Upload trainer_state.json
Browse files- trainer_state.json +61 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 98176,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.61,
|
| 12 |
+
"learning_rate": 4.168681083504881e-06,
|
| 13 |
+
"loss": 0.3631,
|
| 14 |
+
"step": 30000
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.0,
|
| 18 |
+
"eval_accuracy": 0.9129903209373408,
|
| 19 |
+
"eval_loss": 0.31294044852256775,
|
| 20 |
+
"eval_runtime": 78.2833,
|
| 21 |
+
"eval_samples_per_second": 125.378,
|
| 22 |
+
"eval_steps_per_second": 15.674,
|
| 23 |
+
"step": 49088
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 1.22,
|
| 27 |
+
"learning_rate": 2.334304873326132e-06,
|
| 28 |
+
"loss": 0.2901,
|
| 29 |
+
"step": 60000
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 1.83,
|
| 33 |
+
"learning_rate": 4.999286631473819e-07,
|
| 34 |
+
"loss": 0.2267,
|
| 35 |
+
"step": 90000
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"epoch": 2.0,
|
| 39 |
+
"eval_accuracy": 0.915333672949567,
|
| 40 |
+
"eval_loss": 0.41572055220603943,
|
| 41 |
+
"eval_runtime": 78.1399,
|
| 42 |
+
"eval_samples_per_second": 125.608,
|
| 43 |
+
"eval_steps_per_second": 15.703,
|
| 44 |
+
"step": 98176
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.0,
|
| 48 |
+
"step": 98176,
|
| 49 |
+
"total_flos": 3.65975017449044e+17,
|
| 50 |
+
"train_loss": 0.2868618026395198,
|
| 51 |
+
"train_runtime": 21298.9707,
|
| 52 |
+
"train_samples_per_second": 36.875,
|
| 53 |
+
"train_steps_per_second": 4.609
|
| 54 |
+
}
|
| 55 |
+
],
|
| 56 |
+
"max_steps": 98176,
|
| 57 |
+
"num_train_epochs": 2,
|
| 58 |
+
"total_flos": 3.65975017449044e+17,
|
| 59 |
+
"trial_name": null,
|
| 60 |
+
"trial_params": null
|
| 61 |
+
}
|