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: 0.3314
Accuracy: 0.9287

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
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 20
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	50	1.7532	0.4263
No log	2.0	100	1.4569	0.535
No log	3.0	150	1.3329	0.5262
No log	4.0	200	1.1306	0.6475
No log	5.0	250	1.0279	0.7275
No log	6.0	300	0.8815	0.7863
No log	7.0	350	0.7592	0.8337
No log	8.0	400	0.7329	0.785
No log	9.0	450	0.6043	0.875
1.1234	10.0	500	0.5688	0.8612
1.1234	11.0	550	0.5193	0.88
1.1234	12.0	600	0.4879	0.8938
1.1234	13.0	650	0.4170	0.9038
1.1234	14.0	700	0.4425	0.8912
1.1234	15.0	750	0.4089	0.905
1.1234	16.0	800	0.3781	0.9263
1.1234	17.0	850	0.3431	0.9225
1.1234	18.0	900	0.3388	0.93
1.1234	19.0	950	0.2973	0.9475
0.3972	20.0	1000	0.3314	0.9287

Framework versions

Transformers 4.35.2
Pytorch 2.1.0+cu121
Datasets 2.16.1
Tokenizers 0.15.1

silvering/vit-emotions-classification-fp16

vit-emotions-fp16

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

silvering/
vit-emotions-classification-fp16