vit-base-patch16-224-dmae-va-U5-100bcont

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5453
• Accuracy: 0.8667

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: 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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9	7	0.5397	0.85
0.3809	1.94	15	0.5212	0.85
0.316	2.97	23	0.5690	0.8
0.2892	4.0	31	0.6506	0.7667
0.2892	4.9	38	0.5529	0.8333
0.2127	5.94	46	0.4987	0.8333
0.1712	6.97	54	0.5859	0.8167
0.1539	8.0	62	0.5937	0.8
0.1539	8.9	69	0.5103	0.8333
0.1378	9.94	77	0.6844	0.7833
0.1144	10.97	85	0.5357	0.85
0.1055	12.0	93	0.6695	0.8
0.1093	12.9	100	0.5593	0.85
0.1093	13.94	108	0.5453	0.8667
0.0956	14.97	116	0.6144	0.85
0.1057	16.0	124	0.5067	0.8333
0.0907	16.9	131	0.6570	0.8
0.0907	17.94	139	0.5343	0.8667
0.1184	18.97	147	0.5516	0.8667
0.1014	20.0	155	0.8173	0.7667
0.0997	20.9	162	0.6839	0.8167
0.1067	21.94	170	0.5552	0.8667
0.1067	22.97	178	0.5475	0.8667
0.082	24.0	186	0.5567	0.85
0.0852	24.9	193	0.6374	0.8167
0.0815	25.94	201	0.6486	0.8167
0.0815	26.97	209	0.6218	0.8167
0.0917	27.1	210	0.6209	0.8167

Framework versions

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