README.md

metadata

license: apache-2.0
base_model: facebook/deit-base-distilled-patch16-224
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: deit-base-distilled-patch16-224-hasta-65-fold5
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5555555555555556

deit-base-distilled-patch16-224-hasta-65-fold5

This model is a fine-tuned version of facebook/deit-base-distilled-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.0163
• Accuracy: 0.5556

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.5714	1	1.1751	0.3889
No log	1.7143	3	1.0341	0.4722
No log	2.8571	5	1.3059	0.2778
No log	4.0	7	1.3255	0.2778
No log	4.5714	8	1.1834	0.2778
1.1839	5.7143	10	1.0357	0.5278
1.1839	6.8571	12	1.0608	0.3889
1.1839	8.0	14	1.2060	0.3333
1.1839	8.5714	15	1.1938	0.3889
1.1839	9.7143	17	1.0825	0.5
1.1839	10.8571	19	1.1488	0.3889
0.9707	12.0	21	1.1268	0.3889
0.9707	12.5714	22	1.0563	0.5
0.9707	13.7143	24	1.0570	0.5278
0.9707	14.8571	26	1.1166	0.4167
0.9707	16.0	28	1.0609	0.4444
0.9707	16.5714	29	1.0379	0.4722
0.8668	17.7143	31	1.0610	0.4444
0.8668	18.8571	33	1.1811	0.4167
0.8668	20.0	35	1.1028	0.4444
0.8668	20.5714	36	1.0950	0.4444
0.8668	21.7143	38	1.1424	0.4722
0.6889	22.8571	40	1.3027	0.4167
0.6889	24.0	42	1.2030	0.4167
0.6889	24.5714	43	1.2148	0.4167
0.6889	25.7143	45	1.3066	0.4167
0.6889	26.8571	47	1.3881	0.3611
0.6889	28.0	49	1.2566	0.4444
0.576	28.5714	50	1.1891	0.4444
0.576	29.7143	52	1.1638	0.4167
0.576	30.8571	54	1.2530	0.4167
0.576	32.0	56	1.1383	0.5
0.576	32.5714	57	1.0968	0.5
0.576	33.7143	59	1.0163	0.5556
0.4773	34.8571	61	1.1107	0.5
0.4773	36.0	63	1.1341	0.5
0.4773	36.5714	64	1.1152	0.5278
0.4773	37.7143	66	1.1158	0.5556
0.4773	38.8571	68	1.1628	0.4722
0.4186	40.0	70	1.2305	0.4444
0.4186	40.5714	71	1.2181	0.4722
0.4186	41.7143	73	1.2164	0.5
0.4186	42.8571	75	1.2225	0.5
0.4186	44.0	77	1.2298	0.5
0.4186	44.5714	78	1.2651	0.4722
0.3318	45.7143	80	1.3628	0.4167
0.3318	46.8571	82	1.3817	0.4167
0.3318	48.0	84	1.3594	0.4167
0.3318	48.5714	85	1.3553	0.4444
0.3318	49.7143	87	1.3548	0.4167
0.3318	50.8571	89	1.4113	0.4167
0.344	52.0	91	1.4433	0.4167
0.344	52.5714	92	1.4449	0.4167
0.344	53.7143	94	1.4514	0.4167
0.344	54.8571	96	1.4685	0.4167
0.344	56.0	98	1.4734	0.4167
0.344	56.5714	99	1.4747	0.4167
0.3305	57.1429	100	1.4732	0.4167

Framework versions

• Transformers 4.41.0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1