emotion_classification_2

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: 1.3274
• Accuracy: 0.5188

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: 4e-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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	20	1.9337	0.3563
No log	2.0	40	1.7116	0.3375
No log	3.0	60	1.5755	0.4562
No log	4.0	80	1.4939	0.45
No log	5.0	100	1.4377	0.5062
No log	6.0	120	1.4363	0.4562
No log	7.0	140	1.3615	0.5125
No log	8.0	160	1.3021	0.5375
No log	9.0	180	1.3307	0.525
No log	10.0	200	1.3085	0.4938
No log	11.0	220	1.2798	0.5813
No log	12.0	240	1.2707	0.525
No log	13.0	260	1.2339	0.55
No log	14.0	280	1.3053	0.5437
No log	15.0	300	1.3038	0.4938
No log	16.0	320	1.3088	0.5375
No log	17.0	340	1.3336	0.5312
No log	18.0	360	1.3053	0.5
No log	19.0	380	1.2206	0.5687
No log	20.0	400	1.2598	0.5312
No log	21.0	420	1.3332	0.5125
No log	22.0	440	1.3388	0.5312
No log	23.0	460	1.3129	0.5563
No log	24.0	480	1.3632	0.5062
0.9153	25.0	500	1.4166	0.4688
0.9153	26.0	520	1.4094	0.5
0.9153	27.0	540	1.4294	0.475
0.9153	28.0	560	1.4937	0.475
0.9153	29.0	580	1.3897	0.4938
0.9153	30.0	600	1.4565	0.475

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3