google-vit-base-patch16-224-cartoon-emotion-detection

This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 0.3706
Accuracy: 0.8807
Precision: 0.8769
Recall: 0.8807
F1: 0.8783

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: 0.00012
train_batch_size: 64
eval_batch_size: 64
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 256
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
No log	0.97	8	0.9902	0.5596	0.5506	0.5596	0.5360
1.242	1.97	16	0.5157	0.8165	0.8195	0.8165	0.8132
0.4438	2.97	24	0.3871	0.8440	0.8516	0.8440	0.8446
0.1768	3.97	32	0.3531	0.8624	0.8653	0.8624	0.8585
0.0661	4.97	40	0.3780	0.8716	0.8693	0.8716	0.8674
0.0661	5.97	48	0.3747	0.8624	0.8649	0.8624	0.8632
0.0375	6.97	56	0.3760	0.8991	0.8961	0.8991	0.8971
0.0362	7.97	64	0.4092	0.8716	0.8684	0.8716	0.8681
0.0322	8.97	72	0.3499	0.8899	0.8880	0.8899	0.8888
0.029	9.97	80	0.3706	0.8807	0.8769	0.8807	0.8783

Framework versions

Transformers 4.25.1
Pytorch 1.13.1+cu117
Datasets 2.8.0
Tokenizers 0.11.0

作者:

paul

数据集大小:

327.42 MB