End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.94,
|
| 3 |
+
"eval_accuracy": 0.9988826815642458,
|
| 4 |
+
"eval_loss": 0.005963917821645737,
|
| 5 |
+
"eval_runtime": 41.8358,
|
| 6 |
+
"eval_samples_per_second": 21.393,
|
| 7 |
+
"eval_steps_per_second": 2.677,
|
| 8 |
+
"total_flos": 1.4186448226450612e+19,
|
| 9 |
+
"train_loss": 0.0581204283070731,
|
| 10 |
+
"train_runtime": 12121.0022,
|
| 11 |
+
"train_samples_per_second": 6.643,
|
| 12 |
+
"train_steps_per_second": 0.118
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.94,
|
| 3 |
+
"eval_accuracy": 0.9988826815642458,
|
| 4 |
+
"eval_loss": 0.005963917821645737,
|
| 5 |
+
"eval_runtime": 41.8358,
|
| 6 |
+
"eval_samples_per_second": 21.393,
|
| 7 |
+
"eval_steps_per_second": 2.677
|
| 8 |
+
}
|
runs/May08_08-04-19_8142974efca5/events.out.tfevents.1683545296.8142974efca5.225.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c9e20c1c5bd58cf905deb0b40df89afa726b0e83205cafe1e83df646ac39597
|
| 3 |
+
size 411
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.94,
|
| 3 |
+
"total_flos": 1.4186448226450612e+19,
|
| 4 |
+
"train_loss": 0.0581204283070731,
|
| 5 |
+
"train_runtime": 12121.0022,
|
| 6 |
+
"train_samples_per_second": 6.643,
|
| 7 |
+
"train_steps_per_second": 0.118
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,973 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9988826815642458,
|
| 3 |
+
"best_model_checkpoint": "convnext-large-224-22k-1k-LongSleeveCleanedData/checkpoint-287",
|
| 4 |
+
"epoch": 9.94041708043694,
|
| 5 |
+
"global_step": 1430,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.07,
|
| 12 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 13 |
+
"loss": 1.126,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.14,
|
| 18 |
+
"learning_rate": 4.982332155477032e-05,
|
| 19 |
+
"loss": 0.4641,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.21,
|
| 24 |
+
"learning_rate": 4.946996466431096e-05,
|
| 25 |
+
"loss": 0.1498,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.28,
|
| 30 |
+
"learning_rate": 4.9116607773851593e-05,
|
| 31 |
+
"loss": 0.0822,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.35,
|
| 36 |
+
"learning_rate": 4.8763250883392234e-05,
|
| 37 |
+
"loss": 0.0971,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.42,
|
| 42 |
+
"learning_rate": 4.840989399293286e-05,
|
| 43 |
+
"loss": 0.1169,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.49,
|
| 48 |
+
"learning_rate": 4.8056537102473495e-05,
|
| 49 |
+
"loss": 0.0601,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.56,
|
| 54 |
+
"learning_rate": 4.7703180212014135e-05,
|
| 55 |
+
"loss": 0.0768,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.63,
|
| 60 |
+
"learning_rate": 4.734982332155477e-05,
|
| 61 |
+
"loss": 0.0824,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.7,
|
| 66 |
+
"learning_rate": 4.699646643109541e-05,
|
| 67 |
+
"loss": 0.1023,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.76,
|
| 72 |
+
"learning_rate": 4.664310954063604e-05,
|
| 73 |
+
"loss": 0.0724,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.83,
|
| 78 |
+
"learning_rate": 4.6289752650176684e-05,
|
| 79 |
+
"loss": 0.0838,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.9,
|
| 84 |
+
"learning_rate": 4.593639575971732e-05,
|
| 85 |
+
"loss": 0.0706,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.97,
|
| 90 |
+
"learning_rate": 4.558303886925796e-05,
|
| 91 |
+
"loss": 0.0805,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.99,
|
| 96 |
+
"eval_accuracy": 0.9977653631284916,
|
| 97 |
+
"eval_loss": 0.009151672944426537,
|
| 98 |
+
"eval_runtime": 385.1025,
|
| 99 |
+
"eval_samples_per_second": 2.324,
|
| 100 |
+
"eval_steps_per_second": 0.291,
|
| 101 |
+
"step": 143
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.04,
|
| 105 |
+
"learning_rate": 4.5229681978798585e-05,
|
| 106 |
+
"loss": 0.082,
|
| 107 |
+
"step": 150
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.11,
|
| 111 |
+
"learning_rate": 4.4876325088339225e-05,
|
| 112 |
+
"loss": 0.0348,
|
| 113 |
+
"step": 160
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 1.18,
|
| 117 |
+
"learning_rate": 4.452296819787986e-05,
|
| 118 |
+
"loss": 0.1542,
|
| 119 |
+
"step": 170
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.25,
|
| 123 |
+
"learning_rate": 4.416961130742049e-05,
|
| 124 |
+
"loss": 0.0675,
|
| 125 |
+
"step": 180
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 1.32,
|
| 129 |
+
"learning_rate": 4.381625441696113e-05,
|
| 130 |
+
"loss": 0.096,
|
| 131 |
+
"step": 190
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 1.39,
|
| 135 |
+
"learning_rate": 4.346289752650177e-05,
|
| 136 |
+
"loss": 0.0616,
|
| 137 |
+
"step": 200
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.46,
|
| 141 |
+
"learning_rate": 4.310954063604241e-05,
|
| 142 |
+
"loss": 0.0636,
|
| 143 |
+
"step": 210
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.53,
|
| 147 |
+
"learning_rate": 4.275618374558304e-05,
|
| 148 |
+
"loss": 0.0508,
|
| 149 |
+
"step": 220
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.6,
|
| 153 |
+
"learning_rate": 4.240282685512368e-05,
|
| 154 |
+
"loss": 0.0825,
|
| 155 |
+
"step": 230
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 1.67,
|
| 159 |
+
"learning_rate": 4.204946996466431e-05,
|
| 160 |
+
"loss": 0.0513,
|
| 161 |
+
"step": 240
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.74,
|
| 165 |
+
"learning_rate": 4.169611307420495e-05,
|
| 166 |
+
"loss": 0.0455,
|
| 167 |
+
"step": 250
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 1.81,
|
| 171 |
+
"learning_rate": 4.134275618374558e-05,
|
| 172 |
+
"loss": 0.105,
|
| 173 |
+
"step": 260
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.88,
|
| 177 |
+
"learning_rate": 4.0989399293286223e-05,
|
| 178 |
+
"loss": 0.0822,
|
| 179 |
+
"step": 270
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.95,
|
| 183 |
+
"learning_rate": 4.063604240282686e-05,
|
| 184 |
+
"loss": 0.0396,
|
| 185 |
+
"step": 280
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.0,
|
| 189 |
+
"eval_accuracy": 0.9988826815642458,
|
| 190 |
+
"eval_loss": 0.005963917821645737,
|
| 191 |
+
"eval_runtime": 41.4755,
|
| 192 |
+
"eval_samples_per_second": 21.579,
|
| 193 |
+
"eval_steps_per_second": 2.7,
|
| 194 |
+
"step": 287
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 2.02,
|
| 198 |
+
"learning_rate": 4.028268551236749e-05,
|
| 199 |
+
"loss": 0.075,
|
| 200 |
+
"step": 290
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 2.09,
|
| 204 |
+
"learning_rate": 3.992932862190813e-05,
|
| 205 |
+
"loss": 0.0661,
|
| 206 |
+
"step": 300
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.15,
|
| 210 |
+
"learning_rate": 3.9575971731448765e-05,
|
| 211 |
+
"loss": 0.0343,
|
| 212 |
+
"step": 310
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 2.22,
|
| 216 |
+
"learning_rate": 3.9222614840989406e-05,
|
| 217 |
+
"loss": 0.0574,
|
| 218 |
+
"step": 320
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 2.29,
|
| 222 |
+
"learning_rate": 3.886925795053003e-05,
|
| 223 |
+
"loss": 0.1035,
|
| 224 |
+
"step": 330
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.36,
|
| 228 |
+
"learning_rate": 3.851590106007067e-05,
|
| 229 |
+
"loss": 0.0878,
|
| 230 |
+
"step": 340
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.43,
|
| 234 |
+
"learning_rate": 3.816254416961131e-05,
|
| 235 |
+
"loss": 0.0474,
|
| 236 |
+
"step": 350
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 2.5,
|
| 240 |
+
"learning_rate": 3.780918727915195e-05,
|
| 241 |
+
"loss": 0.0357,
|
| 242 |
+
"step": 360
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 2.57,
|
| 246 |
+
"learning_rate": 3.745583038869258e-05,
|
| 247 |
+
"loss": 0.0574,
|
| 248 |
+
"step": 370
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 2.64,
|
| 252 |
+
"learning_rate": 3.710247349823322e-05,
|
| 253 |
+
"loss": 0.0671,
|
| 254 |
+
"step": 380
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 2.71,
|
| 258 |
+
"learning_rate": 3.6749116607773855e-05,
|
| 259 |
+
"loss": 0.0608,
|
| 260 |
+
"step": 390
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 2.78,
|
| 264 |
+
"learning_rate": 3.639575971731449e-05,
|
| 265 |
+
"loss": 0.0706,
|
| 266 |
+
"step": 400
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 2.85,
|
| 270 |
+
"learning_rate": 3.604240282685513e-05,
|
| 271 |
+
"loss": 0.0404,
|
| 272 |
+
"step": 410
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.92,
|
| 276 |
+
"learning_rate": 3.5689045936395756e-05,
|
| 277 |
+
"loss": 0.0663,
|
| 278 |
+
"step": 420
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 2.99,
|
| 282 |
+
"learning_rate": 3.53356890459364e-05,
|
| 283 |
+
"loss": 0.0382,
|
| 284 |
+
"step": 430
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 3.0,
|
| 288 |
+
"eval_accuracy": 0.9988826815642458,
|
| 289 |
+
"eval_loss": 0.0049131326377391815,
|
| 290 |
+
"eval_runtime": 40.5246,
|
| 291 |
+
"eval_samples_per_second": 22.085,
|
| 292 |
+
"eval_steps_per_second": 2.764,
|
| 293 |
+
"step": 431
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 3.06,
|
| 297 |
+
"learning_rate": 3.498233215547703e-05,
|
| 298 |
+
"loss": 0.0474,
|
| 299 |
+
"step": 440
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 3.13,
|
| 303 |
+
"learning_rate": 3.462897526501767e-05,
|
| 304 |
+
"loss": 0.0526,
|
| 305 |
+
"step": 450
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.2,
|
| 309 |
+
"learning_rate": 3.4275618374558305e-05,
|
| 310 |
+
"loss": 0.053,
|
| 311 |
+
"step": 460
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 3.27,
|
| 315 |
+
"learning_rate": 3.3922261484098945e-05,
|
| 316 |
+
"loss": 0.064,
|
| 317 |
+
"step": 470
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 3.34,
|
| 321 |
+
"learning_rate": 3.356890459363958e-05,
|
| 322 |
+
"loss": 0.0414,
|
| 323 |
+
"step": 480
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 3.41,
|
| 327 |
+
"learning_rate": 3.321554770318021e-05,
|
| 328 |
+
"loss": 0.0448,
|
| 329 |
+
"step": 490
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 3.48,
|
| 333 |
+
"learning_rate": 3.286219081272085e-05,
|
| 334 |
+
"loss": 0.0379,
|
| 335 |
+
"step": 500
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 3.55,
|
| 339 |
+
"learning_rate": 3.250883392226148e-05,
|
| 340 |
+
"loss": 0.0506,
|
| 341 |
+
"step": 510
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 3.61,
|
| 345 |
+
"learning_rate": 3.215547703180212e-05,
|
| 346 |
+
"loss": 0.0412,
|
| 347 |
+
"step": 520
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 3.68,
|
| 351 |
+
"learning_rate": 3.1802120141342755e-05,
|
| 352 |
+
"loss": 0.0684,
|
| 353 |
+
"step": 530
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 3.75,
|
| 357 |
+
"learning_rate": 3.1448763250883395e-05,
|
| 358 |
+
"loss": 0.0738,
|
| 359 |
+
"step": 540
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 3.82,
|
| 363 |
+
"learning_rate": 3.109540636042403e-05,
|
| 364 |
+
"loss": 0.0408,
|
| 365 |
+
"step": 550
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 3.89,
|
| 369 |
+
"learning_rate": 3.074204946996467e-05,
|
| 370 |
+
"loss": 0.0344,
|
| 371 |
+
"step": 560
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 3.96,
|
| 375 |
+
"learning_rate": 3.03886925795053e-05,
|
| 376 |
+
"loss": 0.0394,
|
| 377 |
+
"step": 570
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 4.0,
|
| 381 |
+
"eval_accuracy": 0.9977653631284916,
|
| 382 |
+
"eval_loss": 0.003506433218717575,
|
| 383 |
+
"eval_runtime": 40.9642,
|
| 384 |
+
"eval_samples_per_second": 21.848,
|
| 385 |
+
"eval_steps_per_second": 2.734,
|
| 386 |
+
"step": 575
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 4.03,
|
| 390 |
+
"learning_rate": 3.003533568904594e-05,
|
| 391 |
+
"loss": 0.0404,
|
| 392 |
+
"step": 580
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 4.1,
|
| 396 |
+
"learning_rate": 2.9681978798586574e-05,
|
| 397 |
+
"loss": 0.0426,
|
| 398 |
+
"step": 590
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 4.17,
|
| 402 |
+
"learning_rate": 2.9328621908127208e-05,
|
| 403 |
+
"loss": 0.0365,
|
| 404 |
+
"step": 600
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 4.24,
|
| 408 |
+
"learning_rate": 2.8975265017667848e-05,
|
| 409 |
+
"loss": 0.0488,
|
| 410 |
+
"step": 610
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 4.31,
|
| 414 |
+
"learning_rate": 2.862190812720848e-05,
|
| 415 |
+
"loss": 0.0381,
|
| 416 |
+
"step": 620
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 4.38,
|
| 420 |
+
"learning_rate": 2.826855123674912e-05,
|
| 421 |
+
"loss": 0.0633,
|
| 422 |
+
"step": 630
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 4.45,
|
| 426 |
+
"learning_rate": 2.7915194346289753e-05,
|
| 427 |
+
"loss": 0.046,
|
| 428 |
+
"step": 640
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 4.52,
|
| 432 |
+
"learning_rate": 2.7561837455830393e-05,
|
| 433 |
+
"loss": 0.0665,
|
| 434 |
+
"step": 650
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 4.59,
|
| 438 |
+
"learning_rate": 2.7208480565371023e-05,
|
| 439 |
+
"loss": 0.0397,
|
| 440 |
+
"step": 660
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 4.66,
|
| 444 |
+
"learning_rate": 2.6855123674911664e-05,
|
| 445 |
+
"loss": 0.0605,
|
| 446 |
+
"step": 670
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 4.73,
|
| 450 |
+
"learning_rate": 2.6501766784452298e-05,
|
| 451 |
+
"loss": 0.019,
|
| 452 |
+
"step": 680
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 4.8,
|
| 456 |
+
"learning_rate": 2.6148409893992938e-05,
|
| 457 |
+
"loss": 0.0394,
|
| 458 |
+
"step": 690
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 4.87,
|
| 462 |
+
"learning_rate": 2.5795053003533572e-05,
|
| 463 |
+
"loss": 0.0488,
|
| 464 |
+
"step": 700
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 4.94,
|
| 468 |
+
"learning_rate": 2.5441696113074202e-05,
|
| 469 |
+
"loss": 0.0269,
|
| 470 |
+
"step": 710
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 5.0,
|
| 474 |
+
"eval_accuracy": 0.9988826815642458,
|
| 475 |
+
"eval_loss": 0.0033426920417696238,
|
| 476 |
+
"eval_runtime": 40.8952,
|
| 477 |
+
"eval_samples_per_second": 21.885,
|
| 478 |
+
"eval_steps_per_second": 2.739,
|
| 479 |
+
"step": 719
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 5.0,
|
| 483 |
+
"learning_rate": 2.5088339222614843e-05,
|
| 484 |
+
"loss": 0.021,
|
| 485 |
+
"step": 720
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 5.07,
|
| 489 |
+
"learning_rate": 2.473498233215548e-05,
|
| 490 |
+
"loss": 0.0635,
|
| 491 |
+
"step": 730
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 5.14,
|
| 495 |
+
"learning_rate": 2.4381625441696117e-05,
|
| 496 |
+
"loss": 0.0391,
|
| 497 |
+
"step": 740
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 5.21,
|
| 501 |
+
"learning_rate": 2.4028268551236747e-05,
|
| 502 |
+
"loss": 0.0583,
|
| 503 |
+
"step": 750
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 5.28,
|
| 507 |
+
"learning_rate": 2.3674911660777384e-05,
|
| 508 |
+
"loss": 0.0248,
|
| 509 |
+
"step": 760
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 5.35,
|
| 513 |
+
"learning_rate": 2.332155477031802e-05,
|
| 514 |
+
"loss": 0.0309,
|
| 515 |
+
"step": 770
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 5.42,
|
| 519 |
+
"learning_rate": 2.296819787985866e-05,
|
| 520 |
+
"loss": 0.0403,
|
| 521 |
+
"step": 780
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 5.49,
|
| 525 |
+
"learning_rate": 2.2614840989399292e-05,
|
| 526 |
+
"loss": 0.0602,
|
| 527 |
+
"step": 790
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 5.56,
|
| 531 |
+
"learning_rate": 2.226148409893993e-05,
|
| 532 |
+
"loss": 0.0152,
|
| 533 |
+
"step": 800
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 5.63,
|
| 537 |
+
"learning_rate": 2.1908127208480567e-05,
|
| 538 |
+
"loss": 0.0581,
|
| 539 |
+
"step": 810
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 5.7,
|
| 543 |
+
"learning_rate": 2.1554770318021204e-05,
|
| 544 |
+
"loss": 0.0292,
|
| 545 |
+
"step": 820
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 5.77,
|
| 549 |
+
"learning_rate": 2.120141342756184e-05,
|
| 550 |
+
"loss": 0.045,
|
| 551 |
+
"step": 830
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 5.84,
|
| 555 |
+
"learning_rate": 2.0848056537102475e-05,
|
| 556 |
+
"loss": 0.0278,
|
| 557 |
+
"step": 840
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 5.91,
|
| 561 |
+
"learning_rate": 2.0494699646643112e-05,
|
| 562 |
+
"loss": 0.0743,
|
| 563 |
+
"step": 850
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 5.98,
|
| 567 |
+
"learning_rate": 2.0141342756183745e-05,
|
| 568 |
+
"loss": 0.0249,
|
| 569 |
+
"step": 860
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 6.0,
|
| 573 |
+
"eval_accuracy": 0.9988826815642458,
|
| 574 |
+
"eval_loss": 0.0035183620639145374,
|
| 575 |
+
"eval_runtime": 39.497,
|
| 576 |
+
"eval_samples_per_second": 22.66,
|
| 577 |
+
"eval_steps_per_second": 2.836,
|
| 578 |
+
"step": 863
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 6.05,
|
| 582 |
+
"learning_rate": 1.9787985865724383e-05,
|
| 583 |
+
"loss": 0.0336,
|
| 584 |
+
"step": 870
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 6.12,
|
| 588 |
+
"learning_rate": 1.9434628975265016e-05,
|
| 589 |
+
"loss": 0.0408,
|
| 590 |
+
"step": 880
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 6.19,
|
| 594 |
+
"learning_rate": 1.9081272084805653e-05,
|
| 595 |
+
"loss": 0.0255,
|
| 596 |
+
"step": 890
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 6.26,
|
| 600 |
+
"learning_rate": 1.872791519434629e-05,
|
| 601 |
+
"loss": 0.0309,
|
| 602 |
+
"step": 900
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 6.33,
|
| 606 |
+
"learning_rate": 1.8374558303886928e-05,
|
| 607 |
+
"loss": 0.0479,
|
| 608 |
+
"step": 910
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 6.4,
|
| 612 |
+
"learning_rate": 1.8021201413427565e-05,
|
| 613 |
+
"loss": 0.0268,
|
| 614 |
+
"step": 920
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 6.46,
|
| 618 |
+
"learning_rate": 1.76678445229682e-05,
|
| 619 |
+
"loss": 0.0553,
|
| 620 |
+
"step": 930
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 6.53,
|
| 624 |
+
"learning_rate": 1.7314487632508836e-05,
|
| 625 |
+
"loss": 0.0169,
|
| 626 |
+
"step": 940
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 6.6,
|
| 630 |
+
"learning_rate": 1.6961130742049473e-05,
|
| 631 |
+
"loss": 0.0245,
|
| 632 |
+
"step": 950
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 6.67,
|
| 636 |
+
"learning_rate": 1.6607773851590106e-05,
|
| 637 |
+
"loss": 0.0242,
|
| 638 |
+
"step": 960
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 6.74,
|
| 642 |
+
"learning_rate": 1.625441696113074e-05,
|
| 643 |
+
"loss": 0.0303,
|
| 644 |
+
"step": 970
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 6.81,
|
| 648 |
+
"learning_rate": 1.5901060070671377e-05,
|
| 649 |
+
"loss": 0.0527,
|
| 650 |
+
"step": 980
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 6.88,
|
| 654 |
+
"learning_rate": 1.5547703180212014e-05,
|
| 655 |
+
"loss": 0.0296,
|
| 656 |
+
"step": 990
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 6.95,
|
| 660 |
+
"learning_rate": 1.519434628975265e-05,
|
| 661 |
+
"loss": 0.0388,
|
| 662 |
+
"step": 1000
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 7.0,
|
| 666 |
+
"eval_accuracy": 0.9988826815642458,
|
| 667 |
+
"eval_loss": 0.005780559964478016,
|
| 668 |
+
"eval_runtime": 41.8587,
|
| 669 |
+
"eval_samples_per_second": 21.381,
|
| 670 |
+
"eval_steps_per_second": 2.676,
|
| 671 |
+
"step": 1007
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 7.02,
|
| 675 |
+
"learning_rate": 1.4840989399293287e-05,
|
| 676 |
+
"loss": 0.0124,
|
| 677 |
+
"step": 1010
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 7.09,
|
| 681 |
+
"learning_rate": 1.4487632508833924e-05,
|
| 682 |
+
"loss": 0.0487,
|
| 683 |
+
"step": 1020
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 7.16,
|
| 687 |
+
"learning_rate": 1.413427561837456e-05,
|
| 688 |
+
"loss": 0.0464,
|
| 689 |
+
"step": 1030
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 7.23,
|
| 693 |
+
"learning_rate": 1.3780918727915197e-05,
|
| 694 |
+
"loss": 0.013,
|
| 695 |
+
"step": 1040
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 7.3,
|
| 699 |
+
"learning_rate": 1.3427561837455832e-05,
|
| 700 |
+
"loss": 0.0411,
|
| 701 |
+
"step": 1050
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 7.37,
|
| 705 |
+
"learning_rate": 1.3074204946996469e-05,
|
| 706 |
+
"loss": 0.0318,
|
| 707 |
+
"step": 1060
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 7.44,
|
| 711 |
+
"learning_rate": 1.2720848056537101e-05,
|
| 712 |
+
"loss": 0.0382,
|
| 713 |
+
"step": 1070
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 7.51,
|
| 717 |
+
"learning_rate": 1.236749116607774e-05,
|
| 718 |
+
"loss": 0.0275,
|
| 719 |
+
"step": 1080
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 7.58,
|
| 723 |
+
"learning_rate": 1.2014134275618374e-05,
|
| 724 |
+
"loss": 0.0168,
|
| 725 |
+
"step": 1090
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 7.65,
|
| 729 |
+
"learning_rate": 1.166077738515901e-05,
|
| 730 |
+
"loss": 0.0452,
|
| 731 |
+
"step": 1100
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 7.72,
|
| 735 |
+
"learning_rate": 1.1307420494699646e-05,
|
| 736 |
+
"loss": 0.0307,
|
| 737 |
+
"step": 1110
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 7.79,
|
| 741 |
+
"learning_rate": 1.0954063604240283e-05,
|
| 742 |
+
"loss": 0.0359,
|
| 743 |
+
"step": 1120
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 7.86,
|
| 747 |
+
"learning_rate": 1.060070671378092e-05,
|
| 748 |
+
"loss": 0.0341,
|
| 749 |
+
"step": 1130
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 7.92,
|
| 753 |
+
"learning_rate": 1.0247349823321556e-05,
|
| 754 |
+
"loss": 0.0587,
|
| 755 |
+
"step": 1140
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 7.99,
|
| 759 |
+
"learning_rate": 9.893992932862191e-06,
|
| 760 |
+
"loss": 0.0175,
|
| 761 |
+
"step": 1150
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 7.99,
|
| 765 |
+
"eval_accuracy": 0.9988826815642458,
|
| 766 |
+
"eval_loss": 0.003908189013600349,
|
| 767 |
+
"eval_runtime": 41.7986,
|
| 768 |
+
"eval_samples_per_second": 21.412,
|
| 769 |
+
"eval_steps_per_second": 2.68,
|
| 770 |
+
"step": 1150
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 8.06,
|
| 774 |
+
"learning_rate": 9.540636042402827e-06,
|
| 775 |
+
"loss": 0.0397,
|
| 776 |
+
"step": 1160
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 8.13,
|
| 780 |
+
"learning_rate": 9.187279151943464e-06,
|
| 781 |
+
"loss": 0.0422,
|
| 782 |
+
"step": 1170
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 8.2,
|
| 786 |
+
"learning_rate": 8.8339222614841e-06,
|
| 787 |
+
"loss": 0.0324,
|
| 788 |
+
"step": 1180
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 8.27,
|
| 792 |
+
"learning_rate": 8.480565371024736e-06,
|
| 793 |
+
"loss": 0.0206,
|
| 794 |
+
"step": 1190
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 8.34,
|
| 798 |
+
"learning_rate": 8.12720848056537e-06,
|
| 799 |
+
"loss": 0.0306,
|
| 800 |
+
"step": 1200
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 8.41,
|
| 804 |
+
"learning_rate": 7.773851590106007e-06,
|
| 805 |
+
"loss": 0.0436,
|
| 806 |
+
"step": 1210
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 8.48,
|
| 810 |
+
"learning_rate": 7.4204946996466435e-06,
|
| 811 |
+
"loss": 0.0125,
|
| 812 |
+
"step": 1220
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 8.55,
|
| 816 |
+
"learning_rate": 7.06713780918728e-06,
|
| 817 |
+
"loss": 0.0265,
|
| 818 |
+
"step": 1230
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 8.62,
|
| 822 |
+
"learning_rate": 6.713780918727916e-06,
|
| 823 |
+
"loss": 0.0232,
|
| 824 |
+
"step": 1240
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 8.69,
|
| 828 |
+
"learning_rate": 6.3604240282685506e-06,
|
| 829 |
+
"loss": 0.039,
|
| 830 |
+
"step": 1250
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 8.76,
|
| 834 |
+
"learning_rate": 6.007067137809187e-06,
|
| 835 |
+
"loss": 0.0245,
|
| 836 |
+
"step": 1260
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 8.83,
|
| 840 |
+
"learning_rate": 5.653710247349823e-06,
|
| 841 |
+
"loss": 0.0447,
|
| 842 |
+
"step": 1270
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 8.9,
|
| 846 |
+
"learning_rate": 5.30035335689046e-06,
|
| 847 |
+
"loss": 0.0179,
|
| 848 |
+
"step": 1280
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 8.97,
|
| 852 |
+
"learning_rate": 4.946996466431096e-06,
|
| 853 |
+
"loss": 0.041,
|
| 854 |
+
"step": 1290
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 9.0,
|
| 858 |
+
"eval_accuracy": 0.9988826815642458,
|
| 859 |
+
"eval_loss": 0.0022225547581911087,
|
| 860 |
+
"eval_runtime": 41.6857,
|
| 861 |
+
"eval_samples_per_second": 21.47,
|
| 862 |
+
"eval_steps_per_second": 2.687,
|
| 863 |
+
"step": 1294
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 9.04,
|
| 867 |
+
"learning_rate": 4.593639575971732e-06,
|
| 868 |
+
"loss": 0.0294,
|
| 869 |
+
"step": 1300
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 9.11,
|
| 873 |
+
"learning_rate": 4.240282685512368e-06,
|
| 874 |
+
"loss": 0.0258,
|
| 875 |
+
"step": 1310
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 9.18,
|
| 879 |
+
"learning_rate": 3.886925795053004e-06,
|
| 880 |
+
"loss": 0.0357,
|
| 881 |
+
"step": 1320
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 9.25,
|
| 885 |
+
"learning_rate": 3.53356890459364e-06,
|
| 886 |
+
"loss": 0.0065,
|
| 887 |
+
"step": 1330
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 9.31,
|
| 891 |
+
"learning_rate": 3.1802120141342753e-06,
|
| 892 |
+
"loss": 0.0199,
|
| 893 |
+
"step": 1340
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 9.38,
|
| 897 |
+
"learning_rate": 2.8268551236749116e-06,
|
| 898 |
+
"loss": 0.0175,
|
| 899 |
+
"step": 1350
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 9.45,
|
| 903 |
+
"learning_rate": 2.473498233215548e-06,
|
| 904 |
+
"loss": 0.023,
|
| 905 |
+
"step": 1360
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 9.52,
|
| 909 |
+
"learning_rate": 2.120141342756184e-06,
|
| 910 |
+
"loss": 0.026,
|
| 911 |
+
"step": 1370
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 9.59,
|
| 915 |
+
"learning_rate": 1.76678445229682e-06,
|
| 916 |
+
"loss": 0.0327,
|
| 917 |
+
"step": 1380
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 9.66,
|
| 921 |
+
"learning_rate": 1.4134275618374558e-06,
|
| 922 |
+
"loss": 0.0374,
|
| 923 |
+
"step": 1390
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 9.73,
|
| 927 |
+
"learning_rate": 1.060070671378092e-06,
|
| 928 |
+
"loss": 0.0272,
|
| 929 |
+
"step": 1400
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 9.8,
|
| 933 |
+
"learning_rate": 7.067137809187279e-07,
|
| 934 |
+
"loss": 0.0303,
|
| 935 |
+
"step": 1410
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 9.87,
|
| 939 |
+
"learning_rate": 3.5335689045936394e-07,
|
| 940 |
+
"loss": 0.0343,
|
| 941 |
+
"step": 1420
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 9.94,
|
| 945 |
+
"learning_rate": 0.0,
|
| 946 |
+
"loss": 0.0432,
|
| 947 |
+
"step": 1430
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 9.94,
|
| 951 |
+
"eval_accuracy": 0.9988826815642458,
|
| 952 |
+
"eval_loss": 0.0024573865812271833,
|
| 953 |
+
"eval_runtime": 40.191,
|
| 954 |
+
"eval_samples_per_second": 22.269,
|
| 955 |
+
"eval_steps_per_second": 2.787,
|
| 956 |
+
"step": 1430
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 9.94,
|
| 960 |
+
"step": 1430,
|
| 961 |
+
"total_flos": 1.4186448226450612e+19,
|
| 962 |
+
"train_loss": 0.0581204283070731,
|
| 963 |
+
"train_runtime": 12121.0022,
|
| 964 |
+
"train_samples_per_second": 6.643,
|
| 965 |
+
"train_steps_per_second": 0.118
|
| 966 |
+
}
|
| 967 |
+
],
|
| 968 |
+
"max_steps": 1430,
|
| 969 |
+
"num_train_epochs": 10,
|
| 970 |
+
"total_flos": 1.4186448226450612e+19,
|
| 971 |
+
"trial_name": null,
|
| 972 |
+
"trial_params": null
|
| 973 |
+
}
|