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-55-fold4
    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: 0.6666666666666666

deit-base-distilled-patch16-224-hasta-55-fold4

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: 1.0211
• Accuracy: 0.6667

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	1.1563	0.4167
No log	1.7143	3	1.1167	0.3611
No log	2.8571	5	1.3251	0.3611
No log	4.0	7	1.1401	0.3611
No log	4.5714	8	1.0346	0.5278
1.1327	5.7143	10	1.0377	0.4444
1.1327	6.8571	12	1.0117	0.5556
1.1327	8.0	14	1.0721	0.3889
1.1327	8.5714	15	1.0312	0.4167
1.1327	9.7143	17	1.0228	0.5
1.1327	10.8571	19	0.9868	0.5556
0.9623	12.0	21	1.1119	0.5
0.9623	12.5714	22	1.1094	0.5
0.9623	13.7143	24	0.9811	0.6389
0.9623	14.8571	26	1.0888	0.4722
0.9623	16.0	28	1.0576	0.5833
0.9623	16.5714	29	1.0583	0.6389
0.8158	17.7143	31	1.0731	0.5833
0.8158	18.8571	33	0.9776	0.5833
0.8158	20.0	35	1.0211	0.6667
0.8158	20.5714	36	1.0035	0.6111
0.8158	21.7143	38	0.9845	0.5556
0.653	22.8571	40	1.0856	0.6389
0.653	24.0	42	1.0787	0.6111
0.653	24.5714	43	1.0536	0.5833
0.653	25.7143	45	1.0995	0.5833
0.653	26.8571	47	1.1207	0.5833
0.653	28.0	49	1.1610	0.5833
0.5371	28.5714	50	1.1164	0.6111
0.5371	29.7143	52	1.1101	0.5833
0.5371	30.8571	54	1.1040	0.5556
0.5371	32.0	56	1.0995	0.5833
0.5371	32.5714	57	1.1076	0.5833
0.5371	33.7143	59	1.1530	0.5833
0.4039	34.8571	61	1.1780	0.5556
0.4039	36.0	63	1.1379	0.6111
0.4039	36.5714	64	1.1324	0.6111
0.4039	37.7143	66	1.1546	0.5556
0.4039	38.8571	68	1.1910	0.5556
0.348	40.0	70	1.1487	0.5833
0.348	40.5714	71	1.1367	0.5833
0.348	41.7143	73	1.1627	0.5833
0.348	42.8571	75	1.2421	0.5833
0.348	44.0	77	1.3073	0.5556
0.348	44.5714	78	1.2773	0.5556
0.321	45.7143	80	1.2160	0.5833
0.321	46.8571	82	1.2220	0.5833
0.321	48.0	84	1.2168	0.5833
0.321	48.5714	85	1.2190	0.5833
0.321	49.7143	87	1.2255	0.5833
0.321	50.8571	89	1.2174	0.5833
0.3168	52.0	91	1.2152	0.5833
0.3168	52.5714	92	1.2308	0.5833
0.3168	53.7143	94	1.2728	0.5556
0.3168	54.8571	96	1.3121	0.5556
0.3168	56.0	98	1.3350	0.5556
0.3168	56.5714	99	1.3354	0.5556
0.2876	57.1429	100	1.3346	0.5556

Framework versions

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