beit-base-patch16-224-RH

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

• Loss: 0.4340
• Accuracy: 0.8037

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: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	8	0.7927	0.5888
0.8183	2.0	16	0.7412	0.5888
0.7414	3.0	24	0.6851	0.5888
0.6837	4.0	32	0.6638	0.5888
0.6621	5.0	40	0.6619	0.5981
0.6621	6.0	48	0.6446	0.6262
0.6538	7.0	56	0.6370	0.6729
0.641	8.0	64	0.6485	0.6636
0.628	9.0	72	0.6393	0.6449
0.6187	10.0	80	0.6409	0.5794
0.6187	11.0	88	0.6360	0.5794
0.6075	12.0	96	0.6209	0.6355
0.6081	13.0	104	0.6377	0.6449
0.5886	14.0	112	0.5931	0.6729
0.5945	15.0	120	0.6108	0.6636
0.5945	16.0	128	0.5846	0.7009
0.5808	17.0	136	0.5945	0.6822
0.5636	18.0	144	0.7402	0.6636
0.5839	19.0	152	0.5661	0.6916
0.5166	20.0	160	0.5360	0.6636
0.5166	21.0	168	0.5621	0.6729
0.5165	22.0	176	0.5509	0.7196
0.5308	23.0	184	0.5602	0.7570
0.4595	24.0	192	0.4735	0.7850
0.4553	25.0	200	0.4696	0.7664
0.4553	26.0	208	0.5306	0.7850
0.4004	27.0	216	0.4819	0.7944
0.3954	28.0	224	0.4831	0.7944
0.3521	29.0	232	0.4340	0.8037
0.3436	30.0	240	0.4790	0.7757
0.3436	31.0	248	0.4720	0.7757
0.34	32.0	256	0.5283	0.7850
0.2995	33.0	264	0.4383	0.7944
0.2951	34.0	272	0.4740	0.7944
0.3094	35.0	280	0.5863	0.7664
0.3094	36.0	288	0.4483	0.7850
0.2963	37.0	296	0.4759	0.7944
0.3045	38.0	304	0.4469	0.7944
0.2739	39.0	312	0.4517	0.7850
0.2717	40.0	320	0.4654	0.7944

Framework versions

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