emotion_classification

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: 1.2493
• Accuracy: 0.5687

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
2.0679	1.0	10	2.0574	0.175
2.0366	2.0	20	2.0083	0.2812
1.9469	3.0	30	1.9119	0.35
1.8166	4.0	40	1.7702	0.4125
1.6821	5.0	50	1.6176	0.45
1.5587	6.0	60	1.5747	0.425
1.4703	7.0	70	1.4444	0.5375
1.4032	8.0	80	1.4226	0.5312
1.3367	9.0	90	1.3937	0.5188
1.2889	10.0	100	1.3186	0.5375
1.2136	11.0	110	1.3313	0.55
1.1745	12.0	120	1.3027	0.5312
1.1477	13.0	130	1.3004	0.5375
1.1414	14.0	140	1.2442	0.55
1.1202	15.0	150	1.2957	0.5062
1.0923	16.0	160	1.3045	0.5125
1.0765	17.0	170	1.2533	0.5563
1.0678	18.0	180	1.2392	0.5437
1.0837	19.0	190	1.2750	0.5375
1.0562	20.0	200	1.2275	0.5625

Framework versions

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