Upload 8 files
Browse files- config.json +43 -0
- optimizer.pt +3 -0
- preprocessor_config.json +18 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1756 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/resnet-50",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ResNetForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"depths": [
|
| 7 |
+
3,
|
| 8 |
+
4,
|
| 9 |
+
6,
|
| 10 |
+
3
|
| 11 |
+
],
|
| 12 |
+
"downsample_in_first_stage": false,
|
| 13 |
+
"embedding_size": 64,
|
| 14 |
+
"hidden_act": "relu",
|
| 15 |
+
"hidden_sizes": [
|
| 16 |
+
256,
|
| 17 |
+
512,
|
| 18 |
+
1024,
|
| 19 |
+
2048
|
| 20 |
+
],
|
| 21 |
+
"id2label": {
|
| 22 |
+
"0": "Cilindro",
|
| 23 |
+
"1": "Cone",
|
| 24 |
+
"2": "Esfera",
|
| 25 |
+
"3": "Nao definido",
|
| 26 |
+
"4": "Paralelepipedo",
|
| 27 |
+
"5": "Superficie"
|
| 28 |
+
},
|
| 29 |
+
"label2id": {
|
| 30 |
+
"Cilindro": 0,
|
| 31 |
+
"Cone": 1,
|
| 32 |
+
"Esfera": 2,
|
| 33 |
+
"Nao definido": 3,
|
| 34 |
+
"Paralelepipedo": 4,
|
| 35 |
+
"Superficie": 5
|
| 36 |
+
},
|
| 37 |
+
"layer_type": "bottleneck",
|
| 38 |
+
"model_type": "resnet",
|
| 39 |
+
"num_channels": 3,
|
| 40 |
+
"problem_type": "single_label_classification",
|
| 41 |
+
"torch_dtype": "float32",
|
| 42 |
+
"transformers_version": "4.20.1"
|
| 43 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3166448721501dd5f9c16a4a995fd23843447d9a1b5bfc66f16305116a87fc81
|
| 3 |
+
size 188256445
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"crop_pct": 0.875,
|
| 3 |
+
"do_normalize": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"feature_extractor_type": "ConvNextFeatureExtractor",
|
| 6 |
+
"image_mean": [
|
| 7 |
+
0.485,
|
| 8 |
+
0.456,
|
| 9 |
+
0.406
|
| 10 |
+
],
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.229,
|
| 13 |
+
0.224,
|
| 14 |
+
0.225
|
| 15 |
+
],
|
| 16 |
+
"resample": 3,
|
| 17 |
+
"size": 224
|
| 18 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:75fe5e778e42fb0245358c60c7a160920933d0ddc3f492f0e6d027f00f704799
|
| 3 |
+
size 94405441
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5b26d84dfb0f134105c948fe88c30bfefe94cd551072824ea42a8230edfb24e8
|
| 3 |
+
size 14503
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5cd4208283981f9e5c5d6ca89a3ddaf890614c7e1041711d9b632b95c53494b6
|
| 3 |
+
size 623
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1756 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7818181818181819,
|
| 3 |
+
"best_model_checkpoint": "resnet-50-finetuned\\checkpoint-1190",
|
| 4 |
+
"epoch": 100.0,
|
| 5 |
+
"global_step": 1400,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.71,
|
| 12 |
+
"learning_rate": 3.5714285714285714e-06,
|
| 13 |
+
"loss": 1.7972,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.0,
|
| 18 |
+
"eval_accuracy": 0.11363636363636363,
|
| 19 |
+
"eval_loss": 1.800904393196106,
|
| 20 |
+
"eval_runtime": 5.5218,
|
| 21 |
+
"eval_samples_per_second": 79.685,
|
| 22 |
+
"eval_steps_per_second": 1.268,
|
| 23 |
+
"step": 14
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 1.43,
|
| 27 |
+
"learning_rate": 7.142857142857143e-06,
|
| 28 |
+
"loss": 1.7991,
|
| 29 |
+
"step": 20
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 2.0,
|
| 33 |
+
"eval_accuracy": 0.16136363636363638,
|
| 34 |
+
"eval_loss": 1.7934340238571167,
|
| 35 |
+
"eval_runtime": 3.2902,
|
| 36 |
+
"eval_samples_per_second": 133.73,
|
| 37 |
+
"eval_steps_per_second": 2.128,
|
| 38 |
+
"step": 28
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 2.14,
|
| 42 |
+
"learning_rate": 1.0714285714285714e-05,
|
| 43 |
+
"loss": 1.7925,
|
| 44 |
+
"step": 30
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 2.86,
|
| 48 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 49 |
+
"loss": 1.7871,
|
| 50 |
+
"step": 40
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.0,
|
| 54 |
+
"eval_accuracy": 0.2,
|
| 55 |
+
"eval_loss": 1.78143310546875,
|
| 56 |
+
"eval_runtime": 3.2024,
|
| 57 |
+
"eval_samples_per_second": 137.395,
|
| 58 |
+
"eval_steps_per_second": 2.186,
|
| 59 |
+
"step": 42
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 3.57,
|
| 63 |
+
"learning_rate": 1.785714285714286e-05,
|
| 64 |
+
"loss": 1.7813,
|
| 65 |
+
"step": 50
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 4.0,
|
| 69 |
+
"eval_accuracy": 0.20909090909090908,
|
| 70 |
+
"eval_loss": 1.7663880586624146,
|
| 71 |
+
"eval_runtime": 3.3082,
|
| 72 |
+
"eval_samples_per_second": 133.005,
|
| 73 |
+
"eval_steps_per_second": 2.116,
|
| 74 |
+
"step": 56
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 4.29,
|
| 78 |
+
"learning_rate": 2.1428571428571428e-05,
|
| 79 |
+
"loss": 1.7673,
|
| 80 |
+
"step": 60
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 5.0,
|
| 84 |
+
"learning_rate": 2.5e-05,
|
| 85 |
+
"loss": 1.7602,
|
| 86 |
+
"step": 70
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 5.0,
|
| 90 |
+
"eval_accuracy": 0.2772727272727273,
|
| 91 |
+
"eval_loss": 1.7464295625686646,
|
| 92 |
+
"eval_runtime": 3.3301,
|
| 93 |
+
"eval_samples_per_second": 132.128,
|
| 94 |
+
"eval_steps_per_second": 2.102,
|
| 95 |
+
"step": 70
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 5.71,
|
| 99 |
+
"learning_rate": 2.857142857142857e-05,
|
| 100 |
+
"loss": 1.7455,
|
| 101 |
+
"step": 80
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 6.0,
|
| 105 |
+
"eval_accuracy": 0.2840909090909091,
|
| 106 |
+
"eval_loss": 1.726538896560669,
|
| 107 |
+
"eval_runtime": 3.2579,
|
| 108 |
+
"eval_samples_per_second": 135.057,
|
| 109 |
+
"eval_steps_per_second": 2.149,
|
| 110 |
+
"step": 84
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 6.43,
|
| 114 |
+
"learning_rate": 3.2142857142857144e-05,
|
| 115 |
+
"loss": 1.7321,
|
| 116 |
+
"step": 90
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 7.0,
|
| 120 |
+
"eval_accuracy": 0.2727272727272727,
|
| 121 |
+
"eval_loss": 1.7056454420089722,
|
| 122 |
+
"eval_runtime": 3.2693,
|
| 123 |
+
"eval_samples_per_second": 134.587,
|
| 124 |
+
"eval_steps_per_second": 2.141,
|
| 125 |
+
"step": 98
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 7.14,
|
| 129 |
+
"learning_rate": 3.571428571428572e-05,
|
| 130 |
+
"loss": 1.7215,
|
| 131 |
+
"step": 100
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 7.86,
|
| 135 |
+
"learning_rate": 3.928571428571429e-05,
|
| 136 |
+
"loss": 1.706,
|
| 137 |
+
"step": 110
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 8.0,
|
| 141 |
+
"eval_accuracy": 0.3068181818181818,
|
| 142 |
+
"eval_loss": 1.6843111515045166,
|
| 143 |
+
"eval_runtime": 3.2827,
|
| 144 |
+
"eval_samples_per_second": 134.038,
|
| 145 |
+
"eval_steps_per_second": 2.132,
|
| 146 |
+
"step": 112
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 8.57,
|
| 150 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 151 |
+
"loss": 1.6891,
|
| 152 |
+
"step": 120
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 9.0,
|
| 156 |
+
"eval_accuracy": 0.325,
|
| 157 |
+
"eval_loss": 1.6573911905288696,
|
| 158 |
+
"eval_runtime": 3.1775,
|
| 159 |
+
"eval_samples_per_second": 138.473,
|
| 160 |
+
"eval_steps_per_second": 2.203,
|
| 161 |
+
"step": 126
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 9.29,
|
| 165 |
+
"learning_rate": 4.642857142857143e-05,
|
| 166 |
+
"loss": 1.6751,
|
| 167 |
+
"step": 130
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 10.0,
|
| 171 |
+
"learning_rate": 5e-05,
|
| 172 |
+
"loss": 1.6524,
|
| 173 |
+
"step": 140
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 10.0,
|
| 177 |
+
"eval_accuracy": 0.3090909090909091,
|
| 178 |
+
"eval_loss": 1.6302080154418945,
|
| 179 |
+
"eval_runtime": 3.3459,
|
| 180 |
+
"eval_samples_per_second": 131.503,
|
| 181 |
+
"eval_steps_per_second": 2.092,
|
| 182 |
+
"step": 140
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 10.71,
|
| 186 |
+
"learning_rate": 4.960317460317461e-05,
|
| 187 |
+
"loss": 1.6348,
|
| 188 |
+
"step": 150
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 11.0,
|
| 192 |
+
"eval_accuracy": 0.3613636363636364,
|
| 193 |
+
"eval_loss": 1.5982272624969482,
|
| 194 |
+
"eval_runtime": 3.2269,
|
| 195 |
+
"eval_samples_per_second": 136.356,
|
| 196 |
+
"eval_steps_per_second": 2.169,
|
| 197 |
+
"step": 154
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 11.43,
|
| 201 |
+
"learning_rate": 4.9206349206349204e-05,
|
| 202 |
+
"loss": 1.611,
|
| 203 |
+
"step": 160
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 12.0,
|
| 207 |
+
"eval_accuracy": 0.39545454545454545,
|
| 208 |
+
"eval_loss": 1.556486964225769,
|
| 209 |
+
"eval_runtime": 3.2229,
|
| 210 |
+
"eval_samples_per_second": 136.521,
|
| 211 |
+
"eval_steps_per_second": 2.172,
|
| 212 |
+
"step": 168
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 12.14,
|
| 216 |
+
"learning_rate": 4.880952380952381e-05,
|
| 217 |
+
"loss": 1.5891,
|
| 218 |
+
"step": 170
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 12.86,
|
| 222 |
+
"learning_rate": 4.841269841269841e-05,
|
| 223 |
+
"loss": 1.5673,
|
| 224 |
+
"step": 180
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 13.0,
|
| 228 |
+
"eval_accuracy": 0.41363636363636364,
|
| 229 |
+
"eval_loss": 1.516886591911316,
|
| 230 |
+
"eval_runtime": 3.2733,
|
| 231 |
+
"eval_samples_per_second": 134.423,
|
| 232 |
+
"eval_steps_per_second": 2.139,
|
| 233 |
+
"step": 182
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 13.57,
|
| 237 |
+
"learning_rate": 4.801587301587302e-05,
|
| 238 |
+
"loss": 1.5239,
|
| 239 |
+
"step": 190
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 14.0,
|
| 243 |
+
"eval_accuracy": 0.509090909090909,
|
| 244 |
+
"eval_loss": 1.4703104496002197,
|
| 245 |
+
"eval_runtime": 3.2812,
|
| 246 |
+
"eval_samples_per_second": 134.096,
|
| 247 |
+
"eval_steps_per_second": 2.133,
|
| 248 |
+
"step": 196
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 14.29,
|
| 252 |
+
"learning_rate": 4.761904761904762e-05,
|
| 253 |
+
"loss": 1.5131,
|
| 254 |
+
"step": 200
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 15.0,
|
| 258 |
+
"learning_rate": 4.722222222222222e-05,
|
| 259 |
+
"loss": 1.4698,
|
| 260 |
+
"step": 210
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 15.0,
|
| 264 |
+
"eval_accuracy": 0.5363636363636364,
|
| 265 |
+
"eval_loss": 1.4071372747421265,
|
| 266 |
+
"eval_runtime": 3.3595,
|
| 267 |
+
"eval_samples_per_second": 130.973,
|
| 268 |
+
"eval_steps_per_second": 2.084,
|
| 269 |
+
"step": 210
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 15.71,
|
| 273 |
+
"learning_rate": 4.682539682539683e-05,
|
| 274 |
+
"loss": 1.4277,
|
| 275 |
+
"step": 220
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 16.0,
|
| 279 |
+
"eval_accuracy": 0.5886363636363636,
|
| 280 |
+
"eval_loss": 1.3305206298828125,
|
| 281 |
+
"eval_runtime": 3.3151,
|
| 282 |
+
"eval_samples_per_second": 132.725,
|
| 283 |
+
"eval_steps_per_second": 2.112,
|
| 284 |
+
"step": 224
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 16.43,
|
| 288 |
+
"learning_rate": 4.642857142857143e-05,
|
| 289 |
+
"loss": 1.3798,
|
| 290 |
+
"step": 230
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 17.0,
|
| 294 |
+
"eval_accuracy": 0.6068181818181818,
|
| 295 |
+
"eval_loss": 1.250695824623108,
|
| 296 |
+
"eval_runtime": 3.348,
|
| 297 |
+
"eval_samples_per_second": 131.42,
|
| 298 |
+
"eval_steps_per_second": 2.091,
|
| 299 |
+
"step": 238
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 17.14,
|
| 303 |
+
"learning_rate": 4.603174603174603e-05,
|
| 304 |
+
"loss": 1.355,
|
| 305 |
+
"step": 240
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 17.86,
|
| 309 |
+
"learning_rate": 4.563492063492064e-05,
|
| 310 |
+
"loss": 1.2858,
|
| 311 |
+
"step": 250
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 18.0,
|
| 315 |
+
"eval_accuracy": 0.6409090909090909,
|
| 316 |
+
"eval_loss": 1.1516846418380737,
|
| 317 |
+
"eval_runtime": 3.35,
|
| 318 |
+
"eval_samples_per_second": 131.342,
|
| 319 |
+
"eval_steps_per_second": 2.09,
|
| 320 |
+
"step": 252
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 18.57,
|
| 324 |
+
"learning_rate": 4.523809523809524e-05,
|
| 325 |
+
"loss": 1.2543,
|
| 326 |
+
"step": 260
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 19.0,
|
| 330 |
+
"eval_accuracy": 0.6659090909090909,
|
| 331 |
+
"eval_loss": 1.0930873155593872,
|
| 332 |
+
"eval_runtime": 3.4597,
|
| 333 |
+
"eval_samples_per_second": 127.177,
|
| 334 |
+
"eval_steps_per_second": 2.023,
|
| 335 |
+
"step": 266
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 19.29,
|
| 339 |
+
"learning_rate": 4.4841269841269846e-05,
|
| 340 |
+
"loss": 1.2006,
|
| 341 |
+
"step": 270
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 20.0,
|
| 345 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 346 |
+
"loss": 1.1696,
|
| 347 |
+
"step": 280
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 20.0,
|
| 351 |
+
"eval_accuracy": 0.675,
|
| 352 |
+
"eval_loss": 1.012363314628601,
|
| 353 |
+
"eval_runtime": 3.3077,
|
| 354 |
+
"eval_samples_per_second": 133.023,
|
| 355 |
+
"eval_steps_per_second": 2.116,
|
| 356 |
+
"step": 280
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 20.71,
|
| 360 |
+
"learning_rate": 4.404761904761905e-05,
|
| 361 |
+
"loss": 1.1124,
|
| 362 |
+
"step": 290
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 21.0,
|
| 366 |
+
"eval_accuracy": 0.6931818181818182,
|
| 367 |
+
"eval_loss": 0.954047679901123,
|
| 368 |
+
"eval_runtime": 3.3411,
|
| 369 |
+
"eval_samples_per_second": 131.693,
|
| 370 |
+
"eval_steps_per_second": 2.095,
|
| 371 |
+
"step": 294
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 21.43,
|
| 375 |
+
"learning_rate": 4.3650793650793655e-05,
|
| 376 |
+
"loss": 1.059,
|
| 377 |
+
"step": 300
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 22.0,
|
| 381 |
+
"eval_accuracy": 0.7022727272727273,
|
| 382 |
+
"eval_loss": 0.9010851979255676,
|
| 383 |
+
"eval_runtime": 3.6911,
|
| 384 |
+
"eval_samples_per_second": 119.205,
|
| 385 |
+
"eval_steps_per_second": 1.896,
|
| 386 |
+
"step": 308
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 22.14,
|
| 390 |
+
"learning_rate": 4.3253968253968256e-05,
|
| 391 |
+
"loss": 1.0318,
|
| 392 |
+
"step": 310
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 22.86,
|
| 396 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 397 |
+
"loss": 1.0006,
|
| 398 |
+
"step": 320
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 23.0,
|
| 402 |
+
"eval_accuracy": 0.7090909090909091,
|
| 403 |
+
"eval_loss": 0.8692610859870911,
|
| 404 |
+
"eval_runtime": 3.4254,
|
| 405 |
+
"eval_samples_per_second": 128.451,
|
| 406 |
+
"eval_steps_per_second": 2.044,
|
| 407 |
+
"step": 322
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 23.57,
|
| 411 |
+
"learning_rate": 4.2460317460317464e-05,
|
| 412 |
+
"loss": 0.9606,
|
| 413 |
+
"step": 330
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 24.0,
|
| 417 |
+
"eval_accuracy": 0.6977272727272728,
|
| 418 |
+
"eval_loss": 0.8458234667778015,
|
| 419 |
+
"eval_runtime": 3.7789,
|
| 420 |
+
"eval_samples_per_second": 116.436,
|
| 421 |
+
"eval_steps_per_second": 1.852,
|
| 422 |
+
"step": 336
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 24.29,
|
| 426 |
+
"learning_rate": 4.2063492063492065e-05,
|
| 427 |
+
"loss": 0.9541,
|
| 428 |
+
"step": 340
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 25.0,
|
| 432 |
+
"learning_rate": 4.166666666666667e-05,
|
| 433 |
+
"loss": 0.9146,
|
| 434 |
+
"step": 350
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 25.0,
|
| 438 |
+
"eval_accuracy": 0.7227272727272728,
|
| 439 |
+
"eval_loss": 0.8226051926612854,
|
| 440 |
+
"eval_runtime": 3.375,
|
| 441 |
+
"eval_samples_per_second": 130.371,
|
| 442 |
+
"eval_steps_per_second": 2.074,
|
| 443 |
+
"step": 350
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 25.71,
|
| 447 |
+
"learning_rate": 4.126984126984127e-05,
|
| 448 |
+
"loss": 0.8864,
|
| 449 |
+
"step": 360
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 26.0,
|
| 453 |
+
"eval_accuracy": 0.7159090909090909,
|
| 454 |
+
"eval_loss": 0.8187159895896912,
|
| 455 |
+
"eval_runtime": 3.4251,
|
| 456 |
+
"eval_samples_per_second": 128.462,
|
| 457 |
+
"eval_steps_per_second": 2.044,
|
| 458 |
+
"step": 364
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 26.43,
|
| 462 |
+
"learning_rate": 4.0873015873015874e-05,
|
| 463 |
+
"loss": 0.8864,
|
| 464 |
+
"step": 370
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 27.0,
|
| 468 |
+
"eval_accuracy": 0.7227272727272728,
|
| 469 |
+
"eval_loss": 0.7862333655357361,
|
| 470 |
+
"eval_runtime": 3.3921,
|
| 471 |
+
"eval_samples_per_second": 129.714,
|
| 472 |
+
"eval_steps_per_second": 2.064,
|
| 473 |
+
"step": 378
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 27.14,
|
| 477 |
+
"learning_rate": 4.047619047619048e-05,
|
| 478 |
+
"loss": 0.8447,
|
| 479 |
+
"step": 380
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 27.86,
|
| 483 |
+
"learning_rate": 4.007936507936508e-05,
|
| 484 |
+
"loss": 0.8458,
|
| 485 |
+
"step": 390
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 28.0,
|
| 489 |
+
"eval_accuracy": 0.7295454545454545,
|
| 490 |
+
"eval_loss": 0.7686573266983032,
|
| 491 |
+
"eval_runtime": 3.4126,
|
| 492 |
+
"eval_samples_per_second": 128.935,
|
| 493 |
+
"eval_steps_per_second": 2.051,
|
| 494 |
+
"step": 392
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 28.57,
|
| 498 |
+
"learning_rate": 3.968253968253968e-05,
|
| 499 |
+
"loss": 0.8337,
|
| 500 |
+
"step": 400
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 29.0,
|
| 504 |
+
"eval_accuracy": 0.7272727272727273,
|
| 505 |
+
"eval_loss": 0.747409999370575,
|
| 506 |
+
"eval_runtime": 3.4047,
|
| 507 |
+
"eval_samples_per_second": 129.233,
|
| 508 |
+
"eval_steps_per_second": 2.056,
|
| 509 |
+
"step": 406
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 29.29,
|
| 513 |
+
"learning_rate": 3.928571428571429e-05,
|
| 514 |
+
"loss": 0.8107,
|
| 515 |
+
"step": 410
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 30.0,
|
| 519 |
+
"learning_rate": 3.888888888888889e-05,
|
| 520 |
+
"loss": 0.7943,
|
| 521 |
+
"step": 420
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 30.0,
|
| 525 |
+
"eval_accuracy": 0.7454545454545455,
|
| 526 |
+
"eval_loss": 0.7368342876434326,
|
| 527 |
+
"eval_runtime": 3.4657,
|
| 528 |
+
"eval_samples_per_second": 126.957,
|
| 529 |
+
"eval_steps_per_second": 2.02,
|
| 530 |
+
"step": 420
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 30.71,
|
| 534 |
+
"learning_rate": 3.84920634920635e-05,
|
| 535 |
+
"loss": 0.7884,
|
| 536 |
+
"step": 430
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 31.0,
|
| 540 |
+
"eval_accuracy": 0.7454545454545455,
|
| 541 |
+
"eval_loss": 0.7331739068031311,
|
| 542 |
+
"eval_runtime": 3.4308,
|
| 543 |
+
"eval_samples_per_second": 128.249,
|
| 544 |
+
"eval_steps_per_second": 2.04,
|
| 545 |
+
"step": 434
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 31.43,
|
| 549 |
+
"learning_rate": 3.809523809523809e-05,
|
| 550 |
+
"loss": 0.7495,
|
| 551 |
+
"step": 440
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 32.0,
|
| 555 |
+
"eval_accuracy": 0.725,
|
| 556 |
+
"eval_loss": 0.7354092597961426,
|
| 557 |
+
"eval_runtime": 3.3899,
|
| 558 |
+
"eval_samples_per_second": 129.796,
|
| 559 |
+
"eval_steps_per_second": 2.065,
|
| 560 |
+
"step": 448
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 32.14,
|
| 564 |
+
"learning_rate": 3.76984126984127e-05,
|
| 565 |
+
"loss": 0.78,
|
| 566 |
+
"step": 450
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 32.86,
|
| 570 |
+
"learning_rate": 3.730158730158731e-05,
|
| 571 |
+
"loss": 0.7503,
|
| 572 |
+
"step": 460
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 33.0,
|
| 576 |
+
"eval_accuracy": 0.7477272727272727,
|
| 577 |
+
"eval_loss": 0.7180919051170349,
|
| 578 |
+
"eval_runtime": 3.3351,
|
| 579 |
+
"eval_samples_per_second": 131.931,
|
| 580 |
+
"eval_steps_per_second": 2.099,
|
| 581 |
+
"step": 462
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 33.57,
|
| 585 |
+
"learning_rate": 3.690476190476191e-05,
|
| 586 |
+
"loss": 0.7358,
|
| 587 |
+
"step": 470
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 34.0,
|
| 591 |
+
"eval_accuracy": 0.7477272727272727,
|
| 592 |
+
"eval_loss": 0.7007948160171509,
|
| 593 |
+
"eval_runtime": 3.4029,
|
| 594 |
+
"eval_samples_per_second": 129.301,
|
| 595 |
+
"eval_steps_per_second": 2.057,
|
| 596 |
+
"step": 476
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 34.29,
|
| 600 |
+
"learning_rate": 3.650793650793651e-05,
|
| 601 |
+
"loss": 0.7605,
|
| 602 |
+
"step": 480
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 35.0,
|
| 606 |
+
"learning_rate": 3.611111111111111e-05,
|
| 607 |
+
"loss": 0.7229,
|
| 608 |
+
"step": 490
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 35.0,
|
| 612 |
+
"eval_accuracy": 0.740909090909091,
|
| 613 |
+
"eval_loss": 0.6959581971168518,
|
| 614 |
+
"eval_runtime": 3.6574,
|
| 615 |
+
"eval_samples_per_second": 120.304,
|
| 616 |
+
"eval_steps_per_second": 1.914,
|
| 617 |
+
"step": 490
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 35.71,
|
| 621 |
+
"learning_rate": 3.571428571428572e-05,
|
| 622 |
+
"loss": 0.7114,
|
| 623 |
+
"step": 500
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 36.0,
|
| 627 |
+
"eval_accuracy": 0.7431818181818182,
|
| 628 |
+
"eval_loss": 0.6837091445922852,
|
| 629 |
+
"eval_runtime": 3.6041,
|
| 630 |
+
"eval_samples_per_second": 122.083,
|
| 631 |
+
"eval_steps_per_second": 1.942,
|
| 632 |
+
"step": 504
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 36.43,
|
| 636 |
+
"learning_rate": 3.5317460317460324e-05,
|
| 637 |
+
"loss": 0.7075,
|
| 638 |
+
"step": 510
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 37.0,
|
| 642 |
+
"eval_accuracy": 0.740909090909091,
|
| 643 |
+
"eval_loss": 0.6922788023948669,
|
| 644 |
+
"eval_runtime": 3.3849,
|
| 645 |
+
"eval_samples_per_second": 129.987,
|
| 646 |
+
"eval_steps_per_second": 2.068,
|
| 647 |
+
"step": 518
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 37.14,
|
| 651 |
+
"learning_rate": 3.492063492063492e-05,
|
| 652 |
+
"loss": 0.6915,
|
| 653 |
+
"step": 520
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 37.86,
|
| 657 |
+
"learning_rate": 3.4523809523809526e-05,
|
| 658 |
+
"loss": 0.682,
|
| 659 |
+
"step": 530
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 38.0,
|
| 663 |
+
"eval_accuracy": 0.7431818181818182,
|
| 664 |
+
"eval_loss": 0.6930937767028809,
|
| 665 |
+
"eval_runtime": 3.4683,
|
| 666 |
+
"eval_samples_per_second": 126.864,
|
| 667 |
+
"eval_steps_per_second": 2.018,
|
| 668 |
+
"step": 532
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 38.57,
|
| 672 |
+
"learning_rate": 3.412698412698413e-05,
|
| 673 |
+
"loss": 0.6626,
|
| 674 |
+
"step": 540
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 39.0,
|
| 678 |
+
"eval_accuracy": 0.740909090909091,
|
| 679 |
+
"eval_loss": 0.6872156262397766,
|
| 680 |
+
"eval_runtime": 3.3949,
|
| 681 |
+
"eval_samples_per_second": 129.605,
|
| 682 |
+
"eval_steps_per_second": 2.062,
|
| 683 |
+
"step": 546
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 39.29,
|
| 687 |
+
"learning_rate": 3.3730158730158734e-05,
|
| 688 |
+
"loss": 0.6708,
|
| 689 |
+
"step": 550
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 40.0,
|
| 693 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 694 |
+
"loss": 0.6688,
|
| 695 |
+
"step": 560
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 40.0,
|
| 699 |
+
"eval_accuracy": 0.75,
|
| 700 |
+
"eval_loss": 0.6758830547332764,
|
| 701 |
+
"eval_runtime": 3.2304,
|
| 702 |
+
"eval_samples_per_second": 136.208,
|
| 703 |
+
"eval_steps_per_second": 2.167,
|
| 704 |
+
"step": 560
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 40.71,
|
| 708 |
+
"learning_rate": 3.2936507936507936e-05,
|
| 709 |
+
"loss": 0.655,
|
| 710 |
+
"step": 570
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 41.0,
|
| 714 |
+
"eval_accuracy": 0.7340909090909091,
|
| 715 |
+
"eval_loss": 0.7006358504295349,
|
| 716 |
+
"eval_runtime": 3.3002,
|
| 717 |
+
"eval_samples_per_second": 133.326,
|
| 718 |
+
"eval_steps_per_second": 2.121,
|
| 719 |
+
"step": 574
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 41.43,
|
| 723 |
+
"learning_rate": 3.253968253968254e-05,
|
| 724 |
+
"loss": 0.6533,
|
| 725 |
+
"step": 580
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 42.0,
|
| 729 |
+
"eval_accuracy": 0.75,
|
| 730 |
+
"eval_loss": 0.6774057149887085,
|
| 731 |
+
"eval_runtime": 3.351,
|
| 732 |
+
"eval_samples_per_second": 131.303,
|
| 733 |
+
"eval_steps_per_second": 2.089,
|
| 734 |
+
"step": 588
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 42.14,
|
| 738 |
+
"learning_rate": 3.2142857142857144e-05,
|
| 739 |
+
"loss": 0.6488,
|
| 740 |
+
"step": 590
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 42.86,
|
| 744 |
+
"learning_rate": 3.1746031746031745e-05,
|
| 745 |
+
"loss": 0.6215,
|
| 746 |
+
"step": 600
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 43.0,
|
| 750 |
+
"eval_accuracy": 0.7454545454545455,
|
| 751 |
+
"eval_loss": 0.6663152575492859,
|
| 752 |
+
"eval_runtime": 3.4833,
|
| 753 |
+
"eval_samples_per_second": 126.317,
|
| 754 |
+
"eval_steps_per_second": 2.01,
|
| 755 |
+
"step": 602
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 43.57,
|
| 759 |
+
"learning_rate": 3.134920634920635e-05,
|
| 760 |
+
"loss": 0.6267,
|
| 761 |
+
"step": 610
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 44.0,
|
| 765 |
+
"eval_accuracy": 0.7545454545454545,
|
| 766 |
+
"eval_loss": 0.6608263850212097,
|
| 767 |
+
"eval_runtime": 3.375,
|
| 768 |
+
"eval_samples_per_second": 130.371,
|
| 769 |
+
"eval_steps_per_second": 2.074,
|
| 770 |
+
"step": 616
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 44.29,
|
| 774 |
+
"learning_rate": 3.095238095238095e-05,
|
| 775 |
+
"loss": 0.6383,
|
| 776 |
+
"step": 620
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 45.0,
|
| 780 |
+
"learning_rate": 3.055555555555556e-05,
|
| 781 |
+
"loss": 0.6122,
|
| 782 |
+
"step": 630
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 45.0,
|
| 786 |
+
"eval_accuracy": 0.7477272727272727,
|
| 787 |
+
"eval_loss": 0.6523904800415039,
|
| 788 |
+
"eval_runtime": 3.3102,
|
| 789 |
+
"eval_samples_per_second": 132.923,
|
| 790 |
+
"eval_steps_per_second": 2.115,
|
| 791 |
+
"step": 630
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 45.71,
|
| 795 |
+
"learning_rate": 3.0158730158730158e-05,
|
| 796 |
+
"loss": 0.6047,
|
| 797 |
+
"step": 640
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 46.0,
|
| 801 |
+
"eval_accuracy": 0.7477272727272727,
|
| 802 |
+
"eval_loss": 0.6757220029830933,
|
| 803 |
+
"eval_runtime": 3.3012,
|
| 804 |
+
"eval_samples_per_second": 133.286,
|
| 805 |
+
"eval_steps_per_second": 2.12,
|
| 806 |
+
"step": 644
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 46.43,
|
| 810 |
+
"learning_rate": 2.9761904761904762e-05,
|
| 811 |
+
"loss": 0.6135,
|
| 812 |
+
"step": 650
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 47.0,
|
| 816 |
+
"eval_accuracy": 0.7431818181818182,
|
| 817 |
+
"eval_loss": 0.6492685675621033,
|
| 818 |
+
"eval_runtime": 3.4379,
|
| 819 |
+
"eval_samples_per_second": 127.986,
|
| 820 |
+
"eval_steps_per_second": 2.036,
|
| 821 |
+
"step": 658
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 47.14,
|
| 825 |
+
"learning_rate": 2.9365079365079366e-05,
|
| 826 |
+
"loss": 0.6145,
|
| 827 |
+
"step": 660
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 47.86,
|
| 831 |
+
"learning_rate": 2.8968253968253974e-05,
|
| 832 |
+
"loss": 0.5805,
|
| 833 |
+
"step": 670
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 48.0,
|
| 837 |
+
"eval_accuracy": 0.75,
|
| 838 |
+
"eval_loss": 0.6503338813781738,
|
| 839 |
+
"eval_runtime": 3.383,
|
| 840 |
+
"eval_samples_per_second": 130.064,
|
| 841 |
+
"eval_steps_per_second": 2.069,
|
| 842 |
+
"step": 672
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 48.57,
|
| 846 |
+
"learning_rate": 2.857142857142857e-05,
|
| 847 |
+
"loss": 0.6124,
|
| 848 |
+
"step": 680
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 49.0,
|
| 852 |
+
"eval_accuracy": 0.7454545454545455,
|
| 853 |
+
"eval_loss": 0.6626009941101074,
|
| 854 |
+
"eval_runtime": 3.5236,
|
| 855 |
+
"eval_samples_per_second": 124.873,
|
| 856 |
+
"eval_steps_per_second": 1.987,
|
| 857 |
+
"step": 686
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 49.29,
|
| 861 |
+
"learning_rate": 2.8174603174603175e-05,
|
| 862 |
+
"loss": 0.6079,
|
| 863 |
+
"step": 690
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 50.0,
|
| 867 |
+
"learning_rate": 2.777777777777778e-05,
|
| 868 |
+
"loss": 0.5826,
|
| 869 |
+
"step": 700
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 50.0,
|
| 873 |
+
"eval_accuracy": 0.7522727272727273,
|
| 874 |
+
"eval_loss": 0.6511955857276917,
|
| 875 |
+
"eval_runtime": 3.2723,
|
| 876 |
+
"eval_samples_per_second": 134.464,
|
| 877 |
+
"eval_steps_per_second": 2.139,
|
| 878 |
+
"step": 700
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 50.71,
|
| 882 |
+
"learning_rate": 2.7380952380952383e-05,
|
| 883 |
+
"loss": 0.5698,
|
| 884 |
+
"step": 710
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 51.0,
|
| 888 |
+
"eval_accuracy": 0.7477272727272727,
|
| 889 |
+
"eval_loss": 0.6589775085449219,
|
| 890 |
+
"eval_runtime": 3.4253,
|
| 891 |
+
"eval_samples_per_second": 128.455,
|
| 892 |
+
"eval_steps_per_second": 2.044,
|
| 893 |
+
"step": 714
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 51.43,
|
| 897 |
+
"learning_rate": 2.6984126984126984e-05,
|
| 898 |
+
"loss": 0.5721,
|
| 899 |
+
"step": 720
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 52.0,
|
| 903 |
+
"eval_accuracy": 0.7454545454545455,
|
| 904 |
+
"eval_loss": 0.6558127403259277,
|
| 905 |
+
"eval_runtime": 3.2646,
|
| 906 |
+
"eval_samples_per_second": 134.777,
|
| 907 |
+
"eval_steps_per_second": 2.144,
|
| 908 |
+
"step": 728
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 52.14,
|
| 912 |
+
"learning_rate": 2.6587301587301588e-05,
|
| 913 |
+
"loss": 0.5912,
|
| 914 |
+
"step": 730
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 52.86,
|
| 918 |
+
"learning_rate": 2.6190476190476192e-05,
|
| 919 |
+
"loss": 0.5569,
|
| 920 |
+
"step": 740
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 53.0,
|
| 924 |
+
"eval_accuracy": 0.7454545454545455,
|
| 925 |
+
"eval_loss": 0.6654322743415833,
|
| 926 |
+
"eval_runtime": 3.2932,
|
| 927 |
+
"eval_samples_per_second": 133.609,
|
| 928 |
+
"eval_steps_per_second": 2.126,
|
| 929 |
+
"step": 742
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 53.57,
|
| 933 |
+
"learning_rate": 2.5793650793650796e-05,
|
| 934 |
+
"loss": 0.573,
|
| 935 |
+
"step": 750
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 54.0,
|
| 939 |
+
"eval_accuracy": 0.7545454545454545,
|
| 940 |
+
"eval_loss": 0.636202335357666,
|
| 941 |
+
"eval_runtime": 3.3271,
|
| 942 |
+
"eval_samples_per_second": 132.247,
|
| 943 |
+
"eval_steps_per_second": 2.104,
|
| 944 |
+
"step": 756
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 54.29,
|
| 948 |
+
"learning_rate": 2.5396825396825397e-05,
|
| 949 |
+
"loss": 0.5816,
|
| 950 |
+
"step": 760
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 55.0,
|
| 954 |
+
"learning_rate": 2.5e-05,
|
| 955 |
+
"loss": 0.543,
|
| 956 |
+
"step": 770
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 55.0,
|
| 960 |
+
"eval_accuracy": 0.7636363636363637,
|
| 961 |
+
"eval_loss": 0.6368376016616821,
|
| 962 |
+
"eval_runtime": 3.2054,
|
| 963 |
+
"eval_samples_per_second": 137.267,
|
| 964 |
+
"eval_steps_per_second": 2.184,
|
| 965 |
+
"step": 770
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 55.71,
|
| 969 |
+
"learning_rate": 2.4603174603174602e-05,
|
| 970 |
+
"loss": 0.5494,
|
| 971 |
+
"step": 780
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 56.0,
|
| 975 |
+
"eval_accuracy": 0.759090909090909,
|
| 976 |
+
"eval_loss": 0.6263792514801025,
|
| 977 |
+
"eval_runtime": 3.2663,
|
| 978 |
+
"eval_samples_per_second": 134.71,
|
| 979 |
+
"eval_steps_per_second": 2.143,
|
| 980 |
+
"step": 784
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 56.43,
|
| 984 |
+
"learning_rate": 2.4206349206349206e-05,
|
| 985 |
+
"loss": 0.5598,
|
| 986 |
+
"step": 790
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 57.0,
|
| 990 |
+
"eval_accuracy": 0.7681818181818182,
|
| 991 |
+
"eval_loss": 0.627074658870697,
|
| 992 |
+
"eval_runtime": 3.4193,
|
| 993 |
+
"eval_samples_per_second": 128.683,
|
| 994 |
+
"eval_steps_per_second": 2.047,
|
| 995 |
+
"step": 798
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 57.14,
|
| 999 |
+
"learning_rate": 2.380952380952381e-05,
|
| 1000 |
+
"loss": 0.5655,
|
| 1001 |
+
"step": 800
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 57.86,
|
| 1005 |
+
"learning_rate": 2.3412698412698414e-05,
|
| 1006 |
+
"loss": 0.5412,
|
| 1007 |
+
"step": 810
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 58.0,
|
| 1011 |
+
"eval_accuracy": 0.7522727272727273,
|
| 1012 |
+
"eval_loss": 0.6506673693656921,
|
| 1013 |
+
"eval_runtime": 3.2684,
|
| 1014 |
+
"eval_samples_per_second": 134.623,
|
| 1015 |
+
"eval_steps_per_second": 2.142,
|
| 1016 |
+
"step": 812
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 58.57,
|
| 1020 |
+
"learning_rate": 2.3015873015873015e-05,
|
| 1021 |
+
"loss": 0.5512,
|
| 1022 |
+
"step": 820
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 59.0,
|
| 1026 |
+
"eval_accuracy": 0.7568181818181818,
|
| 1027 |
+
"eval_loss": 0.6446354985237122,
|
| 1028 |
+
"eval_runtime": 3.3303,
|
| 1029 |
+
"eval_samples_per_second": 132.119,
|
| 1030 |
+
"eval_steps_per_second": 2.102,
|
| 1031 |
+
"step": 826
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 59.29,
|
| 1035 |
+
"learning_rate": 2.261904761904762e-05,
|
| 1036 |
+
"loss": 0.5217,
|
| 1037 |
+
"step": 830
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 60.0,
|
| 1041 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 1042 |
+
"loss": 0.5504,
|
| 1043 |
+
"step": 840
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 60.0,
|
| 1047 |
+
"eval_accuracy": 0.7727272727272727,
|
| 1048 |
+
"eval_loss": 0.6208813190460205,
|
| 1049 |
+
"eval_runtime": 3.2503,
|
| 1050 |
+
"eval_samples_per_second": 135.372,
|
| 1051 |
+
"eval_steps_per_second": 2.154,
|
| 1052 |
+
"step": 840
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 60.71,
|
| 1056 |
+
"learning_rate": 2.1825396825396827e-05,
|
| 1057 |
+
"loss": 0.5304,
|
| 1058 |
+
"step": 850
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 61.0,
|
| 1062 |
+
"eval_accuracy": 0.7568181818181818,
|
| 1063 |
+
"eval_loss": 0.6428102850914001,
|
| 1064 |
+
"eval_runtime": 3.4103,
|
| 1065 |
+
"eval_samples_per_second": 129.021,
|
| 1066 |
+
"eval_steps_per_second": 2.053,
|
| 1067 |
+
"step": 854
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 61.43,
|
| 1071 |
+
"learning_rate": 2.1428571428571428e-05,
|
| 1072 |
+
"loss": 0.5238,
|
| 1073 |
+
"step": 860
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 62.0,
|
| 1077 |
+
"eval_accuracy": 0.7613636363636364,
|
| 1078 |
+
"eval_loss": 0.6323724985122681,
|
| 1079 |
+
"eval_runtime": 3.2991,
|
| 1080 |
+
"eval_samples_per_second": 133.368,
|
| 1081 |
+
"eval_steps_per_second": 2.122,
|
| 1082 |
+
"step": 868
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 62.14,
|
| 1086 |
+
"learning_rate": 2.1031746031746032e-05,
|
| 1087 |
+
"loss": 0.5344,
|
| 1088 |
+
"step": 870
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 62.86,
|
| 1092 |
+
"learning_rate": 2.0634920634920636e-05,
|
| 1093 |
+
"loss": 0.5298,
|
| 1094 |
+
"step": 880
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 63.0,
|
| 1098 |
+
"eval_accuracy": 0.7568181818181818,
|
| 1099 |
+
"eval_loss": 0.6347512006759644,
|
| 1100 |
+
"eval_runtime": 3.2796,
|
| 1101 |
+
"eval_samples_per_second": 134.162,
|
| 1102 |
+
"eval_steps_per_second": 2.134,
|
| 1103 |
+
"step": 882
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 63.57,
|
| 1107 |
+
"learning_rate": 2.023809523809524e-05,
|
| 1108 |
+
"loss": 0.4965,
|
| 1109 |
+
"step": 890
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 64.0,
|
| 1113 |
+
"eval_accuracy": 0.7545454545454545,
|
| 1114 |
+
"eval_loss": 0.6351690292358398,
|
| 1115 |
+
"eval_runtime": 3.394,
|
| 1116 |
+
"eval_samples_per_second": 129.639,
|
| 1117 |
+
"eval_steps_per_second": 2.062,
|
| 1118 |
+
"step": 896
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 64.29,
|
| 1122 |
+
"learning_rate": 1.984126984126984e-05,
|
| 1123 |
+
"loss": 0.5338,
|
| 1124 |
+
"step": 900
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 65.0,
|
| 1128 |
+
"learning_rate": 1.9444444444444445e-05,
|
| 1129 |
+
"loss": 0.5188,
|
| 1130 |
+
"step": 910
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 65.0,
|
| 1134 |
+
"eval_accuracy": 0.7704545454545455,
|
| 1135 |
+
"eval_loss": 0.6352004408836365,
|
| 1136 |
+
"eval_runtime": 3.3712,
|
| 1137 |
+
"eval_samples_per_second": 130.519,
|
| 1138 |
+
"eval_steps_per_second": 2.076,
|
| 1139 |
+
"step": 910
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 65.71,
|
| 1143 |
+
"learning_rate": 1.9047619047619046e-05,
|
| 1144 |
+
"loss": 0.5081,
|
| 1145 |
+
"step": 920
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 66.0,
|
| 1149 |
+
"eval_accuracy": 0.7681818181818182,
|
| 1150 |
+
"eval_loss": 0.6369162201881409,
|
| 1151 |
+
"eval_runtime": 3.3987,
|
| 1152 |
+
"eval_samples_per_second": 129.462,
|
| 1153 |
+
"eval_steps_per_second": 2.06,
|
| 1154 |
+
"step": 924
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 66.43,
|
| 1158 |
+
"learning_rate": 1.8650793650793654e-05,
|
| 1159 |
+
"loss": 0.4725,
|
| 1160 |
+
"step": 930
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 67.0,
|
| 1164 |
+
"eval_accuracy": 0.775,
|
| 1165 |
+
"eval_loss": 0.6235827207565308,
|
| 1166 |
+
"eval_runtime": 3.2862,
|
| 1167 |
+
"eval_samples_per_second": 133.893,
|
| 1168 |
+
"eval_steps_per_second": 2.13,
|
| 1169 |
+
"step": 938
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 67.14,
|
| 1173 |
+
"learning_rate": 1.8253968253968254e-05,
|
| 1174 |
+
"loss": 0.52,
|
| 1175 |
+
"step": 940
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 67.86,
|
| 1179 |
+
"learning_rate": 1.785714285714286e-05,
|
| 1180 |
+
"loss": 0.4977,
|
| 1181 |
+
"step": 950
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 68.0,
|
| 1185 |
+
"eval_accuracy": 0.7681818181818182,
|
| 1186 |
+
"eval_loss": 0.621067225933075,
|
| 1187 |
+
"eval_runtime": 3.3453,
|
| 1188 |
+
"eval_samples_per_second": 131.527,
|
| 1189 |
+
"eval_steps_per_second": 2.092,
|
| 1190 |
+
"step": 952
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 68.57,
|
| 1194 |
+
"learning_rate": 1.746031746031746e-05,
|
| 1195 |
+
"loss": 0.5118,
|
| 1196 |
+
"step": 960
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 69.0,
|
| 1200 |
+
"eval_accuracy": 0.7568181818181818,
|
| 1201 |
+
"eval_loss": 0.6289528012275696,
|
| 1202 |
+
"eval_runtime": 3.3375,
|
| 1203 |
+
"eval_samples_per_second": 131.834,
|
| 1204 |
+
"eval_steps_per_second": 2.097,
|
| 1205 |
+
"step": 966
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 69.29,
|
| 1209 |
+
"learning_rate": 1.7063492063492063e-05,
|
| 1210 |
+
"loss": 0.5001,
|
| 1211 |
+
"step": 970
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 70.0,
|
| 1215 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 1216 |
+
"loss": 0.4876,
|
| 1217 |
+
"step": 980
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 70.0,
|
| 1221 |
+
"eval_accuracy": 0.7636363636363637,
|
| 1222 |
+
"eval_loss": 0.6366029381752014,
|
| 1223 |
+
"eval_runtime": 3.2398,
|
| 1224 |
+
"eval_samples_per_second": 135.809,
|
| 1225 |
+
"eval_steps_per_second": 2.161,
|
| 1226 |
+
"step": 980
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 70.71,
|
| 1230 |
+
"learning_rate": 1.626984126984127e-05,
|
| 1231 |
+
"loss": 0.49,
|
| 1232 |
+
"step": 990
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 71.0,
|
| 1236 |
+
"eval_accuracy": 0.7727272727272727,
|
| 1237 |
+
"eval_loss": 0.6185997128486633,
|
| 1238 |
+
"eval_runtime": 3.2603,
|
| 1239 |
+
"eval_samples_per_second": 134.958,
|
| 1240 |
+
"eval_steps_per_second": 2.147,
|
| 1241 |
+
"step": 994
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 71.43,
|
| 1245 |
+
"learning_rate": 1.5873015873015872e-05,
|
| 1246 |
+
"loss": 0.4912,
|
| 1247 |
+
"step": 1000
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 72.0,
|
| 1251 |
+
"eval_accuracy": 0.7727272727272727,
|
| 1252 |
+
"eval_loss": 0.6244505643844604,
|
| 1253 |
+
"eval_runtime": 3.3698,
|
| 1254 |
+
"eval_samples_per_second": 130.572,
|
| 1255 |
+
"eval_steps_per_second": 2.077,
|
| 1256 |
+
"step": 1008
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 72.14,
|
| 1260 |
+
"learning_rate": 1.5476190476190476e-05,
|
| 1261 |
+
"loss": 0.4971,
|
| 1262 |
+
"step": 1010
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 72.86,
|
| 1266 |
+
"learning_rate": 1.5079365079365079e-05,
|
| 1267 |
+
"loss": 0.4774,
|
| 1268 |
+
"step": 1020
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 73.0,
|
| 1272 |
+
"eval_accuracy": 0.7727272727272727,
|
| 1273 |
+
"eval_loss": 0.6223783493041992,
|
| 1274 |
+
"eval_runtime": 3.4233,
|
| 1275 |
+
"eval_samples_per_second": 128.533,
|
| 1276 |
+
"eval_steps_per_second": 2.045,
|
| 1277 |
+
"step": 1022
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 73.57,
|
| 1281 |
+
"learning_rate": 1.4682539682539683e-05,
|
| 1282 |
+
"loss": 0.4972,
|
| 1283 |
+
"step": 1030
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 74.0,
|
| 1287 |
+
"eval_accuracy": 0.7454545454545455,
|
| 1288 |
+
"eval_loss": 0.6407915353775024,
|
| 1289 |
+
"eval_runtime": 3.2294,
|
| 1290 |
+
"eval_samples_per_second": 136.25,
|
| 1291 |
+
"eval_steps_per_second": 2.168,
|
| 1292 |
+
"step": 1036
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 74.29,
|
| 1296 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 1297 |
+
"loss": 0.4663,
|
| 1298 |
+
"step": 1040
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 75.0,
|
| 1302 |
+
"learning_rate": 1.388888888888889e-05,
|
| 1303 |
+
"loss": 0.4973,
|
| 1304 |
+
"step": 1050
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 75.0,
|
| 1308 |
+
"eval_accuracy": 0.7795454545454545,
|
| 1309 |
+
"eval_loss": 0.6241269707679749,
|
| 1310 |
+
"eval_runtime": 3.4141,
|
| 1311 |
+
"eval_samples_per_second": 128.879,
|
| 1312 |
+
"eval_steps_per_second": 2.05,
|
| 1313 |
+
"step": 1050
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 75.71,
|
| 1317 |
+
"learning_rate": 1.3492063492063492e-05,
|
| 1318 |
+
"loss": 0.4807,
|
| 1319 |
+
"step": 1060
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 76.0,
|
| 1323 |
+
"eval_accuracy": 0.7727272727272727,
|
| 1324 |
+
"eval_loss": 0.6188752055168152,
|
| 1325 |
+
"eval_runtime": 3.372,
|
| 1326 |
+
"eval_samples_per_second": 130.487,
|
| 1327 |
+
"eval_steps_per_second": 2.076,
|
| 1328 |
+
"step": 1064
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 76.43,
|
| 1332 |
+
"learning_rate": 1.3095238095238096e-05,
|
| 1333 |
+
"loss": 0.4845,
|
| 1334 |
+
"step": 1070
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 77.0,
|
| 1338 |
+
"eval_accuracy": 0.775,
|
| 1339 |
+
"eval_loss": 0.6290163397789001,
|
| 1340 |
+
"eval_runtime": 3.37,
|
| 1341 |
+
"eval_samples_per_second": 130.564,
|
| 1342 |
+
"eval_steps_per_second": 2.077,
|
| 1343 |
+
"step": 1078
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 77.14,
|
| 1347 |
+
"learning_rate": 1.2698412698412699e-05,
|
| 1348 |
+
"loss": 0.4699,
|
| 1349 |
+
"step": 1080
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 77.86,
|
| 1353 |
+
"learning_rate": 1.2301587301587301e-05,
|
| 1354 |
+
"loss": 0.4874,
|
| 1355 |
+
"step": 1090
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 78.0,
|
| 1359 |
+
"eval_accuracy": 0.7772727272727272,
|
| 1360 |
+
"eval_loss": 0.6192808747291565,
|
| 1361 |
+
"eval_runtime": 3.365,
|
| 1362 |
+
"eval_samples_per_second": 130.758,
|
| 1363 |
+
"eval_steps_per_second": 2.08,
|
| 1364 |
+
"step": 1092
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 78.57,
|
| 1368 |
+
"learning_rate": 1.1904761904761905e-05,
|
| 1369 |
+
"loss": 0.4898,
|
| 1370 |
+
"step": 1100
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 79.0,
|
| 1374 |
+
"eval_accuracy": 0.775,
|
| 1375 |
+
"eval_loss": 0.643506646156311,
|
| 1376 |
+
"eval_runtime": 3.4199,
|
| 1377 |
+
"eval_samples_per_second": 128.66,
|
| 1378 |
+
"eval_steps_per_second": 2.047,
|
| 1379 |
+
"step": 1106
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 79.29,
|
| 1383 |
+
"learning_rate": 1.1507936507936508e-05,
|
| 1384 |
+
"loss": 0.4844,
|
| 1385 |
+
"step": 1110
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 80.0,
|
| 1389 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 1390 |
+
"loss": 0.4904,
|
| 1391 |
+
"step": 1120
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 80.0,
|
| 1395 |
+
"eval_accuracy": 0.7613636363636364,
|
| 1396 |
+
"eval_loss": 0.6483902931213379,
|
| 1397 |
+
"eval_runtime": 3.373,
|
| 1398 |
+
"eval_samples_per_second": 130.448,
|
| 1399 |
+
"eval_steps_per_second": 2.075,
|
| 1400 |
+
"step": 1120
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 80.71,
|
| 1404 |
+
"learning_rate": 1.0714285714285714e-05,
|
| 1405 |
+
"loss": 0.4702,
|
| 1406 |
+
"step": 1130
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 81.0,
|
| 1410 |
+
"eval_accuracy": 0.7727272727272727,
|
| 1411 |
+
"eval_loss": 0.6238275170326233,
|
| 1412 |
+
"eval_runtime": 3.5076,
|
| 1413 |
+
"eval_samples_per_second": 125.441,
|
| 1414 |
+
"eval_steps_per_second": 1.996,
|
| 1415 |
+
"step": 1134
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 81.43,
|
| 1419 |
+
"learning_rate": 1.0317460317460318e-05,
|
| 1420 |
+
"loss": 0.4802,
|
| 1421 |
+
"step": 1140
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 82.0,
|
| 1425 |
+
"eval_accuracy": 0.7727272727272727,
|
| 1426 |
+
"eval_loss": 0.6386602520942688,
|
| 1427 |
+
"eval_runtime": 3.3979,
|
| 1428 |
+
"eval_samples_per_second": 129.491,
|
| 1429 |
+
"eval_steps_per_second": 2.06,
|
| 1430 |
+
"step": 1148
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 82.14,
|
| 1434 |
+
"learning_rate": 9.92063492063492e-06,
|
| 1435 |
+
"loss": 0.4529,
|
| 1436 |
+
"step": 1150
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 82.86,
|
| 1440 |
+
"learning_rate": 9.523809523809523e-06,
|
| 1441 |
+
"loss": 0.4635,
|
| 1442 |
+
"step": 1160
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 83.0,
|
| 1446 |
+
"eval_accuracy": 0.7704545454545455,
|
| 1447 |
+
"eval_loss": 0.6206506490707397,
|
| 1448 |
+
"eval_runtime": 3.4119,
|
| 1449 |
+
"eval_samples_per_second": 128.961,
|
| 1450 |
+
"eval_steps_per_second": 2.052,
|
| 1451 |
+
"step": 1162
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 83.57,
|
| 1455 |
+
"learning_rate": 9.126984126984127e-06,
|
| 1456 |
+
"loss": 0.4457,
|
| 1457 |
+
"step": 1170
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 84.0,
|
| 1461 |
+
"eval_accuracy": 0.7659090909090909,
|
| 1462 |
+
"eval_loss": 0.6249757409095764,
|
| 1463 |
+
"eval_runtime": 3.4328,
|
| 1464 |
+
"eval_samples_per_second": 128.174,
|
| 1465 |
+
"eval_steps_per_second": 2.039,
|
| 1466 |
+
"step": 1176
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 84.29,
|
| 1470 |
+
"learning_rate": 8.73015873015873e-06,
|
| 1471 |
+
"loss": 0.4665,
|
| 1472 |
+
"step": 1180
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 85.0,
|
| 1476 |
+
"learning_rate": 8.333333333333334e-06,
|
| 1477 |
+
"loss": 0.4697,
|
| 1478 |
+
"step": 1190
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 85.0,
|
| 1482 |
+
"eval_accuracy": 0.7818181818181819,
|
| 1483 |
+
"eval_loss": 0.611347496509552,
|
| 1484 |
+
"eval_runtime": 3.4099,
|
| 1485 |
+
"eval_samples_per_second": 129.037,
|
| 1486 |
+
"eval_steps_per_second": 2.053,
|
| 1487 |
+
"step": 1190
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 85.71,
|
| 1491 |
+
"learning_rate": 7.936507936507936e-06,
|
| 1492 |
+
"loss": 0.4359,
|
| 1493 |
+
"step": 1200
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 86.0,
|
| 1497 |
+
"eval_accuracy": 0.7659090909090909,
|
| 1498 |
+
"eval_loss": 0.637219250202179,
|
| 1499 |
+
"eval_runtime": 3.369,
|
| 1500 |
+
"eval_samples_per_second": 130.603,
|
| 1501 |
+
"eval_steps_per_second": 2.078,
|
| 1502 |
+
"step": 1204
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 86.43,
|
| 1506 |
+
"learning_rate": 7.5396825396825394e-06,
|
| 1507 |
+
"loss": 0.4876,
|
| 1508 |
+
"step": 1210
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 87.0,
|
| 1512 |
+
"eval_accuracy": 0.759090909090909,
|
| 1513 |
+
"eval_loss": 0.6338838338851929,
|
| 1514 |
+
"eval_runtime": 3.4298,
|
| 1515 |
+
"eval_samples_per_second": 128.286,
|
| 1516 |
+
"eval_steps_per_second": 2.041,
|
| 1517 |
+
"step": 1218
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 87.14,
|
| 1521 |
+
"learning_rate": 7.142857142857143e-06,
|
| 1522 |
+
"loss": 0.4816,
|
| 1523 |
+
"step": 1220
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 87.86,
|
| 1527 |
+
"learning_rate": 6.746031746031746e-06,
|
| 1528 |
+
"loss": 0.4574,
|
| 1529 |
+
"step": 1230
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 88.0,
|
| 1533 |
+
"eval_accuracy": 0.7545454545454545,
|
| 1534 |
+
"eval_loss": 0.6318604946136475,
|
| 1535 |
+
"eval_runtime": 3.5206,
|
| 1536 |
+
"eval_samples_per_second": 124.979,
|
| 1537 |
+
"eval_steps_per_second": 1.988,
|
| 1538 |
+
"step": 1232
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 88.57,
|
| 1542 |
+
"learning_rate": 6.349206349206349e-06,
|
| 1543 |
+
"loss": 0.4633,
|
| 1544 |
+
"step": 1240
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 89.0,
|
| 1548 |
+
"eval_accuracy": 0.7727272727272727,
|
| 1549 |
+
"eval_loss": 0.6072612404823303,
|
| 1550 |
+
"eval_runtime": 3.6356,
|
| 1551 |
+
"eval_samples_per_second": 121.025,
|
| 1552 |
+
"eval_steps_per_second": 1.925,
|
| 1553 |
+
"step": 1246
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 89.29,
|
| 1557 |
+
"learning_rate": 5.9523809523809525e-06,
|
| 1558 |
+
"loss": 0.4459,
|
| 1559 |
+
"step": 1250
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 90.0,
|
| 1563 |
+
"learning_rate": 5.555555555555556e-06,
|
| 1564 |
+
"loss": 0.4548,
|
| 1565 |
+
"step": 1260
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 90.0,
|
| 1569 |
+
"eval_accuracy": 0.7704545454545455,
|
| 1570 |
+
"eval_loss": 0.6304746866226196,
|
| 1571 |
+
"eval_runtime": 3.4518,
|
| 1572 |
+
"eval_samples_per_second": 127.471,
|
| 1573 |
+
"eval_steps_per_second": 2.028,
|
| 1574 |
+
"step": 1260
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 90.71,
|
| 1578 |
+
"learning_rate": 5.158730158730159e-06,
|
| 1579 |
+
"loss": 0.4527,
|
| 1580 |
+
"step": 1270
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 91.0,
|
| 1584 |
+
"eval_accuracy": 0.7681818181818182,
|
| 1585 |
+
"eval_loss": 0.6324154734611511,
|
| 1586 |
+
"eval_runtime": 3.5884,
|
| 1587 |
+
"eval_samples_per_second": 122.617,
|
| 1588 |
+
"eval_steps_per_second": 1.951,
|
| 1589 |
+
"step": 1274
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 91.43,
|
| 1593 |
+
"learning_rate": 4.7619047619047615e-06,
|
| 1594 |
+
"loss": 0.4311,
|
| 1595 |
+
"step": 1280
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 92.0,
|
| 1599 |
+
"eval_accuracy": 0.7795454545454545,
|
| 1600 |
+
"eval_loss": 0.6228455901145935,
|
| 1601 |
+
"eval_runtime": 3.543,
|
| 1602 |
+
"eval_samples_per_second": 124.187,
|
| 1603 |
+
"eval_steps_per_second": 1.976,
|
| 1604 |
+
"step": 1288
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 92.14,
|
| 1608 |
+
"learning_rate": 4.365079365079365e-06,
|
| 1609 |
+
"loss": 0.4641,
|
| 1610 |
+
"step": 1290
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 92.86,
|
| 1614 |
+
"learning_rate": 3.968253968253968e-06,
|
| 1615 |
+
"loss": 0.4499,
|
| 1616 |
+
"step": 1300
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 93.0,
|
| 1620 |
+
"eval_accuracy": 0.7636363636363637,
|
| 1621 |
+
"eval_loss": 0.6188679337501526,
|
| 1622 |
+
"eval_runtime": 3.6596,
|
| 1623 |
+
"eval_samples_per_second": 120.231,
|
| 1624 |
+
"eval_steps_per_second": 1.913,
|
| 1625 |
+
"step": 1302
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 93.57,
|
| 1629 |
+
"learning_rate": 3.5714285714285714e-06,
|
| 1630 |
+
"loss": 0.4333,
|
| 1631 |
+
"step": 1310
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 94.0,
|
| 1635 |
+
"eval_accuracy": 0.7727272727272727,
|
| 1636 |
+
"eval_loss": 0.6302651166915894,
|
| 1637 |
+
"eval_runtime": 3.5367,
|
| 1638 |
+
"eval_samples_per_second": 124.41,
|
| 1639 |
+
"eval_steps_per_second": 1.979,
|
| 1640 |
+
"step": 1316
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 94.29,
|
| 1644 |
+
"learning_rate": 3.1746031746031746e-06,
|
| 1645 |
+
"loss": 0.4647,
|
| 1646 |
+
"step": 1320
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 95.0,
|
| 1650 |
+
"learning_rate": 2.777777777777778e-06,
|
| 1651 |
+
"loss": 0.4526,
|
| 1652 |
+
"step": 1330
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 95.0,
|
| 1656 |
+
"eval_accuracy": 0.7704545454545455,
|
| 1657 |
+
"eval_loss": 0.6163169145584106,
|
| 1658 |
+
"eval_runtime": 3.5271,
|
| 1659 |
+
"eval_samples_per_second": 124.747,
|
| 1660 |
+
"eval_steps_per_second": 1.985,
|
| 1661 |
+
"step": 1330
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 95.71,
|
| 1665 |
+
"learning_rate": 2.3809523809523808e-06,
|
| 1666 |
+
"loss": 0.4709,
|
| 1667 |
+
"step": 1340
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 96.0,
|
| 1671 |
+
"eval_accuracy": 0.7659090909090909,
|
| 1672 |
+
"eval_loss": 0.618194580078125,
|
| 1673 |
+
"eval_runtime": 3.4856,
|
| 1674 |
+
"eval_samples_per_second": 126.234,
|
| 1675 |
+
"eval_steps_per_second": 2.008,
|
| 1676 |
+
"step": 1344
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 96.43,
|
| 1680 |
+
"learning_rate": 1.984126984126984e-06,
|
| 1681 |
+
"loss": 0.4451,
|
| 1682 |
+
"step": 1350
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 97.0,
|
| 1686 |
+
"eval_accuracy": 0.7818181818181819,
|
| 1687 |
+
"eval_loss": 0.6048444509506226,
|
| 1688 |
+
"eval_runtime": 3.4925,
|
| 1689 |
+
"eval_samples_per_second": 125.984,
|
| 1690 |
+
"eval_steps_per_second": 2.004,
|
| 1691 |
+
"step": 1358
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 97.14,
|
| 1695 |
+
"learning_rate": 1.5873015873015873e-06,
|
| 1696 |
+
"loss": 0.4617,
|
| 1697 |
+
"step": 1360
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 97.86,
|
| 1701 |
+
"learning_rate": 1.1904761904761904e-06,
|
| 1702 |
+
"loss": 0.4174,
|
| 1703 |
+
"step": 1370
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 98.0,
|
| 1707 |
+
"eval_accuracy": 0.7659090909090909,
|
| 1708 |
+
"eval_loss": 0.6300343871116638,
|
| 1709 |
+
"eval_runtime": 3.4255,
|
| 1710 |
+
"eval_samples_per_second": 128.45,
|
| 1711 |
+
"eval_steps_per_second": 2.044,
|
| 1712 |
+
"step": 1372
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 98.57,
|
| 1716 |
+
"learning_rate": 7.936507936507937e-07,
|
| 1717 |
+
"loss": 0.4832,
|
| 1718 |
+
"step": 1380
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 99.0,
|
| 1722 |
+
"eval_accuracy": 0.7704545454545455,
|
| 1723 |
+
"eval_loss": 0.6194254159927368,
|
| 1724 |
+
"eval_runtime": 3.5142,
|
| 1725 |
+
"eval_samples_per_second": 125.206,
|
| 1726 |
+
"eval_steps_per_second": 1.992,
|
| 1727 |
+
"step": 1386
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 99.29,
|
| 1731 |
+
"learning_rate": 3.9682539682539683e-07,
|
| 1732 |
+
"loss": 0.4501,
|
| 1733 |
+
"step": 1390
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 100.0,
|
| 1737 |
+
"learning_rate": 0.0,
|
| 1738 |
+
"loss": 0.4508,
|
| 1739 |
+
"step": 1400
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 100.0,
|
| 1743 |
+
"eval_accuracy": 0.7795454545454545,
|
| 1744 |
+
"eval_loss": 0.6144731640815735,
|
| 1745 |
+
"eval_runtime": 3.5121,
|
| 1746 |
+
"eval_samples_per_second": 125.28,
|
| 1747 |
+
"eval_steps_per_second": 1.993,
|
| 1748 |
+
"step": 1400
|
| 1749 |
+
}
|
| 1750 |
+
],
|
| 1751 |
+
"max_steps": 1400,
|
| 1752 |
+
"num_train_epochs": 100,
|
| 1753 |
+
"total_flos": 7.496586164267827e+18,
|
| 1754 |
+
"trial_name": null,
|
| 1755 |
+
"trial_params": null
|
| 1756 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f31d578f4c563627af04f71ba1c4d097d5f97d224870f72c1ec58ba3658a88b8
|
| 3 |
+
size 3311
|