image_classification

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.1140
• Accuracy: 0.6125

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.9952	1.0	10	2.0130	0.3063
1.9609	2.0	20	1.9619	0.3563
1.8939	3.0	30	1.8425	0.4188
1.7588	4.0	40	1.6837	0.45
1.6045	5.0	50	1.5389	0.4688
1.4959	6.0	60	1.4618	0.5062
1.3876	7.0	70	1.3693	0.5375
1.295	8.0	80	1.3286	0.575
1.2328	9.0	90	1.3112	0.5563
1.1447	10.0	100	1.2627	0.5813
1.0791	11.0	110	1.2462	0.5813
1.0378	12.0	120	1.2410	0.6
1.0013	13.0	130	1.2353	0.5687
0.9512	14.0	140	1.2324	0.5625
0.8505	15.0	150	1.2216	0.575
0.8193	16.0	160	1.2061	0.6
0.7379	17.0	170	1.1829	0.5563
0.7133	18.0	180	1.2131	0.5625
0.6582	19.0	190	1.1882	0.5625
0.6663	20.0	200	1.0910	0.6188
0.589	21.0	210	1.1769	0.5687
0.5865	22.0	220	1.1242	0.6375
0.5336	23.0	230	1.1933	0.5375
0.5168	24.0	240	1.1956	0.575
0.4937	25.0	250	1.1943	0.6
0.487	26.0	260	1.1298	0.575
0.4582	27.0	270	1.1004	0.6312
0.4611	28.0	280	1.1108	0.5875
0.4386	29.0	290	1.2242	0.5813
0.4255	30.0	300	1.1560	0.5875
0.4136	31.0	310	1.2545	0.5437
0.4204	32.0	320	1.1661	0.6125
0.3959	33.0	330	1.1248	0.5875
0.3661	34.0	340	1.1475	0.6062
0.3603	35.0	350	1.1463	0.6125
0.3617	36.0	360	1.2161	0.5563
0.3761	37.0	370	1.2575	0.5312
0.3452	38.0	380	1.1720	0.575
0.3665	39.0	390	1.1165	0.6
0.3471	40.0	400	1.2233	0.5375

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0