End of training
Browse files- all_results.json +13 -0
- eval_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +1009 -0
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.98,
|
| 3 |
+
"eval_accuracy": 0.998719590268886,
|
| 4 |
+
"eval_loss": 0.06100322678685188,
|
| 5 |
+
"eval_runtime": 64.552,
|
| 6 |
+
"eval_samples_per_second": 108.889,
|
| 7 |
+
"eval_steps_per_second": 1.704,
|
| 8 |
+
"total_flos": 3.041208679733851e+19,
|
| 9 |
+
"train_loss": 0.46931519971240676,
|
| 10 |
+
"train_runtime": 11436.94,
|
| 11 |
+
"train_samples_per_second": 34.419,
|
| 12 |
+
"train_steps_per_second": 0.134
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.98,
|
| 3 |
+
"eval_accuracy": 0.998719590268886,
|
| 4 |
+
"eval_loss": 0.06100322678685188,
|
| 5 |
+
"eval_runtime": 64.552,
|
| 6 |
+
"eval_samples_per_second": 108.889,
|
| 7 |
+
"eval_steps_per_second": 1.704
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.98,
|
| 3 |
+
"total_flos": 3.041208679733851e+19,
|
| 4 |
+
"train_loss": 0.46931519971240676,
|
| 5 |
+
"train_runtime": 11436.94,
|
| 6 |
+
"train_samples_per_second": 34.419,
|
| 7 |
+
"train_steps_per_second": 0.134
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1009 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.998719590268886,
|
| 3 |
+
"best_model_checkpoint": "PlantDiseaseDetectorV2/checkpoint-1533",
|
| 4 |
+
"epoch": 6.976109215017065,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1533,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.05,
|
| 13 |
+
"learning_rate": 3.2467532467532465e-06,
|
| 14 |
+
"loss": 3.6354,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.09,
|
| 19 |
+
"learning_rate": 6.493506493506493e-06,
|
| 20 |
+
"loss": 3.6146,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.14,
|
| 25 |
+
"learning_rate": 9.740259740259742e-06,
|
| 26 |
+
"loss": 3.5784,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.18,
|
| 31 |
+
"learning_rate": 1.2987012987012986e-05,
|
| 32 |
+
"loss": 3.5183,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.23,
|
| 37 |
+
"learning_rate": 1.6233766233766234e-05,
|
| 38 |
+
"loss": 3.435,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.27,
|
| 43 |
+
"learning_rate": 1.9480519480519483e-05,
|
| 44 |
+
"loss": 3.3236,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.32,
|
| 49 |
+
"learning_rate": 2.272727272727273e-05,
|
| 50 |
+
"loss": 3.1774,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.36,
|
| 55 |
+
"learning_rate": 2.5974025974025972e-05,
|
| 56 |
+
"loss": 3.0042,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.41,
|
| 61 |
+
"learning_rate": 2.922077922077922e-05,
|
| 62 |
+
"loss": 2.8095,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.46,
|
| 67 |
+
"learning_rate": 3.246753246753247e-05,
|
| 68 |
+
"loss": 2.6034,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.5,
|
| 73 |
+
"learning_rate": 3.571428571428572e-05,
|
| 74 |
+
"loss": 2.3817,
|
| 75 |
+
"step": 110
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.55,
|
| 79 |
+
"learning_rate": 3.8961038961038966e-05,
|
| 80 |
+
"loss": 2.1751,
|
| 81 |
+
"step": 120
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.59,
|
| 85 |
+
"learning_rate": 4.220779220779221e-05,
|
| 86 |
+
"loss": 1.9796,
|
| 87 |
+
"step": 130
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.64,
|
| 91 |
+
"learning_rate": 4.545454545454546e-05,
|
| 92 |
+
"loss": 1.7805,
|
| 93 |
+
"step": 140
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.68,
|
| 97 |
+
"learning_rate": 4.87012987012987e-05,
|
| 98 |
+
"loss": 1.6209,
|
| 99 |
+
"step": 150
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.73,
|
| 103 |
+
"learning_rate": 4.978245105148659e-05,
|
| 104 |
+
"loss": 1.4515,
|
| 105 |
+
"step": 160
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.77,
|
| 109 |
+
"learning_rate": 4.941986947063089e-05,
|
| 110 |
+
"loss": 1.3197,
|
| 111 |
+
"step": 170
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.82,
|
| 115 |
+
"learning_rate": 4.90572878897752e-05,
|
| 116 |
+
"loss": 1.1848,
|
| 117 |
+
"step": 180
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.86,
|
| 121 |
+
"learning_rate": 4.8694706308919505e-05,
|
| 122 |
+
"loss": 1.0976,
|
| 123 |
+
"step": 190
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.91,
|
| 127 |
+
"learning_rate": 4.833212472806382e-05,
|
| 128 |
+
"loss": 0.9857,
|
| 129 |
+
"step": 200
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.96,
|
| 133 |
+
"learning_rate": 4.7969543147208126e-05,
|
| 134 |
+
"loss": 0.9051,
|
| 135 |
+
"step": 210
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 1.0,
|
| 139 |
+
"eval_accuracy": 0.986057760705648,
|
| 140 |
+
"eval_loss": 0.8024668097496033,
|
| 141 |
+
"eval_runtime": 70.6659,
|
| 142 |
+
"eval_samples_per_second": 99.468,
|
| 143 |
+
"eval_steps_per_second": 1.557,
|
| 144 |
+
"step": 219
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.0,
|
| 148 |
+
"learning_rate": 4.760696156635243e-05,
|
| 149 |
+
"loss": 0.8431,
|
| 150 |
+
"step": 220
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 1.05,
|
| 154 |
+
"learning_rate": 4.724437998549674e-05,
|
| 155 |
+
"loss": 0.7784,
|
| 156 |
+
"step": 230
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 1.09,
|
| 160 |
+
"learning_rate": 4.6881798404641046e-05,
|
| 161 |
+
"loss": 0.7198,
|
| 162 |
+
"step": 240
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 1.14,
|
| 166 |
+
"learning_rate": 4.651921682378535e-05,
|
| 167 |
+
"loss": 0.6712,
|
| 168 |
+
"step": 250
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.18,
|
| 172 |
+
"learning_rate": 4.6156635242929666e-05,
|
| 173 |
+
"loss": 0.6365,
|
| 174 |
+
"step": 260
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 1.23,
|
| 178 |
+
"learning_rate": 4.579405366207397e-05,
|
| 179 |
+
"loss": 0.5892,
|
| 180 |
+
"step": 270
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 1.27,
|
| 184 |
+
"learning_rate": 4.543147208121827e-05,
|
| 185 |
+
"loss": 0.5563,
|
| 186 |
+
"step": 280
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 1.32,
|
| 190 |
+
"learning_rate": 4.506889050036258e-05,
|
| 191 |
+
"loss": 0.5254,
|
| 192 |
+
"step": 290
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 1.37,
|
| 196 |
+
"learning_rate": 4.4706308919506886e-05,
|
| 197 |
+
"loss": 0.4982,
|
| 198 |
+
"step": 300
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 1.41,
|
| 202 |
+
"learning_rate": 4.43437273386512e-05,
|
| 203 |
+
"loss": 0.4667,
|
| 204 |
+
"step": 310
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 1.46,
|
| 208 |
+
"learning_rate": 4.398114575779551e-05,
|
| 209 |
+
"loss": 0.4481,
|
| 210 |
+
"step": 320
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 1.5,
|
| 214 |
+
"learning_rate": 4.3618564176939813e-05,
|
| 215 |
+
"loss": 0.4199,
|
| 216 |
+
"step": 330
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 1.55,
|
| 220 |
+
"learning_rate": 4.325598259608412e-05,
|
| 221 |
+
"loss": 0.4117,
|
| 222 |
+
"step": 340
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 1.59,
|
| 226 |
+
"learning_rate": 4.289340101522843e-05,
|
| 227 |
+
"loss": 0.3867,
|
| 228 |
+
"step": 350
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 1.64,
|
| 232 |
+
"learning_rate": 4.253081943437274e-05,
|
| 233 |
+
"loss": 0.3736,
|
| 234 |
+
"step": 360
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 1.68,
|
| 238 |
+
"learning_rate": 4.216823785351705e-05,
|
| 239 |
+
"loss": 0.3668,
|
| 240 |
+
"step": 370
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.73,
|
| 244 |
+
"learning_rate": 4.1805656272661354e-05,
|
| 245 |
+
"loss": 0.3369,
|
| 246 |
+
"step": 380
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 1.77,
|
| 250 |
+
"learning_rate": 4.1443074691805654e-05,
|
| 251 |
+
"loss": 0.3416,
|
| 252 |
+
"step": 390
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.82,
|
| 256 |
+
"learning_rate": 4.108049311094996e-05,
|
| 257 |
+
"loss": 0.3287,
|
| 258 |
+
"step": 400
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 1.87,
|
| 262 |
+
"learning_rate": 4.0717911530094274e-05,
|
| 263 |
+
"loss": 0.2989,
|
| 264 |
+
"step": 410
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 1.91,
|
| 268 |
+
"learning_rate": 4.035532994923858e-05,
|
| 269 |
+
"loss": 0.2934,
|
| 270 |
+
"step": 420
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 1.96,
|
| 274 |
+
"learning_rate": 3.999274836838289e-05,
|
| 275 |
+
"loss": 0.2801,
|
| 276 |
+
"step": 430
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 2.0,
|
| 280 |
+
"eval_accuracy": 0.9958742353108551,
|
| 281 |
+
"eval_loss": 0.26061177253723145,
|
| 282 |
+
"eval_runtime": 71.3214,
|
| 283 |
+
"eval_samples_per_second": 98.554,
|
| 284 |
+
"eval_steps_per_second": 1.542,
|
| 285 |
+
"step": 439
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 2.0,
|
| 289 |
+
"learning_rate": 3.9630166787527195e-05,
|
| 290 |
+
"loss": 0.2742,
|
| 291 |
+
"step": 440
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 2.05,
|
| 295 |
+
"learning_rate": 3.92675852066715e-05,
|
| 296 |
+
"loss": 0.2615,
|
| 297 |
+
"step": 450
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 2.09,
|
| 301 |
+
"learning_rate": 3.890500362581581e-05,
|
| 302 |
+
"loss": 0.2512,
|
| 303 |
+
"step": 460
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 2.14,
|
| 307 |
+
"learning_rate": 3.854242204496012e-05,
|
| 308 |
+
"loss": 0.2442,
|
| 309 |
+
"step": 470
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 2.18,
|
| 313 |
+
"learning_rate": 3.817984046410443e-05,
|
| 314 |
+
"loss": 0.2369,
|
| 315 |
+
"step": 480
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 2.23,
|
| 319 |
+
"learning_rate": 3.7817258883248735e-05,
|
| 320 |
+
"loss": 0.2288,
|
| 321 |
+
"step": 490
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 2.28,
|
| 325 |
+
"learning_rate": 3.7454677302393035e-05,
|
| 326 |
+
"loss": 0.2226,
|
| 327 |
+
"step": 500
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 2.32,
|
| 331 |
+
"learning_rate": 3.709209572153735e-05,
|
| 332 |
+
"loss": 0.2162,
|
| 333 |
+
"step": 510
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 2.37,
|
| 337 |
+
"learning_rate": 3.6729514140681655e-05,
|
| 338 |
+
"loss": 0.2055,
|
| 339 |
+
"step": 520
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 2.41,
|
| 343 |
+
"learning_rate": 3.636693255982596e-05,
|
| 344 |
+
"loss": 0.199,
|
| 345 |
+
"step": 530
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 2.46,
|
| 349 |
+
"learning_rate": 3.600435097897027e-05,
|
| 350 |
+
"loss": 0.2128,
|
| 351 |
+
"step": 540
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 2.5,
|
| 355 |
+
"learning_rate": 3.5641769398114576e-05,
|
| 356 |
+
"loss": 0.1885,
|
| 357 |
+
"step": 550
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 2.55,
|
| 361 |
+
"learning_rate": 3.527918781725888e-05,
|
| 362 |
+
"loss": 0.1854,
|
| 363 |
+
"step": 560
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 2.59,
|
| 367 |
+
"learning_rate": 3.4916606236403196e-05,
|
| 368 |
+
"loss": 0.1845,
|
| 369 |
+
"step": 570
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 2.64,
|
| 373 |
+
"learning_rate": 3.45540246555475e-05,
|
| 374 |
+
"loss": 0.1814,
|
| 375 |
+
"step": 580
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 2.68,
|
| 379 |
+
"learning_rate": 3.419144307469181e-05,
|
| 380 |
+
"loss": 0.1816,
|
| 381 |
+
"step": 590
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 2.73,
|
| 385 |
+
"learning_rate": 3.3828861493836116e-05,
|
| 386 |
+
"loss": 0.1689,
|
| 387 |
+
"step": 600
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 2.78,
|
| 391 |
+
"learning_rate": 3.3466279912980416e-05,
|
| 392 |
+
"loss": 0.1658,
|
| 393 |
+
"step": 610
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 2.82,
|
| 397 |
+
"learning_rate": 3.310369833212473e-05,
|
| 398 |
+
"loss": 0.1612,
|
| 399 |
+
"step": 620
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 2.87,
|
| 403 |
+
"learning_rate": 3.2741116751269036e-05,
|
| 404 |
+
"loss": 0.1572,
|
| 405 |
+
"step": 630
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 2.91,
|
| 409 |
+
"learning_rate": 3.237853517041334e-05,
|
| 410 |
+
"loss": 0.1552,
|
| 411 |
+
"step": 640
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 2.96,
|
| 415 |
+
"learning_rate": 3.201595358955765e-05,
|
| 416 |
+
"loss": 0.1455,
|
| 417 |
+
"step": 650
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 3.0,
|
| 421 |
+
"eval_accuracy": 0.9972969127898705,
|
| 422 |
+
"eval_loss": 0.14022748172283173,
|
| 423 |
+
"eval_runtime": 71.3997,
|
| 424 |
+
"eval_samples_per_second": 98.446,
|
| 425 |
+
"eval_steps_per_second": 1.541,
|
| 426 |
+
"step": 659
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 3.0,
|
| 430 |
+
"learning_rate": 3.165337200870196e-05,
|
| 431 |
+
"loss": 0.1446,
|
| 432 |
+
"step": 660
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 3.05,
|
| 436 |
+
"learning_rate": 3.129079042784627e-05,
|
| 437 |
+
"loss": 0.1422,
|
| 438 |
+
"step": 670
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 3.09,
|
| 442 |
+
"learning_rate": 3.092820884699058e-05,
|
| 443 |
+
"loss": 0.1411,
|
| 444 |
+
"step": 680
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 3.14,
|
| 448 |
+
"learning_rate": 3.0565627266134884e-05,
|
| 449 |
+
"loss": 0.1334,
|
| 450 |
+
"step": 690
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 3.19,
|
| 454 |
+
"learning_rate": 3.020304568527919e-05,
|
| 455 |
+
"loss": 0.1333,
|
| 456 |
+
"step": 700
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 3.23,
|
| 460 |
+
"learning_rate": 2.98404641044235e-05,
|
| 461 |
+
"loss": 0.1325,
|
| 462 |
+
"step": 710
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 3.28,
|
| 466 |
+
"learning_rate": 2.9477882523567807e-05,
|
| 467 |
+
"loss": 0.1278,
|
| 468 |
+
"step": 720
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 3.32,
|
| 472 |
+
"learning_rate": 2.911530094271211e-05,
|
| 473 |
+
"loss": 0.127,
|
| 474 |
+
"step": 730
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 3.37,
|
| 478 |
+
"learning_rate": 2.8752719361856417e-05,
|
| 479 |
+
"loss": 0.1272,
|
| 480 |
+
"step": 740
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 3.41,
|
| 484 |
+
"learning_rate": 2.8390137781000724e-05,
|
| 485 |
+
"loss": 0.1289,
|
| 486 |
+
"step": 750
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 3.46,
|
| 490 |
+
"learning_rate": 2.8027556200145034e-05,
|
| 491 |
+
"loss": 0.1235,
|
| 492 |
+
"step": 760
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 3.5,
|
| 496 |
+
"learning_rate": 2.766497461928934e-05,
|
| 497 |
+
"loss": 0.1194,
|
| 498 |
+
"step": 770
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 3.55,
|
| 502 |
+
"learning_rate": 2.7302393038433648e-05,
|
| 503 |
+
"loss": 0.1179,
|
| 504 |
+
"step": 780
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 3.59,
|
| 508 |
+
"learning_rate": 2.6939811457577958e-05,
|
| 509 |
+
"loss": 0.11,
|
| 510 |
+
"step": 790
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 3.64,
|
| 514 |
+
"learning_rate": 2.6577229876722265e-05,
|
| 515 |
+
"loss": 0.1126,
|
| 516 |
+
"step": 800
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 3.69,
|
| 520 |
+
"learning_rate": 2.621464829586657e-05,
|
| 521 |
+
"loss": 0.1106,
|
| 522 |
+
"step": 810
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 3.73,
|
| 526 |
+
"learning_rate": 2.585206671501088e-05,
|
| 527 |
+
"loss": 0.1073,
|
| 528 |
+
"step": 820
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 3.78,
|
| 532 |
+
"learning_rate": 2.548948513415519e-05,
|
| 533 |
+
"loss": 0.1024,
|
| 534 |
+
"step": 830
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 3.82,
|
| 538 |
+
"learning_rate": 2.5126903553299492e-05,
|
| 539 |
+
"loss": 0.1001,
|
| 540 |
+
"step": 840
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 3.87,
|
| 544 |
+
"learning_rate": 2.4764321972443802e-05,
|
| 545 |
+
"loss": 0.1013,
|
| 546 |
+
"step": 850
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 3.91,
|
| 550 |
+
"learning_rate": 2.440174039158811e-05,
|
| 551 |
+
"loss": 0.1003,
|
| 552 |
+
"step": 860
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 3.96,
|
| 556 |
+
"learning_rate": 2.4039158810732415e-05,
|
| 557 |
+
"loss": 0.0949,
|
| 558 |
+
"step": 870
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 4.0,
|
| 562 |
+
"eval_accuracy": 0.9985773225209845,
|
| 563 |
+
"eval_loss": 0.09417087584733963,
|
| 564 |
+
"eval_runtime": 71.2514,
|
| 565 |
+
"eval_samples_per_second": 98.651,
|
| 566 |
+
"eval_steps_per_second": 1.544,
|
| 567 |
+
"step": 879
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 4.0,
|
| 571 |
+
"learning_rate": 2.3676577229876722e-05,
|
| 572 |
+
"loss": 0.0965,
|
| 573 |
+
"step": 880
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 4.05,
|
| 577 |
+
"learning_rate": 2.3313995649021032e-05,
|
| 578 |
+
"loss": 0.0936,
|
| 579 |
+
"step": 890
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 4.1,
|
| 583 |
+
"learning_rate": 2.295141406816534e-05,
|
| 584 |
+
"loss": 0.0938,
|
| 585 |
+
"step": 900
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 4.14,
|
| 589 |
+
"learning_rate": 2.2588832487309646e-05,
|
| 590 |
+
"loss": 0.0907,
|
| 591 |
+
"step": 910
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 4.19,
|
| 595 |
+
"learning_rate": 2.2226250906453953e-05,
|
| 596 |
+
"loss": 0.0889,
|
| 597 |
+
"step": 920
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 4.23,
|
| 601 |
+
"learning_rate": 2.186366932559826e-05,
|
| 602 |
+
"loss": 0.0889,
|
| 603 |
+
"step": 930
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 4.28,
|
| 607 |
+
"learning_rate": 2.150108774474257e-05,
|
| 608 |
+
"loss": 0.0905,
|
| 609 |
+
"step": 940
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 4.32,
|
| 613 |
+
"learning_rate": 2.1138506163886876e-05,
|
| 614 |
+
"loss": 0.0888,
|
| 615 |
+
"step": 950
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 4.37,
|
| 619 |
+
"learning_rate": 2.0775924583031183e-05,
|
| 620 |
+
"loss": 0.0849,
|
| 621 |
+
"step": 960
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 4.41,
|
| 625 |
+
"learning_rate": 2.0413343002175493e-05,
|
| 626 |
+
"loss": 0.0852,
|
| 627 |
+
"step": 970
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 4.46,
|
| 631 |
+
"learning_rate": 2.0050761421319797e-05,
|
| 632 |
+
"loss": 0.0843,
|
| 633 |
+
"step": 980
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 4.51,
|
| 637 |
+
"learning_rate": 1.9688179840464103e-05,
|
| 638 |
+
"loss": 0.0851,
|
| 639 |
+
"step": 990
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 4.55,
|
| 643 |
+
"learning_rate": 1.9325598259608413e-05,
|
| 644 |
+
"loss": 0.0884,
|
| 645 |
+
"step": 1000
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 4.6,
|
| 649 |
+
"learning_rate": 1.896301667875272e-05,
|
| 650 |
+
"loss": 0.0899,
|
| 651 |
+
"step": 1010
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 4.64,
|
| 655 |
+
"learning_rate": 1.860043509789703e-05,
|
| 656 |
+
"loss": 0.0848,
|
| 657 |
+
"step": 1020
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 4.69,
|
| 661 |
+
"learning_rate": 1.8237853517041334e-05,
|
| 662 |
+
"loss": 0.0794,
|
| 663 |
+
"step": 1030
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 4.73,
|
| 667 |
+
"learning_rate": 1.787527193618564e-05,
|
| 668 |
+
"loss": 0.0802,
|
| 669 |
+
"step": 1040
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 4.78,
|
| 673 |
+
"learning_rate": 1.751269035532995e-05,
|
| 674 |
+
"loss": 0.0808,
|
| 675 |
+
"step": 1050
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 4.82,
|
| 679 |
+
"learning_rate": 1.7150108774474257e-05,
|
| 680 |
+
"loss": 0.0763,
|
| 681 |
+
"step": 1060
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 4.87,
|
| 685 |
+
"learning_rate": 1.6787527193618564e-05,
|
| 686 |
+
"loss": 0.0801,
|
| 687 |
+
"step": 1070
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 4.91,
|
| 691 |
+
"learning_rate": 1.6424945612762874e-05,
|
| 692 |
+
"loss": 0.0745,
|
| 693 |
+
"step": 1080
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 4.96,
|
| 697 |
+
"learning_rate": 1.6062364031907178e-05,
|
| 698 |
+
"loss": 0.0741,
|
| 699 |
+
"step": 1090
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 5.0,
|
| 703 |
+
"eval_accuracy": 0.9984350547730829,
|
| 704 |
+
"eval_loss": 0.07485666126012802,
|
| 705 |
+
"eval_runtime": 71.7237,
|
| 706 |
+
"eval_samples_per_second": 98.001,
|
| 707 |
+
"eval_steps_per_second": 1.534,
|
| 708 |
+
"step": 1098
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 5.01,
|
| 712 |
+
"learning_rate": 1.5699782451051488e-05,
|
| 713 |
+
"loss": 0.0738,
|
| 714 |
+
"step": 1100
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 5.05,
|
| 718 |
+
"learning_rate": 1.5337200870195794e-05,
|
| 719 |
+
"loss": 0.0729,
|
| 720 |
+
"step": 1110
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 5.1,
|
| 724 |
+
"learning_rate": 1.4974619289340103e-05,
|
| 725 |
+
"loss": 0.0748,
|
| 726 |
+
"step": 1120
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 5.14,
|
| 730 |
+
"learning_rate": 1.4612037708484411e-05,
|
| 731 |
+
"loss": 0.0698,
|
| 732 |
+
"step": 1130
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 5.19,
|
| 736 |
+
"learning_rate": 1.4249456127628716e-05,
|
| 737 |
+
"loss": 0.0721,
|
| 738 |
+
"step": 1140
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 5.23,
|
| 742 |
+
"learning_rate": 1.3886874546773023e-05,
|
| 743 |
+
"loss": 0.0748,
|
| 744 |
+
"step": 1150
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 5.28,
|
| 748 |
+
"learning_rate": 1.3524292965917332e-05,
|
| 749 |
+
"loss": 0.0695,
|
| 750 |
+
"step": 1160
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 5.32,
|
| 754 |
+
"learning_rate": 1.316171138506164e-05,
|
| 755 |
+
"loss": 0.0688,
|
| 756 |
+
"step": 1170
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 5.37,
|
| 760 |
+
"learning_rate": 1.2799129804205947e-05,
|
| 761 |
+
"loss": 0.0723,
|
| 762 |
+
"step": 1180
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 5.42,
|
| 766 |
+
"learning_rate": 1.2436548223350254e-05,
|
| 767 |
+
"loss": 0.0683,
|
| 768 |
+
"step": 1190
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 5.46,
|
| 772 |
+
"learning_rate": 1.2073966642494562e-05,
|
| 773 |
+
"loss": 0.0709,
|
| 774 |
+
"step": 1200
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 5.51,
|
| 778 |
+
"learning_rate": 1.1711385061638869e-05,
|
| 779 |
+
"loss": 0.0674,
|
| 780 |
+
"step": 1210
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 5.55,
|
| 784 |
+
"learning_rate": 1.1348803480783177e-05,
|
| 785 |
+
"loss": 0.0685,
|
| 786 |
+
"step": 1220
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 5.6,
|
| 790 |
+
"learning_rate": 1.0986221899927484e-05,
|
| 791 |
+
"loss": 0.0679,
|
| 792 |
+
"step": 1230
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 5.64,
|
| 796 |
+
"learning_rate": 1.062364031907179e-05,
|
| 797 |
+
"loss": 0.0672,
|
| 798 |
+
"step": 1240
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 5.69,
|
| 802 |
+
"learning_rate": 1.02610587382161e-05,
|
| 803 |
+
"loss": 0.0648,
|
| 804 |
+
"step": 1250
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 5.73,
|
| 808 |
+
"learning_rate": 9.898477157360408e-06,
|
| 809 |
+
"loss": 0.0659,
|
| 810 |
+
"step": 1260
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 5.78,
|
| 814 |
+
"learning_rate": 9.535895576504714e-06,
|
| 815 |
+
"loss": 0.0629,
|
| 816 |
+
"step": 1270
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 5.82,
|
| 820 |
+
"learning_rate": 9.173313995649021e-06,
|
| 821 |
+
"loss": 0.0627,
|
| 822 |
+
"step": 1280
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 5.87,
|
| 826 |
+
"learning_rate": 8.81073241479333e-06,
|
| 827 |
+
"loss": 0.0643,
|
| 828 |
+
"step": 1290
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 5.92,
|
| 832 |
+
"learning_rate": 8.448150833937636e-06,
|
| 833 |
+
"loss": 0.065,
|
| 834 |
+
"step": 1300
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 5.96,
|
| 838 |
+
"learning_rate": 8.085569253081945e-06,
|
| 839 |
+
"loss": 0.0623,
|
| 840 |
+
"step": 1310
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 6.0,
|
| 844 |
+
"eval_accuracy": 0.9984350547730829,
|
| 845 |
+
"eval_loss": 0.06421820819377899,
|
| 846 |
+
"eval_runtime": 71.8803,
|
| 847 |
+
"eval_samples_per_second": 97.788,
|
| 848 |
+
"eval_steps_per_second": 1.53,
|
| 849 |
+
"step": 1318
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 6.01,
|
| 853 |
+
"learning_rate": 7.72298767222625e-06,
|
| 854 |
+
"loss": 0.0634,
|
| 855 |
+
"step": 1320
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 6.05,
|
| 859 |
+
"learning_rate": 7.360406091370558e-06,
|
| 860 |
+
"loss": 0.0625,
|
| 861 |
+
"step": 1330
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 6.1,
|
| 865 |
+
"learning_rate": 6.997824510514867e-06,
|
| 866 |
+
"loss": 0.0612,
|
| 867 |
+
"step": 1340
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 6.14,
|
| 871 |
+
"learning_rate": 6.635242929659173e-06,
|
| 872 |
+
"loss": 0.0607,
|
| 873 |
+
"step": 1350
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 6.19,
|
| 877 |
+
"learning_rate": 6.272661348803481e-06,
|
| 878 |
+
"loss": 0.0595,
|
| 879 |
+
"step": 1360
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 6.23,
|
| 883 |
+
"learning_rate": 5.910079767947789e-06,
|
| 884 |
+
"loss": 0.0602,
|
| 885 |
+
"step": 1370
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 6.28,
|
| 889 |
+
"learning_rate": 5.547498187092096e-06,
|
| 890 |
+
"loss": 0.0606,
|
| 891 |
+
"step": 1380
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 6.33,
|
| 895 |
+
"learning_rate": 5.184916606236404e-06,
|
| 896 |
+
"loss": 0.0608,
|
| 897 |
+
"step": 1390
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 6.37,
|
| 901 |
+
"learning_rate": 4.822335025380711e-06,
|
| 902 |
+
"loss": 0.0621,
|
| 903 |
+
"step": 1400
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 6.42,
|
| 907 |
+
"learning_rate": 4.459753444525018e-06,
|
| 908 |
+
"loss": 0.0608,
|
| 909 |
+
"step": 1410
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 6.46,
|
| 913 |
+
"learning_rate": 4.097171863669326e-06,
|
| 914 |
+
"loss": 0.0615,
|
| 915 |
+
"step": 1420
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 6.51,
|
| 919 |
+
"learning_rate": 3.734590282813633e-06,
|
| 920 |
+
"loss": 0.0604,
|
| 921 |
+
"step": 1430
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 6.55,
|
| 925 |
+
"learning_rate": 3.3720087019579407e-06,
|
| 926 |
+
"loss": 0.0612,
|
| 927 |
+
"step": 1440
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 6.6,
|
| 931 |
+
"learning_rate": 3.0094271211022483e-06,
|
| 932 |
+
"loss": 0.058,
|
| 933 |
+
"step": 1450
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 6.64,
|
| 937 |
+
"learning_rate": 2.646845540246556e-06,
|
| 938 |
+
"loss": 0.0598,
|
| 939 |
+
"step": 1460
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 6.69,
|
| 943 |
+
"learning_rate": 2.284263959390863e-06,
|
| 944 |
+
"loss": 0.0592,
|
| 945 |
+
"step": 1470
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 6.73,
|
| 949 |
+
"learning_rate": 1.9216823785351706e-06,
|
| 950 |
+
"loss": 0.058,
|
| 951 |
+
"step": 1480
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 6.78,
|
| 955 |
+
"learning_rate": 1.559100797679478e-06,
|
| 956 |
+
"loss": 0.0573,
|
| 957 |
+
"step": 1490
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 6.83,
|
| 961 |
+
"learning_rate": 1.1965192168237854e-06,
|
| 962 |
+
"loss": 0.058,
|
| 963 |
+
"step": 1500
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 6.87,
|
| 967 |
+
"learning_rate": 8.339376359680928e-07,
|
| 968 |
+
"loss": 0.0603,
|
| 969 |
+
"step": 1510
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 6.92,
|
| 973 |
+
"learning_rate": 4.713560551124003e-07,
|
| 974 |
+
"loss": 0.0576,
|
| 975 |
+
"step": 1520
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 6.96,
|
| 979 |
+
"learning_rate": 1.0877447425670776e-07,
|
| 980 |
+
"loss": 0.0586,
|
| 981 |
+
"step": 1530
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 6.98,
|
| 985 |
+
"eval_accuracy": 0.998719590268886,
|
| 986 |
+
"eval_loss": 0.06100322678685188,
|
| 987 |
+
"eval_runtime": 70.9668,
|
| 988 |
+
"eval_samples_per_second": 99.046,
|
| 989 |
+
"eval_steps_per_second": 1.55,
|
| 990 |
+
"step": 1533
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 6.98,
|
| 994 |
+
"step": 1533,
|
| 995 |
+
"total_flos": 3.041208679733851e+19,
|
| 996 |
+
"train_loss": 0.46931519971240676,
|
| 997 |
+
"train_runtime": 11436.94,
|
| 998 |
+
"train_samples_per_second": 34.419,
|
| 999 |
+
"train_steps_per_second": 0.134
|
| 1000 |
+
}
|
| 1001 |
+
],
|
| 1002 |
+
"logging_steps": 10,
|
| 1003 |
+
"max_steps": 1533,
|
| 1004 |
+
"num_train_epochs": 7,
|
| 1005 |
+
"save_steps": 500,
|
| 1006 |
+
"total_flos": 3.041208679733851e+19,
|
| 1007 |
+
"trial_name": null,
|
| 1008 |
+
"trial_params": null
|
| 1009 |
+
}
|