surface_grade-swin-tiny-patch4-window7-224-finetuned-v1

This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.4768
• Accuracy: 0.4094

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: 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: 0.1
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.6612	1.0	734	1.6258	0.2908
1.6119	2.0	1468	1.5665	0.3149
1.5692	3.0	2202	1.5426	0.3301
1.5691	4.0	2936	1.4929	0.3599
1.5401	5.0	3670	1.4630	0.3702
1.5245	6.0	4404	1.4586	0.3728
1.5342	7.0	5138	1.4018	0.4008
1.5268	8.0	5872	1.3966	0.4040
1.4918	9.0	6606	1.4097	0.3960
1.4447	10.0	7340	1.3942	0.3997
1.468	11.0	8074	1.3802	0.4164
1.4379	12.0	8808	1.3927	0.4091
1.4152	13.0	9542	1.3916	0.4091
1.3845	14.0	10276	1.3901	0.4063
1.3659	15.0	11010	1.3846	0.4121
1.3429	16.0	11744	1.4010	0.4099
1.3534	17.0	12478	1.3968	0.4115
1.3026	18.0	13212	1.4060	0.4106
1.2955	19.0	13946	1.4469	0.4008
1.2849	20.0	14680	1.4081	0.4136
1.2331	21.0	15414	1.4188	0.4089
1.2313	22.0	16148	1.4256	0.4101
1.2	23.0	16882	1.4414	0.4100
1.2271	24.0	17616	1.4540	0.4088
1.2142	25.0	18350	1.4528	0.4064
1.1986	26.0	19084	1.4566	0.4090
1.134	27.0	19818	1.4648	0.4104
1.1756	28.0	20552	1.4700	0.4089
1.1415	29.0	21286	1.4791	0.4082
1.1411	30.0	22020	1.4768	0.4094

Framework versions

• Transformers 4.37.2
• Pytorch 2.0.1+cu117
• Datasets 2.17.0
• Tokenizers 0.15.2