521434307098389,
|
| 22 |
-
"learning_rate": 9.819888106839921e-05,
|
| 23 |
-
"loss": 3.5097,
|
| 24 |
-
"step": 500
|
| 25 |
-
},
|
| 26 |
-
{
|
| 27 |
-
"epoch": 0.05414185165132648,
|
| 28 |
-
"eval_loss": 2.95473575592041,
|
| 29 |
-
"eval_runtime": 87.829,
|
| 30 |
-
"eval_samples_per_second": 112.15,
|
| 31 |
-
"eval_steps_per_second": 7.014,
|
| 32 |
-
"step": 500
|
| 33 |
-
},
|
| 34 |
-
{
|
| 35 |
-
"epoch": 0.08121277747698971,
|
| 36 |
-
"grad_norm": 0.9334385395050049,
|
| 37 |
-
"learning_rate": 9.729651687421044e-05,
|
| 38 |
-
"loss": 3.4922,
|
| 39 |
-
"step": 750
|
| 40 |
-
},
|
| 41 |
-
{
|
| 42 |
-
"epoch": 0.10828370330265295,
|
| 43 |
-
"grad_norm": 1.2239603996276855,
|
| 44 |
-
"learning_rate": 9.639415268002166e-05,
|
| 45 |
-
"loss": 3.477,
|
| 46 |
-
"step": 1000
|
| 47 |
-
},
|
| 48 |
-
{
|
| 49 |
-
"epoch": 0.10828370330265295,
|
| 50 |
-
"eval_loss": 2.9319565296173096,
|
| 51 |
-
"eval_runtime": 87.8964,
|
| 52 |
-
"eval_samples_per_second": 112.064,
|
| 53 |
-
"eval_steps_per_second": 7.008,
|
| 54 |
-
"step": 1000
|
| 55 |
-
}
|
| 56 |
-
],
|
| 57 |
-
"logging_steps": 250,
|
| 58 |
-
"max_steps": 27705,
|
| 59 |
-
"num_input_tokens_seen": 0,
|
| 60 |
-
"num_train_epochs": 3,
|
| 61 |
-
"save_steps": 1000,
|
| 62 |
-
"stateful_callbacks": {
|
| 63 |
-
"TrainerControl": {
|
| 64 |
-
"args": {
|
| 65 |
-
"should_epoch_stop": false,
|
| 66 |
-
"should_evaluate": false,
|
| 67 |
-
"should_log": false,
|
| 68 |
-
"should_save": true,
|
| 69 |
-
"should_training_stop": false
|
| 70 |
-
},
|
| 71 |
-
"attributes": {}
|
| 72 |
-
}
|
| 73 |
-
},
|
| 74 |
-
"total_flos": 361054076928000.0,
|
| 75 |
-
"train_batch_size": 16,
|
| 76 |
-
"trial_name": null,
|
| 77 |
-
"trial_params": null
|
| 78 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|