README.md

metadata

library_name: transformers
license: apache-2.0
base_model: facebook/deit-small-patch16-224
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
  - recall
model-index:
  - name: deit_fold_5
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: None
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9619377162629758
          - name: Recall
            type: recall
            value: 0.9645564195827585

deit_fold_5

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

• Loss: 1.1404
• Accuracy: 0.9619
• F1 Score: 0.9640
• Recall: 0.9646

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: 1e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 150
• num_epochs: 100
• label_smoothing_factor: 0.15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Score	Recall
2.8530	1.0	19	2.8472	0.2457	0.1618	0.2585
2.7569	2.0	38	2.6803	0.3460	0.2744	0.3182
2.5095	3.0	57	2.4824	0.5329	0.4757	0.4751
2.2538	4.0	76	2.2623	0.6436	0.6377	0.6053
1.8808	5.0	95	1.9165	0.7855	0.7984	0.7868
1.5094	6.0	114	1.5722	0.8720	0.8806	0.8797
1.2994	7.0	133	1.3893	0.8927	0.8983	0.8981
1.1850	8.0	152	1.3009	0.9135	0.9173	0.9189
1.0969	9.0	171	1.2608	0.9239	0.9300	0.9365
1.0729	10.0	190	1.2305	0.9377	0.9417	0.9437
1.0772	11.0	209	1.2142	0.9412	0.9445	0.9474
1.0385	12.0	228	1.1961	0.9481	0.9508	0.9524
1.0451	13.0	247	1.1773	0.9481	0.9514	0.9524
1.0382	14.0	266	1.1736	0.9516	0.9550	0.9561
1.0027	15.0	285	1.1668	0.9446	0.9486	0.9512
1.0166	16.0	304	1.1624	0.9446	0.9489	0.9525
1.0117	17.0	323	1.1576	0.9481	0.9506	0.9511
0.9923	18.0	342	1.1607	0.9446	0.9481	0.9486
0.9858	19.0	361	1.1551	0.9446	0.9489	0.9499
0.9924	20.0	380	1.1926	0.9377	0.9414	0.9384
0.9967	21.0	399	1.1682	0.9481	0.9507	0.9496
0.9727	22.0	418	1.1736	0.9446	0.9482	0.9461
0.9704	23.0	437	1.1639	0.9516	0.9544	0.9522
0.9733	24.0	456	1.1616	0.9481	0.9516	0.9536
0.9721	25.0	475	1.1662	0.9446	0.9482	0.9461
0.9680	26.0	494	1.1546	0.9481	0.9516	0.9511
0.9649	27.0	513	1.1476	0.9550	0.9582	0.9571
0.9737	28.0	532	1.1391	0.9550	0.9584	0.9584
0.9751	29.0	551	1.1496	0.9550	0.9580	0.9558
0.9702	30.0	570	1.1402	0.9550	0.9581	0.9597
0.9657	31.0	589	1.1368	0.9585	0.9610	0.9621
0.9674	32.0	608	1.1334	0.9550	0.9581	0.9597
0.9623	33.0	627	1.1573	0.9550	0.9574	0.9558
0.9718	34.0	646	1.1462	0.9585	0.9606	0.9595
0.9699	35.0	665	1.1404	0.9619	0.9640	0.9646
0.9699	36.0	684	1.1588	0.9516	0.9551	0.9572
0.9673	37.0	703	1.1423	0.9585	0.9615	0.9621
0.9611	38.0	722	1.1642	0.9481	0.9518	0.9485
0.9566	39.0	741	1.1492	0.9550	0.9580	0.9585
0.9627	40.0	760	1.1508	0.9585	0.9610	0.9621
0.9546	41.0	779	1.1408	0.9619	0.9640	0.9646
0.9604	42.0	798	1.1421	0.9619	0.9640	0.9646

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 5.0.0
• Tokenizers 0.22.2