results

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.8129
• Accuracy: 0.5969

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
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.9759	1.0	37	0.9392	0.5408
0.8313	2.0	74	0.8845	0.6122
0.8032	3.0	111	0.8459	0.6122
0.7375	4.0	148	0.8693	0.5782
0.635	5.0	185	0.8724	0.6344
0.578	6.0	222	0.9932	0.5629
0.3875	7.0	259	1.0738	0.5952
0.3544	8.0	296	1.1359	0.6156
0.407	9.0	333	1.3020	0.5493
0.2329	10.0	370	1.2567	0.6020
0.2305	11.0	407	1.3148	0.6156
0.2098	12.0	444	1.2928	0.6241
0.1595	13.0	481	1.5325	0.5629
0.1515	14.0	518	1.4402	0.6156
0.1429	15.0	555	1.4456	0.6276
0.1812	16.0	592	1.5088	0.5663
0.1169	17.0	629	1.6266	0.5850
0.1375	18.0	666	1.5252	0.6173
0.0907	19.0	703	1.6055	0.6088
0.1003	20.0	740	1.5785	0.6003
0.0756	21.0	777	1.6485	0.5850
0.0641	22.0	814	1.6257	0.6190
0.0387	23.0	851	1.6758	0.6105
0.0341	24.0	888	1.7239	0.6088
0.0227	25.0	925	1.7956	0.6020
0.0247	26.0	962	1.7542	0.6037
0.014	27.0	999	1.7693	0.6139
0.0152	28.0	1036	1.8133	0.5969
0.0125	29.0	1073	1.8082	0.6037
0.0116	30.0	1110	1.8129	0.5969

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu117
• Datasets 2.13.1
• Tokenizers 0.13.3