vit-base-patch16-224-RXL1-24

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: 0.6158
• Accuracy: 0.8431

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: 5.5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 24

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.3745	0.95	13	1.3056	0.4706
1.2896	1.96	27	1.1039	0.6471
0.9896	2.98	41	0.9413	0.6471
0.8472	4.0	55	0.9059	0.6275
0.7375	4.95	68	0.6520	0.8039
0.458	5.96	82	0.6754	0.8039
0.3807	6.98	96	0.6158	0.8431
0.3003	8.0	110	0.5666	0.8039
0.2337	8.95	123	0.5409	0.8039
0.2252	9.96	137	0.7382	0.7647
0.1644	10.98	151	0.6363	0.8039
0.1608	12.0	165	0.6941	0.8039
0.1354	12.95	178	0.6985	0.7843
0.1298	13.96	192	0.6610	0.8039
0.1333	14.98	206	0.6751	0.8039
0.1209	16.0	220	0.7723	0.7843
0.1057	16.95	233	0.8038	0.7255
0.0972	17.96	247	0.8375	0.7647
0.0789	18.98	261	0.6971	0.8235
0.0833	20.0	275	0.7507	0.7843
0.0813	20.95	288	0.7085	0.7843
0.0803	21.96	302	0.7566	0.7647
0.0693	22.69	312	0.7772	0.7647

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0