Commit ·
82e9a2a
1
Parent(s): 323e6f8
Upload trainer_state.json
Browse files- trainer_state.json +94 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 4.685408299866131,
|
| 5 |
+
"global_step": 3500,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.67,
|
| 12 |
+
"learning_rate": 1.7322623828647926e-05,
|
| 13 |
+
"loss": 0.0412,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.0,
|
| 18 |
+
"eval_accuracy": 0.993359375,
|
| 19 |
+
"eval_loss": 0.03145532310009003,
|
| 20 |
+
"eval_runtime": 22.6963,
|
| 21 |
+
"eval_samples_per_second": 225.587,
|
| 22 |
+
"eval_steps_per_second": 14.099,
|
| 23 |
+
"step": 747
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 1.34,
|
| 27 |
+
"learning_rate": 1.4645247657295851e-05,
|
| 28 |
+
"loss": 0.0047,
|
| 29 |
+
"step": 1000
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 2.0,
|
| 33 |
+
"eval_accuracy": 0.998828125,
|
| 34 |
+
"eval_loss": 0.007673000451177359,
|
| 35 |
+
"eval_runtime": 22.4025,
|
| 36 |
+
"eval_samples_per_second": 228.546,
|
| 37 |
+
"eval_steps_per_second": 14.284,
|
| 38 |
+
"step": 1494
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 2.01,
|
| 42 |
+
"learning_rate": 1.1967871485943776e-05,
|
| 43 |
+
"loss": 0.0057,
|
| 44 |
+
"step": 1500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.68,
|
| 48 |
+
"learning_rate": 9.2904953145917e-06,
|
| 49 |
+
"loss": 0.0015,
|
| 50 |
+
"step": 2000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.0,
|
| 54 |
+
"eval_accuracy": 0.998046875,
|
| 55 |
+
"eval_loss": 0.010105514898896217,
|
| 56 |
+
"eval_runtime": 22.3654,
|
| 57 |
+
"eval_samples_per_second": 228.925,
|
| 58 |
+
"eval_steps_per_second": 14.308,
|
| 59 |
+
"step": 2241
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 3.35,
|
| 63 |
+
"learning_rate": 6.6131191432396255e-06,
|
| 64 |
+
"loss": 0.0001,
|
| 65 |
+
"step": 2500
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 4.0,
|
| 69 |
+
"eval_accuracy": 0.9978515625,
|
| 70 |
+
"eval_loss": 0.012340093031525612,
|
| 71 |
+
"eval_runtime": 22.3303,
|
| 72 |
+
"eval_samples_per_second": 229.285,
|
| 73 |
+
"eval_steps_per_second": 14.33,
|
| 74 |
+
"step": 2988
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 4.02,
|
| 78 |
+
"learning_rate": 3.93574297188755e-06,
|
| 79 |
+
"loss": 0.0,
|
| 80 |
+
"step": 3000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 4.69,
|
| 84 |
+
"learning_rate": 1.2583668005354755e-06,
|
| 85 |
+
"loss": 0.0,
|
| 86 |
+
"step": 3500
|
| 87 |
+
}
|
| 88 |
+
],
|
| 89 |
+
"max_steps": 3735,
|
| 90 |
+
"num_train_epochs": 5,
|
| 91 |
+
"total_flos": 746705064001440.0,
|
| 92 |
+
"trial_name": null,
|
| 93 |
+
"trial_params": null
|
| 94 |
+
}
|