Training in progress, epoch 0
Browse files- all_results.json +8 -0
- runs/Jul09_17-24-18_IISC143/events.out.tfevents.1720517092.IISC143.18316.0 +3 -0
- train_results.json +8 -0
- trained_weights_biases.pkl +3 -0
- trainer_state.json +1014 -0
- training_args.bin +1 -1
- weights_biases.pkl +1 -1
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9996370235934664,
|
| 3 |
+
"total_flos": 2.1905085973455176e+18,
|
| 4 |
+
"train_loss": 0.13800792182885657,
|
| 5 |
+
"train_runtime": 867.8791,
|
| 6 |
+
"train_samples_per_second": 101.567,
|
| 7 |
+
"train_steps_per_second": 1.587
|
| 8 |
+
}
|
runs/Jul09_17-24-18_IISC143/events.out.tfevents.1720517092.IISC143.18316.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fd972ff5ddceeac9e7a0c93c94d48557ac29499e074db83e7ac54f5c95d97061
|
| 3 |
+
size 35077
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9996370235934664,
|
| 3 |
+
"total_flos": 2.1905085973455176e+18,
|
| 4 |
+
"train_loss": 0.13800792182885657,
|
| 5 |
+
"train_runtime": 867.8791,
|
| 6 |
+
"train_samples_per_second": 101.567,
|
| 7 |
+
"train_steps_per_second": 1.587
|
| 8 |
+
}
|
trained_weights_biases.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:47e79bdc04eb3d262b8917c5667bd65831e653c37e8c2455cebb15e4f6c5a6a4
|
| 3 |
+
size 6858
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1014 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9835503925216681,
|
| 3 |
+
"best_model_checkpoint": "batch-size16_FFPP-raw_opencv-1FPS_faces-expand0-aligned_unaugmentation_seeds-42_143_4090\\checkpoint-1377",
|
| 4 |
+
"epoch": 0.9996370235934664,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1377,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.007259528130671506,
|
| 13 |
+
"grad_norm": 2.73757266998291,
|
| 14 |
+
"learning_rate": 3.6231884057971017e-06,
|
| 15 |
+
"loss": 0.5456,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.014519056261343012,
|
| 20 |
+
"grad_norm": 5.650032043457031,
|
| 21 |
+
"learning_rate": 7.246376811594203e-06,
|
| 22 |
+
"loss": 0.5189,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.021778584392014518,
|
| 27 |
+
"grad_norm": 3.772991895675659,
|
| 28 |
+
"learning_rate": 1.0869565217391305e-05,
|
| 29 |
+
"loss": 0.5151,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.029038112522686024,
|
| 34 |
+
"grad_norm": 3.6677093505859375,
|
| 35 |
+
"learning_rate": 1.4492753623188407e-05,
|
| 36 |
+
"loss": 0.4985,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.036297640653357534,
|
| 41 |
+
"grad_norm": 4.538272380828857,
|
| 42 |
+
"learning_rate": 1.8115942028985507e-05,
|
| 43 |
+
"loss": 0.4928,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.043557168784029036,
|
| 48 |
+
"grad_norm": 5.5191192626953125,
|
| 49 |
+
"learning_rate": 2.173913043478261e-05,
|
| 50 |
+
"loss": 0.5259,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.050816696914700546,
|
| 55 |
+
"grad_norm": 4.596761226654053,
|
| 56 |
+
"learning_rate": 2.5362318840579714e-05,
|
| 57 |
+
"loss": 0.4591,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.05807622504537205,
|
| 62 |
+
"grad_norm": 5.684937477111816,
|
| 63 |
+
"learning_rate": 2.8985507246376814e-05,
|
| 64 |
+
"loss": 0.4572,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06533575317604355,
|
| 69 |
+
"grad_norm": 22.83547592163086,
|
| 70 |
+
"learning_rate": 3.260869565217392e-05,
|
| 71 |
+
"loss": 0.3957,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.07259528130671507,
|
| 76 |
+
"grad_norm": 7.260858058929443,
|
| 77 |
+
"learning_rate": 3.6231884057971014e-05,
|
| 78 |
+
"loss": 0.3996,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.07985480943738657,
|
| 83 |
+
"grad_norm": 19.33077621459961,
|
| 84 |
+
"learning_rate": 3.985507246376812e-05,
|
| 85 |
+
"loss": 0.3463,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.08711433756805807,
|
| 90 |
+
"grad_norm": 18.935882568359375,
|
| 91 |
+
"learning_rate": 4.347826086956522e-05,
|
| 92 |
+
"loss": 0.3207,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.09437386569872959,
|
| 97 |
+
"grad_norm": 34.95930099487305,
|
| 98 |
+
"learning_rate": 4.710144927536232e-05,
|
| 99 |
+
"loss": 0.2941,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.10163339382940109,
|
| 104 |
+
"grad_norm": 10.575639724731445,
|
| 105 |
+
"learning_rate": 4.9919289749798225e-05,
|
| 106 |
+
"loss": 0.2765,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.1088929219600726,
|
| 111 |
+
"grad_norm": 18.34967041015625,
|
| 112 |
+
"learning_rate": 4.951573849878935e-05,
|
| 113 |
+
"loss": 0.1929,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.1161524500907441,
|
| 118 |
+
"grad_norm": 12.154097557067871,
|
| 119 |
+
"learning_rate": 4.911218724778047e-05,
|
| 120 |
+
"loss": 0.2253,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.12341197822141561,
|
| 125 |
+
"grad_norm": 11.486285209655762,
|
| 126 |
+
"learning_rate": 4.870863599677159e-05,
|
| 127 |
+
"loss": 0.2611,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1306715063520871,
|
| 132 |
+
"grad_norm": 33.42046356201172,
|
| 133 |
+
"learning_rate": 4.8305084745762714e-05,
|
| 134 |
+
"loss": 0.2789,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.13793103448275862,
|
| 139 |
+
"grad_norm": 6.870066165924072,
|
| 140 |
+
"learning_rate": 4.7901533494753834e-05,
|
| 141 |
+
"loss": 0.2954,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.14519056261343014,
|
| 146 |
+
"grad_norm": 16.494075775146484,
|
| 147 |
+
"learning_rate": 4.749798224374496e-05,
|
| 148 |
+
"loss": 0.1558,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.15245009074410162,
|
| 153 |
+
"grad_norm": 37.45346450805664,
|
| 154 |
+
"learning_rate": 4.7094430992736075e-05,
|
| 155 |
+
"loss": 0.2254,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.15970961887477314,
|
| 160 |
+
"grad_norm": 15.169726371765137,
|
| 161 |
+
"learning_rate": 4.66908797417272e-05,
|
| 162 |
+
"loss": 0.219,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.16696914700544466,
|
| 167 |
+
"grad_norm": 9.790349006652832,
|
| 168 |
+
"learning_rate": 4.628732849071832e-05,
|
| 169 |
+
"loss": 0.1902,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.17422867513611615,
|
| 174 |
+
"grad_norm": 8.544652938842773,
|
| 175 |
+
"learning_rate": 4.588377723970945e-05,
|
| 176 |
+
"loss": 0.1901,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.18148820326678766,
|
| 181 |
+
"grad_norm": 6.982186317443848,
|
| 182 |
+
"learning_rate": 4.548022598870056e-05,
|
| 183 |
+
"loss": 0.1465,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.18874773139745918,
|
| 188 |
+
"grad_norm": 11.616936683654785,
|
| 189 |
+
"learning_rate": 4.507667473769169e-05,
|
| 190 |
+
"loss": 0.1217,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.19600725952813067,
|
| 195 |
+
"grad_norm": 9.371502876281738,
|
| 196 |
+
"learning_rate": 4.467312348668281e-05,
|
| 197 |
+
"loss": 0.148,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.20326678765880218,
|
| 202 |
+
"grad_norm": 12.38410472869873,
|
| 203 |
+
"learning_rate": 4.426957223567393e-05,
|
| 204 |
+
"loss": 0.149,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.21052631578947367,
|
| 209 |
+
"grad_norm": 9.507818222045898,
|
| 210 |
+
"learning_rate": 4.386602098466506e-05,
|
| 211 |
+
"loss": 0.2108,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.2177858439201452,
|
| 216 |
+
"grad_norm": 5.370192527770996,
|
| 217 |
+
"learning_rate": 4.346246973365617e-05,
|
| 218 |
+
"loss": 0.1677,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.2250453720508167,
|
| 223 |
+
"grad_norm": 19.771787643432617,
|
| 224 |
+
"learning_rate": 4.30589184826473e-05,
|
| 225 |
+
"loss": 0.1273,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.2323049001814882,
|
| 230 |
+
"grad_norm": 25.864179611206055,
|
| 231 |
+
"learning_rate": 4.265536723163842e-05,
|
| 232 |
+
"loss": 0.1916,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.2395644283121597,
|
| 237 |
+
"grad_norm": 22.332120895385742,
|
| 238 |
+
"learning_rate": 4.225181598062955e-05,
|
| 239 |
+
"loss": 0.1454,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.24682395644283123,
|
| 244 |
+
"grad_norm": 9.30060863494873,
|
| 245 |
+
"learning_rate": 4.184826472962066e-05,
|
| 246 |
+
"loss": 0.1321,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.2540834845735027,
|
| 251 |
+
"grad_norm": 5.451467037200928,
|
| 252 |
+
"learning_rate": 4.144471347861179e-05,
|
| 253 |
+
"loss": 0.1558,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.2613430127041742,
|
| 258 |
+
"grad_norm": 11.918793678283691,
|
| 259 |
+
"learning_rate": 4.104116222760291e-05,
|
| 260 |
+
"loss": 0.1416,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.26860254083484575,
|
| 265 |
+
"grad_norm": 6.430271148681641,
|
| 266 |
+
"learning_rate": 4.063761097659403e-05,
|
| 267 |
+
"loss": 0.1414,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.27586206896551724,
|
| 272 |
+
"grad_norm": 7.710476875305176,
|
| 273 |
+
"learning_rate": 4.023405972558515e-05,
|
| 274 |
+
"loss": 0.1411,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.2831215970961887,
|
| 279 |
+
"grad_norm": 12.249679565429688,
|
| 280 |
+
"learning_rate": 3.983050847457627e-05,
|
| 281 |
+
"loss": 0.1272,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.29038112522686027,
|
| 286 |
+
"grad_norm": 4.084755897521973,
|
| 287 |
+
"learning_rate": 3.94269572235674e-05,
|
| 288 |
+
"loss": 0.1411,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.29764065335753176,
|
| 293 |
+
"grad_norm": 12.142387390136719,
|
| 294 |
+
"learning_rate": 3.902340597255852e-05,
|
| 295 |
+
"loss": 0.1302,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.30490018148820325,
|
| 300 |
+
"grad_norm": 14.800415992736816,
|
| 301 |
+
"learning_rate": 3.861985472154964e-05,
|
| 302 |
+
"loss": 0.1327,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.3121597096188748,
|
| 307 |
+
"grad_norm": 15.094321250915527,
|
| 308 |
+
"learning_rate": 3.821630347054076e-05,
|
| 309 |
+
"loss": 0.1303,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.3194192377495463,
|
| 314 |
+
"grad_norm": 15.302112579345703,
|
| 315 |
+
"learning_rate": 3.7812752219531885e-05,
|
| 316 |
+
"loss": 0.1354,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.32667876588021777,
|
| 321 |
+
"grad_norm": 6.560603141784668,
|
| 322 |
+
"learning_rate": 3.7409200968523006e-05,
|
| 323 |
+
"loss": 0.1156,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.3339382940108893,
|
| 328 |
+
"grad_norm": 18.94932746887207,
|
| 329 |
+
"learning_rate": 3.7005649717514126e-05,
|
| 330 |
+
"loss": 0.1532,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.3411978221415608,
|
| 335 |
+
"grad_norm": 11.589672088623047,
|
| 336 |
+
"learning_rate": 3.6602098466505247e-05,
|
| 337 |
+
"loss": 0.14,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3484573502722323,
|
| 342 |
+
"grad_norm": 13.704521179199219,
|
| 343 |
+
"learning_rate": 3.619854721549637e-05,
|
| 344 |
+
"loss": 0.115,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.35571687840290384,
|
| 349 |
+
"grad_norm": 7.8261332511901855,
|
| 350 |
+
"learning_rate": 3.5794995964487494e-05,
|
| 351 |
+
"loss": 0.1447,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.3629764065335753,
|
| 356 |
+
"grad_norm": 6.8338704109191895,
|
| 357 |
+
"learning_rate": 3.539144471347861e-05,
|
| 358 |
+
"loss": 0.1367,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.3702359346642468,
|
| 363 |
+
"grad_norm": 9.027732849121094,
|
| 364 |
+
"learning_rate": 3.4987893462469735e-05,
|
| 365 |
+
"loss": 0.0941,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.37749546279491836,
|
| 370 |
+
"grad_norm": 18.083940505981445,
|
| 371 |
+
"learning_rate": 3.4584342211460856e-05,
|
| 372 |
+
"loss": 0.1338,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.38475499092558985,
|
| 377 |
+
"grad_norm": 10.027618408203125,
|
| 378 |
+
"learning_rate": 3.418079096045198e-05,
|
| 379 |
+
"loss": 0.1337,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.39201451905626133,
|
| 384 |
+
"grad_norm": 8.791827201843262,
|
| 385 |
+
"learning_rate": 3.3777239709443096e-05,
|
| 386 |
+
"loss": 0.1287,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3992740471869328,
|
| 391 |
+
"grad_norm": 10.15536880493164,
|
| 392 |
+
"learning_rate": 3.3373688458434224e-05,
|
| 393 |
+
"loss": 0.1349,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.40653357531760437,
|
| 398 |
+
"grad_norm": 14.157705307006836,
|
| 399 |
+
"learning_rate": 3.2970137207425344e-05,
|
| 400 |
+
"loss": 0.1044,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.41379310344827586,
|
| 405 |
+
"grad_norm": 1.7369716167449951,
|
| 406 |
+
"learning_rate": 3.2566585956416464e-05,
|
| 407 |
+
"loss": 0.1169,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.42105263157894735,
|
| 412 |
+
"grad_norm": 9.967548370361328,
|
| 413 |
+
"learning_rate": 3.216303470540759e-05,
|
| 414 |
+
"loss": 0.08,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.4283121597096189,
|
| 419 |
+
"grad_norm": 10.823721885681152,
|
| 420 |
+
"learning_rate": 3.1759483454398705e-05,
|
| 421 |
+
"loss": 0.1172,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.4355716878402904,
|
| 426 |
+
"grad_norm": 16.309240341186523,
|
| 427 |
+
"learning_rate": 3.135593220338983e-05,
|
| 428 |
+
"loss": 0.116,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.44283121597096187,
|
| 433 |
+
"grad_norm": 5.755407810211182,
|
| 434 |
+
"learning_rate": 3.095238095238095e-05,
|
| 435 |
+
"loss": 0.1316,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.4500907441016334,
|
| 440 |
+
"grad_norm": 3.8949596881866455,
|
| 441 |
+
"learning_rate": 3.054882970137208e-05,
|
| 442 |
+
"loss": 0.0987,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.4573502722323049,
|
| 447 |
+
"grad_norm": 1.9218209981918335,
|
| 448 |
+
"learning_rate": 3.0145278450363197e-05,
|
| 449 |
+
"loss": 0.0787,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.4646098003629764,
|
| 454 |
+
"grad_norm": 14.866432189941406,
|
| 455 |
+
"learning_rate": 2.9741727199354318e-05,
|
| 456 |
+
"loss": 0.0952,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.47186932849364793,
|
| 461 |
+
"grad_norm": 14.051959037780762,
|
| 462 |
+
"learning_rate": 2.933817594834544e-05,
|
| 463 |
+
"loss": 0.1073,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.4791288566243194,
|
| 468 |
+
"grad_norm": 7.117464542388916,
|
| 469 |
+
"learning_rate": 2.8934624697336565e-05,
|
| 470 |
+
"loss": 0.0984,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.4863883847549909,
|
| 475 |
+
"grad_norm": 6.493090629577637,
|
| 476 |
+
"learning_rate": 2.8531073446327682e-05,
|
| 477 |
+
"loss": 0.0849,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.49364791288566245,
|
| 482 |
+
"grad_norm": 6.476171970367432,
|
| 483 |
+
"learning_rate": 2.8127522195318806e-05,
|
| 484 |
+
"loss": 0.0896,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.5009074410163339,
|
| 489 |
+
"grad_norm": 1.109189748764038,
|
| 490 |
+
"learning_rate": 2.772397094430993e-05,
|
| 491 |
+
"loss": 0.0899,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.5081669691470054,
|
| 496 |
+
"grad_norm": 6.053189277648926,
|
| 497 |
+
"learning_rate": 2.7320419693301054e-05,
|
| 498 |
+
"loss": 0.1058,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.515426497277677,
|
| 503 |
+
"grad_norm": 2.7531440258026123,
|
| 504 |
+
"learning_rate": 2.691686844229217e-05,
|
| 505 |
+
"loss": 0.0835,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.5226860254083484,
|
| 510 |
+
"grad_norm": 1.9153120517730713,
|
| 511 |
+
"learning_rate": 2.6513317191283295e-05,
|
| 512 |
+
"loss": 0.0911,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.52994555353902,
|
| 517 |
+
"grad_norm": 18.87368392944336,
|
| 518 |
+
"learning_rate": 2.6109765940274415e-05,
|
| 519 |
+
"loss": 0.0866,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.5372050816696915,
|
| 524 |
+
"grad_norm": 6.364779949188232,
|
| 525 |
+
"learning_rate": 2.570621468926554e-05,
|
| 526 |
+
"loss": 0.1047,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.5444646098003629,
|
| 531 |
+
"grad_norm": 8.270808219909668,
|
| 532 |
+
"learning_rate": 2.5302663438256656e-05,
|
| 533 |
+
"loss": 0.1033,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.5517241379310345,
|
| 538 |
+
"grad_norm": 8.82036304473877,
|
| 539 |
+
"learning_rate": 2.489911218724778e-05,
|
| 540 |
+
"loss": 0.0766,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.558983666061706,
|
| 545 |
+
"grad_norm": 12.227109909057617,
|
| 546 |
+
"learning_rate": 2.4495560936238903e-05,
|
| 547 |
+
"loss": 0.0992,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.5662431941923775,
|
| 552 |
+
"grad_norm": 16.895111083984375,
|
| 553 |
+
"learning_rate": 2.4092009685230024e-05,
|
| 554 |
+
"loss": 0.0815,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.573502722323049,
|
| 559 |
+
"grad_norm": 8.905038833618164,
|
| 560 |
+
"learning_rate": 2.3688458434221148e-05,
|
| 561 |
+
"loss": 0.0939,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.5807622504537205,
|
| 566 |
+
"grad_norm": 32.21564483642578,
|
| 567 |
+
"learning_rate": 2.3284907183212268e-05,
|
| 568 |
+
"loss": 0.115,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.588021778584392,
|
| 573 |
+
"grad_norm": 20.237606048583984,
|
| 574 |
+
"learning_rate": 2.2881355932203392e-05,
|
| 575 |
+
"loss": 0.1157,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.5952813067150635,
|
| 580 |
+
"grad_norm": 10.920185089111328,
|
| 581 |
+
"learning_rate": 2.2477804681194512e-05,
|
| 582 |
+
"loss": 0.1166,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.6025408348457351,
|
| 587 |
+
"grad_norm": 16.288949966430664,
|
| 588 |
+
"learning_rate": 2.2074253430185636e-05,
|
| 589 |
+
"loss": 0.072,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.6098003629764065,
|
| 594 |
+
"grad_norm": 9.229625701904297,
|
| 595 |
+
"learning_rate": 2.1670702179176757e-05,
|
| 596 |
+
"loss": 0.0989,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.617059891107078,
|
| 601 |
+
"grad_norm": 4.963629722595215,
|
| 602 |
+
"learning_rate": 2.1267150928167877e-05,
|
| 603 |
+
"loss": 0.0715,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.6243194192377496,
|
| 608 |
+
"grad_norm": 1.9014785289764404,
|
| 609 |
+
"learning_rate": 2.0863599677159e-05,
|
| 610 |
+
"loss": 0.0959,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.631578947368421,
|
| 615 |
+
"grad_norm": 12.746657371520996,
|
| 616 |
+
"learning_rate": 2.046004842615012e-05,
|
| 617 |
+
"loss": 0.0962,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.6388384754990926,
|
| 622 |
+
"grad_norm": 3.6424975395202637,
|
| 623 |
+
"learning_rate": 2.0056497175141245e-05,
|
| 624 |
+
"loss": 0.0644,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.6460980036297641,
|
| 629 |
+
"grad_norm": 15.015966415405273,
|
| 630 |
+
"learning_rate": 1.9652945924132365e-05,
|
| 631 |
+
"loss": 0.0942,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.6533575317604355,
|
| 636 |
+
"grad_norm": 3.1783370971679688,
|
| 637 |
+
"learning_rate": 1.924939467312349e-05,
|
| 638 |
+
"loss": 0.0583,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.6606170598911071,
|
| 643 |
+
"grad_norm": 14.949738502502441,
|
| 644 |
+
"learning_rate": 1.884584342211461e-05,
|
| 645 |
+
"loss": 0.0559,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.6678765880217786,
|
| 650 |
+
"grad_norm": 25.341707229614258,
|
| 651 |
+
"learning_rate": 1.8442292171105734e-05,
|
| 652 |
+
"loss": 0.1261,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.6751361161524501,
|
| 657 |
+
"grad_norm": 9.584409713745117,
|
| 658 |
+
"learning_rate": 1.8038740920096854e-05,
|
| 659 |
+
"loss": 0.1446,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.6823956442831216,
|
| 664 |
+
"grad_norm": 2.6669745445251465,
|
| 665 |
+
"learning_rate": 1.7635189669087974e-05,
|
| 666 |
+
"loss": 0.0853,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.6896551724137931,
|
| 671 |
+
"grad_norm": 11.011751174926758,
|
| 672 |
+
"learning_rate": 1.7231638418079095e-05,
|
| 673 |
+
"loss": 0.0754,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.6969147005444646,
|
| 678 |
+
"grad_norm": 6.686949729919434,
|
| 679 |
+
"learning_rate": 1.682808716707022e-05,
|
| 680 |
+
"loss": 0.0512,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.7041742286751361,
|
| 685 |
+
"grad_norm": 11.589463233947754,
|
| 686 |
+
"learning_rate": 1.642453591606134e-05,
|
| 687 |
+
"loss": 0.0512,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.7114337568058077,
|
| 692 |
+
"grad_norm": 0.608360767364502,
|
| 693 |
+
"learning_rate": 1.6020984665052463e-05,
|
| 694 |
+
"loss": 0.0847,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.7186932849364791,
|
| 699 |
+
"grad_norm": 5.557631015777588,
|
| 700 |
+
"learning_rate": 1.5617433414043583e-05,
|
| 701 |
+
"loss": 0.0789,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.7259528130671506,
|
| 706 |
+
"grad_norm": 7.303544044494629,
|
| 707 |
+
"learning_rate": 1.5213882163034707e-05,
|
| 708 |
+
"loss": 0.0978,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.7332123411978222,
|
| 713 |
+
"grad_norm": 5.494322299957275,
|
| 714 |
+
"learning_rate": 1.4810330912025828e-05,
|
| 715 |
+
"loss": 0.0551,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.7404718693284936,
|
| 720 |
+
"grad_norm": 9.331663131713867,
|
| 721 |
+
"learning_rate": 1.440677966101695e-05,
|
| 722 |
+
"loss": 0.0621,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.7477313974591652,
|
| 727 |
+
"grad_norm": 19.471540451049805,
|
| 728 |
+
"learning_rate": 1.400322841000807e-05,
|
| 729 |
+
"loss": 0.0721,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.7549909255898367,
|
| 734 |
+
"grad_norm": 5.023273468017578,
|
| 735 |
+
"learning_rate": 1.3599677158999194e-05,
|
| 736 |
+
"loss": 0.1148,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.7622504537205081,
|
| 741 |
+
"grad_norm": 5.059109687805176,
|
| 742 |
+
"learning_rate": 1.3196125907990314e-05,
|
| 743 |
+
"loss": 0.0644,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.7695099818511797,
|
| 748 |
+
"grad_norm": 7.810068607330322,
|
| 749 |
+
"learning_rate": 1.2792574656981438e-05,
|
| 750 |
+
"loss": 0.0452,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.7767695099818511,
|
| 755 |
+
"grad_norm": 7.319639682769775,
|
| 756 |
+
"learning_rate": 1.2389023405972559e-05,
|
| 757 |
+
"loss": 0.0626,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.7840290381125227,
|
| 762 |
+
"grad_norm": 2.656923294067383,
|
| 763 |
+
"learning_rate": 1.198547215496368e-05,
|
| 764 |
+
"loss": 0.0872,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.7912885662431942,
|
| 769 |
+
"grad_norm": 15.217109680175781,
|
| 770 |
+
"learning_rate": 1.1581920903954803e-05,
|
| 771 |
+
"loss": 0.0546,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.7985480943738656,
|
| 776 |
+
"grad_norm": 4.025232315063477,
|
| 777 |
+
"learning_rate": 1.1178369652945925e-05,
|
| 778 |
+
"loss": 0.0612,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.8058076225045372,
|
| 783 |
+
"grad_norm": 12.107078552246094,
|
| 784 |
+
"learning_rate": 1.0774818401937047e-05,
|
| 785 |
+
"loss": 0.0466,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.8130671506352087,
|
| 790 |
+
"grad_norm": 16.272573471069336,
|
| 791 |
+
"learning_rate": 1.0371267150928169e-05,
|
| 792 |
+
"loss": 0.0948,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.8203266787658802,
|
| 797 |
+
"grad_norm": 6.521091461181641,
|
| 798 |
+
"learning_rate": 9.96771589991929e-06,
|
| 799 |
+
"loss": 0.0428,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.8275862068965517,
|
| 804 |
+
"grad_norm": 19.286582946777344,
|
| 805 |
+
"learning_rate": 9.564164648910412e-06,
|
| 806 |
+
"loss": 0.05,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.8348457350272233,
|
| 811 |
+
"grad_norm": 5.874216079711914,
|
| 812 |
+
"learning_rate": 9.160613397901534e-06,
|
| 813 |
+
"loss": 0.0933,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.8421052631578947,
|
| 818 |
+
"grad_norm": 8.450905799865723,
|
| 819 |
+
"learning_rate": 8.757062146892656e-06,
|
| 820 |
+
"loss": 0.0639,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.8493647912885662,
|
| 825 |
+
"grad_norm": 13.851323127746582,
|
| 826 |
+
"learning_rate": 8.353510895883778e-06,
|
| 827 |
+
"loss": 0.0759,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.8566243194192378,
|
| 832 |
+
"grad_norm": 13.392071723937988,
|
| 833 |
+
"learning_rate": 7.949959644874898e-06,
|
| 834 |
+
"loss": 0.0605,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.8638838475499092,
|
| 839 |
+
"grad_norm": 4.568936347961426,
|
| 840 |
+
"learning_rate": 7.546408393866021e-06,
|
| 841 |
+
"loss": 0.08,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.8711433756805808,
|
| 846 |
+
"grad_norm": 1.7593340873718262,
|
| 847 |
+
"learning_rate": 7.142857142857143e-06,
|
| 848 |
+
"loss": 0.0423,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.8784029038112523,
|
| 853 |
+
"grad_norm": 4.213983058929443,
|
| 854 |
+
"learning_rate": 6.739305891848265e-06,
|
| 855 |
+
"loss": 0.0374,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.8856624319419237,
|
| 860 |
+
"grad_norm": 10.53001594543457,
|
| 861 |
+
"learning_rate": 6.335754640839386e-06,
|
| 862 |
+
"loss": 0.0531,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.8929219600725953,
|
| 867 |
+
"grad_norm": 11.019963264465332,
|
| 868 |
+
"learning_rate": 5.932203389830509e-06,
|
| 869 |
+
"loss": 0.0511,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.9001814882032668,
|
| 874 |
+
"grad_norm": 6.9437785148620605,
|
| 875 |
+
"learning_rate": 5.52865213882163e-06,
|
| 876 |
+
"loss": 0.0559,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.9074410163339383,
|
| 881 |
+
"grad_norm": 2.238781213760376,
|
| 882 |
+
"learning_rate": 5.1251008878127525e-06,
|
| 883 |
+
"loss": 0.051,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.9147005444646098,
|
| 888 |
+
"grad_norm": 10.967581748962402,
|
| 889 |
+
"learning_rate": 4.721549636803875e-06,
|
| 890 |
+
"loss": 0.0731,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.9219600725952813,
|
| 895 |
+
"grad_norm": 7.211132526397705,
|
| 896 |
+
"learning_rate": 4.317998385794996e-06,
|
| 897 |
+
"loss": 0.0438,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.9292196007259528,
|
| 902 |
+
"grad_norm": 4.4996209144592285,
|
| 903 |
+
"learning_rate": 3.914447134786118e-06,
|
| 904 |
+
"loss": 0.0338,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.9364791288566243,
|
| 909 |
+
"grad_norm": 13.61806869506836,
|
| 910 |
+
"learning_rate": 3.5108958837772397e-06,
|
| 911 |
+
"loss": 0.0584,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.9437386569872959,
|
| 916 |
+
"grad_norm": 7.869454860687256,
|
| 917 |
+
"learning_rate": 3.107344632768362e-06,
|
| 918 |
+
"loss": 0.0497,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.9509981851179673,
|
| 923 |
+
"grad_norm": 4.4050469398498535,
|
| 924 |
+
"learning_rate": 2.7037933817594835e-06,
|
| 925 |
+
"loss": 0.0274,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.9582577132486388,
|
| 930 |
+
"grad_norm": 15.032096862792969,
|
| 931 |
+
"learning_rate": 2.3002421307506056e-06,
|
| 932 |
+
"loss": 0.0523,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.9655172413793104,
|
| 937 |
+
"grad_norm": 13.652729034423828,
|
| 938 |
+
"learning_rate": 1.8966908797417273e-06,
|
| 939 |
+
"loss": 0.0562,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.9727767695099818,
|
| 944 |
+
"grad_norm": 3.0128390789031982,
|
| 945 |
+
"learning_rate": 1.4931396287328492e-06,
|
| 946 |
+
"loss": 0.043,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.9800362976406534,
|
| 951 |
+
"grad_norm": 7.441168785095215,
|
| 952 |
+
"learning_rate": 1.089588377723971e-06,
|
| 953 |
+
"loss": 0.049,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.9872958257713249,
|
| 958 |
+
"grad_norm": 8.190865516662598,
|
| 959 |
+
"learning_rate": 6.860371267150928e-07,
|
| 960 |
+
"loss": 0.0546,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.9945553539019963,
|
| 965 |
+
"grad_norm": 14.838714599609375,
|
| 966 |
+
"learning_rate": 2.8248587570621473e-07,
|
| 967 |
+
"loss": 0.0406,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.9996370235934664,
|
| 972 |
+
"eval_accuracy": 0.9835503925216681,
|
| 973 |
+
"eval_f1": 0.9895570759812747,
|
| 974 |
+
"eval_loss": 0.04456344619393349,
|
| 975 |
+
"eval_precision": 0.9834235162758739,
|
| 976 |
+
"eval_recall": 0.995767625231911,
|
| 977 |
+
"eval_roc_auc": 0.9989405310921683,
|
| 978 |
+
"eval_runtime": 188.542,
|
| 979 |
+
"eval_samples_per_second": 467.524,
|
| 980 |
+
"eval_steps_per_second": 29.224,
|
| 981 |
+
"step": 1377
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 0.9996370235934664,
|
| 985 |
+
"step": 1377,
|
| 986 |
+
"total_flos": 2.1905085973455176e+18,
|
| 987 |
+
"train_loss": 0.13800792182885657,
|
| 988 |
+
"train_runtime": 867.8791,
|
| 989 |
+
"train_samples_per_second": 101.567,
|
| 990 |
+
"train_steps_per_second": 1.587
|
| 991 |
+
}
|
| 992 |
+
],
|
| 993 |
+
"logging_steps": 10,
|
| 994 |
+
"max_steps": 1377,
|
| 995 |
+
"num_input_tokens_seen": 0,
|
| 996 |
+
"num_train_epochs": 1,
|
| 997 |
+
"save_steps": 500,
|
| 998 |
+
"stateful_callbacks": {
|
| 999 |
+
"TrainerControl": {
|
| 1000 |
+
"args": {
|
| 1001 |
+
"should_epoch_stop": false,
|
| 1002 |
+
"should_evaluate": false,
|
| 1003 |
+
"should_log": false,
|
| 1004 |
+
"should_save": true,
|
| 1005 |
+
"should_training_stop": true
|
| 1006 |
+
},
|
| 1007 |
+
"attributes": {}
|
| 1008 |
+
}
|
| 1009 |
+
},
|
| 1010 |
+
"total_flos": 2.1905085973455176e+18,
|
| 1011 |
+
"train_batch_size": 16,
|
| 1012 |
+
"trial_name": null,
|
| 1013 |
+
"trial_params": null
|
| 1014 |
+
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5240
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b501c95f9de0f433a655beb47c129d9fc96b2bb19a8daf0a2123cf3a2a856cd
|
| 3 |
size 5240
|
weights_biases.pkl
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6858
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:add71734889f1fa9cfb280ce1539480e7f7d692ab9fb524285260b510f151ae9
|
| 3 |
size 6858
|