classifier-posterior-glare-removal

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the classifier_posterior_glare_removal_256_crop_s1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4990
• Accuracy: 0.8593

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: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.626	0.8065	50	0.5622	0.7582
0.4848	1.6129	100	0.5952	0.6675
0.2195	2.4194	150	0.5258	0.8325
0.1967	3.2258	200	0.5911	0.7960
0.2945	4.0323	250	0.4966	0.8300
0.1866	4.8387	300	0.5222	0.8350
0.1211	5.6452	350	0.5328	0.8426
0.1666	6.4516	400	0.5545	0.8426
0.0737	7.2581	450	0.5327	0.8526
0.0314	8.0645	500	0.5208	0.8526
0.0329	8.8710	550	0.5773	0.8489
0.0497	9.6774	600	0.5994	0.8489

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1