Hugging Face's logo

sharren
/
vit-lr-cosine-restarts

Image Classification
Transformers
TensorBoard
Safetensors
vit
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
sharren commited on
Commit
dd9342e
•
1 Parent(s): f94a46b

Model save

Browse files
Files changed (2) hide show
  1. README.md +25 -26
  2. model.safetensors +1 -1
README.md CHANGED
@@ -2,7 +2,6 @@
2
  license: apache-2.0
3
  base_model: google/vit-base-patch16-224
4
  tags:
5
- - image-classification
6
  - generated_from_trainer
7
  metrics:
8
  - accuracy
@@ -19,13 +18,13 @@ should probably proofread and complete it, then remove this comment. -->
19
 
20
  # vit-lr-cosine-restarts
21
 
22
- This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the skin-cancer dataset.
23
  It achieves the following results on the evaluation set:
24
- - Loss: 0.4524
25
- - Accuracy: 0.8350
26
- - Precision: 0.8390
27
- - Recall: 0.8350
28
- - F1: 0.8329
29
 
30
  ## Model description
31
 
@@ -50,7 +49,7 @@ The following hyperparameters were used during training:
50
  - seed: 42
51
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
52
  - lr_scheduler_type: cosine_with_restarts
53
- - lr_scheduler_warmup_steps: 1600
54
  - num_epochs: 100
55
  - mixed_precision_training: Native AMP
56
 
@@ -58,24 +57,24 @@ The following hyperparameters were used during training:
58
 
59
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
60
  |:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
61
- | 0.9572 | 0.31 | 100 | 0.9378 | 0.6786 | 0.5347 | 0.6786 | 0.5768 |
62
- | 0.7576 | 0.62 | 200 | 0.7868 | 0.7171 | 0.6469 | 0.7171 | 0.6482 |
63
- | 0.793 | 0.93 | 300 | 0.6203 | 0.7812 | 0.7718 | 0.7812 | 0.7641 |
64
- | 0.4895 | 1.25 | 400 | 0.5442 | 0.7982 | 0.7915 | 0.7982 | 0.7908 |
65
- | 0.416 | 1.56 | 500 | 0.5408 | 0.8110 | 0.8223 | 0.8110 | 0.7957 |
66
- | 0.4463 | 1.87 | 600 | 0.6205 | 0.7656 | 0.8404 | 0.7656 | 0.7763 |
67
- | 0.206 | 2.18 | 700 | 0.4994 | 0.8304 | 0.8348 | 0.8304 | 0.8244 |
68
- | 0.4006 | 2.49 | 800 | 0.4524 | 0.8350 | 0.8390 | 0.8350 | 0.8329 |
69
- | 0.3208 | 2.8 | 900 | 0.5084 | 0.8291 | 0.8511 | 0.8291 | 0.8357 |
70
- | 0.1916 | 3.12 | 1000 | 0.5120 | 0.8200 | 0.8465 | 0.8200 | 0.8255 |
71
- | 0.2015 | 3.43 | 1100 | 0.6912 | 0.8183 | 0.8453 | 0.8183 | 0.7983 |
72
- | 0.2384 | 3.74 | 1200 | 0.7052 | 0.8131 | 0.8165 | 0.8131 | 0.8046 |
73
- | 0.1694 | 4.05 | 1300 | 0.4923 | 0.8308 | 0.8425 | 0.8308 | 0.8345 |
74
- | 0.2445 | 4.36 | 1400 | 0.5869 | 0.8308 | 0.8472 | 0.8308 | 0.8344 |
75
- | 0.1757 | 4.67 | 1500 | 0.6699 | 0.8259 | 0.8380 | 0.8259 | 0.8236 |
76
- | 0.2443 | 4.98 | 1600 | 0.7501 | 0.8031 | 0.8172 | 0.8031 | 0.7857 |
77
- | 0.1498 | 5.3 | 1700 | 0.8649 | 0.7972 | 0.8395 | 0.7972 | 0.8016 |
78
- | 0.1072 | 5.61 | 1800 | 0.6284 | 0.8336 | 0.8511 | 0.8336 | 0.8386 |
79
 
80
 
81
  ### Framework versions
 
2
  license: apache-2.0
3
  base_model: google/vit-base-patch16-224
4
  tags:
 
5
  - generated_from_trainer
6
  metrics:
7
  - accuracy
 
18
 
19
  # vit-lr-cosine-restarts
20
 
21
+ This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on an unknown dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 0.7405
24
+ - Accuracy: 0.8533
25
+ - Precision: 0.8523
26
+ - Recall: 0.8533
27
+ - F1: 0.8511
28
 
29
  ## Model description
30
 
 
49
  - seed: 42
50
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
51
  - lr_scheduler_type: cosine_with_restarts
52
+ - lr_scheduler_warmup_steps: 80
53
  - num_epochs: 100
54
  - mixed_precision_training: Native AMP
55
 
 
57
 
58
  | Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
59
  |:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
60
+ | 0.686 | 0.31 | 100 | 0.6707 | 0.7517 | 0.7624 | 0.7517 | 0.7445 |
61
+ | 0.4852 | 0.62 | 200 | 0.7022 | 0.7705 | 0.7858 | 0.7705 | 0.7231 |
62
+ | 0.7098 | 0.93 | 300 | 0.5637 | 0.7996 | 0.8181 | 0.7996 | 0.7973 |
63
+ | 0.4226 | 1.25 | 400 | 0.6494 | 0.7621 | 0.8137 | 0.7621 | 0.7735 |
64
+ | 0.3599 | 1.56 | 500 | 0.5214 | 0.8235 | 0.8207 | 0.8235 | 0.8109 |
65
+ | 0.3533 | 1.87 | 600 | 0.5347 | 0.8273 | 0.8392 | 0.8273 | 0.8193 |
66
+ | 0.1178 | 2.18 | 700 | 0.5425 | 0.8284 | 0.8381 | 0.8284 | 0.8277 |
67
+ | 0.2719 | 2.49 | 800 | 0.4453 | 0.8464 | 0.8464 | 0.8464 | 0.8438 |
68
+ | 0.1559 | 2.8 | 900 | 0.6127 | 0.8325 | 0.8567 | 0.8325 | 0.8284 |
69
+ | 0.1328 | 3.12 | 1000 | 0.5303 | 0.8509 | 0.8456 | 0.8509 | 0.8451 |
70
+ | 0.1756 | 3.43 | 1100 | 0.7960 | 0.8322 | 0.8366 | 0.8322 | 0.8151 |
71
+ | 0.3582 | 3.74 | 1200 | 0.6676 | 0.8343 | 0.8284 | 0.8343 | 0.8249 |
72
+ | 0.025 | 4.05 | 1300 | 0.5981 | 0.8474 | 0.8599 | 0.8474 | 0.8477 |
73
+ | 0.042 | 4.36 | 1400 | 0.8096 | 0.8162 | 0.8477 | 0.8162 | 0.8241 |
74
+ | 0.05 | 4.67 | 1500 | 0.7948 | 0.8419 | 0.8474 | 0.8419 | 0.8341 |
75
+ | 0.028 | 4.98 | 1600 | 0.6742 | 0.8457 | 0.8558 | 0.8457 | 0.8476 |
76
+ | 0.0048 | 5.3 | 1700 | 0.7833 | 0.8485 | 0.8577 | 0.8485 | 0.8500 |
77
+ | 0.0324 | 5.61 | 1800 | 0.7405 | 0.8533 | 0.8523 | 0.8533 | 0.8511 |
78
 
79
 
80
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e9d88d8814bed30222e210f8724c2230f64875a6f032746207b05127c0a742cb
3
  size 343239356
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d46ee26556ebc3162f0cc0a986c8df85ac9ef6bd790b99ab2e838c25e42b9e2b
3
  size 343239356