Daniil Larionov
commited on
Commit
·
3d6becb
1
Parent(s):
e45f954
End of training
Browse files- all_results.json +14 -0
- eval_results.json +9 -0
- pytorch_model.bin +1 -1
- runs/events.out.tfevents.1636571350.5690110ef819.12031.0 +2 -2
- runs/events.out.tfevents.1636571613.5690110ef819.12031.2 +3 -0
- train_results.json +8 -0
- trainer_state.json +805 -0
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.9306930899620056,
|
| 4 |
+
"eval_loss": 0.39733296632766724,
|
| 5 |
+
"eval_runtime": 0.4549,
|
| 6 |
+
"eval_samples": 101,
|
| 7 |
+
"eval_samples_per_second": 222.004,
|
| 8 |
+
"eval_steps_per_second": 28.575,
|
| 9 |
+
"train_loss": 0.1360975634233144,
|
| 10 |
+
"train_runtime": 211.9781,
|
| 11 |
+
"train_samples": 908,
|
| 12 |
+
"train_samples_per_second": 42.835,
|
| 13 |
+
"train_steps_per_second": 2.689
|
| 14 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.9306930899620056,
|
| 4 |
+
"eval_loss": 0.39733296632766724,
|
| 5 |
+
"eval_runtime": 0.4549,
|
| 6 |
+
"eval_samples": 101,
|
| 7 |
+
"eval_samples_per_second": 222.004,
|
| 8 |
+
"eval_steps_per_second": 28.575
|
| 9 |
+
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 713322669
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6eb89ee804cc02c1c4c21c5890fc3226645b832719184d7fb5cf4e1c123736ac
|
| 3 |
size 713322669
|
runs/events.out.tfevents.1636571350.5690110ef819.12031.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:87c197d2235d8ffd5d993a98eea6166d153c051ff9586290ceeb376a05fceb9a
|
| 3 |
+
size 25255
|
runs/events.out.tfevents.1636571613.5690110ef819.12031.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:200a6591b465c775454bb582f74f4cfaf88dfac2c18b68f92e570bad196f57f0
|
| 3 |
+
size 363
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"train_loss": 0.1360975634233144,
|
| 4 |
+
"train_runtime": 211.9781,
|
| 5 |
+
"train_samples": 908,
|
| 6 |
+
"train_samples_per_second": 42.835,
|
| 7 |
+
"train_steps_per_second": 2.689
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,805 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"global_step": 570,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 1.4285714285714286e-06,
|
| 13 |
+
"loss": 1.0416,
|
| 14 |
+
"step": 1
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.09,
|
| 18 |
+
"learning_rate": 7.142857142857143e-06,
|
| 19 |
+
"loss": 1.0953,
|
| 20 |
+
"step": 5
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.18,
|
| 24 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 25 |
+
"loss": 0.9549,
|
| 26 |
+
"step": 10
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.26,
|
| 30 |
+
"learning_rate": 2.1428571428571428e-05,
|
| 31 |
+
"loss": 0.8765,
|
| 32 |
+
"step": 15
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.35,
|
| 36 |
+
"learning_rate": 2.857142857142857e-05,
|
| 37 |
+
"loss": 0.7896,
|
| 38 |
+
"step": 20
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.44,
|
| 42 |
+
"learning_rate": 3.571428571428572e-05,
|
| 43 |
+
"loss": 0.8036,
|
| 44 |
+
"step": 25
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.53,
|
| 48 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 49 |
+
"loss": 0.6975,
|
| 50 |
+
"step": 30
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.61,
|
| 54 |
+
"learning_rate": 5e-05,
|
| 55 |
+
"loss": 0.7023,
|
| 56 |
+
"step": 35
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.7,
|
| 60 |
+
"learning_rate": 4.998922515567496e-05,
|
| 61 |
+
"loss": 0.6305,
|
| 62 |
+
"step": 40
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.79,
|
| 66 |
+
"learning_rate": 4.995690991048146e-05,
|
| 67 |
+
"loss": 0.7052,
|
| 68 |
+
"step": 45
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.88,
|
| 72 |
+
"learning_rate": 4.99030821197584e-05,
|
| 73 |
+
"loss": 0.6331,
|
| 74 |
+
"step": 50
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.96,
|
| 78 |
+
"learning_rate": 4.982778818239101e-05,
|
| 79 |
+
"loss": 0.4892,
|
| 80 |
+
"step": 55
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.0,
|
| 84 |
+
"eval_accuracy": 0.7524752616882324,
|
| 85 |
+
"eval_loss": 0.5470823049545288,
|
| 86 |
+
"eval_runtime": 0.4397,
|
| 87 |
+
"eval_samples_per_second": 229.701,
|
| 88 |
+
"eval_steps_per_second": 29.566,
|
| 89 |
+
"step": 57
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.05,
|
| 93 |
+
"learning_rate": 4.97310930008156e-05,
|
| 94 |
+
"loss": 0.4139,
|
| 95 |
+
"step": 60
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 1.14,
|
| 99 |
+
"learning_rate": 4.961307992507443e-05,
|
| 100 |
+
"loss": 0.3817,
|
| 101 |
+
"step": 65
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.23,
|
| 105 |
+
"learning_rate": 4.947385068096907e-05,
|
| 106 |
+
"loss": 0.2255,
|
| 107 |
+
"step": 70
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.32,
|
| 111 |
+
"learning_rate": 4.9313525282373974e-05,
|
| 112 |
+
"loss": 0.3455,
|
| 113 |
+
"step": 75
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 1.4,
|
| 117 |
+
"learning_rate": 4.9132241927786035e-05,
|
| 118 |
+
"loss": 0.3717,
|
| 119 |
+
"step": 80
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.49,
|
| 123 |
+
"learning_rate": 4.893015688119921e-05,
|
| 124 |
+
"loss": 0.4358,
|
| 125 |
+
"step": 85
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 1.58,
|
| 129 |
+
"learning_rate": 4.870744433740688e-05,
|
| 130 |
+
"loss": 0.3544,
|
| 131 |
+
"step": 90
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 1.67,
|
| 135 |
+
"learning_rate": 4.8464296271848155e-05,
|
| 136 |
+
"loss": 0.1876,
|
| 137 |
+
"step": 95
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.75,
|
| 141 |
+
"learning_rate": 4.8200922275127355e-05,
|
| 142 |
+
"loss": 0.1961,
|
| 143 |
+
"step": 100
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.84,
|
| 147 |
+
"learning_rate": 4.7917549372349616e-05,
|
| 148 |
+
"loss": 0.2548,
|
| 149 |
+
"step": 105
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.93,
|
| 153 |
+
"learning_rate": 4.761442182742799e-05,
|
| 154 |
+
"loss": 0.2302,
|
| 155 |
+
"step": 110
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 2.0,
|
| 159 |
+
"eval_accuracy": 0.8613861203193665,
|
| 160 |
+
"eval_loss": 0.42704179883003235,
|
| 161 |
+
"eval_runtime": 0.4285,
|
| 162 |
+
"eval_samples_per_second": 235.733,
|
| 163 |
+
"eval_steps_per_second": 30.342,
|
| 164 |
+
"step": 114
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.02,
|
| 168 |
+
"learning_rate": 4.7291800932531064e-05,
|
| 169 |
+
"loss": 0.3154,
|
| 170 |
+
"step": 115
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 2.11,
|
| 174 |
+
"learning_rate": 4.694996478285231e-05,
|
| 175 |
+
"loss": 0.2103,
|
| 176 |
+
"step": 120
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 2.19,
|
| 180 |
+
"learning_rate": 4.6589208036895535e-05,
|
| 181 |
+
"loss": 0.1499,
|
| 182 |
+
"step": 125
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 2.28,
|
| 186 |
+
"learning_rate": 4.620984166248288e-05,
|
| 187 |
+
"loss": 0.1851,
|
| 188 |
+
"step": 130
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 2.37,
|
| 192 |
+
"learning_rate": 4.581219266870446e-05,
|
| 193 |
+
"loss": 0.0972,
|
| 194 |
+
"step": 135
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 2.46,
|
| 198 |
+
"learning_rate": 4.53966038240406e-05,
|
| 199 |
+
"loss": 0.2495,
|
| 200 |
+
"step": 140
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 2.54,
|
| 204 |
+
"learning_rate": 4.496343336089965e-05,
|
| 205 |
+
"loss": 0.0528,
|
| 206 |
+
"step": 145
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.63,
|
| 210 |
+
"learning_rate": 4.4513054666826146e-05,
|
| 211 |
+
"loss": 0.127,
|
| 212 |
+
"step": 150
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 2.72,
|
| 216 |
+
"learning_rate": 4.4045855962645363e-05,
|
| 217 |
+
"loss": 0.2142,
|
| 218 |
+
"step": 155
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 2.81,
|
| 222 |
+
"learning_rate": 4.3562239967821805e-05,
|
| 223 |
+
"loss": 0.1373,
|
| 224 |
+
"step": 160
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.89,
|
| 228 |
+
"learning_rate": 4.306262355332006e-05,
|
| 229 |
+
"loss": 0.0635,
|
| 230 |
+
"step": 165
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.98,
|
| 234 |
+
"learning_rate": 4.254743738226721e-05,
|
| 235 |
+
"loss": 0.1744,
|
| 236 |
+
"step": 170
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 3.0,
|
| 240 |
+
"eval_accuracy": 0.9306930899620056,
|
| 241 |
+
"eval_loss": 0.24064408242702484,
|
| 242 |
+
"eval_runtime": 1.6792,
|
| 243 |
+
"eval_samples_per_second": 60.146,
|
| 244 |
+
"eval_steps_per_second": 7.742,
|
| 245 |
+
"step": 171
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 3.07,
|
| 249 |
+
"learning_rate": 4.201712553872658e-05,
|
| 250 |
+
"loss": 0.0273,
|
| 251 |
+
"step": 175
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 3.16,
|
| 255 |
+
"learning_rate": 4.147214514490278e-05,
|
| 256 |
+
"loss": 0.0421,
|
| 257 |
+
"step": 180
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 3.25,
|
| 261 |
+
"learning_rate": 4.0912965967108125e-05,
|
| 262 |
+
"loss": 0.0495,
|
| 263 |
+
"step": 185
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 3.33,
|
| 267 |
+
"learning_rate": 4.034007001082985e-05,
|
| 268 |
+
"loss": 0.0192,
|
| 269 |
+
"step": 190
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 3.42,
|
| 273 |
+
"learning_rate": 3.975395110524742e-05,
|
| 274 |
+
"loss": 0.0252,
|
| 275 |
+
"step": 195
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 3.51,
|
| 279 |
+
"learning_rate": 3.9155114477557933e-05,
|
| 280 |
+
"loss": 0.0318,
|
| 281 |
+
"step": 200
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 3.6,
|
| 285 |
+
"learning_rate": 3.854407631747654e-05,
|
| 286 |
+
"loss": 0.0873,
|
| 287 |
+
"step": 205
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 3.68,
|
| 291 |
+
"learning_rate": 3.792136333228735e-05,
|
| 292 |
+
"loss": 0.0091,
|
| 293 |
+
"step": 210
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 3.77,
|
| 297 |
+
"learning_rate": 3.728751229282836e-05,
|
| 298 |
+
"loss": 0.0402,
|
| 299 |
+
"step": 215
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 3.86,
|
| 303 |
+
"learning_rate": 3.664306957080159e-05,
|
| 304 |
+
"loss": 0.1102,
|
| 305 |
+
"step": 220
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 3.95,
|
| 309 |
+
"learning_rate": 3.598859066780754e-05,
|
| 310 |
+
"loss": 0.0497,
|
| 311 |
+
"step": 225
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 4.0,
|
| 315 |
+
"eval_accuracy": 0.9009901285171509,
|
| 316 |
+
"eval_loss": 0.5820234417915344,
|
| 317 |
+
"eval_runtime": 0.4317,
|
| 318 |
+
"eval_samples_per_second": 233.942,
|
| 319 |
+
"eval_steps_per_second": 30.111,
|
| 320 |
+
"step": 228
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 4.04,
|
| 324 |
+
"learning_rate": 3.5324639736509714e-05,
|
| 325 |
+
"loss": 0.1968,
|
| 326 |
+
"step": 230
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 4.12,
|
| 330 |
+
"learning_rate": 3.4651789094342044e-05,
|
| 331 |
+
"loss": 0.0267,
|
| 332 |
+
"step": 235
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 4.21,
|
| 336 |
+
"learning_rate": 3.39706187301784e-05,
|
| 337 |
+
"loss": 0.0339,
|
| 338 |
+
"step": 240
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 4.3,
|
| 342 |
+
"learning_rate": 3.3281715804389403e-05,
|
| 343 |
+
"loss": 0.0614,
|
| 344 |
+
"step": 245
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 4.39,
|
| 348 |
+
"learning_rate": 3.258567414271748e-05,
|
| 349 |
+
"loss": 0.0682,
|
| 350 |
+
"step": 250
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 4.47,
|
| 354 |
+
"learning_rate": 3.18830937244065e-05,
|
| 355 |
+
"loss": 0.0068,
|
| 356 |
+
"step": 255
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 4.56,
|
| 360 |
+
"learning_rate": 3.117458016502711e-05,
|
| 361 |
+
"loss": 0.0028,
|
| 362 |
+
"step": 260
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 4.65,
|
| 366 |
+
"learning_rate": 3.046074419444366e-05,
|
| 367 |
+
"loss": 0.0069,
|
| 368 |
+
"step": 265
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 4.74,
|
| 372 |
+
"learning_rate": 2.9742201130372693e-05,
|
| 373 |
+
"loss": 0.0019,
|
| 374 |
+
"step": 270
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 4.82,
|
| 378 |
+
"learning_rate": 2.901957034798671e-05,
|
| 379 |
+
"loss": 0.0868,
|
| 380 |
+
"step": 275
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 4.91,
|
| 384 |
+
"learning_rate": 2.8293474746020472e-05,
|
| 385 |
+
"loss": 0.002,
|
| 386 |
+
"step": 280
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 5.0,
|
| 390 |
+
"learning_rate": 2.756454020984009e-05,
|
| 391 |
+
"loss": 0.0207,
|
| 392 |
+
"step": 285
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 5.0,
|
| 396 |
+
"eval_accuracy": 0.9306930899620056,
|
| 397 |
+
"eval_loss": 0.3814234733581543,
|
| 398 |
+
"eval_runtime": 0.4243,
|
| 399 |
+
"eval_samples_per_second": 238.063,
|
| 400 |
+
"eval_steps_per_second": 30.642,
|
| 401 |
+
"step": 285
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 5.09,
|
| 405 |
+
"learning_rate": 2.68333950719376e-05,
|
| 406 |
+
"loss": 0.0021,
|
| 407 |
+
"step": 290
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 5.18,
|
| 411 |
+
"learning_rate": 2.6100669570316195e-05,
|
| 412 |
+
"loss": 0.0019,
|
| 413 |
+
"step": 295
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 5.26,
|
| 417 |
+
"learning_rate": 2.5366995305232916e-05,
|
| 418 |
+
"loss": 0.0016,
|
| 419 |
+
"step": 300
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 5.35,
|
| 423 |
+
"learning_rate": 2.463300469476709e-05,
|
| 424 |
+
"loss": 0.0798,
|
| 425 |
+
"step": 305
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 5.44,
|
| 429 |
+
"learning_rate": 2.3899330429683807e-05,
|
| 430 |
+
"loss": 0.071,
|
| 431 |
+
"step": 310
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 5.53,
|
| 435 |
+
"learning_rate": 2.3166604928062406e-05,
|
| 436 |
+
"loss": 0.0023,
|
| 437 |
+
"step": 315
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 5.61,
|
| 441 |
+
"learning_rate": 2.243545979015992e-05,
|
| 442 |
+
"loss": 0.0024,
|
| 443 |
+
"step": 320
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 5.7,
|
| 447 |
+
"learning_rate": 2.1706525253979534e-05,
|
| 448 |
+
"loss": 0.002,
|
| 449 |
+
"step": 325
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 5.79,
|
| 453 |
+
"learning_rate": 2.0980429652013297e-05,
|
| 454 |
+
"loss": 0.0019,
|
| 455 |
+
"step": 330
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 5.88,
|
| 459 |
+
"learning_rate": 2.025779886962731e-05,
|
| 460 |
+
"loss": 0.0138,
|
| 461 |
+
"step": 335
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 5.96,
|
| 465 |
+
"learning_rate": 1.9539255805556344e-05,
|
| 466 |
+
"loss": 0.0019,
|
| 467 |
+
"step": 340
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 6.0,
|
| 471 |
+
"eval_accuracy": 0.9306930899620056,
|
| 472 |
+
"eval_loss": 0.39340826869010925,
|
| 473 |
+
"eval_runtime": 0.4347,
|
| 474 |
+
"eval_samples_per_second": 232.361,
|
| 475 |
+
"eval_steps_per_second": 29.908,
|
| 476 |
+
"step": 342
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 6.05,
|
| 480 |
+
"learning_rate": 1.8825419834972902e-05,
|
| 481 |
+
"loss": 0.1167,
|
| 482 |
+
"step": 345
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 6.14,
|
| 486 |
+
"learning_rate": 1.811690627559351e-05,
|
| 487 |
+
"loss": 0.0018,
|
| 488 |
+
"step": 350
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 6.23,
|
| 492 |
+
"learning_rate": 1.7414325857282526e-05,
|
| 493 |
+
"loss": 0.0019,
|
| 494 |
+
"step": 355
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 6.32,
|
| 498 |
+
"learning_rate": 1.6718284195610606e-05,
|
| 499 |
+
"loss": 0.0521,
|
| 500 |
+
"step": 360
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 6.4,
|
| 504 |
+
"learning_rate": 1.6029381269821604e-05,
|
| 505 |
+
"loss": 0.0957,
|
| 506 |
+
"step": 365
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 6.49,
|
| 510 |
+
"learning_rate": 1.534821090565796e-05,
|
| 511 |
+
"loss": 0.0021,
|
| 512 |
+
"step": 370
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 6.58,
|
| 516 |
+
"learning_rate": 1.4675360263490295e-05,
|
| 517 |
+
"loss": 0.0017,
|
| 518 |
+
"step": 375
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 6.67,
|
| 522 |
+
"learning_rate": 1.4011409332192472e-05,
|
| 523 |
+
"loss": 0.0015,
|
| 524 |
+
"step": 380
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 6.75,
|
| 528 |
+
"learning_rate": 1.335693042919841e-05,
|
| 529 |
+
"loss": 0.0029,
|
| 530 |
+
"step": 385
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 6.84,
|
| 534 |
+
"learning_rate": 1.2712487707171645e-05,
|
| 535 |
+
"loss": 0.0782,
|
| 536 |
+
"step": 390
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 6.93,
|
| 540 |
+
"learning_rate": 1.2078636667712649e-05,
|
| 541 |
+
"loss": 0.0053,
|
| 542 |
+
"step": 395
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 7.0,
|
| 546 |
+
"eval_accuracy": 0.9306930899620056,
|
| 547 |
+
"eval_loss": 0.3684821128845215,
|
| 548 |
+
"eval_runtime": 0.4335,
|
| 549 |
+
"eval_samples_per_second": 232.965,
|
| 550 |
+
"eval_steps_per_second": 29.986,
|
| 551 |
+
"step": 399
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 7.02,
|
| 555 |
+
"learning_rate": 1.1455923682523475e-05,
|
| 556 |
+
"loss": 0.003,
|
| 557 |
+
"step": 400
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 7.11,
|
| 561 |
+
"learning_rate": 1.0844885522442074e-05,
|
| 562 |
+
"loss": 0.0016,
|
| 563 |
+
"step": 405
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 7.19,
|
| 567 |
+
"learning_rate": 1.0246048894752589e-05,
|
| 568 |
+
"loss": 0.0014,
|
| 569 |
+
"step": 410
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 7.28,
|
| 573 |
+
"learning_rate": 9.659929989170154e-06,
|
| 574 |
+
"loss": 0.0014,
|
| 575 |
+
"step": 415
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 7.37,
|
| 579 |
+
"learning_rate": 9.087034032891883e-06,
|
| 580 |
+
"loss": 0.0012,
|
| 581 |
+
"step": 420
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 7.46,
|
| 585 |
+
"learning_rate": 8.527854855097225e-06,
|
| 586 |
+
"loss": 0.0016,
|
| 587 |
+
"step": 425
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 7.54,
|
| 591 |
+
"learning_rate": 7.982874461273438e-06,
|
| 592 |
+
"loss": 0.0838,
|
| 593 |
+
"step": 430
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 7.63,
|
| 597 |
+
"learning_rate": 7.452562617732794e-06,
|
| 598 |
+
"loss": 0.0012,
|
| 599 |
+
"step": 435
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 7.72,
|
| 603 |
+
"learning_rate": 6.93737644667995e-06,
|
| 604 |
+
"loss": 0.0013,
|
| 605 |
+
"step": 440
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 7.81,
|
| 609 |
+
"learning_rate": 6.4377600321782e-06,
|
| 610 |
+
"loss": 0.0013,
|
| 611 |
+
"step": 445
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 7.89,
|
| 615 |
+
"learning_rate": 5.954144037354645e-06,
|
| 616 |
+
"loss": 0.0015,
|
| 617 |
+
"step": 450
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 7.98,
|
| 621 |
+
"learning_rate": 5.486945333173851e-06,
|
| 622 |
+
"loss": 0.0012,
|
| 623 |
+
"step": 455
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 8.0,
|
| 627 |
+
"eval_accuracy": 0.9306930899620056,
|
| 628 |
+
"eval_loss": 0.3639465570449829,
|
| 629 |
+
"eval_runtime": 0.4448,
|
| 630 |
+
"eval_samples_per_second": 227.087,
|
| 631 |
+
"eval_steps_per_second": 29.229,
|
| 632 |
+
"step": 456
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 8.07,
|
| 636 |
+
"learning_rate": 5.036566639100351e-06,
|
| 637 |
+
"loss": 0.0012,
|
| 638 |
+
"step": 460
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 8.16,
|
| 642 |
+
"learning_rate": 4.603396175959404e-06,
|
| 643 |
+
"loss": 0.0013,
|
| 644 |
+
"step": 465
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 8.25,
|
| 648 |
+
"learning_rate": 4.187807331295549e-06,
|
| 649 |
+
"loss": 0.0013,
|
| 650 |
+
"step": 470
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 8.33,
|
| 654 |
+
"learning_rate": 3.7901583375171273e-06,
|
| 655 |
+
"loss": 0.001,
|
| 656 |
+
"step": 475
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 8.42,
|
| 660 |
+
"learning_rate": 3.4107919631044732e-06,
|
| 661 |
+
"loss": 0.001,
|
| 662 |
+
"step": 480
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 8.51,
|
| 666 |
+
"learning_rate": 3.0500352171476897e-06,
|
| 667 |
+
"loss": 0.001,
|
| 668 |
+
"step": 485
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 8.6,
|
| 672 |
+
"learning_rate": 2.708199067468939e-06,
|
| 673 |
+
"loss": 0.0009,
|
| 674 |
+
"step": 490
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 8.68,
|
| 678 |
+
"learning_rate": 2.385578172572009e-06,
|
| 679 |
+
"loss": 0.0008,
|
| 680 |
+
"step": 495
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 8.77,
|
| 684 |
+
"learning_rate": 2.0824506276503897e-06,
|
| 685 |
+
"loss": 0.0594,
|
| 686 |
+
"step": 500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 8.86,
|
| 690 |
+
"learning_rate": 1.7990777248726442e-06,
|
| 691 |
+
"loss": 0.0011,
|
| 692 |
+
"step": 505
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 8.95,
|
| 696 |
+
"learning_rate": 1.5357037281518522e-06,
|
| 697 |
+
"loss": 0.001,
|
| 698 |
+
"step": 510
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 9.0,
|
| 702 |
+
"eval_accuracy": 0.9306930899620056,
|
| 703 |
+
"eval_loss": 0.39552733302116394,
|
| 704 |
+
"eval_runtime": 0.4382,
|
| 705 |
+
"eval_samples_per_second": 230.47,
|
| 706 |
+
"eval_steps_per_second": 29.664,
|
| 707 |
+
"step": 513
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 9.04,
|
| 711 |
+
"learning_rate": 1.2925556625931173e-06,
|
| 712 |
+
"loss": 0.0019,
|
| 713 |
+
"step": 515
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 9.12,
|
| 717 |
+
"learning_rate": 1.0698431188007952e-06,
|
| 718 |
+
"loss": 0.0009,
|
| 719 |
+
"step": 520
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 9.21,
|
| 723 |
+
"learning_rate": 8.677580722139672e-07,
|
| 724 |
+
"loss": 0.001,
|
| 725 |
+
"step": 525
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 9.3,
|
| 729 |
+
"learning_rate": 6.864747176260289e-07,
|
| 730 |
+
"loss": 0.0009,
|
| 731 |
+
"step": 530
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 9.39,
|
| 735 |
+
"learning_rate": 5.261493190309303e-07,
|
| 736 |
+
"loss": 0.0008,
|
| 737 |
+
"step": 535
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 9.47,
|
| 741 |
+
"learning_rate": 3.8692007492557024e-07,
|
| 742 |
+
"loss": 0.0011,
|
| 743 |
+
"step": 540
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 9.56,
|
| 747 |
+
"learning_rate": 2.6890699918440676e-07,
|
| 748 |
+
"loss": 0.0016,
|
| 749 |
+
"step": 545
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 9.65,
|
| 753 |
+
"learning_rate": 1.7221181760899152e-07,
|
| 754 |
+
"loss": 0.0009,
|
| 755 |
+
"step": 550
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 9.74,
|
| 759 |
+
"learning_rate": 9.691788024160376e-08,
|
| 760 |
+
"loss": 0.001,
|
| 761 |
+
"step": 555
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 9.82,
|
| 765 |
+
"learning_rate": 4.3090089518540987e-08,
|
| 766 |
+
"loss": 0.045,
|
| 767 |
+
"step": 560
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 9.91,
|
| 771 |
+
"learning_rate": 1.0774844325039946e-08,
|
| 772 |
+
"loss": 0.0009,
|
| 773 |
+
"step": 565
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 10.0,
|
| 777 |
+
"learning_rate": 0.0,
|
| 778 |
+
"loss": 0.0009,
|
| 779 |
+
"step": 570
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 10.0,
|
| 783 |
+
"eval_accuracy": 0.9306930899620056,
|
| 784 |
+
"eval_loss": 0.39733296632766724,
|
| 785 |
+
"eval_runtime": 1.6316,
|
| 786 |
+
"eval_samples_per_second": 61.901,
|
| 787 |
+
"eval_steps_per_second": 7.968,
|
| 788 |
+
"step": 570
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 10.0,
|
| 792 |
+
"step": 570,
|
| 793 |
+
"total_flos": 197222006749320.0,
|
| 794 |
+
"train_loss": 0.1360975634233144,
|
| 795 |
+
"train_runtime": 211.9781,
|
| 796 |
+
"train_samples_per_second": 42.835,
|
| 797 |
+
"train_steps_per_second": 2.689
|
| 798 |
+
}
|
| 799 |
+
],
|
| 800 |
+
"max_steps": 570,
|
| 801 |
+
"num_train_epochs": 10,
|
| 802 |
+
"total_flos": 197222006749320.0,
|
| 803 |
+
"trial_name": null,
|
| 804 |
+
"trial_params": null
|
| 805 |
+
}
|