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: 1.5599
• Accuracy: 0.475

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	5	2.0884	0.1125
2.08	2.0	10	2.0750	0.1437
2.08	3.0	15	2.0519	0.2125
2.0091	4.0	20	2.0177	0.225
2.0091	5.0	25	1.9777	0.2625
1.8779	6.0	30	1.9381	0.3125
1.8779	7.0	35	1.8990	0.3438
1.7355	8.0	40	1.8592	0.3688
1.7355	9.0	45	1.8217	0.3812
1.598	10.0	50	1.7844	0.4
1.598	11.0	55	1.7536	0.4062
1.4689	12.0	60	1.7217	0.4188
1.4689	13.0	65	1.7019	0.4188
1.3534	14.0	70	1.6773	0.4188
1.3534	15.0	75	1.6614	0.425
1.2526	16.0	80	1.6448	0.4562
1.2526	17.0	85	1.6306	0.45
1.1657	18.0	90	1.6201	0.4562
1.1657	19.0	95	1.6067	0.4562
1.0918	20.0	100	1.5992	0.45
1.0918	21.0	105	1.5889	0.4562
1.0311	22.0	110	1.5852	0.4562
1.0311	23.0	115	1.5767	0.4625
0.9814	24.0	120	1.5733	0.45
0.9814	25.0	125	1.5688	0.4625
0.9439	26.0	130	1.5643	0.4562
0.9439	27.0	135	1.5620	0.4625
0.918	28.0	140	1.5599	0.475
0.918	29.0	145	1.5586	0.4625
0.9044	30.0	150	1.5582	0.4562

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3