deit-rice-disease

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: 1.1873
• Accuracy: 0.9490
• F1 Score: 0.9528
• Recall: 0.9596

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: 50
• mixed_precision_training: Native AMP
• label_smoothing_factor: 0.15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Score	Recall
2.7759	1.0	20	2.7706	0.3057	0.3020	0.3099
2.6344	2.0	40	2.6186	0.4777	0.4633	0.4620
2.3890	3.0	60	2.3909	0.6242	0.6270	0.6041
2.0561	4.0	80	2.0708	0.7452	0.7566	0.7454
1.6978	5.0	100	1.7234	0.8153	0.8236	0.8209
1.4285	6.0	120	1.5000	0.8408	0.8465	0.8518
1.2876	7.0	140	1.3515	0.8917	0.8952	0.9039
1.2463	8.0	160	1.3041	0.9108	0.9136	0.9222
1.1444	9.0	180	1.2831	0.9236	0.9256	0.9364
1.1086	10.0	200	1.2490	0.9299	0.9327	0.9425
1.0744	11.0	220	1.2496	0.9299	0.9327	0.9425
1.0598	12.0	240	1.2804	0.9299	0.9327	0.9425
1.0738	13.0	260	1.2100	0.9363	0.9398	0.9486
1.0744	14.0	280	1.2112	0.9299	0.9327	0.9405
1.0476	15.0	300	1.2396	0.9363	0.9398	0.9486
1.0307	16.0	320	1.1843	0.9363	0.9396	0.9461
1.0167	17.0	340	1.1528	0.9490	0.9513	0.9559
1.0089	18.0	360	1.1433	0.9554	0.9586	0.9620
1.0003	19.0	380	1.1530	0.9427	0.9442	0.9498
1.0012	20.0	400	1.1984	0.9299	0.9327	0.9400
0.9774	21.0	420	1.1554	0.9554	0.9584	0.9620
0.9777	22.0	440	1.1511	0.9490	0.9527	0.9571
0.9818	23.0	460	1.1955	0.9490	0.9528	0.9596
0.9811	24.0	480	1.2225	0.9427	0.9458	0.9515
0.9757	25.0	500	1.1872	0.9427	0.9457	0.9535
0.9845	26.0	520	1.2139	0.9363	0.9410	0.9498
0.9657	27.0	540	1.1985	0.9427	0.9469	0.9547
0.9737	28.0	560	1.1873	0.9490	0.9528	0.9596

Framework versions

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