Training in progress, epoch 0
Browse files- config.json +65 -0
- model.safetensors +3 -0
- preprocessor_config.json +22 -0
- tmp-checkpoint-557/config.json +65 -0
- tmp-checkpoint-557/model.safetensors +3 -0
- tmp-checkpoint-557/optimizer.pt +3 -0
- tmp-checkpoint-557/preprocessor_config.json +22 -0
- tmp-checkpoint-557/rng_state.pth +3 -0
- tmp-checkpoint-557/scheduler.pt +3 -0
- tmp-checkpoint-557/trainer_state.json +371 -0
- tmp-checkpoint-557/training_args.bin +3 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/swin-tiny-patch4-window7-224",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"SwinForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"depths": [
|
| 8 |
+
2,
|
| 9 |
+
2,
|
| 10 |
+
6,
|
| 11 |
+
2
|
| 12 |
+
],
|
| 13 |
+
"drop_path_rate": 0.1,
|
| 14 |
+
"embed_dim": 96,
|
| 15 |
+
"encoder_stride": 32,
|
| 16 |
+
"hidden_act": "gelu",
|
| 17 |
+
"hidden_dropout_prob": 0.0,
|
| 18 |
+
"hidden_size": 768,
|
| 19 |
+
"id2label": {
|
| 20 |
+
"0": "noncrack",
|
| 21 |
+
"1": "sealedcrack",
|
| 22 |
+
"2": "sealedpatch",
|
| 23 |
+
"3": "thincrack"
|
| 24 |
+
},
|
| 25 |
+
"image_size": 224,
|
| 26 |
+
"initializer_range": 0.02,
|
| 27 |
+
"label2id": {
|
| 28 |
+
"noncrack": 0,
|
| 29 |
+
"sealedcrack": 1,
|
| 30 |
+
"sealedpatch": 2,
|
| 31 |
+
"thincrack": 3
|
| 32 |
+
},
|
| 33 |
+
"layer_norm_eps": 1e-05,
|
| 34 |
+
"mlp_ratio": 4.0,
|
| 35 |
+
"model_type": "swin",
|
| 36 |
+
"num_channels": 3,
|
| 37 |
+
"num_heads": [
|
| 38 |
+
3,
|
| 39 |
+
6,
|
| 40 |
+
12,
|
| 41 |
+
24
|
| 42 |
+
],
|
| 43 |
+
"num_layers": 4,
|
| 44 |
+
"out_features": [
|
| 45 |
+
"stage4"
|
| 46 |
+
],
|
| 47 |
+
"out_indices": [
|
| 48 |
+
4
|
| 49 |
+
],
|
| 50 |
+
"patch_size": 4,
|
| 51 |
+
"path_norm": true,
|
| 52 |
+
"problem_type": "single_label_classification",
|
| 53 |
+
"qkv_bias": true,
|
| 54 |
+
"stage_names": [
|
| 55 |
+
"stem",
|
| 56 |
+
"stage1",
|
| 57 |
+
"stage2",
|
| 58 |
+
"stage3",
|
| 59 |
+
"stage4"
|
| 60 |
+
],
|
| 61 |
+
"torch_dtype": "float32",
|
| 62 |
+
"transformers_version": "4.37.2",
|
| 63 |
+
"use_absolute_embeddings": false,
|
| 64 |
+
"window_size": 7
|
| 65 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3b58cf0056f7df5b5217ea96f826a585b2a599920ca01ffb44ef4cbcc49662a5
|
| 3 |
+
size 110348984
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"image_mean": [
|
| 6 |
+
0.485,
|
| 7 |
+
0.456,
|
| 8 |
+
0.406
|
| 9 |
+
],
|
| 10 |
+
"image_processor_type": "ViTImageProcessor",
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.229,
|
| 13 |
+
0.224,
|
| 14 |
+
0.225
|
| 15 |
+
],
|
| 16 |
+
"resample": 3,
|
| 17 |
+
"rescale_factor": 0.00392156862745098,
|
| 18 |
+
"size": {
|
| 19 |
+
"height": 224,
|
| 20 |
+
"width": 224
|
| 21 |
+
}
|
| 22 |
+
}
|
tmp-checkpoint-557/config.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "microsoft/swin-tiny-patch4-window7-224",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"SwinForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"depths": [
|
| 8 |
+
2,
|
| 9 |
+
2,
|
| 10 |
+
6,
|
| 11 |
+
2
|
| 12 |
+
],
|
| 13 |
+
"drop_path_rate": 0.1,
|
| 14 |
+
"embed_dim": 96,
|
| 15 |
+
"encoder_stride": 32,
|
| 16 |
+
"hidden_act": "gelu",
|
| 17 |
+
"hidden_dropout_prob": 0.0,
|
| 18 |
+
"hidden_size": 768,
|
| 19 |
+
"id2label": {
|
| 20 |
+
"0": "noncrack",
|
| 21 |
+
"1": "sealedcrack",
|
| 22 |
+
"2": "sealedpatch",
|
| 23 |
+
"3": "thincrack"
|
| 24 |
+
},
|
| 25 |
+
"image_size": 224,
|
| 26 |
+
"initializer_range": 0.02,
|
| 27 |
+
"label2id": {
|
| 28 |
+
"noncrack": 0,
|
| 29 |
+
"sealedcrack": 1,
|
| 30 |
+
"sealedpatch": 2,
|
| 31 |
+
"thincrack": 3
|
| 32 |
+
},
|
| 33 |
+
"layer_norm_eps": 1e-05,
|
| 34 |
+
"mlp_ratio": 4.0,
|
| 35 |
+
"model_type": "swin",
|
| 36 |
+
"num_channels": 3,
|
| 37 |
+
"num_heads": [
|
| 38 |
+
3,
|
| 39 |
+
6,
|
| 40 |
+
12,
|
| 41 |
+
24
|
| 42 |
+
],
|
| 43 |
+
"num_layers": 4,
|
| 44 |
+
"out_features": [
|
| 45 |
+
"stage4"
|
| 46 |
+
],
|
| 47 |
+
"out_indices": [
|
| 48 |
+
4
|
| 49 |
+
],
|
| 50 |
+
"patch_size": 4,
|
| 51 |
+
"path_norm": true,
|
| 52 |
+
"problem_type": "single_label_classification",
|
| 53 |
+
"qkv_bias": true,
|
| 54 |
+
"stage_names": [
|
| 55 |
+
"stem",
|
| 56 |
+
"stage1",
|
| 57 |
+
"stage2",
|
| 58 |
+
"stage3",
|
| 59 |
+
"stage4"
|
| 60 |
+
],
|
| 61 |
+
"torch_dtype": "float32",
|
| 62 |
+
"transformers_version": "4.37.2",
|
| 63 |
+
"use_absolute_embeddings": false,
|
| 64 |
+
"window_size": 7
|
| 65 |
+
}
|
tmp-checkpoint-557/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3b58cf0056f7df5b5217ea96f826a585b2a599920ca01ffb44ef4cbcc49662a5
|
| 3 |
+
size 110348984
|
tmp-checkpoint-557/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:355b05855a304b00d2832b6090e991677c67f1bf62319bac82d83ddd68b3a868
|
| 3 |
+
size 220362874
|
tmp-checkpoint-557/preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"image_mean": [
|
| 6 |
+
0.485,
|
| 7 |
+
0.456,
|
| 8 |
+
0.406
|
| 9 |
+
],
|
| 10 |
+
"image_processor_type": "ViTImageProcessor",
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.229,
|
| 13 |
+
0.224,
|
| 14 |
+
0.225
|
| 15 |
+
],
|
| 16 |
+
"resample": 3,
|
| 17 |
+
"rescale_factor": 0.00392156862745098,
|
| 18 |
+
"size": {
|
| 19 |
+
"height": 224,
|
| 20 |
+
"width": 224
|
| 21 |
+
}
|
| 22 |
+
}
|
tmp-checkpoint-557/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8d615268b99d6b2b1c404e92148480a3473a846eb543ef952738586f2ee5a5cb
|
| 3 |
+
size 14244
|
tmp-checkpoint-557/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a7b878d0d15f3684a2027634fc49696fc1670f0ac208f7c62706d478f4ee51b4
|
| 3 |
+
size 1064
|
tmp-checkpoint-557/trainer_state.json
ADDED
|
@@ -0,0 +1,371 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.43474748730659485,
|
| 3 |
+
"best_model_checkpoint": "swin-tiny-patch4-window7-224-crack-detectorMAIN50epochsFINAL\\checkpoint-557",
|
| 4 |
+
"epoch": 0.9991031390134529,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 557,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 1.7953321364452426e-07,
|
| 14 |
+
"loss": 1.4257,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.04,
|
| 19 |
+
"learning_rate": 3.590664272890485e-07,
|
| 20 |
+
"loss": 1.4227,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.05,
|
| 25 |
+
"learning_rate": 5.385996409335728e-07,
|
| 26 |
+
"loss": 1.4112,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.07,
|
| 31 |
+
"learning_rate": 7.18132854578097e-07,
|
| 32 |
+
"loss": 1.4052,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.09,
|
| 37 |
+
"learning_rate": 8.976660682226213e-07,
|
| 38 |
+
"loss": 1.3954,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.11,
|
| 43 |
+
"learning_rate": 1.0771992818671456e-06,
|
| 44 |
+
"loss": 1.3978,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.13,
|
| 49 |
+
"learning_rate": 1.2567324955116697e-06,
|
| 50 |
+
"loss": 1.3862,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.14,
|
| 55 |
+
"learning_rate": 1.436265709156194e-06,
|
| 56 |
+
"loss": 1.3651,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.16,
|
| 61 |
+
"learning_rate": 1.6157989228007182e-06,
|
| 62 |
+
"loss": 1.359,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.18,
|
| 67 |
+
"learning_rate": 1.7953321364452425e-06,
|
| 68 |
+
"loss": 1.3476,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.2,
|
| 73 |
+
"learning_rate": 1.9748653500897667e-06,
|
| 74 |
+
"loss": 1.3303,
|
| 75 |
+
"step": 110
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.22,
|
| 79 |
+
"learning_rate": 2.1543985637342912e-06,
|
| 80 |
+
"loss": 1.3047,
|
| 81 |
+
"step": 120
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.23,
|
| 85 |
+
"learning_rate": 2.333931777378815e-06,
|
| 86 |
+
"loss": 1.2944,
|
| 87 |
+
"step": 130
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.25,
|
| 91 |
+
"learning_rate": 2.5134649910233395e-06,
|
| 92 |
+
"loss": 1.284,
|
| 93 |
+
"step": 140
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.27,
|
| 97 |
+
"learning_rate": 2.6929982046678636e-06,
|
| 98 |
+
"loss": 1.23,
|
| 99 |
+
"step": 150
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.29,
|
| 103 |
+
"learning_rate": 2.872531418312388e-06,
|
| 104 |
+
"loss": 1.2103,
|
| 105 |
+
"step": 160
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.3,
|
| 109 |
+
"learning_rate": 3.0520646319569123e-06,
|
| 110 |
+
"loss": 1.187,
|
| 111 |
+
"step": 170
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.32,
|
| 115 |
+
"learning_rate": 3.2315978456014364e-06,
|
| 116 |
+
"loss": 1.1591,
|
| 117 |
+
"step": 180
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.34,
|
| 121 |
+
"learning_rate": 3.411131059245961e-06,
|
| 122 |
+
"loss": 1.144,
|
| 123 |
+
"step": 190
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.36,
|
| 127 |
+
"learning_rate": 3.590664272890485e-06,
|
| 128 |
+
"loss": 1.0904,
|
| 129 |
+
"step": 200
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.38,
|
| 133 |
+
"learning_rate": 3.770197486535009e-06,
|
| 134 |
+
"loss": 1.0401,
|
| 135 |
+
"step": 210
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.39,
|
| 139 |
+
"learning_rate": 3.949730700179533e-06,
|
| 140 |
+
"loss": 1.0281,
|
| 141 |
+
"step": 220
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.41,
|
| 145 |
+
"learning_rate": 4.129263913824058e-06,
|
| 146 |
+
"loss": 0.9577,
|
| 147 |
+
"step": 230
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.43,
|
| 151 |
+
"learning_rate": 4.3087971274685824e-06,
|
| 152 |
+
"loss": 0.9251,
|
| 153 |
+
"step": 240
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.45,
|
| 157 |
+
"learning_rate": 4.488330341113106e-06,
|
| 158 |
+
"loss": 0.8964,
|
| 159 |
+
"step": 250
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.47,
|
| 163 |
+
"learning_rate": 4.66786355475763e-06,
|
| 164 |
+
"loss": 0.8897,
|
| 165 |
+
"step": 260
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.48,
|
| 169 |
+
"learning_rate": 4.847396768402154e-06,
|
| 170 |
+
"loss": 0.8481,
|
| 171 |
+
"step": 270
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.5,
|
| 175 |
+
"learning_rate": 5.026929982046679e-06,
|
| 176 |
+
"loss": 0.8048,
|
| 177 |
+
"step": 280
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.52,
|
| 181 |
+
"learning_rate": 5.206463195691203e-06,
|
| 182 |
+
"loss": 0.8014,
|
| 183 |
+
"step": 290
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.54,
|
| 187 |
+
"learning_rate": 5.385996409335727e-06,
|
| 188 |
+
"loss": 0.7511,
|
| 189 |
+
"step": 300
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.56,
|
| 193 |
+
"learning_rate": 5.565529622980251e-06,
|
| 194 |
+
"loss": 0.741,
|
| 195 |
+
"step": 310
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.57,
|
| 199 |
+
"learning_rate": 5.745062836624776e-06,
|
| 200 |
+
"loss": 0.7327,
|
| 201 |
+
"step": 320
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.59,
|
| 205 |
+
"learning_rate": 5.9245960502693004e-06,
|
| 206 |
+
"loss": 0.7492,
|
| 207 |
+
"step": 330
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.61,
|
| 211 |
+
"learning_rate": 6.1041292639138246e-06,
|
| 212 |
+
"loss": 0.7613,
|
| 213 |
+
"step": 340
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.63,
|
| 217 |
+
"learning_rate": 6.283662477558349e-06,
|
| 218 |
+
"loss": 0.6902,
|
| 219 |
+
"step": 350
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.65,
|
| 223 |
+
"learning_rate": 6.463195691202873e-06,
|
| 224 |
+
"loss": 0.6687,
|
| 225 |
+
"step": 360
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.66,
|
| 229 |
+
"learning_rate": 6.642728904847396e-06,
|
| 230 |
+
"loss": 0.6623,
|
| 231 |
+
"step": 370
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.68,
|
| 235 |
+
"learning_rate": 6.822262118491922e-06,
|
| 236 |
+
"loss": 0.6901,
|
| 237 |
+
"step": 380
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.7,
|
| 241 |
+
"learning_rate": 7.001795332136446e-06,
|
| 242 |
+
"loss": 0.6638,
|
| 243 |
+
"step": 390
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.72,
|
| 247 |
+
"learning_rate": 7.18132854578097e-06,
|
| 248 |
+
"loss": 0.6414,
|
| 249 |
+
"step": 400
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.74,
|
| 253 |
+
"learning_rate": 7.360861759425494e-06,
|
| 254 |
+
"loss": 0.6877,
|
| 255 |
+
"step": 410
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.75,
|
| 259 |
+
"learning_rate": 7.540394973070018e-06,
|
| 260 |
+
"loss": 0.6249,
|
| 261 |
+
"step": 420
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.77,
|
| 265 |
+
"learning_rate": 7.719928186714543e-06,
|
| 266 |
+
"loss": 0.6099,
|
| 267 |
+
"step": 430
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.79,
|
| 271 |
+
"learning_rate": 7.899461400359067e-06,
|
| 272 |
+
"loss": 0.6321,
|
| 273 |
+
"step": 440
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.81,
|
| 277 |
+
"learning_rate": 8.07899461400359e-06,
|
| 278 |
+
"loss": 0.6431,
|
| 279 |
+
"step": 450
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.83,
|
| 283 |
+
"learning_rate": 8.258527827648117e-06,
|
| 284 |
+
"loss": 0.6404,
|
| 285 |
+
"step": 460
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.84,
|
| 289 |
+
"learning_rate": 8.43806104129264e-06,
|
| 290 |
+
"loss": 0.6148,
|
| 291 |
+
"step": 470
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.86,
|
| 295 |
+
"learning_rate": 8.617594254937165e-06,
|
| 296 |
+
"loss": 0.5844,
|
| 297 |
+
"step": 480
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.88,
|
| 301 |
+
"learning_rate": 8.797127468581689e-06,
|
| 302 |
+
"loss": 0.6316,
|
| 303 |
+
"step": 490
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.9,
|
| 307 |
+
"learning_rate": 8.976660682226211e-06,
|
| 308 |
+
"loss": 0.6018,
|
| 309 |
+
"step": 500
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.91,
|
| 313 |
+
"learning_rate": 9.156193895870736e-06,
|
| 314 |
+
"loss": 0.5843,
|
| 315 |
+
"step": 510
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.93,
|
| 319 |
+
"learning_rate": 9.33572710951526e-06,
|
| 320 |
+
"loss": 0.5986,
|
| 321 |
+
"step": 520
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.95,
|
| 325 |
+
"learning_rate": 9.515260323159784e-06,
|
| 326 |
+
"loss": 0.5718,
|
| 327 |
+
"step": 530
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.97,
|
| 331 |
+
"learning_rate": 9.694793536804308e-06,
|
| 332 |
+
"loss": 0.5163,
|
| 333 |
+
"step": 540
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.99,
|
| 337 |
+
"learning_rate": 9.874326750448834e-06,
|
| 338 |
+
"loss": 0.5448,
|
| 339 |
+
"step": 550
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.0,
|
| 343 |
+
"eval_accuracy": {
|
| 344 |
+
"accuracy": 0.8387530836510428
|
| 345 |
+
},
|
| 346 |
+
"eval_f1": {
|
| 347 |
+
"f1": 0.8242528613153339
|
| 348 |
+
},
|
| 349 |
+
"eval_loss": 0.43474748730659485,
|
| 350 |
+
"eval_precision": {
|
| 351 |
+
"precision": 0.8238035334510331
|
| 352 |
+
},
|
| 353 |
+
"eval_recall": {
|
| 354 |
+
"recall": 0.8266051811271123
|
| 355 |
+
},
|
| 356 |
+
"eval_runtime": 171.551,
|
| 357 |
+
"eval_samples_per_second": 103.969,
|
| 358 |
+
"eval_steps_per_second": 3.253,
|
| 359 |
+
"step": 557
|
| 360 |
+
}
|
| 361 |
+
],
|
| 362 |
+
"logging_steps": 10,
|
| 363 |
+
"max_steps": 27850,
|
| 364 |
+
"num_input_tokens_seen": 0,
|
| 365 |
+
"num_train_epochs": 50,
|
| 366 |
+
"save_steps": 500,
|
| 367 |
+
"total_flos": 1.7733253506748416e+18,
|
| 368 |
+
"train_batch_size": 32,
|
| 369 |
+
"trial_name": null,
|
| 370 |
+
"trial_params": null
|
| 371 |
+
}
|
tmp-checkpoint-557/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c081bc3c7c7c6c246b97a8d5de1f01d4b69caf186a7167574020cb94e76e323a
|
| 3 |
+
size 4792
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c081bc3c7c7c6c246b97a8d5de1f01d4b69caf186a7167574020cb94e76e323a
|
| 3 |
+
size 4792
|