Upload checkpoint-30/trainer_state.json with huggingface_hub
Browse files
checkpoint-30/trainer_state.json
ADDED
|
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.9375,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 30,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.15625,
|
| 14 |
+
"grad_norm": 0.21668776869773865,
|
| 15 |
+
"learning_rate": 0.00017333333333333334,
|
| 16 |
+
"loss": 5.2044,
|
| 17 |
+
"step": 5
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.3125,
|
| 21 |
+
"grad_norm": 0.24632510542869568,
|
| 22 |
+
"learning_rate": 0.00014,
|
| 23 |
+
"loss": 5.2178,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.46875,
|
| 28 |
+
"grad_norm": 0.26152700185775757,
|
| 29 |
+
"learning_rate": 0.00010666666666666667,
|
| 30 |
+
"loss": 5.2657,
|
| 31 |
+
"step": 15
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.625,
|
| 35 |
+
"grad_norm": 0.24584853649139404,
|
| 36 |
+
"learning_rate": 7.333333333333333e-05,
|
| 37 |
+
"loss": 5.2687,
|
| 38 |
+
"step": 20
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.78125,
|
| 42 |
+
"grad_norm": 0.25921785831451416,
|
| 43 |
+
"learning_rate": 4e-05,
|
| 44 |
+
"loss": 5.1945,
|
| 45 |
+
"step": 25
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.9375,
|
| 49 |
+
"grad_norm": 0.25720950961112976,
|
| 50 |
+
"learning_rate": 6.666666666666667e-06,
|
| 51 |
+
"loss": 5.207,
|
| 52 |
+
"step": 30
|
| 53 |
+
}
|
| 54 |
+
],
|
| 55 |
+
"logging_steps": 5,
|
| 56 |
+
"max_steps": 30,
|
| 57 |
+
"num_input_tokens_seen": 0,
|
| 58 |
+
"num_train_epochs": 1,
|
| 59 |
+
"save_steps": 50,
|
| 60 |
+
"stateful_callbacks": {
|
| 61 |
+
"TrainerControl": {
|
| 62 |
+
"args": {
|
| 63 |
+
"should_epoch_stop": false,
|
| 64 |
+
"should_evaluate": false,
|
| 65 |
+
"should_log": false,
|
| 66 |
+
"should_save": true,
|
| 67 |
+
"should_training_stop": true
|
| 68 |
+
},
|
| 69 |
+
"attributes": {}
|
| 70 |
+
}
|
| 71 |
+
},
|
| 72 |
+
"total_flos": 11620030464.0,
|
| 73 |
+
"train_batch_size": 2,
|
| 74 |
+
"trial_name": null,
|
| 75 |
+
"trial_params": null
|
| 76 |
+
}
|