categorAI_img

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: 0.7080
• Accuracy: 0.8378

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: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9091	5	1.8872	0.3784
7.7979	1.9091	10	1.7777	0.6419
7.7979	2.9091	15	1.6224	0.6622
6.9519	3.9091	20	1.4667	0.6959
6.9519	4.9091	25	1.3353	0.7365
5.7562	5.9091	30	1.2522	0.7703
5.7562	6.9091	35	1.1617	0.7838
4.7446	7.9091	40	1.0967	0.7635
4.7446	8.9091	45	1.0362	0.7568
4.0655	9.9091	50	0.9349	0.8108
4.0655	10.9091	55	0.9393	0.7905
3.5041	11.9091	60	0.8859	0.7838
3.5041	12.9091	65	0.9039	0.7770
3.0788	13.9091	70	0.8123	0.8041
3.0788	14.9091	75	0.7946	0.8243
2.7461	15.9091	80	0.8003	0.8311
2.7461	16.9091	85	0.8101	0.7703
2.4988	17.9091	90	0.7111	0.8176
2.4988	18.9091	95	0.7439	0.8243
2.3122	19.9091	100	0.7542	0.7905
2.3122	20.9091	105	0.7323	0.8311
2.3408	21.9091	110	0.7175	0.8243
2.3408	22.9091	115	0.7652	0.8041
2.2846	23.9091	120	0.7211	0.8176
2.2846	24.9091	125	0.7080	0.8378

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1.post306
• Datasets 3.2.0
• Tokenizers 0.21.0