vit-pretraining-2024_04_02-atelectasis-classifier

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

• Loss: 0.5020
• Accuracy: 0.7644

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
0.6304	1.0	537	0.6342	0.6709
0.5931	2.0	1074	0.5669	0.7207
0.5027	3.0	1611	0.5397	0.7393
0.5659	4.0	2148	0.5341	0.7458
0.5115	5.0	2685	0.5433	0.7346
0.5108	6.0	3222	0.5454	0.7309
0.5187	7.0	3759	0.5136	0.7621
0.4435	8.0	4296	0.5057	0.7677
0.583	9.0	4833	0.5042	0.7584
0.5256	10.0	5370	0.5249	0.7495
0.4818	11.0	5907	0.5212	0.7481
0.5575	12.0	6444	0.5061	0.7481
0.3572	13.0	6981	0.5042	0.7602
0.489	14.0	7518	0.5004	0.7709
0.4773	15.0	8055	0.5074	0.7700
0.4577	16.0	8592	0.5054	0.7677
0.4619	17.0	9129	0.5021	0.7686
0.3865	18.0	9666	0.5074	0.7644
0.4889	19.0	10203	0.5113	0.7598
0.4637	20.0	10740	0.5020	0.7644

Framework versions

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