emotion_recog

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

• Loss: 1.2377
• Accuracy: 0.5375

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	10	1.6772	0.4062
No log	2.0	20	1.5802	0.4437
No log	3.0	30	1.4877	0.4875
No log	4.0	40	1.4649	0.475
No log	5.0	50	1.4092	0.5
No log	6.0	60	1.3454	0.5188
No log	7.0	70	1.3469	0.5312
No log	8.0	80	1.3010	0.5375
No log	9.0	90	1.2688	0.5563
No log	10.0	100	1.2854	0.5563
No log	11.0	110	1.2516	0.5437
No log	12.0	120	1.2819	0.5312
No log	13.0	130	1.2228	0.5875
No log	14.0	140	1.2250	0.5813
No log	15.0	150	1.2177	0.5563
No log	16.0	160	1.2172	0.55
No log	17.0	170	1.2198	0.6
No log	18.0	180	1.2341	0.5563
No log	19.0	190	1.2206	0.6
No log	20.0	200	1.1635	0.5813

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1