vit-base-patch16-224-RX1-24

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: 0.5687
• 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
No log	0.93	7	1.3485	0.4706
1.3674	2.0	15	1.2284	0.5490
1.2414	2.93	22	1.1307	0.6471
1.1146	4.0	30	1.0230	0.6471
1.1146	4.93	37	0.9251	0.6863
0.9522	6.0	45	0.9122	0.6471
0.8247	6.93	52	0.9374	0.6275
0.6825	8.0	60	0.8320	0.6863
0.6825	8.93	67	0.8286	0.6667
0.6191	10.0	75	0.8418	0.6667
0.5312	10.93	82	0.7836	0.8235
0.454	12.0	90	0.7356	0.8039
0.454	12.93	97	0.6117	0.8235
0.3752	14.0	105	0.6014	0.8235
0.3269	14.93	112	0.6102	0.8039
0.2733	16.0	120	0.6404	0.8039
0.2733	16.93	127	0.5687	0.8431
0.2711	18.0	135	0.6120	0.8235
0.2519	18.93	142	0.6250	0.8431
0.2484	20.0	150	0.6086	0.7843
0.2484	20.93	157	0.6229	0.8235
0.2258	22.0	165	0.6390	0.7843
0.2258	22.4	168	0.6337	0.8039

Framework versions

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