vit-base-patch16-224-FV-20epochs-finetuned-memes

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: 0.6532
• Accuracy: 0.8632
• Precision: 0.8617
• Recall: 0.8632
• F1: 0.8621

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.1709	0.99	20	0.9393	0.6971	0.6896	0.6971	0.6890
0.5295	1.99	40	0.5024	0.8091	0.8210	0.8091	0.8133
0.2909	2.99	60	0.4070	0.8539	0.8529	0.8539	0.8529
0.1435	3.99	80	0.4136	0.8539	0.8522	0.8539	0.8522
0.0928	4.99	100	0.4495	0.8478	0.8548	0.8478	0.8507
0.0643	5.99	120	0.4897	0.8594	0.8572	0.8594	0.8573
0.061	6.99	140	0.5040	0.8423	0.8490	0.8423	0.8453
0.0519	7.99	160	0.5266	0.8524	0.8502	0.8524	0.8510
0.0546	8.99	180	0.5200	0.8586	0.8632	0.8586	0.8605
0.0478	9.99	200	0.5654	0.8555	0.8548	0.8555	0.8548
0.0509	10.99	220	0.5774	0.8609	0.8626	0.8609	0.8616
0.0467	11.99	240	0.5847	0.8594	0.8602	0.8594	0.8594
0.0468	12.99	260	0.5909	0.8601	0.8597	0.8601	0.8596
0.0469	13.99	280	0.5970	0.8563	0.8560	0.8563	0.8561
0.0438	14.99	300	0.6234	0.8594	0.8583	0.8594	0.8586
0.0441	15.99	320	0.6190	0.8563	0.8582	0.8563	0.8570
0.0431	16.99	340	0.6419	0.8570	0.8584	0.8570	0.8574
0.0454	17.99	360	0.6528	0.8563	0.8556	0.8563	0.8558
0.0417	18.99	380	0.6688	0.8578	0.8575	0.8578	0.8574
0.0432	19.99	400	0.6532	0.8632	0.8617	0.8632	0.8621

Framework versions

• Transformers 4.24.0.dev0
• Pytorch 1.11.0+cu102
• Datasets 2.6.1.dev0
• Tokenizers 0.13.1