emotion_recognition

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: 1.2014
• Accuracy: 0.6125

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0842	1.0	10	2.0668	0.175
2.039	2.0	20	2.0070	0.2875
1.9285	3.0	30	1.8789	0.4062
1.7699	4.0	40	1.6942	0.425
1.6135	5.0	50	1.5758	0.4313
1.5056	6.0	60	1.4884	0.55
1.3896	7.0	70	1.3999	0.5437
1.2804	8.0	80	1.3563	0.5437
1.2043	9.0	90	1.3244	0.55
1.1231	10.0	100	1.2775	0.6062
1.0652	11.0	110	1.2567	0.575
1.0005	12.0	120	1.2833	0.5563
0.9878	13.0	130	1.2277	0.5687
0.9714	14.0	140	1.2557	0.5563
0.9057	15.0	150	1.2187	0.6125
0.8854	16.0	160	1.2612	0.5437
0.8478	17.0	170	1.2450	0.5437
0.8601	18.0	180	1.2456	0.5375
0.8498	19.0	190	1.2413	0.5875
0.8775	20.0	200	1.1928	0.6

Framework versions

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