vit-base-patch16-224-dmae-va-U5-20-45-5e-05

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

• Loss: 0.5800
• Accuracy: 0.8833

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: 5.5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9	7	1.3607	0.4833
1.3752	1.94	15	1.2550	0.5833
1.2809	2.97	23	1.1436	0.65
1.1273	4.0	31	1.0381	0.5833
1.1273	4.9	38	0.9620	0.6833
0.9919	5.94	46	0.9154	0.65
0.8971	6.97	54	0.8502	0.7667
0.8049	8.0	62	0.8644	0.75
0.8049	8.9	69	0.8010	0.7833
0.7119	9.94	77	0.7276	0.8333
0.6172	10.97	85	0.6699	0.8167
0.5294	12.0	93	0.6532	0.8167
0.4696	12.9	100	0.6265	0.85
0.4696	13.94	108	0.6012	0.85
0.4074	14.97	116	0.5800	0.8833
0.3822	16.0	124	0.5692	0.8667
0.3651	16.9	131	0.6065	0.8
0.3651	17.94	139	0.5681	0.8667
0.3731	18.06	140	0.5675	0.8667

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0