vit-pretraining-2024_03_25-classifier

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

• Loss: 0.5083
• Accuracy: 0.7649

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-06
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6422	1.0	537	0.6409	0.6560
0.5509	2.0	1074	0.5966	0.6862
0.5123	3.0	1611	0.5743	0.7044
0.5237	4.0	2148	0.5523	0.7188
0.5589	5.0	2685	0.5352	0.7370
0.5671	6.0	3222	0.5317	0.7407
0.5247	7.0	3759	0.5228	0.7486
0.4855	8.0	4296	0.5422	0.7374
0.5122	9.0	4833	0.5195	0.7477
0.5381	10.0	5370	0.5277	0.7398
0.5465	11.0	5907	0.5213	0.7514
0.4552	12.0	6444	0.5300	0.7495
0.5188	13.0	6981	0.5107	0.7505
0.5056	14.0	7518	0.5075	0.7579
0.4759	15.0	8055	0.5077	0.7644
0.6042	16.0	8592	0.5143	0.7602
0.4002	17.0	9129	0.5184	0.7612
0.4664	18.0	9666	0.5072	0.7630
0.4653	19.0	10203	0.5103	0.7626
0.4096	20.0	10740	0.5083	0.7649

Framework versions

• Transformers 4.39.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2