ViT_face

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

• Loss: 0.6941

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: 2e-05
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	10	1.0691
No log	2.0	20	1.0378
No log	3.0	30	0.9958
No log	4.0	40	0.9437
No log	5.0	50	0.8915
No log	6.0	60	0.8396
No log	7.0	70	0.7950
No log	8.0	80	0.7602
No log	9.0	90	0.7246
No log	10.0	100	0.7009
No log	11.0	110	0.6882
No log	12.0	120	0.6700
No log	13.0	130	0.6629
No log	14.0	140	0.6646
No log	15.0	150	0.6558
No log	16.0	160	0.6679
No log	17.0	170	0.6637
No log	18.0	180	0.6689
No log	19.0	190	0.6690
No log	20.0	200	0.6744
No log	21.0	210	0.6787
No log	22.0	220	0.6823
No log	23.0	230	0.6832
No log	24.0	240	0.6866
No log	25.0	250	0.6883
No log	26.0	260	0.6912
No log	27.0	270	0.6923
No log	28.0	280	0.6935
No log	29.0	290	0.6939
No log	30.0	300	0.6941

Framework versions

• Transformers 4.42.4
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1