better evaluation and only trained the classifier layer
Browse files- README.md +11 -10
- all_results.json +5 -5
- config.json +70 -70
- model.safetensors +1 -1
- runs/Mar31_10-44-13_0de03c5233fc/events.out.tfevents.1711881913.0de03c5233fc.34.0 +3 -0
- runs/Mar31_10-44-13_0de03c5233fc/events.out.tfevents.1711882353.0de03c5233fc.34.1 +3 -0
- train_results.json +5 -5
- trainer_state.json +148 -300
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the pcuenq/oxford-pets dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss: 0.
|
| 22 |
-
- Accuracy: 0.
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
@@ -39,27 +39,28 @@ More information needed
|
|
| 39 |
|
| 40 |
The following hyperparameters were used during training:
|
| 41 |
- learning_rate: 0.0003
|
| 42 |
-
- train_batch_size:
|
| 43 |
-
- eval_batch_size:
|
| 44 |
- seed: 42
|
| 45 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
- lr_scheduler_type: linear
|
| 47 |
-
- num_epochs:
|
| 48 |
- mixed_precision_training: Native AMP
|
| 49 |
|
| 50 |
### Training results
|
| 51 |
|
| 52 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 53 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 54 |
-
|
|
| 55 |
-
| 0.
|
| 56 |
-
| 0.
|
| 57 |
-
| 0.
|
|
|
|
| 58 |
|
| 59 |
|
| 60 |
### Framework versions
|
| 61 |
|
| 62 |
- Transformers 4.39.2
|
| 63 |
-
- Pytorch 2.2
|
| 64 |
- Datasets 2.16.0
|
| 65 |
- Tokenizers 0.15.2
|
|
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the pcuenq/oxford-pets dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.3168
|
| 22 |
+
- Accuracy: 0.9432
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
|
|
| 39 |
|
| 40 |
The following hyperparameters were used during training:
|
| 41 |
- learning_rate: 0.0003
|
| 42 |
+
- train_batch_size: 128
|
| 43 |
+
- eval_batch_size: 16
|
| 44 |
- seed: 42
|
| 45 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
- lr_scheduler_type: linear
|
| 47 |
+
- num_epochs: 5
|
| 48 |
- mixed_precision_training: Native AMP
|
| 49 |
|
| 50 |
### Training results
|
| 51 |
|
| 52 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 53 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 54 |
+
| 1.5136 | 1.0 | 47 | 1.1031 | 0.8430 |
|
| 55 |
+
| 0.5547 | 2.0 | 94 | 0.5232 | 0.9269 |
|
| 56 |
+
| 0.4111 | 3.0 | 141 | 0.3988 | 0.9310 |
|
| 57 |
+
| 0.3438 | 4.0 | 188 | 0.3553 | 0.9337 |
|
| 58 |
+
| 0.298 | 5.0 | 235 | 0.3448 | 0.9296 |
|
| 59 |
|
| 60 |
|
| 61 |
### Framework versions
|
| 62 |
|
| 63 |
- Transformers 4.39.2
|
| 64 |
+
- Pytorch 2.1.2
|
| 65 |
- Datasets 2.16.0
|
| 66 |
- Tokenizers 0.15.2
|
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
"total_flos": 2.2913817801515827e+18,
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
"total_flos": 2.2913817801515827e+18,
|
| 4 |
+
"train_loss": 0.8009341437765892,
|
| 5 |
+
"train_runtime": 407.3253,
|
| 6 |
+
"train_samples_per_second": 72.571,
|
| 7 |
+
"train_steps_per_second": 0.577
|
| 8 |
}
|
config.json
CHANGED
|
@@ -9,85 +9,85 @@
|
|
| 9 |
"hidden_dropout_prob": 0.0,
|
| 10 |
"hidden_size": 768,
|
| 11 |
"id2label": {
|
| 12 |
-
"0": "
|
| 13 |
-
"1": "
|
| 14 |
-
"10": "
|
| 15 |
-
"11": "
|
| 16 |
-
"12": "
|
| 17 |
-
"13": "
|
| 18 |
-
"14": "
|
| 19 |
-
"15": "
|
| 20 |
-
"16": "
|
| 21 |
-
"17": "
|
| 22 |
-
"18": "
|
| 23 |
-
"19": "
|
| 24 |
-
"2": "
|
| 25 |
-
"20": "
|
| 26 |
-
"21": "
|
| 27 |
-
"22": "
|
| 28 |
-
"23": "
|
| 29 |
-
"24": "
|
| 30 |
-
"25": "
|
| 31 |
-
"26": "
|
| 32 |
-
"27": "
|
| 33 |
-
"28": "
|
| 34 |
-
"29": "
|
| 35 |
-
"3": "
|
| 36 |
-
"30": "
|
| 37 |
-
"31": "
|
| 38 |
-
"32": "
|
| 39 |
-
"33": "
|
| 40 |
-
"34": "
|
| 41 |
-
"35": "
|
| 42 |
"36": "boxer dog",
|
| 43 |
-
"4": "
|
| 44 |
-
"5": "
|
| 45 |
-
"6": "
|
| 46 |
-
"7": "
|
| 47 |
"8": "British Shorthair cat",
|
| 48 |
-
"9": "
|
| 49 |
},
|
| 50 |
"image_size": 224,
|
| 51 |
"initializer_range": 0.02,
|
| 52 |
"intermediate_size": 3072,
|
| 53 |
"label2id": {
|
| 54 |
-
"Abyssinian cat": "
|
| 55 |
-
"Bengal cat": "
|
| 56 |
-
"Birman cat": "
|
| 57 |
-
"Bombay cat": "
|
| 58 |
"British Shorthair cat": "8",
|
| 59 |
-
"Egyptian Mau cat": "
|
| 60 |
-
"Maine Coon cat": "
|
| 61 |
-
"Persian cat": "
|
| 62 |
-
"Ragdoll cat": "
|
| 63 |
-
"Russian Blue cat": "
|
| 64 |
-
"Siamese cat": "
|
| 65 |
-
"Sphynx cat": "
|
| 66 |
-
"american bulldog dog": "
|
| 67 |
-
"american pit bull terrier dog": "
|
| 68 |
-
"basset hound dog": "
|
| 69 |
-
"beagle dog": "
|
| 70 |
"boxer dog": "36",
|
| 71 |
-
"chihuahua dog": "
|
| 72 |
-
"english cocker spaniel dog": "
|
| 73 |
-
"english setter dog": "
|
| 74 |
-
"german shorthaired dog": "
|
| 75 |
-
"great pyrenees dog": "
|
| 76 |
-
"havanese dog": "
|
| 77 |
-
"japanese chin dog": "
|
| 78 |
-
"keeshond dog": "
|
| 79 |
-
"leonberger dog": "
|
| 80 |
-
"miniature pinscher dog": "
|
| 81 |
-
"newfoundland dog": "
|
| 82 |
-
"pomeranian dog": "
|
| 83 |
-
"pug dog": "
|
| 84 |
-
"saint bernard dog": "
|
| 85 |
-
"samoyed dog": "
|
| 86 |
-
"scottish terrier dog": "
|
| 87 |
-
"shiba inu dog": "
|
| 88 |
-
"staffordshire bull terrier dog": "
|
| 89 |
-
"wheaten terrier dog": "
|
| 90 |
-
"yorkshire terrier dog": "
|
| 91 |
},
|
| 92 |
"layer_norm_eps": 1e-12,
|
| 93 |
"model_type": "vit",
|
|
|
|
| 9 |
"hidden_dropout_prob": 0.0,
|
| 10 |
"hidden_size": 768,
|
| 11 |
"id2label": {
|
| 12 |
+
"0": "american pit bull terrier dog",
|
| 13 |
+
"1": "Maine Coon cat",
|
| 14 |
+
"10": "german shorthaired dog",
|
| 15 |
+
"11": "Abyssinian cat",
|
| 16 |
+
"12": "havanese dog",
|
| 17 |
+
"13": "pomeranian dog",
|
| 18 |
+
"14": "newfoundland dog",
|
| 19 |
+
"15": "Egyptian Mau cat",
|
| 20 |
+
"16": "Russian Blue cat",
|
| 21 |
+
"17": "Bombay cat",
|
| 22 |
+
"18": "leonberger dog",
|
| 23 |
+
"19": "Persian cat",
|
| 24 |
+
"2": "english setter dog",
|
| 25 |
+
"20": "samoyed dog",
|
| 26 |
+
"21": "keeshond dog",
|
| 27 |
+
"22": "scottish terrier dog",
|
| 28 |
+
"23": "shiba inu dog",
|
| 29 |
+
"24": "staffordshire bull terrier dog",
|
| 30 |
+
"25": "basset hound dog",
|
| 31 |
+
"26": "Siamese cat",
|
| 32 |
+
"27": "pug dog",
|
| 33 |
+
"28": "miniature pinscher dog",
|
| 34 |
+
"29": "japanese chin dog",
|
| 35 |
+
"3": "Birman cat",
|
| 36 |
+
"30": "chihuahua dog",
|
| 37 |
+
"31": "Bengal cat",
|
| 38 |
+
"32": "beagle dog",
|
| 39 |
+
"33": "Sphynx cat",
|
| 40 |
+
"34": "saint bernard dog",
|
| 41 |
+
"35": "Ragdoll cat",
|
| 42 |
"36": "boxer dog",
|
| 43 |
+
"4": "wheaten terrier dog",
|
| 44 |
+
"5": "great pyrenees dog",
|
| 45 |
+
"6": "american bulldog dog",
|
| 46 |
+
"7": "english cocker spaniel dog",
|
| 47 |
"8": "British Shorthair cat",
|
| 48 |
+
"9": "yorkshire terrier dog"
|
| 49 |
},
|
| 50 |
"image_size": 224,
|
| 51 |
"initializer_range": 0.02,
|
| 52 |
"intermediate_size": 3072,
|
| 53 |
"label2id": {
|
| 54 |
+
"Abyssinian cat": "11",
|
| 55 |
+
"Bengal cat": "31",
|
| 56 |
+
"Birman cat": "3",
|
| 57 |
+
"Bombay cat": "17",
|
| 58 |
"British Shorthair cat": "8",
|
| 59 |
+
"Egyptian Mau cat": "15",
|
| 60 |
+
"Maine Coon cat": "1",
|
| 61 |
+
"Persian cat": "19",
|
| 62 |
+
"Ragdoll cat": "35",
|
| 63 |
+
"Russian Blue cat": "16",
|
| 64 |
+
"Siamese cat": "26",
|
| 65 |
+
"Sphynx cat": "33",
|
| 66 |
+
"american bulldog dog": "6",
|
| 67 |
+
"american pit bull terrier dog": "0",
|
| 68 |
+
"basset hound dog": "25",
|
| 69 |
+
"beagle dog": "32",
|
| 70 |
"boxer dog": "36",
|
| 71 |
+
"chihuahua dog": "30",
|
| 72 |
+
"english cocker spaniel dog": "7",
|
| 73 |
+
"english setter dog": "2",
|
| 74 |
+
"german shorthaired dog": "10",
|
| 75 |
+
"great pyrenees dog": "5",
|
| 76 |
+
"havanese dog": "12",
|
| 77 |
+
"japanese chin dog": "29",
|
| 78 |
+
"keeshond dog": "21",
|
| 79 |
+
"leonberger dog": "18",
|
| 80 |
+
"miniature pinscher dog": "28",
|
| 81 |
+
"newfoundland dog": "14",
|
| 82 |
+
"pomeranian dog": "13",
|
| 83 |
+
"pug dog": "27",
|
| 84 |
+
"saint bernard dog": "34",
|
| 85 |
+
"samoyed dog": "20",
|
| 86 |
+
"scottish terrier dog": "22",
|
| 87 |
+
"shiba inu dog": "23",
|
| 88 |
+
"staffordshire bull terrier dog": "24",
|
| 89 |
+
"wheaten terrier dog": "4",
|
| 90 |
+
"yorkshire terrier dog": "9"
|
| 91 |
},
|
| 92 |
"layer_norm_eps": 1e-12,
|
| 93 |
"model_type": "vit",
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 343331644
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c481d24458fdc1fd579d745131513f9125c8e3b8fb0c33d566356465f8370611
|
| 3 |
size 343331644
|
runs/Mar31_10-44-13_0de03c5233fc/events.out.tfevents.1711881913.0de03c5233fc.34.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:04377611248f9ad9357d199307a6d9002f0980bb18705e47089639b077f7b8a5
|
| 3 |
+
size 13763
|
runs/Mar31_10-44-13_0de03c5233fc/events.out.tfevents.1711882353.0de03c5233fc.34.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a81816eb4f226e038038b0ed56ffb97533dac0ae477cba379146a719849d54c7
|
| 3 |
+
size 411
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
"total_flos": 2.2913817801515827e+18,
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
"total_flos": 2.2913817801515827e+18,
|
| 4 |
+
"train_loss": 0.8009341437765892,
|
| 5 |
+
"train_runtime": 407.3253,
|
| 6 |
+
"train_samples_per_second": 72.571,
|
| 7 |
+
"train_steps_per_second": 0.577
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,388 +1,236 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "./vit-base-pets/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
-
"eval_steps":
|
| 6 |
-
"global_step":
|
| 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.
|
| 13 |
-
"grad_norm":
|
| 14 |
-
"learning_rate": 0.
|
| 15 |
-
"loss":
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
-
"epoch": 0.
|
| 20 |
-
"grad_norm":
|
| 21 |
-
"learning_rate": 0.
|
| 22 |
-
"loss":
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"grad_norm":
|
| 28 |
-
"learning_rate": 0.
|
| 29 |
-
"loss":
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
-
"epoch": 0.
|
| 34 |
-
"grad_norm":
|
| 35 |
-
"learning_rate": 0.
|
| 36 |
-
"loss":
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
-
"epoch": 0
|
| 41 |
-
"
|
| 42 |
-
"
|
| 43 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
"step": 50
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch":
|
| 48 |
-
"grad_norm":
|
| 49 |
-
"learning_rate": 0.
|
| 50 |
-
"loss": 0.
|
| 51 |
"step": 60
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch":
|
| 55 |
-
"grad_norm":
|
| 56 |
-
"learning_rate": 0.
|
| 57 |
-
"loss": 0.
|
| 58 |
"step": 70
|
| 59 |
},
|
| 60 |
{
|
| 61 |
-
"epoch":
|
| 62 |
-
"grad_norm":
|
| 63 |
-
"learning_rate": 0.
|
| 64 |
-
"loss": 0.
|
| 65 |
"step": 80
|
| 66 |
},
|
| 67 |
{
|
| 68 |
-
"epoch":
|
| 69 |
-
"grad_norm":
|
| 70 |
-
"learning_rate": 0.
|
| 71 |
-
"loss": 0.
|
| 72 |
"step": 90
|
| 73 |
},
|
| 74 |
{
|
| 75 |
-
"epoch": 0
|
| 76 |
-
"
|
| 77 |
-
"
|
| 78 |
-
"
|
| 79 |
-
"
|
|
|
|
|
|
|
| 80 |
},
|
| 81 |
{
|
| 82 |
-
"epoch":
|
| 83 |
-
"
|
| 84 |
-
"
|
| 85 |
-
"
|
| 86 |
-
"eval_samples_per_second": 108.346,
|
| 87 |
-
"eval_steps_per_second": 13.547,
|
| 88 |
"step": 100
|
| 89 |
},
|
| 90 |
{
|
| 91 |
-
"epoch":
|
| 92 |
-
"grad_norm":
|
| 93 |
-
"learning_rate": 0.
|
| 94 |
-
"loss": 0.
|
| 95 |
"step": 110
|
| 96 |
},
|
| 97 |
{
|
| 98 |
-
"epoch":
|
| 99 |
-
"grad_norm":
|
| 100 |
-
"learning_rate": 0.
|
| 101 |
-
"loss": 0.
|
| 102 |
"step": 120
|
| 103 |
},
|
| 104 |
{
|
| 105 |
-
"epoch":
|
| 106 |
-
"grad_norm":
|
| 107 |
-
"learning_rate": 0.
|
| 108 |
-
"loss": 0.
|
| 109 |
"step": 130
|
| 110 |
},
|
| 111 |
{
|
| 112 |
-
"epoch":
|
| 113 |
-
"grad_norm":
|
| 114 |
-
"learning_rate": 0.
|
| 115 |
-
"loss": 0.
|
| 116 |
"step": 140
|
| 117 |
},
|
| 118 |
{
|
| 119 |
-
"epoch":
|
| 120 |
-
"
|
| 121 |
-
"
|
| 122 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 123 |
"step": 150
|
| 124 |
},
|
| 125 |
{
|
| 126 |
-
"epoch":
|
| 127 |
-
"grad_norm":
|
| 128 |
-
"learning_rate":
|
| 129 |
-
"loss": 0.
|
| 130 |
"step": 160
|
| 131 |
},
|
| 132 |
{
|
| 133 |
-
"epoch":
|
| 134 |
-
"grad_norm":
|
| 135 |
-
"learning_rate":
|
| 136 |
-
"loss": 0.
|
| 137 |
"step": 170
|
| 138 |
},
|
| 139 |
{
|
| 140 |
-
"epoch":
|
| 141 |
-
"grad_norm":
|
| 142 |
-
"learning_rate":
|
| 143 |
-
"loss": 0.
|
| 144 |
"step": 180
|
| 145 |
},
|
| 146 |
{
|
| 147 |
-
"epoch":
|
| 148 |
-
"
|
| 149 |
-
"
|
| 150 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 151 |
"step": 190
|
| 152 |
},
|
| 153 |
{
|
| 154 |
-
"epoch":
|
| 155 |
-
"grad_norm":
|
| 156 |
-
"learning_rate":
|
| 157 |
-
"loss": 0.
|
| 158 |
-
"step": 200
|
| 159 |
-
},
|
| 160 |
-
{
|
| 161 |
-
"epoch": 1.72,
|
| 162 |
-
"eval_accuracy": 0.976319350473613,
|
| 163 |
-
"eval_loss": 0.07739943265914917,
|
| 164 |
-
"eval_runtime": 67.8967,
|
| 165 |
-
"eval_samples_per_second": 108.842,
|
| 166 |
-
"eval_steps_per_second": 13.609,
|
| 167 |
"step": 200
|
| 168 |
},
|
| 169 |
{
|
| 170 |
-
"epoch":
|
| 171 |
-
"grad_norm":
|
| 172 |
-
"learning_rate":
|
| 173 |
-
"loss": 0.
|
| 174 |
"step": 210
|
| 175 |
},
|
| 176 |
{
|
| 177 |
-
"epoch":
|
| 178 |
-
"grad_norm":
|
| 179 |
-
"learning_rate":
|
| 180 |
-
"loss": 0.
|
| 181 |
"step": 220
|
| 182 |
},
|
| 183 |
{
|
| 184 |
-
"epoch":
|
| 185 |
-
"grad_norm":
|
| 186 |
-
"learning_rate":
|
| 187 |
-
"loss": 0.
|
| 188 |
"step": 230
|
| 189 |
},
|
| 190 |
{
|
| 191 |
-
"epoch":
|
| 192 |
-
"
|
| 193 |
-
"
|
| 194 |
-
"
|
| 195 |
-
"
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
"epoch": 2.16,
|
| 199 |
-
"grad_norm": 2.4121060371398926,
|
| 200 |
-
"learning_rate": 0.00013836206896551724,
|
| 201 |
-
"loss": 0.0715,
|
| 202 |
-
"step": 250
|
| 203 |
-
},
|
| 204 |
-
{
|
| 205 |
-
"epoch": 2.24,
|
| 206 |
-
"grad_norm": 0.9658297300338745,
|
| 207 |
-
"learning_rate": 0.00013189655172413792,
|
| 208 |
-
"loss": 0.0444,
|
| 209 |
-
"step": 260
|
| 210 |
-
},
|
| 211 |
-
{
|
| 212 |
-
"epoch": 2.33,
|
| 213 |
-
"grad_norm": 0.24860858917236328,
|
| 214 |
-
"learning_rate": 0.0001254310344827586,
|
| 215 |
-
"loss": 0.0668,
|
| 216 |
-
"step": 270
|
| 217 |
-
},
|
| 218 |
-
{
|
| 219 |
-
"epoch": 2.41,
|
| 220 |
-
"grad_norm": 1.50627601146698,
|
| 221 |
-
"learning_rate": 0.00011896551724137931,
|
| 222 |
-
"loss": 0.034,
|
| 223 |
-
"step": 280
|
| 224 |
-
},
|
| 225 |
-
{
|
| 226 |
-
"epoch": 2.5,
|
| 227 |
-
"grad_norm": 1.2053415775299072,
|
| 228 |
-
"learning_rate": 0.0001125,
|
| 229 |
-
"loss": 0.0489,
|
| 230 |
-
"step": 290
|
| 231 |
-
},
|
| 232 |
-
{
|
| 233 |
-
"epoch": 2.59,
|
| 234 |
-
"grad_norm": 1.2974027395248413,
|
| 235 |
-
"learning_rate": 0.00010603448275862067,
|
| 236 |
-
"loss": 0.0612,
|
| 237 |
-
"step": 300
|
| 238 |
-
},
|
| 239 |
-
{
|
| 240 |
-
"epoch": 2.59,
|
| 241 |
-
"eval_accuracy": 0.9947225981055481,
|
| 242 |
-
"eval_loss": 0.021214015781879425,
|
| 243 |
-
"eval_runtime": 67.3034,
|
| 244 |
-
"eval_samples_per_second": 109.801,
|
| 245 |
-
"eval_steps_per_second": 13.729,
|
| 246 |
-
"step": 300
|
| 247 |
-
},
|
| 248 |
-
{
|
| 249 |
-
"epoch": 2.67,
|
| 250 |
-
"grad_norm": 0.22053079307079315,
|
| 251 |
-
"learning_rate": 9.956896551724137e-05,
|
| 252 |
-
"loss": 0.0308,
|
| 253 |
-
"step": 310
|
| 254 |
-
},
|
| 255 |
-
{
|
| 256 |
-
"epoch": 2.76,
|
| 257 |
-
"grad_norm": 0.8180058002471924,
|
| 258 |
-
"learning_rate": 9.310344827586206e-05,
|
| 259 |
-
"loss": 0.03,
|
| 260 |
-
"step": 320
|
| 261 |
-
},
|
| 262 |
-
{
|
| 263 |
-
"epoch": 2.84,
|
| 264 |
-
"grad_norm": 1.5855587720870972,
|
| 265 |
-
"learning_rate": 8.663793103448275e-05,
|
| 266 |
-
"loss": 0.0235,
|
| 267 |
-
"step": 330
|
| 268 |
-
},
|
| 269 |
-
{
|
| 270 |
-
"epoch": 2.93,
|
| 271 |
-
"grad_norm": 1.7537671327590942,
|
| 272 |
-
"learning_rate": 8.017241379310344e-05,
|
| 273 |
-
"loss": 0.0225,
|
| 274 |
-
"step": 340
|
| 275 |
-
},
|
| 276 |
-
{
|
| 277 |
-
"epoch": 3.02,
|
| 278 |
-
"grad_norm": 0.10956920683383942,
|
| 279 |
-
"learning_rate": 7.370689655172413e-05,
|
| 280 |
-
"loss": 0.0156,
|
| 281 |
-
"step": 350
|
| 282 |
-
},
|
| 283 |
-
{
|
| 284 |
-
"epoch": 3.1,
|
| 285 |
-
"grad_norm": 1.1361974477767944,
|
| 286 |
-
"learning_rate": 6.724137931034483e-05,
|
| 287 |
-
"loss": 0.0094,
|
| 288 |
-
"step": 360
|
| 289 |
-
},
|
| 290 |
-
{
|
| 291 |
-
"epoch": 3.19,
|
| 292 |
-
"grad_norm": 0.9139267802238464,
|
| 293 |
-
"learning_rate": 6.077586206896551e-05,
|
| 294 |
-
"loss": 0.0128,
|
| 295 |
-
"step": 370
|
| 296 |
-
},
|
| 297 |
-
{
|
| 298 |
-
"epoch": 3.28,
|
| 299 |
-
"grad_norm": 0.12278908491134644,
|
| 300 |
-
"learning_rate": 5.4310344827586204e-05,
|
| 301 |
-
"loss": 0.0087,
|
| 302 |
-
"step": 380
|
| 303 |
-
},
|
| 304 |
-
{
|
| 305 |
-
"epoch": 3.36,
|
| 306 |
-
"grad_norm": 0.0675448328256607,
|
| 307 |
-
"learning_rate": 4.78448275862069e-05,
|
| 308 |
-
"loss": 0.0048,
|
| 309 |
-
"step": 390
|
| 310 |
-
},
|
| 311 |
-
{
|
| 312 |
-
"epoch": 3.45,
|
| 313 |
-
"grad_norm": 0.05717047303915024,
|
| 314 |
-
"learning_rate": 4.137931034482758e-05,
|
| 315 |
-
"loss": 0.007,
|
| 316 |
-
"step": 400
|
| 317 |
},
|
| 318 |
{
|
| 319 |
-
"epoch":
|
| 320 |
-
"
|
| 321 |
-
"eval_loss": 0.005833666305989027,
|
| 322 |
-
"eval_runtime": 68.7353,
|
| 323 |
-
"eval_samples_per_second": 107.514,
|
| 324 |
-
"eval_steps_per_second": 13.443,
|
| 325 |
-
"step": 400
|
| 326 |
-
},
|
| 327 |
-
{
|
| 328 |
-
"epoch": 3.53,
|
| 329 |
-
"grad_norm": 0.04572397843003273,
|
| 330 |
-
"learning_rate": 3.4913793103448275e-05,
|
| 331 |
-
"loss": 0.0051,
|
| 332 |
-
"step": 410
|
| 333 |
-
},
|
| 334 |
-
{
|
| 335 |
-
"epoch": 3.62,
|
| 336 |
-
"grad_norm": 0.16931650042533875,
|
| 337 |
-
"learning_rate": 2.8448275862068963e-05,
|
| 338 |
-
"loss": 0.006,
|
| 339 |
-
"step": 420
|
| 340 |
-
},
|
| 341 |
-
{
|
| 342 |
-
"epoch": 3.71,
|
| 343 |
-
"grad_norm": 0.041402418166399,
|
| 344 |
-
"learning_rate": 2.198275862068965e-05,
|
| 345 |
-
"loss": 0.0054,
|
| 346 |
-
"step": 430
|
| 347 |
-
},
|
| 348 |
-
{
|
| 349 |
-
"epoch": 3.79,
|
| 350 |
-
"grad_norm": 0.02362515777349472,
|
| 351 |
-
"learning_rate": 1.5517241379310342e-05,
|
| 352 |
-
"loss": 0.0086,
|
| 353 |
-
"step": 440
|
| 354 |
-
},
|
| 355 |
-
{
|
| 356 |
-
"epoch": 3.88,
|
| 357 |
-
"grad_norm": 0.3355587422847748,
|
| 358 |
-
"learning_rate": 9.051724137931034e-06,
|
| 359 |
-
"loss": 0.0039,
|
| 360 |
-
"step": 450
|
| 361 |
-
},
|
| 362 |
-
{
|
| 363 |
-
"epoch": 3.97,
|
| 364 |
-
"grad_norm": 0.03504301235079765,
|
| 365 |
-
"learning_rate": 2.5862068965517237e-06,
|
| 366 |
-
"loss": 0.0135,
|
| 367 |
-
"step": 460
|
| 368 |
-
},
|
| 369 |
-
{
|
| 370 |
-
"epoch": 4.0,
|
| 371 |
-
"step": 464,
|
| 372 |
"total_flos": 2.2913817801515827e+18,
|
| 373 |
-
"train_loss": 0.
|
| 374 |
-
"train_runtime":
|
| 375 |
-
"train_samples_per_second":
|
| 376 |
-
"train_steps_per_second": 0.
|
| 377 |
}
|
| 378 |
],
|
| 379 |
"logging_steps": 10,
|
| 380 |
-
"max_steps":
|
| 381 |
"num_input_tokens_seen": 0,
|
| 382 |
-
"num_train_epochs":
|
| 383 |
-
"save_steps":
|
| 384 |
"total_flos": 2.2913817801515827e+18,
|
| 385 |
-
"train_batch_size":
|
| 386 |
"trial_name": null,
|
| 387 |
"trial_params": null
|
| 388 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.344835102558136,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-pets/checkpoint-235",
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 235,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.21,
|
| 13 |
+
"grad_norm": 216203.125,
|
| 14 |
+
"learning_rate": 0.0002872340425531915,
|
| 15 |
+
"loss": 3.3311,
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
+
"epoch": 0.43,
|
| 20 |
+
"grad_norm": 198156.9375,
|
| 21 |
+
"learning_rate": 0.000274468085106383,
|
| 22 |
+
"loss": 2.5921,
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.64,
|
| 27 |
+
"grad_norm": 176661.6875,
|
| 28 |
+
"learning_rate": 0.0002617021276595745,
|
| 29 |
+
"loss": 1.9823,
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
+
"epoch": 0.85,
|
| 34 |
+
"grad_norm": 142275.75,
|
| 35 |
+
"learning_rate": 0.0002489361702127659,
|
| 36 |
+
"loss": 1.5136,
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
+
"epoch": 1.0,
|
| 41 |
+
"eval_accuracy": 0.8430311231393776,
|
| 42 |
+
"eval_loss": 1.1030857563018799,
|
| 43 |
+
"eval_runtime": 8.7542,
|
| 44 |
+
"eval_samples_per_second": 84.417,
|
| 45 |
+
"eval_steps_per_second": 5.369,
|
| 46 |
+
"step": 47
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 1.06,
|
| 50 |
+
"grad_norm": 124899.421875,
|
| 51 |
+
"learning_rate": 0.00023617021276595742,
|
| 52 |
+
"loss": 1.1858,
|
| 53 |
"step": 50
|
| 54 |
},
|
| 55 |
{
|
| 56 |
+
"epoch": 1.28,
|
| 57 |
+
"grad_norm": 98946.875,
|
| 58 |
+
"learning_rate": 0.0002234042553191489,
|
| 59 |
+
"loss": 0.9313,
|
| 60 |
"step": 60
|
| 61 |
},
|
| 62 |
{
|
| 63 |
+
"epoch": 1.49,
|
| 64 |
+
"grad_norm": 92924.6484375,
|
| 65 |
+
"learning_rate": 0.0002106382978723404,
|
| 66 |
+
"loss": 0.7466,
|
| 67 |
"step": 70
|
| 68 |
},
|
| 69 |
{
|
| 70 |
+
"epoch": 1.7,
|
| 71 |
+
"grad_norm": 90032.1484375,
|
| 72 |
+
"learning_rate": 0.00019787234042553187,
|
| 73 |
+
"loss": 0.6475,
|
| 74 |
"step": 80
|
| 75 |
},
|
| 76 |
{
|
| 77 |
+
"epoch": 1.91,
|
| 78 |
+
"grad_norm": 68696.1875,
|
| 79 |
+
"learning_rate": 0.0001851063829787234,
|
| 80 |
+
"loss": 0.5547,
|
| 81 |
"step": 90
|
| 82 |
},
|
| 83 |
{
|
| 84 |
+
"epoch": 2.0,
|
| 85 |
+
"eval_accuracy": 0.9269282814614344,
|
| 86 |
+
"eval_loss": 0.5232290625572205,
|
| 87 |
+
"eval_runtime": 8.9185,
|
| 88 |
+
"eval_samples_per_second": 82.861,
|
| 89 |
+
"eval_steps_per_second": 5.27,
|
| 90 |
+
"step": 94
|
| 91 |
},
|
| 92 |
{
|
| 93 |
+
"epoch": 2.13,
|
| 94 |
+
"grad_norm": 67699.609375,
|
| 95 |
+
"learning_rate": 0.0001723404255319149,
|
| 96 |
+
"loss": 0.5311,
|
|
|
|
|
|
|
| 97 |
"step": 100
|
| 98 |
},
|
| 99 |
{
|
| 100 |
+
"epoch": 2.34,
|
| 101 |
+
"grad_norm": 72000.0234375,
|
| 102 |
+
"learning_rate": 0.00015957446808510637,
|
| 103 |
+
"loss": 0.4636,
|
| 104 |
"step": 110
|
| 105 |
},
|
| 106 |
{
|
| 107 |
+
"epoch": 2.55,
|
| 108 |
+
"grad_norm": 54618.05078125,
|
| 109 |
+
"learning_rate": 0.00014680851063829785,
|
| 110 |
+
"loss": 0.4171,
|
| 111 |
"step": 120
|
| 112 |
},
|
| 113 |
{
|
| 114 |
+
"epoch": 2.77,
|
| 115 |
+
"grad_norm": 57285.890625,
|
| 116 |
+
"learning_rate": 0.00013404255319148935,
|
| 117 |
+
"loss": 0.3946,
|
| 118 |
"step": 130
|
| 119 |
},
|
| 120 |
{
|
| 121 |
+
"epoch": 2.98,
|
| 122 |
+
"grad_norm": 73116.6171875,
|
| 123 |
+
"learning_rate": 0.00012127659574468084,
|
| 124 |
+
"loss": 0.4111,
|
| 125 |
"step": 140
|
| 126 |
},
|
| 127 |
{
|
| 128 |
+
"epoch": 3.0,
|
| 129 |
+
"eval_accuracy": 0.9309878213802436,
|
| 130 |
+
"eval_loss": 0.39878538250923157,
|
| 131 |
+
"eval_runtime": 9.1048,
|
| 132 |
+
"eval_samples_per_second": 81.166,
|
| 133 |
+
"eval_steps_per_second": 5.162,
|
| 134 |
+
"step": 141
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.19,
|
| 138 |
+
"grad_norm": 64126.3828125,
|
| 139 |
+
"learning_rate": 0.00010851063829787234,
|
| 140 |
+
"loss": 0.3607,
|
| 141 |
"step": 150
|
| 142 |
},
|
| 143 |
{
|
| 144 |
+
"epoch": 3.4,
|
| 145 |
+
"grad_norm": 69913.5390625,
|
| 146 |
+
"learning_rate": 9.574468085106382e-05,
|
| 147 |
+
"loss": 0.3387,
|
| 148 |
"step": 160
|
| 149 |
},
|
| 150 |
{
|
| 151 |
+
"epoch": 3.62,
|
| 152 |
+
"grad_norm": 54300.03125,
|
| 153 |
+
"learning_rate": 8.297872340425531e-05,
|
| 154 |
+
"loss": 0.3568,
|
| 155 |
"step": 170
|
| 156 |
},
|
| 157 |
{
|
| 158 |
+
"epoch": 3.83,
|
| 159 |
+
"grad_norm": 60929.75390625,
|
| 160 |
+
"learning_rate": 7.02127659574468e-05,
|
| 161 |
+
"loss": 0.3438,
|
| 162 |
"step": 180
|
| 163 |
},
|
| 164 |
{
|
| 165 |
+
"epoch": 4.0,
|
| 166 |
+
"eval_accuracy": 0.9336941813261164,
|
| 167 |
+
"eval_loss": 0.35527506470680237,
|
| 168 |
+
"eval_runtime": 9.4286,
|
| 169 |
+
"eval_samples_per_second": 78.378,
|
| 170 |
+
"eval_steps_per_second": 4.985,
|
| 171 |
+
"step": 188
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 4.04,
|
| 175 |
+
"grad_norm": 69279.328125,
|
| 176 |
+
"learning_rate": 5.7446808510638294e-05,
|
| 177 |
+
"loss": 0.3087,
|
| 178 |
"step": 190
|
| 179 |
},
|
| 180 |
{
|
| 181 |
+
"epoch": 4.26,
|
| 182 |
+
"grad_norm": 49415.69140625,
|
| 183 |
+
"learning_rate": 4.468085106382978e-05,
|
| 184 |
+
"loss": 0.328,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 185 |
"step": 200
|
| 186 |
},
|
| 187 |
{
|
| 188 |
+
"epoch": 4.47,
|
| 189 |
+
"grad_norm": 62788.3359375,
|
| 190 |
+
"learning_rate": 3.1914893617021275e-05,
|
| 191 |
+
"loss": 0.3199,
|
| 192 |
"step": 210
|
| 193 |
},
|
| 194 |
{
|
| 195 |
+
"epoch": 4.68,
|
| 196 |
+
"grad_norm": 61852.7421875,
|
| 197 |
+
"learning_rate": 1.9148936170212762e-05,
|
| 198 |
+
"loss": 0.3244,
|
| 199 |
"step": 220
|
| 200 |
},
|
| 201 |
{
|
| 202 |
+
"epoch": 4.89,
|
| 203 |
+
"grad_norm": 64687.453125,
|
| 204 |
+
"learning_rate": 6.382978723404255e-06,
|
| 205 |
+
"loss": 0.298,
|
| 206 |
"step": 230
|
| 207 |
},
|
| 208 |
{
|
| 209 |
+
"epoch": 5.0,
|
| 210 |
+
"eval_accuracy": 0.9296346414073072,
|
| 211 |
+
"eval_loss": 0.344835102558136,
|
| 212 |
+
"eval_runtime": 9.1923,
|
| 213 |
+
"eval_samples_per_second": 80.393,
|
| 214 |
+
"eval_steps_per_second": 5.113,
|
| 215 |
+
"step": 235
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 216 |
},
|
| 217 |
{
|
| 218 |
+
"epoch": 5.0,
|
| 219 |
+
"step": 235,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 220 |
"total_flos": 2.2913817801515827e+18,
|
| 221 |
+
"train_loss": 0.8009341437765892,
|
| 222 |
+
"train_runtime": 407.3253,
|
| 223 |
+
"train_samples_per_second": 72.571,
|
| 224 |
+
"train_steps_per_second": 0.577
|
| 225 |
}
|
| 226 |
],
|
| 227 |
"logging_steps": 10,
|
| 228 |
+
"max_steps": 235,
|
| 229 |
"num_input_tokens_seen": 0,
|
| 230 |
+
"num_train_epochs": 5,
|
| 231 |
+
"save_steps": 500,
|
| 232 |
"total_flos": 2.2913817801515827e+18,
|
| 233 |
+
"train_batch_size": 128,
|
| 234 |
"trial_name": null,
|
| 235 |
"trial_params": null
|
| 236 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4920
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33d4c76de48bc1cfcdacd9758735c59ce75dd9bf02657f219a30684fa44e0d63
|
| 3 |
size 4920
|