Upload 8 files
Browse files- config.json +64 -0
- optimizer.pt +3 -0
- preprocessor_config.json +22 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +649 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "AnnualCrop",
|
| 23 |
+
"1": "Forest",
|
| 24 |
+
"2": "HerbaceousVegetation",
|
| 25 |
+
"3": "Highway",
|
| 26 |
+
"4": "Industrial",
|
| 27 |
+
"5": "Pasture",
|
| 28 |
+
"6": "PermanentCrop",
|
| 29 |
+
"7": "Residential",
|
| 30 |
+
"8": "River",
|
| 31 |
+
"9": "SeaLake"
|
| 32 |
+
},
|
| 33 |
+
"label2id": {
|
| 34 |
+
"AnnualCrop": "0",
|
| 35 |
+
"Forest": "1",
|
| 36 |
+
"HerbaceousVegetation": "2",
|
| 37 |
+
"Highway": "3",
|
| 38 |
+
"Industrial": "4",
|
| 39 |
+
"Pasture": "5",
|
| 40 |
+
"PermanentCrop": "6",
|
| 41 |
+
"Residential": "7",
|
| 42 |
+
"River": "8",
|
| 43 |
+
"SeaLake": "9"
|
| 44 |
+
},
|
| 45 |
+
"layer_type": "bottleneck",
|
| 46 |
+
"model_type": "resnet",
|
| 47 |
+
"num_channels": 3,
|
| 48 |
+
"out_features": [
|
| 49 |
+
"stage4"
|
| 50 |
+
],
|
| 51 |
+
"out_indices": [
|
| 52 |
+
4
|
| 53 |
+
],
|
| 54 |
+
"problem_type": "single_label_classification",
|
| 55 |
+
"stage_names": [
|
| 56 |
+
"stem",
|
| 57 |
+
"stage1",
|
| 58 |
+
"stage2",
|
| 59 |
+
"stage3",
|
| 60 |
+
"stage4"
|
| 61 |
+
],
|
| 62 |
+
"torch_dtype": "float32",
|
| 63 |
+
"transformers_version": "4.30.1"
|
| 64 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:018e27f5b0885c1c4985f02797d320393a32b52399eef10d85fe9a002f598672
|
| 3 |
+
size 188318725
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"crop_pct": 0.875,
|
| 3 |
+
"do_normalize": true,
|
| 4 |
+
"do_rescale": true,
|
| 5 |
+
"do_resize": true,
|
| 6 |
+
"image_mean": [
|
| 7 |
+
0.485,
|
| 8 |
+
0.456,
|
| 9 |
+
0.406
|
| 10 |
+
],
|
| 11 |
+
"image_processor_type": "ConvNextImageProcessor",
|
| 12 |
+
"image_std": [
|
| 13 |
+
0.229,
|
| 14 |
+
0.224,
|
| 15 |
+
0.225
|
| 16 |
+
],
|
| 17 |
+
"resample": 3,
|
| 18 |
+
"rescale_factor": 0.00392156862745098,
|
| 19 |
+
"size": {
|
| 20 |
+
"shortest_edge": 224
|
| 21 |
+
}
|
| 22 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a7f63ab17724f1d57b9bd3dd9bbdb1cd0a75f06094b5effe59455694f9ea46a9
|
| 3 |
+
size 94436045
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d394d4a45ce3245db19a9ce9c877074336b472c99b268171f1378e7ccaeb97ff
|
| 3 |
+
size 13553
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fd127ec90851335115fe0239761a0b4a7c9a5bf4294e284b968bf0fc255302c0
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,649 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8238888888888889,
|
| 3 |
+
"best_model_checkpoint": "/data/anyue/huggingface_project/resnet/my_resnet50_model/checkpoint-1011",
|
| 4 |
+
"epoch": 2.9955555555555557,
|
| 5 |
+
"global_step": 1011,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.03,
|
| 12 |
+
"learning_rate": 4.901960784313726e-06,
|
| 13 |
+
"loss": 2.3083,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.06,
|
| 18 |
+
"learning_rate": 9.803921568627451e-06,
|
| 19 |
+
"loss": 2.3087,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.09,
|
| 24 |
+
"learning_rate": 1.4705882352941177e-05,
|
| 25 |
+
"loss": 2.3024,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.12,
|
| 30 |
+
"learning_rate": 1.9607843137254903e-05,
|
| 31 |
+
"loss": 2.2955,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.15,
|
| 36 |
+
"learning_rate": 2.4509803921568626e-05,
|
| 37 |
+
"loss": 2.2913,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.18,
|
| 42 |
+
"learning_rate": 2.9411764705882354e-05,
|
| 43 |
+
"loss": 2.281,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.21,
|
| 48 |
+
"learning_rate": 3.431372549019608e-05,
|
| 49 |
+
"loss": 2.2739,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.24,
|
| 54 |
+
"learning_rate": 3.9215686274509805e-05,
|
| 55 |
+
"loss": 2.2618,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.27,
|
| 60 |
+
"learning_rate": 4.411764705882353e-05,
|
| 61 |
+
"loss": 2.2528,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.3,
|
| 66 |
+
"learning_rate": 4.901960784313725e-05,
|
| 67 |
+
"loss": 2.2287,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.33,
|
| 72 |
+
"learning_rate": 4.955995599559956e-05,
|
| 73 |
+
"loss": 2.2181,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.36,
|
| 78 |
+
"learning_rate": 4.9009900990099014e-05,
|
| 79 |
+
"loss": 2.1954,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.39,
|
| 84 |
+
"learning_rate": 4.845984598459846e-05,
|
| 85 |
+
"loss": 2.1613,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.41,
|
| 90 |
+
"learning_rate": 4.790979097909791e-05,
|
| 91 |
+
"loss": 2.1431,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.44,
|
| 96 |
+
"learning_rate": 4.7359735973597365e-05,
|
| 97 |
+
"loss": 2.1183,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.47,
|
| 102 |
+
"learning_rate": 4.680968096809682e-05,
|
| 103 |
+
"loss": 2.082,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.5,
|
| 108 |
+
"learning_rate": 4.6259625962596256e-05,
|
| 109 |
+
"loss": 2.0654,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.53,
|
| 114 |
+
"learning_rate": 4.570957095709571e-05,
|
| 115 |
+
"loss": 2.0345,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.56,
|
| 120 |
+
"learning_rate": 4.515951595159516e-05,
|
| 121 |
+
"loss": 2.0022,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.59,
|
| 126 |
+
"learning_rate": 4.4609460946094614e-05,
|
| 127 |
+
"loss": 1.9356,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.62,
|
| 132 |
+
"learning_rate": 4.405940594059406e-05,
|
| 133 |
+
"loss": 1.8859,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.65,
|
| 138 |
+
"learning_rate": 4.350935093509351e-05,
|
| 139 |
+
"loss": 1.8669,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.68,
|
| 144 |
+
"learning_rate": 4.2959295929592964e-05,
|
| 145 |
+
"loss": 1.8257,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.71,
|
| 150 |
+
"learning_rate": 4.240924092409242e-05,
|
| 151 |
+
"loss": 1.7847,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.74,
|
| 156 |
+
"learning_rate": 4.1859185918591856e-05,
|
| 157 |
+
"loss": 1.7329,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.77,
|
| 162 |
+
"learning_rate": 4.130913091309131e-05,
|
| 163 |
+
"loss": 1.6941,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.8,
|
| 168 |
+
"learning_rate": 4.075907590759076e-05,
|
| 169 |
+
"loss": 1.637,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.83,
|
| 174 |
+
"learning_rate": 4.020902090209021e-05,
|
| 175 |
+
"loss": 1.5448,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.86,
|
| 180 |
+
"learning_rate": 3.965896589658966e-05,
|
| 181 |
+
"loss": 1.5003,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.89,
|
| 186 |
+
"learning_rate": 3.910891089108911e-05,
|
| 187 |
+
"loss": 1.4605,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.92,
|
| 192 |
+
"learning_rate": 3.8558855885588564e-05,
|
| 193 |
+
"loss": 1.4562,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.95,
|
| 198 |
+
"learning_rate": 3.800880088008801e-05,
|
| 199 |
+
"loss": 1.4234,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.98,
|
| 204 |
+
"learning_rate": 3.745874587458746e-05,
|
| 205 |
+
"loss": 1.462,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.0,
|
| 210 |
+
"eval_accuracy": 0.6887037037037037,
|
| 211 |
+
"eval_loss": 1.2266517877578735,
|
| 212 |
+
"eval_runtime": 104.4528,
|
| 213 |
+
"eval_samples_per_second": 51.698,
|
| 214 |
+
"eval_steps_per_second": 3.236,
|
| 215 |
+
"step": 337
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 1.01,
|
| 219 |
+
"learning_rate": 3.690869086908691e-05,
|
| 220 |
+
"loss": 1.3493,
|
| 221 |
+
"step": 340
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.04,
|
| 225 |
+
"learning_rate": 3.635863586358636e-05,
|
| 226 |
+
"loss": 1.2632,
|
| 227 |
+
"step": 350
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 1.07,
|
| 231 |
+
"learning_rate": 3.580858085808581e-05,
|
| 232 |
+
"loss": 1.2919,
|
| 233 |
+
"step": 360
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.1,
|
| 237 |
+
"learning_rate": 3.525852585258526e-05,
|
| 238 |
+
"loss": 1.285,
|
| 239 |
+
"step": 370
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 1.13,
|
| 243 |
+
"learning_rate": 3.470847084708471e-05,
|
| 244 |
+
"loss": 1.2564,
|
| 245 |
+
"step": 380
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.16,
|
| 249 |
+
"learning_rate": 3.415841584158416e-05,
|
| 250 |
+
"loss": 1.2319,
|
| 251 |
+
"step": 390
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 1.19,
|
| 255 |
+
"learning_rate": 3.360836083608361e-05,
|
| 256 |
+
"loss": 1.2217,
|
| 257 |
+
"step": 400
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 1.21,
|
| 261 |
+
"learning_rate": 3.305830583058306e-05,
|
| 262 |
+
"loss": 1.1776,
|
| 263 |
+
"step": 410
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.24,
|
| 267 |
+
"learning_rate": 3.250825082508251e-05,
|
| 268 |
+
"loss": 1.2127,
|
| 269 |
+
"step": 420
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.27,
|
| 273 |
+
"learning_rate": 3.195819581958196e-05,
|
| 274 |
+
"loss": 1.1776,
|
| 275 |
+
"step": 430
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.3,
|
| 279 |
+
"learning_rate": 3.140814081408141e-05,
|
| 280 |
+
"loss": 1.1651,
|
| 281 |
+
"step": 440
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.33,
|
| 285 |
+
"learning_rate": 3.0858085808580864e-05,
|
| 286 |
+
"loss": 1.1678,
|
| 287 |
+
"step": 450
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.36,
|
| 291 |
+
"learning_rate": 3.0308030803080313e-05,
|
| 292 |
+
"loss": 1.1173,
|
| 293 |
+
"step": 460
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.39,
|
| 297 |
+
"learning_rate": 2.9757975797579756e-05,
|
| 298 |
+
"loss": 1.1142,
|
| 299 |
+
"step": 470
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.42,
|
| 303 |
+
"learning_rate": 2.9207920792079208e-05,
|
| 304 |
+
"loss": 1.0391,
|
| 305 |
+
"step": 480
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.45,
|
| 309 |
+
"learning_rate": 2.8657865786578657e-05,
|
| 310 |
+
"loss": 1.0514,
|
| 311 |
+
"step": 490
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.48,
|
| 315 |
+
"learning_rate": 2.810781078107811e-05,
|
| 316 |
+
"loss": 1.0391,
|
| 317 |
+
"step": 500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.51,
|
| 321 |
+
"learning_rate": 2.755775577557756e-05,
|
| 322 |
+
"loss": 1.0546,
|
| 323 |
+
"step": 510
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.54,
|
| 327 |
+
"learning_rate": 2.700770077007701e-05,
|
| 328 |
+
"loss": 1.0664,
|
| 329 |
+
"step": 520
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.57,
|
| 333 |
+
"learning_rate": 2.645764576457646e-05,
|
| 334 |
+
"loss": 0.9683,
|
| 335 |
+
"step": 530
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.6,
|
| 339 |
+
"learning_rate": 2.5907590759075913e-05,
|
| 340 |
+
"loss": 0.9964,
|
| 341 |
+
"step": 540
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.63,
|
| 345 |
+
"learning_rate": 2.5357535753575355e-05,
|
| 346 |
+
"loss": 0.9846,
|
| 347 |
+
"step": 550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.66,
|
| 351 |
+
"learning_rate": 2.480748074807481e-05,
|
| 352 |
+
"loss": 0.9348,
|
| 353 |
+
"step": 560
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.69,
|
| 357 |
+
"learning_rate": 2.4257425742574257e-05,
|
| 358 |
+
"loss": 0.9687,
|
| 359 |
+
"step": 570
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.72,
|
| 363 |
+
"learning_rate": 2.370737073707371e-05,
|
| 364 |
+
"loss": 0.9263,
|
| 365 |
+
"step": 580
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.75,
|
| 369 |
+
"learning_rate": 2.3157315731573158e-05,
|
| 370 |
+
"loss": 0.948,
|
| 371 |
+
"step": 590
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.78,
|
| 375 |
+
"learning_rate": 2.2607260726072607e-05,
|
| 376 |
+
"loss": 0.8746,
|
| 377 |
+
"step": 600
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.81,
|
| 381 |
+
"learning_rate": 2.2057205720572056e-05,
|
| 382 |
+
"loss": 1.0205,
|
| 383 |
+
"step": 610
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.84,
|
| 387 |
+
"learning_rate": 2.150715071507151e-05,
|
| 388 |
+
"loss": 0.9219,
|
| 389 |
+
"step": 620
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.87,
|
| 393 |
+
"learning_rate": 2.0957095709570958e-05,
|
| 394 |
+
"loss": 0.9119,
|
| 395 |
+
"step": 630
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.9,
|
| 399 |
+
"learning_rate": 2.0407040704070407e-05,
|
| 400 |
+
"loss": 0.9224,
|
| 401 |
+
"step": 640
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.93,
|
| 405 |
+
"learning_rate": 1.9856985698569856e-05,
|
| 406 |
+
"loss": 0.9249,
|
| 407 |
+
"step": 650
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 1.96,
|
| 411 |
+
"learning_rate": 1.930693069306931e-05,
|
| 412 |
+
"loss": 0.9007,
|
| 413 |
+
"step": 660
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.99,
|
| 417 |
+
"learning_rate": 1.8756875687568758e-05,
|
| 418 |
+
"loss": 0.8732,
|
| 419 |
+
"step": 670
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 2.0,
|
| 423 |
+
"eval_accuracy": 0.8009259259259259,
|
| 424 |
+
"eval_loss": 0.7092505693435669,
|
| 425 |
+
"eval_runtime": 108.4854,
|
| 426 |
+
"eval_samples_per_second": 49.776,
|
| 427 |
+
"eval_steps_per_second": 3.116,
|
| 428 |
+
"step": 675
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.01,
|
| 432 |
+
"learning_rate": 1.8206820682068207e-05,
|
| 433 |
+
"loss": 0.9019,
|
| 434 |
+
"step": 680
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.04,
|
| 438 |
+
"learning_rate": 1.765676567656766e-05,
|
| 439 |
+
"loss": 0.8636,
|
| 440 |
+
"step": 690
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.07,
|
| 444 |
+
"learning_rate": 1.7106710671067108e-05,
|
| 445 |
+
"loss": 0.8994,
|
| 446 |
+
"step": 700
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.1,
|
| 450 |
+
"learning_rate": 1.6556655665566557e-05,
|
| 451 |
+
"loss": 0.8992,
|
| 452 |
+
"step": 710
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.13,
|
| 456 |
+
"learning_rate": 1.6006600660066006e-05,
|
| 457 |
+
"loss": 0.8894,
|
| 458 |
+
"step": 720
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.16,
|
| 462 |
+
"learning_rate": 1.545654565456546e-05,
|
| 463 |
+
"loss": 0.904,
|
| 464 |
+
"step": 730
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.19,
|
| 468 |
+
"learning_rate": 1.4906490649064908e-05,
|
| 469 |
+
"loss": 0.7925,
|
| 470 |
+
"step": 740
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.22,
|
| 474 |
+
"learning_rate": 1.4356435643564355e-05,
|
| 475 |
+
"loss": 0.7882,
|
| 476 |
+
"step": 750
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.25,
|
| 480 |
+
"learning_rate": 1.3806380638063806e-05,
|
| 481 |
+
"loss": 0.9119,
|
| 482 |
+
"step": 760
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.28,
|
| 486 |
+
"learning_rate": 1.3256325632563257e-05,
|
| 487 |
+
"loss": 0.8652,
|
| 488 |
+
"step": 770
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.31,
|
| 492 |
+
"learning_rate": 1.2706270627062708e-05,
|
| 493 |
+
"loss": 0.8538,
|
| 494 |
+
"step": 780
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.34,
|
| 498 |
+
"learning_rate": 1.2156215621562157e-05,
|
| 499 |
+
"loss": 0.8371,
|
| 500 |
+
"step": 790
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.37,
|
| 504 |
+
"learning_rate": 1.1606160616061606e-05,
|
| 505 |
+
"loss": 0.8746,
|
| 506 |
+
"step": 800
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.4,
|
| 510 |
+
"learning_rate": 1.1056105610561057e-05,
|
| 511 |
+
"loss": 0.7919,
|
| 512 |
+
"step": 810
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 2.43,
|
| 516 |
+
"learning_rate": 1.0506050605060507e-05,
|
| 517 |
+
"loss": 0.7718,
|
| 518 |
+
"step": 820
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.46,
|
| 522 |
+
"learning_rate": 9.955995599559956e-06,
|
| 523 |
+
"loss": 0.8264,
|
| 524 |
+
"step": 830
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.49,
|
| 528 |
+
"learning_rate": 9.405940594059407e-06,
|
| 529 |
+
"loss": 0.8595,
|
| 530 |
+
"step": 840
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.52,
|
| 534 |
+
"learning_rate": 8.855885588558856e-06,
|
| 535 |
+
"loss": 0.8308,
|
| 536 |
+
"step": 850
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.55,
|
| 540 |
+
"learning_rate": 8.305830583058307e-06,
|
| 541 |
+
"loss": 0.8107,
|
| 542 |
+
"step": 860
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.58,
|
| 546 |
+
"learning_rate": 7.755775577557756e-06,
|
| 547 |
+
"loss": 0.782,
|
| 548 |
+
"step": 870
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.61,
|
| 552 |
+
"learning_rate": 7.205720572057206e-06,
|
| 553 |
+
"loss": 0.8989,
|
| 554 |
+
"step": 880
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.64,
|
| 558 |
+
"learning_rate": 6.655665566556656e-06,
|
| 559 |
+
"loss": 0.7427,
|
| 560 |
+
"step": 890
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.67,
|
| 564 |
+
"learning_rate": 6.105610561056106e-06,
|
| 565 |
+
"loss": 0.8377,
|
| 566 |
+
"step": 900
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.7,
|
| 570 |
+
"learning_rate": 5.555555555555556e-06,
|
| 571 |
+
"loss": 0.7893,
|
| 572 |
+
"step": 910
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.73,
|
| 576 |
+
"learning_rate": 5.005500550055006e-06,
|
| 577 |
+
"loss": 0.7916,
|
| 578 |
+
"step": 920
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 2.76,
|
| 582 |
+
"learning_rate": 4.455445544554456e-06,
|
| 583 |
+
"loss": 0.7662,
|
| 584 |
+
"step": 930
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.79,
|
| 588 |
+
"learning_rate": 3.905390539053906e-06,
|
| 589 |
+
"loss": 0.7687,
|
| 590 |
+
"step": 940
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.81,
|
| 594 |
+
"learning_rate": 3.3553355335533555e-06,
|
| 595 |
+
"loss": 0.8293,
|
| 596 |
+
"step": 950
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.84,
|
| 600 |
+
"learning_rate": 2.8052805280528055e-06,
|
| 601 |
+
"loss": 0.8026,
|
| 602 |
+
"step": 960
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 2.87,
|
| 606 |
+
"learning_rate": 2.2552255225522554e-06,
|
| 607 |
+
"loss": 0.8025,
|
| 608 |
+
"step": 970
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 2.9,
|
| 612 |
+
"learning_rate": 1.7051705170517053e-06,
|
| 613 |
+
"loss": 0.8072,
|
| 614 |
+
"step": 980
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 2.93,
|
| 618 |
+
"learning_rate": 1.155115511551155e-06,
|
| 619 |
+
"loss": 0.8195,
|
| 620 |
+
"step": 990
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.96,
|
| 624 |
+
"learning_rate": 6.050605060506051e-07,
|
| 625 |
+
"loss": 0.8394,
|
| 626 |
+
"step": 1000
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 2.99,
|
| 630 |
+
"learning_rate": 5.5005500550055006e-08,
|
| 631 |
+
"loss": 0.7706,
|
| 632 |
+
"step": 1010
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.0,
|
| 636 |
+
"eval_accuracy": 0.8238888888888889,
|
| 637 |
+
"eval_loss": 0.6203305125236511,
|
| 638 |
+
"eval_runtime": 102.309,
|
| 639 |
+
"eval_samples_per_second": 52.781,
|
| 640 |
+
"eval_steps_per_second": 3.304,
|
| 641 |
+
"step": 1011
|
| 642 |
+
}
|
| 643 |
+
],
|
| 644 |
+
"max_steps": 1011,
|
| 645 |
+
"num_train_epochs": 3,
|
| 646 |
+
"total_flos": 1.374973468635562e+18,
|
| 647 |
+
"trial_name": null,
|
| 648 |
+
"trial_params": null
|
| 649 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c01d9a12e75c620b61cdf661055530a109b297c41e61e06fde96acfa7bd783c
|
| 3 |
+
size 4027
|