vit-base-patch16-224-ethos

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

• Loss: 0.2506
• Accuracy: 0.96

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8696	5	0.4608	0.87
0.5337	1.9130	11	0.2743	0.91
0.5337	2.9565	17	0.2239	0.94
0.2275	4.0	23	0.3780	0.88
0.2275	4.8696	28	0.3501	0.88
0.1107	5.9130	34	0.2420	0.92
0.0528	6.9565	40	0.2752	0.94
0.0528	8.0	46	0.3932	0.9
0.0465	8.8696	51	0.2496	0.94
0.0465	9.9130	57	0.3151	0.93
0.0516	10.9565	63	0.1837	0.96
0.0516	12.0	69	0.1885	0.95
0.0317	12.8696	74	0.3941	0.92
0.0463	13.9130	80	0.2577	0.95
0.0463	14.9565	86	0.2128	0.95
0.018	16.0	92	0.2342	0.96
0.018	16.8696	97	0.2483	0.96
0.0179	17.3913	100	0.2506	0.96

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1