End of training
Browse files- all_results.json +12 -0
- eval_results.json +8 -0
- train_results.json +7 -0
- trainer_state.json +910 -0
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9949622166246851,
|
| 4 |
+
"eval_loss": 0.0237874872982502,
|
| 5 |
+
"eval_runtime": 68.8899,
|
| 6 |
+
"eval_samples_per_second": 5.763,
|
| 7 |
+
"eval_steps_per_second": 0.726,
|
| 8 |
+
"train_loss": 0.08033438997336553,
|
| 9 |
+
"train_runtime": 6338.3889,
|
| 10 |
+
"train_samples_per_second": 1.77,
|
| 11 |
+
"train_steps_per_second": 0.222
|
| 12 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9949622166246851,
|
| 4 |
+
"eval_loss": 0.0237874872982502,
|
| 5 |
+
"eval_runtime": 68.8899,
|
| 6 |
+
"eval_samples_per_second": 5.763,
|
| 7 |
+
"eval_steps_per_second": 0.726
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 0.08033438997336553,
|
| 4 |
+
"train_runtime": 6338.3889,
|
| 5 |
+
"train_samples_per_second": 1.77,
|
| 6 |
+
"train_steps_per_second": 0.222
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,910 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.0237874872982502,
|
| 3 |
+
"best_model_checkpoint": "./outputs/checkpoint-1405",
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"global_step": 1405,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"learning_rate": 1.9857651245551603e-05,
|
| 13 |
+
"loss": 0.6071,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.07,
|
| 18 |
+
"learning_rate": 1.9715302491103204e-05,
|
| 19 |
+
"loss": 0.5349,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.11,
|
| 24 |
+
"learning_rate": 1.9572953736654805e-05,
|
| 25 |
+
"loss": 0.3979,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.14,
|
| 30 |
+
"learning_rate": 1.9430604982206406e-05,
|
| 31 |
+
"loss": 0.338,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.18,
|
| 36 |
+
"learning_rate": 1.9288256227758007e-05,
|
| 37 |
+
"loss": 0.314,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.21,
|
| 42 |
+
"learning_rate": 1.9145907473309612e-05,
|
| 43 |
+
"loss": 0.2133,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.25,
|
| 48 |
+
"learning_rate": 1.9003558718861213e-05,
|
| 49 |
+
"loss": 0.197,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.28,
|
| 54 |
+
"learning_rate": 1.886120996441281e-05,
|
| 55 |
+
"loss": 0.1823,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.32,
|
| 60 |
+
"learning_rate": 1.8718861209964415e-05,
|
| 61 |
+
"loss": 0.159,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.36,
|
| 66 |
+
"learning_rate": 1.8576512455516017e-05,
|
| 67 |
+
"loss": 0.162,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.39,
|
| 72 |
+
"learning_rate": 1.8434163701067618e-05,
|
| 73 |
+
"loss": 0.1466,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.43,
|
| 78 |
+
"learning_rate": 1.829181494661922e-05,
|
| 79 |
+
"loss": 0.1919,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.46,
|
| 84 |
+
"learning_rate": 1.814946619217082e-05,
|
| 85 |
+
"loss": 0.0872,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.5,
|
| 90 |
+
"learning_rate": 1.800711743772242e-05,
|
| 91 |
+
"loss": 0.1353,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.53,
|
| 96 |
+
"learning_rate": 1.7864768683274022e-05,
|
| 97 |
+
"loss": 0.1294,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.57,
|
| 102 |
+
"learning_rate": 1.7722419928825624e-05,
|
| 103 |
+
"loss": 0.0948,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.6,
|
| 108 |
+
"learning_rate": 1.7580071174377225e-05,
|
| 109 |
+
"loss": 0.1228,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.64,
|
| 114 |
+
"learning_rate": 1.7437722419928826e-05,
|
| 115 |
+
"loss": 0.1502,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.68,
|
| 120 |
+
"learning_rate": 1.7295373665480427e-05,
|
| 121 |
+
"loss": 0.1073,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.71,
|
| 126 |
+
"learning_rate": 1.715302491103203e-05,
|
| 127 |
+
"loss": 0.0488,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.75,
|
| 132 |
+
"learning_rate": 1.7010676156583633e-05,
|
| 133 |
+
"loss": 0.18,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.78,
|
| 138 |
+
"learning_rate": 1.686832740213523e-05,
|
| 139 |
+
"loss": 0.1082,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.82,
|
| 144 |
+
"learning_rate": 1.6725978647686835e-05,
|
| 145 |
+
"loss": 0.0977,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.85,
|
| 150 |
+
"learning_rate": 1.6583629893238436e-05,
|
| 151 |
+
"loss": 0.1078,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.89,
|
| 156 |
+
"learning_rate": 1.6441281138790037e-05,
|
| 157 |
+
"loss": 0.0355,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.93,
|
| 162 |
+
"learning_rate": 1.629893238434164e-05,
|
| 163 |
+
"loss": 0.1555,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.96,
|
| 168 |
+
"learning_rate": 1.615658362989324e-05,
|
| 169 |
+
"loss": 0.0614,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 1.0,
|
| 174 |
+
"learning_rate": 1.601423487544484e-05,
|
| 175 |
+
"loss": 0.0548,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.0,
|
| 180 |
+
"eval_accuracy": 0.9798488664987406,
|
| 181 |
+
"eval_loss": 0.06160992756485939,
|
| 182 |
+
"eval_runtime": 74.4135,
|
| 183 |
+
"eval_samples_per_second": 5.335,
|
| 184 |
+
"eval_steps_per_second": 0.672,
|
| 185 |
+
"step": 281
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.03,
|
| 189 |
+
"learning_rate": 1.5871886120996442e-05,
|
| 190 |
+
"loss": 0.1327,
|
| 191 |
+
"step": 290
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.07,
|
| 195 |
+
"learning_rate": 1.5729537366548043e-05,
|
| 196 |
+
"loss": 0.0859,
|
| 197 |
+
"step": 300
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.1,
|
| 201 |
+
"learning_rate": 1.5587188612099648e-05,
|
| 202 |
+
"loss": 0.0389,
|
| 203 |
+
"step": 310
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.14,
|
| 207 |
+
"learning_rate": 1.5444839857651245e-05,
|
| 208 |
+
"loss": 0.1412,
|
| 209 |
+
"step": 320
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 1.17,
|
| 213 |
+
"learning_rate": 1.5302491103202847e-05,
|
| 214 |
+
"loss": 0.0414,
|
| 215 |
+
"step": 330
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 1.21,
|
| 219 |
+
"learning_rate": 1.516014234875445e-05,
|
| 220 |
+
"loss": 0.0825,
|
| 221 |
+
"step": 340
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.25,
|
| 225 |
+
"learning_rate": 1.5017793594306052e-05,
|
| 226 |
+
"loss": 0.1417,
|
| 227 |
+
"step": 350
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 1.28,
|
| 231 |
+
"learning_rate": 1.4875444839857654e-05,
|
| 232 |
+
"loss": 0.0896,
|
| 233 |
+
"step": 360
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.32,
|
| 237 |
+
"learning_rate": 1.4733096085409253e-05,
|
| 238 |
+
"loss": 0.1049,
|
| 239 |
+
"step": 370
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 1.35,
|
| 243 |
+
"learning_rate": 1.4590747330960854e-05,
|
| 244 |
+
"loss": 0.1624,
|
| 245 |
+
"step": 380
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.39,
|
| 249 |
+
"learning_rate": 1.4448398576512457e-05,
|
| 250 |
+
"loss": 0.0422,
|
| 251 |
+
"step": 390
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 1.42,
|
| 255 |
+
"learning_rate": 1.4306049822064058e-05,
|
| 256 |
+
"loss": 0.0757,
|
| 257 |
+
"step": 400
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 1.46,
|
| 261 |
+
"learning_rate": 1.4163701067615661e-05,
|
| 262 |
+
"loss": 0.0572,
|
| 263 |
+
"step": 410
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.49,
|
| 267 |
+
"learning_rate": 1.402135231316726e-05,
|
| 268 |
+
"loss": 0.0887,
|
| 269 |
+
"step": 420
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.53,
|
| 273 |
+
"learning_rate": 1.3879003558718862e-05,
|
| 274 |
+
"loss": 0.0711,
|
| 275 |
+
"step": 430
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.57,
|
| 279 |
+
"learning_rate": 1.3736654804270464e-05,
|
| 280 |
+
"loss": 0.0399,
|
| 281 |
+
"step": 440
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.6,
|
| 285 |
+
"learning_rate": 1.3594306049822066e-05,
|
| 286 |
+
"loss": 0.1151,
|
| 287 |
+
"step": 450
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.64,
|
| 291 |
+
"learning_rate": 1.3451957295373668e-05,
|
| 292 |
+
"loss": 0.0224,
|
| 293 |
+
"step": 460
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.67,
|
| 297 |
+
"learning_rate": 1.3309608540925268e-05,
|
| 298 |
+
"loss": 0.0326,
|
| 299 |
+
"step": 470
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.71,
|
| 303 |
+
"learning_rate": 1.3167259786476869e-05,
|
| 304 |
+
"loss": 0.0245,
|
| 305 |
+
"step": 480
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.74,
|
| 309 |
+
"learning_rate": 1.302491103202847e-05,
|
| 310 |
+
"loss": 0.0582,
|
| 311 |
+
"step": 490
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.78,
|
| 315 |
+
"learning_rate": 1.2882562277580073e-05,
|
| 316 |
+
"loss": 0.0277,
|
| 317 |
+
"step": 500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.81,
|
| 321 |
+
"learning_rate": 1.2740213523131673e-05,
|
| 322 |
+
"loss": 0.029,
|
| 323 |
+
"step": 510
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.85,
|
| 327 |
+
"learning_rate": 1.2597864768683274e-05,
|
| 328 |
+
"loss": 0.1294,
|
| 329 |
+
"step": 520
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.89,
|
| 333 |
+
"learning_rate": 1.2455516014234877e-05,
|
| 334 |
+
"loss": 0.0685,
|
| 335 |
+
"step": 530
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.92,
|
| 339 |
+
"learning_rate": 1.2313167259786478e-05,
|
| 340 |
+
"loss": 0.0502,
|
| 341 |
+
"step": 540
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.96,
|
| 345 |
+
"learning_rate": 1.217081850533808e-05,
|
| 346 |
+
"loss": 0.0737,
|
| 347 |
+
"step": 550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.99,
|
| 351 |
+
"learning_rate": 1.202846975088968e-05,
|
| 352 |
+
"loss": 0.1366,
|
| 353 |
+
"step": 560
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 2.0,
|
| 357 |
+
"eval_accuracy": 0.9899244332493703,
|
| 358 |
+
"eval_loss": 0.033983081579208374,
|
| 359 |
+
"eval_runtime": 68.9693,
|
| 360 |
+
"eval_samples_per_second": 5.756,
|
| 361 |
+
"eval_steps_per_second": 0.725,
|
| 362 |
+
"step": 562
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.03,
|
| 366 |
+
"learning_rate": 1.1886120996441281e-05,
|
| 367 |
+
"loss": 0.0415,
|
| 368 |
+
"step": 570
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.06,
|
| 372 |
+
"learning_rate": 1.1743772241992884e-05,
|
| 373 |
+
"loss": 0.0728,
|
| 374 |
+
"step": 580
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 2.1,
|
| 378 |
+
"learning_rate": 1.1601423487544485e-05,
|
| 379 |
+
"loss": 0.0908,
|
| 380 |
+
"step": 590
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.14,
|
| 384 |
+
"learning_rate": 1.1459074733096086e-05,
|
| 385 |
+
"loss": 0.0763,
|
| 386 |
+
"step": 600
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.17,
|
| 390 |
+
"learning_rate": 1.1316725978647688e-05,
|
| 391 |
+
"loss": 0.0673,
|
| 392 |
+
"step": 610
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.21,
|
| 396 |
+
"learning_rate": 1.1174377224199289e-05,
|
| 397 |
+
"loss": 0.0342,
|
| 398 |
+
"step": 620
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.24,
|
| 402 |
+
"learning_rate": 1.103202846975089e-05,
|
| 403 |
+
"loss": 0.1702,
|
| 404 |
+
"step": 630
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.28,
|
| 408 |
+
"learning_rate": 1.0889679715302493e-05,
|
| 409 |
+
"loss": 0.1066,
|
| 410 |
+
"step": 640
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.31,
|
| 414 |
+
"learning_rate": 1.0747330960854094e-05,
|
| 415 |
+
"loss": 0.0286,
|
| 416 |
+
"step": 650
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.35,
|
| 420 |
+
"learning_rate": 1.0604982206405693e-05,
|
| 421 |
+
"loss": 0.0598,
|
| 422 |
+
"step": 660
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.38,
|
| 426 |
+
"learning_rate": 1.0462633451957296e-05,
|
| 427 |
+
"loss": 0.0357,
|
| 428 |
+
"step": 670
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.42,
|
| 432 |
+
"learning_rate": 1.0320284697508897e-05,
|
| 433 |
+
"loss": 0.0357,
|
| 434 |
+
"step": 680
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.46,
|
| 438 |
+
"learning_rate": 1.01779359430605e-05,
|
| 439 |
+
"loss": 0.0513,
|
| 440 |
+
"step": 690
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.49,
|
| 444 |
+
"learning_rate": 1.0035587188612101e-05,
|
| 445 |
+
"loss": 0.102,
|
| 446 |
+
"step": 700
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.53,
|
| 450 |
+
"learning_rate": 9.893238434163703e-06,
|
| 451 |
+
"loss": 0.0238,
|
| 452 |
+
"step": 710
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.56,
|
| 456 |
+
"learning_rate": 9.750889679715304e-06,
|
| 457 |
+
"loss": 0.125,
|
| 458 |
+
"step": 720
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.6,
|
| 462 |
+
"learning_rate": 9.608540925266905e-06,
|
| 463 |
+
"loss": 0.0175,
|
| 464 |
+
"step": 730
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.63,
|
| 468 |
+
"learning_rate": 9.466192170818506e-06,
|
| 469 |
+
"loss": 0.08,
|
| 470 |
+
"step": 740
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.67,
|
| 474 |
+
"learning_rate": 9.323843416370107e-06,
|
| 475 |
+
"loss": 0.0481,
|
| 476 |
+
"step": 750
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.7,
|
| 480 |
+
"learning_rate": 9.18149466192171e-06,
|
| 481 |
+
"loss": 0.0472,
|
| 482 |
+
"step": 760
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.74,
|
| 486 |
+
"learning_rate": 9.03914590747331e-06,
|
| 487 |
+
"loss": 0.104,
|
| 488 |
+
"step": 770
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.78,
|
| 492 |
+
"learning_rate": 8.896797153024912e-06,
|
| 493 |
+
"loss": 0.0658,
|
| 494 |
+
"step": 780
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.81,
|
| 498 |
+
"learning_rate": 8.754448398576513e-06,
|
| 499 |
+
"loss": 0.0218,
|
| 500 |
+
"step": 790
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.85,
|
| 504 |
+
"learning_rate": 8.612099644128115e-06,
|
| 505 |
+
"loss": 0.0675,
|
| 506 |
+
"step": 800
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.88,
|
| 510 |
+
"learning_rate": 8.469750889679716e-06,
|
| 511 |
+
"loss": 0.0443,
|
| 512 |
+
"step": 810
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.92,
|
| 516 |
+
"learning_rate": 8.327402135231317e-06,
|
| 517 |
+
"loss": 0.0131,
|
| 518 |
+
"step": 820
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.95,
|
| 522 |
+
"learning_rate": 8.18505338078292e-06,
|
| 523 |
+
"loss": 0.0516,
|
| 524 |
+
"step": 830
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.99,
|
| 528 |
+
"learning_rate": 8.04270462633452e-06,
|
| 529 |
+
"loss": 0.0218,
|
| 530 |
+
"step": 840
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 3.0,
|
| 534 |
+
"eval_accuracy": 0.9874055415617129,
|
| 535 |
+
"eval_loss": 0.043008677661418915,
|
| 536 |
+
"eval_runtime": 71.679,
|
| 537 |
+
"eval_samples_per_second": 5.539,
|
| 538 |
+
"eval_steps_per_second": 0.698,
|
| 539 |
+
"step": 843
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 3.02,
|
| 543 |
+
"learning_rate": 7.900355871886122e-06,
|
| 544 |
+
"loss": 0.0286,
|
| 545 |
+
"step": 850
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 3.06,
|
| 549 |
+
"learning_rate": 7.758007117437723e-06,
|
| 550 |
+
"loss": 0.0847,
|
| 551 |
+
"step": 860
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 3.1,
|
| 555 |
+
"learning_rate": 7.6156583629893245e-06,
|
| 556 |
+
"loss": 0.0136,
|
| 557 |
+
"step": 870
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 3.13,
|
| 561 |
+
"learning_rate": 7.4733096085409265e-06,
|
| 562 |
+
"loss": 0.0341,
|
| 563 |
+
"step": 880
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 3.17,
|
| 567 |
+
"learning_rate": 7.330960854092527e-06,
|
| 568 |
+
"loss": 0.008,
|
| 569 |
+
"step": 890
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 3.2,
|
| 573 |
+
"learning_rate": 7.188612099644129e-06,
|
| 574 |
+
"loss": 0.0331,
|
| 575 |
+
"step": 900
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 3.24,
|
| 579 |
+
"learning_rate": 7.04626334519573e-06,
|
| 580 |
+
"loss": 0.0187,
|
| 581 |
+
"step": 910
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 3.27,
|
| 585 |
+
"learning_rate": 6.903914590747331e-06,
|
| 586 |
+
"loss": 0.0155,
|
| 587 |
+
"step": 920
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 3.31,
|
| 591 |
+
"learning_rate": 6.761565836298933e-06,
|
| 592 |
+
"loss": 0.0344,
|
| 593 |
+
"step": 930
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 3.35,
|
| 597 |
+
"learning_rate": 6.619217081850534e-06,
|
| 598 |
+
"loss": 0.0745,
|
| 599 |
+
"step": 940
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 3.38,
|
| 603 |
+
"learning_rate": 6.476868327402136e-06,
|
| 604 |
+
"loss": 0.024,
|
| 605 |
+
"step": 950
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 3.42,
|
| 609 |
+
"learning_rate": 6.334519572953737e-06,
|
| 610 |
+
"loss": 0.0764,
|
| 611 |
+
"step": 960
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 3.45,
|
| 615 |
+
"learning_rate": 6.192170818505339e-06,
|
| 616 |
+
"loss": 0.0163,
|
| 617 |
+
"step": 970
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 3.49,
|
| 621 |
+
"learning_rate": 6.049822064056941e-06,
|
| 622 |
+
"loss": 0.0163,
|
| 623 |
+
"step": 980
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 3.52,
|
| 627 |
+
"learning_rate": 5.907473309608541e-06,
|
| 628 |
+
"loss": 0.0265,
|
| 629 |
+
"step": 990
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 3.56,
|
| 633 |
+
"learning_rate": 5.765124555160143e-06,
|
| 634 |
+
"loss": 0.0268,
|
| 635 |
+
"step": 1000
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 3.59,
|
| 639 |
+
"learning_rate": 5.622775800711744e-06,
|
| 640 |
+
"loss": 0.0242,
|
| 641 |
+
"step": 1010
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 3.63,
|
| 645 |
+
"learning_rate": 5.480427046263346e-06,
|
| 646 |
+
"loss": 0.0103,
|
| 647 |
+
"step": 1020
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 3.67,
|
| 651 |
+
"learning_rate": 5.338078291814946e-06,
|
| 652 |
+
"loss": 0.111,
|
| 653 |
+
"step": 1030
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 3.7,
|
| 657 |
+
"learning_rate": 5.195729537366548e-06,
|
| 658 |
+
"loss": 0.0368,
|
| 659 |
+
"step": 1040
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 3.74,
|
| 663 |
+
"learning_rate": 5.05338078291815e-06,
|
| 664 |
+
"loss": 0.0599,
|
| 665 |
+
"step": 1050
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 3.77,
|
| 669 |
+
"learning_rate": 4.911032028469751e-06,
|
| 670 |
+
"loss": 0.0894,
|
| 671 |
+
"step": 1060
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 3.81,
|
| 675 |
+
"learning_rate": 4.768683274021353e-06,
|
| 676 |
+
"loss": 0.0144,
|
| 677 |
+
"step": 1070
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 3.84,
|
| 681 |
+
"learning_rate": 4.626334519572954e-06,
|
| 682 |
+
"loss": 0.0337,
|
| 683 |
+
"step": 1080
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 3.88,
|
| 687 |
+
"learning_rate": 4.483985765124556e-06,
|
| 688 |
+
"loss": 0.0204,
|
| 689 |
+
"step": 1090
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 3.91,
|
| 693 |
+
"learning_rate": 4.341637010676157e-06,
|
| 694 |
+
"loss": 0.0226,
|
| 695 |
+
"step": 1100
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 3.95,
|
| 699 |
+
"learning_rate": 4.199288256227758e-06,
|
| 700 |
+
"loss": 0.0826,
|
| 701 |
+
"step": 1110
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 3.99,
|
| 705 |
+
"learning_rate": 4.05693950177936e-06,
|
| 706 |
+
"loss": 0.0403,
|
| 707 |
+
"step": 1120
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 4.0,
|
| 711 |
+
"eval_accuracy": 0.9874055415617129,
|
| 712 |
+
"eval_loss": 0.04059485346078873,
|
| 713 |
+
"eval_runtime": 69.4464,
|
| 714 |
+
"eval_samples_per_second": 5.717,
|
| 715 |
+
"eval_steps_per_second": 0.72,
|
| 716 |
+
"step": 1124
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 4.02,
|
| 720 |
+
"learning_rate": 3.914590747330961e-06,
|
| 721 |
+
"loss": 0.0569,
|
| 722 |
+
"step": 1130
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 4.06,
|
| 726 |
+
"learning_rate": 3.7722419928825625e-06,
|
| 727 |
+
"loss": 0.0302,
|
| 728 |
+
"step": 1140
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 4.09,
|
| 732 |
+
"learning_rate": 3.629893238434164e-06,
|
| 733 |
+
"loss": 0.0795,
|
| 734 |
+
"step": 1150
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 4.13,
|
| 738 |
+
"learning_rate": 3.4875444839857653e-06,
|
| 739 |
+
"loss": 0.0749,
|
| 740 |
+
"step": 1160
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 4.16,
|
| 744 |
+
"learning_rate": 3.3451957295373664e-06,
|
| 745 |
+
"loss": 0.0384,
|
| 746 |
+
"step": 1170
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 4.2,
|
| 750 |
+
"learning_rate": 3.2028469750889684e-06,
|
| 751 |
+
"loss": 0.043,
|
| 752 |
+
"step": 1180
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 4.23,
|
| 756 |
+
"learning_rate": 3.0604982206405696e-06,
|
| 757 |
+
"loss": 0.0306,
|
| 758 |
+
"step": 1190
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.27,
|
| 762 |
+
"learning_rate": 2.918149466192171e-06,
|
| 763 |
+
"loss": 0.0181,
|
| 764 |
+
"step": 1200
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.31,
|
| 768 |
+
"learning_rate": 2.7758007117437723e-06,
|
| 769 |
+
"loss": 0.0289,
|
| 770 |
+
"step": 1210
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.34,
|
| 774 |
+
"learning_rate": 2.633451957295374e-06,
|
| 775 |
+
"loss": 0.0631,
|
| 776 |
+
"step": 1220
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.38,
|
| 780 |
+
"learning_rate": 2.4911032028469755e-06,
|
| 781 |
+
"loss": 0.0454,
|
| 782 |
+
"step": 1230
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.41,
|
| 786 |
+
"learning_rate": 2.3487544483985766e-06,
|
| 787 |
+
"loss": 0.0461,
|
| 788 |
+
"step": 1240
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.45,
|
| 792 |
+
"learning_rate": 2.2064056939501782e-06,
|
| 793 |
+
"loss": 0.0233,
|
| 794 |
+
"step": 1250
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.48,
|
| 798 |
+
"learning_rate": 2.0640569395017794e-06,
|
| 799 |
+
"loss": 0.0123,
|
| 800 |
+
"step": 1260
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.52,
|
| 804 |
+
"learning_rate": 1.921708185053381e-06,
|
| 805 |
+
"loss": 0.075,
|
| 806 |
+
"step": 1270
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.56,
|
| 810 |
+
"learning_rate": 1.7793594306049826e-06,
|
| 811 |
+
"loss": 0.0131,
|
| 812 |
+
"step": 1280
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.59,
|
| 816 |
+
"learning_rate": 1.6370106761565837e-06,
|
| 817 |
+
"loss": 0.0068,
|
| 818 |
+
"step": 1290
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.63,
|
| 822 |
+
"learning_rate": 1.494661921708185e-06,
|
| 823 |
+
"loss": 0.0212,
|
| 824 |
+
"step": 1300
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.66,
|
| 828 |
+
"learning_rate": 1.3523131672597867e-06,
|
| 829 |
+
"loss": 0.0062,
|
| 830 |
+
"step": 1310
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.7,
|
| 834 |
+
"learning_rate": 1.209964412811388e-06,
|
| 835 |
+
"loss": 0.075,
|
| 836 |
+
"step": 1320
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.73,
|
| 840 |
+
"learning_rate": 1.0676156583629894e-06,
|
| 841 |
+
"loss": 0.0776,
|
| 842 |
+
"step": 1330
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.77,
|
| 846 |
+
"learning_rate": 9.252669039145908e-07,
|
| 847 |
+
"loss": 0.0238,
|
| 848 |
+
"step": 1340
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.8,
|
| 852 |
+
"learning_rate": 7.829181494661923e-07,
|
| 853 |
+
"loss": 0.0174,
|
| 854 |
+
"step": 1350
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 4.84,
|
| 858 |
+
"learning_rate": 6.405693950177936e-07,
|
| 859 |
+
"loss": 0.1009,
|
| 860 |
+
"step": 1360
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 4.88,
|
| 864 |
+
"learning_rate": 4.982206405693951e-07,
|
| 865 |
+
"loss": 0.0823,
|
| 866 |
+
"step": 1370
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 4.91,
|
| 870 |
+
"learning_rate": 3.558718861209965e-07,
|
| 871 |
+
"loss": 0.091,
|
| 872 |
+
"step": 1380
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 4.95,
|
| 876 |
+
"learning_rate": 2.135231316725979e-07,
|
| 877 |
+
"loss": 0.0182,
|
| 878 |
+
"step": 1390
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.98,
|
| 882 |
+
"learning_rate": 7.117437722419929e-08,
|
| 883 |
+
"loss": 0.0184,
|
| 884 |
+
"step": 1400
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 5.0,
|
| 888 |
+
"eval_accuracy": 0.9949622166246851,
|
| 889 |
+
"eval_loss": 0.0237874872982502,
|
| 890 |
+
"eval_runtime": 70.1815,
|
| 891 |
+
"eval_samples_per_second": 5.657,
|
| 892 |
+
"eval_steps_per_second": 0.712,
|
| 893 |
+
"step": 1405
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 5.0,
|
| 897 |
+
"step": 1405,
|
| 898 |
+
"total_flos": 8.694601234757222e+17,
|
| 899 |
+
"train_loss": 0.08033438997336553,
|
| 900 |
+
"train_runtime": 6338.3889,
|
| 901 |
+
"train_samples_per_second": 1.77,
|
| 902 |
+
"train_steps_per_second": 0.222
|
| 903 |
+
}
|
| 904 |
+
],
|
| 905 |
+
"max_steps": 1405,
|
| 906 |
+
"num_train_epochs": 5,
|
| 907 |
+
"total_flos": 8.694601234757222e+17,
|
| 908 |
+
"trial_name": null,
|
| 909 |
+
"trial_params": null
|
| 910 |
+
}
|