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: 1.2723
• Accuracy: 0.5938

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	20	2.0185	0.25
No log	2.0	40	1.9216	0.35
No log	3.0	60	1.8084	0.3875
No log	4.0	80	1.6901	0.4375
No log	5.0	100	1.6031	0.4562
No log	6.0	120	1.5323	0.4688
No log	7.0	140	1.4855	0.4813
No log	8.0	160	1.4305	0.525
No log	9.0	180	1.3853	0.4938
No log	10.0	200	1.3556	0.5312
No log	11.0	220	1.3141	0.5625
No log	12.0	240	1.2958	0.5563
No log	13.0	260	1.2810	0.5437
No log	14.0	280	1.2629	0.6
No log	15.0	300	1.2533	0.5938
No log	16.0	320	1.2728	0.5813
No log	17.0	340	1.2311	0.5437
No log	18.0	360	1.2094	0.5938
No log	19.0	380	1.2584	0.5687
No log	20.0	400	1.2113	0.6125
No log	21.0	420	1.2002	0.5938
No log	22.0	440	1.2211	0.6062
No log	23.0	460	1.2424	0.5875
No log	24.0	480	1.2357	0.5813
0.9674	25.0	500	1.1765	0.5938
0.9674	26.0	520	1.2338	0.5875
0.9674	27.0	540	1.2333	0.5875
0.9674	28.0	560	1.2671	0.5563
0.9674	29.0	580	1.2011	0.6
0.9674	30.0	600	1.2008	0.6062
0.9674	31.0	620	1.2582	0.5687
0.9674	32.0	640	1.2820	0.5813
0.9674	33.0	660	1.2435	0.6
0.9674	34.0	680	1.2691	0.5875
0.9674	35.0	700	1.2324	0.6188
0.9674	36.0	720	1.2008	0.625
0.9674	37.0	740	1.2381	0.6125
0.9674	38.0	760	1.2494	0.5813
0.9674	39.0	780	1.2303	0.5938
0.9674	40.0	800	1.1828	0.6188

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2