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: 1.2677
• Accuracy: 0.575

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 3
• total_train_batch_size: 48
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.9379	0.97	13	1.2947	0.4875
0.9235	1.95	26	1.3397	0.475
0.8298	3.0	40	1.2971	0.5563
0.8883	3.98	53	1.3434	0.4875
0.8547	4.95	66	1.3226	0.475
0.8129	6.0	80	1.3077	0.5062
0.8095	6.97	93	1.2503	0.525
0.7764	7.95	106	1.2989	0.5312
0.7004	9.0	120	1.3383	0.4813
0.7013	9.97	133	1.3370	0.5125
0.6416	10.95	146	1.3073	0.5125
0.5831	12.0	160	1.3192	0.5
0.5968	12.97	173	1.2394	0.5375
0.5434	13.95	186	1.3389	0.5188
0.4605	15.0	200	1.2951	0.525
0.4674	15.97	213	1.2038	0.5687
0.3953	16.95	226	1.4019	0.5062
0.3595	18.0	240	1.4442	0.4813
0.3619	18.98	253	1.4213	0.525
0.3304	19.95	266	1.2937	0.5437
0.34	21.0	280	1.3024	0.5687
0.4215	21.98	293	1.4018	0.5375
0.3606	22.95	306	1.4221	0.5375
0.3402	24.0	320	1.4987	0.4313
0.3058	24.98	333	1.5120	0.5125
0.3047	25.95	346	1.5749	0.5
0.3616	27.0	360	1.4293	0.5188
0.3315	27.98	373	1.5326	0.5312
0.3535	28.95	386	1.5095	0.5188
0.3056	29.25	390	1.5366	0.5

Framework versions

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