README.md

metadata

library_name: transformers
license: apache-2.0
base_model: google/vit-base-patch16-224
tags:
  - generated_from_trainer
datasets:
  - webdataset
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: frost-vision-v2-google_vit-base-patch16-224-v2024-11-14
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: webdataset
          type: webdataset
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9388888888888889
          - name: F1
            type: f1
            value: 0.8436018957345972
          - name: Precision
            type: precision
            value: 0.8654781199351702
          - name: Recall
            type: recall
            value: 0.8228043143297381

frost-vision-v2-google_vit-base-patch16-224-v2024-11-14

This model is a fine-tuned version of google/vit-base-patch16-224 on the webdataset dataset. It achieves the following results on the evaluation set:

• Loss: 0.1577
• Accuracy: 0.9389
• F1: 0.8436
• Precision: 0.8655
• Recall: 0.8228

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.3381	1.2346	100	0.3271	0.8660	0.5669	0.8045	0.4376
0.2067	2.4691	200	0.2080	0.9194	0.7827	0.8514	0.7242
0.1745	3.7037	300	0.1864	0.9228	0.8003	0.8308	0.7720
0.1724	4.9383	400	0.1792	0.9299	0.8188	0.8493	0.7904
0.128	6.1728	500	0.1736	0.9327	0.8292	0.8437	0.8151
0.1034	7.4074	600	0.1672	0.9355	0.8348	0.8571	0.8136
0.0944	8.6420	700	0.1579	0.9392	0.8452	0.8622	0.8290
0.0919	9.8765	800	0.1631	0.9364	0.8347	0.8710	0.8012
0.0791	11.1111	900	0.1592	0.9380	0.8383	0.8771	0.8028
0.0684	12.3457	1000	0.1577	0.9389	0.8436	0.8655	0.8228
0.0737	13.5802	1100	0.1678	0.9380	0.8416	0.8613	0.8228
0.0625	14.8148	1200	0.1646	0.9426	0.8542	0.8692	0.8398
0.0591	16.0494	1300	0.1625	0.9432	0.8549	0.8756	0.8351
0.0464	17.2840	1400	0.1722	0.9386	0.8422	0.8676	0.8182
0.048	18.5185	1500	0.1694	0.9401	0.8472	0.8663	0.8290
0.0353	19.7531	1600	0.1715	0.9392	0.8462	0.8576	0.8351
0.0434	20.9877	1700	0.1817	0.9370	0.8386	0.8618	0.8166
0.0332	22.2222	1800	0.1797	0.9383	0.8423	0.8627	0.8228
0.0283	23.4568	1900	0.1810	0.9401	0.8482	0.8617	0.8351
0.0474	24.6914	2000	0.1765	0.9398	0.8454	0.8709	0.8213
0.0365	25.9259	2100	0.1835	0.9414	0.8516	0.8637	0.8398
0.0244	27.1605	2200	0.1822	0.9404	0.8479	0.8677	0.8290
0.0242	28.3951	2300	0.1808	0.9407	0.8483	0.8703	0.8274
0.0296	29.6296	2400	0.1817	0.9401	0.8477	0.864	0.8320

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.0+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3