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-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: 0.6388888888888888

deit-base-distilled-patch16-224-hasta-65-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: 1.1109
• Accuracy: 0.6389

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	1.2683	0.2778
No log	1.7143	3	1.1169	0.3056
No log	2.8571	5	1.2963	0.2778
No log	4.0	7	1.2729	0.2778
No log	4.5714	8	1.1327	0.2778
1.1554	5.7143	10	1.0658	0.3889
1.1554	6.8571	12	1.1513	0.3611
1.1554	8.0	14	1.1799	0.3889
1.1554	8.5714	15	1.1289	0.3611
1.1554	9.7143	17	1.0167	0.4167
1.1554	10.8571	19	1.0074	0.5
0.9967	12.0	21	0.9982	0.5
0.9967	12.5714	22	0.9707	0.5278
0.9967	13.7143	24	0.9401	0.6111
0.9967	14.8571	26	1.0155	0.6111
0.9967	16.0	28	1.1357	0.4444
0.9967	16.5714	29	1.1284	0.5
0.8472	17.7143	31	1.1840	0.4167
0.8472	18.8571	33	1.3015	0.4444
0.8472	20.0	35	0.9755	0.5
0.8472	20.5714	36	0.9602	0.5278
0.8472	21.7143	38	1.0950	0.4444
0.7133	22.8571	40	1.0607	0.4722
0.7133	24.0	42	0.9963	0.5833
0.7133	24.5714	43	1.0235	0.5833
0.7133	25.7143	45	1.0872	0.5556
0.7133	26.8571	47	1.0526	0.5833
0.7133	28.0	49	1.1579	0.5278
0.5648	28.5714	50	1.2434	0.4444
0.5648	29.7143	52	1.1653	0.5556
0.5648	30.8571	54	1.0947	0.5556
0.5648	32.0	56	1.1531	0.5833
0.5648	32.5714	57	1.1268	0.5833
0.5648	33.7143	59	1.0664	0.5833
0.4429	34.8571	61	1.1503	0.5556
0.4429	36.0	63	1.3473	0.5
0.4429	36.5714	64	1.2786	0.5
0.4429	37.7143	66	1.0905	0.6111
0.4429	38.8571	68	1.0917	0.6111
0.4313	40.0	70	1.2079	0.5556
0.4313	40.5714	71	1.2501	0.5556
0.4313	41.7143	73	1.1789	0.6111
0.4313	42.8571	75	1.1126	0.5833
0.4313	44.0	77	1.1109	0.6389
0.4313	44.5714	78	1.1236	0.6389
0.3764	45.7143	80	1.2211	0.6111
0.3764	46.8571	82	1.3021	0.5278
0.3764	48.0	84	1.3182	0.5556
0.3764	48.5714	85	1.2830	0.5556
0.3764	49.7143	87	1.2503	0.5833
0.3764	50.8571	89	1.1926	0.6389
0.3441	52.0	91	1.2092	0.6111
0.3441	52.5714	92	1.2266	0.6111
0.3441	53.7143	94	1.2685	0.5278
0.3441	54.8571	96	1.3098	0.4722
0.3441	56.0	98	1.3069	0.4722
0.3441	56.5714	99	1.3021	0.4722
0.322	57.1429	100	1.3012	0.5

Framework versions

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