swin-tiny-patch4-window7-224-ve-U13-b-80

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: 0.9190
• Accuracy: 0.8043

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: 80

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.92	6	1.3859	0.1304
1.3859	2.0	13	1.3828	0.2826
1.3859	2.92	19	1.3769	0.3261
1.379	4.0	26	1.3566	0.2826
1.3356	4.92	32	1.3162	0.2391
1.3356	6.0	39	1.2093	0.3478
1.2023	6.92	45	1.1349	0.4565
1.0274	8.0	52	1.0414	0.4783
1.0274	8.92	58	0.9788	0.5217
0.9125	10.0	65	1.0071	0.4348
0.7688	10.92	71	1.0416	0.5217
0.7688	12.0	78	1.0480	0.4130
0.6891	12.92	84	0.9351	0.5870
0.5795	14.0	91	1.0683	0.6304
0.5795	14.92	97	1.0698	0.6087
0.5337	16.0	104	0.9603	0.6304
0.4337	16.92	110	0.7188	0.6957
0.4337	18.0	117	0.7620	0.6739
0.4258	18.92	123	0.9433	0.6739
0.4045	20.0	130	1.0823	0.6522
0.4045	20.92	136	0.7059	0.7174
0.4135	22.0	143	0.7467	0.7391
0.4135	22.92	149	0.7637	0.7391
0.3525	24.0	156	0.8157	0.7391
0.263	24.92	162	0.9995	0.7174
0.263	26.0	169	0.8719	0.7609
0.272	26.92	175	0.9939	0.6957
0.262	28.0	182	0.8639	0.7174
0.262	28.92	188	1.0737	0.6522
0.2282	30.0	195	0.8416	0.7174
0.2098	30.92	201	0.9744	0.6739
0.2098	32.0	208	1.0593	0.6087
0.2141	32.92	214	1.0997	0.7174
0.1759	34.0	221	0.9735	0.5870
0.1759	34.92	227	1.0789	0.6957
0.2042	36.0	234	1.0664	0.6957
0.1591	36.92	240	0.9417	0.7609
0.1591	38.0	247	1.1042	0.6739
0.1579	38.92	253	0.9732	0.7609
0.1626	40.0	260	0.9960	0.6957
0.1626	40.92	266	0.9763	0.7391
0.1458	42.0	273	0.9790	0.7391
0.1458	42.92	279	1.0952	0.7174
0.1317	44.0	286	0.9190	0.8043
0.1255	44.92	292	0.9420	0.7391
0.1255	46.0	299	0.9085	0.7391
0.1352	46.92	305	0.9184	0.7174
0.1311	48.0	312	1.0567	0.7609
0.1311	48.92	318	1.1507	0.7174
0.1501	50.0	325	1.2068	0.7174
0.1088	50.92	331	1.4607	0.6957
0.1088	52.0	338	1.1036	0.6739
0.1152	52.92	344	1.1081	0.6957
0.1141	54.0	351	1.1006	0.6957
0.1141	54.92	357	1.1470	0.7174
0.1307	56.0	364	1.0715	0.7609
0.1273	56.92	370	1.1021	0.7174
0.1273	58.0	377	1.1176	0.6957
0.1066	58.92	383	1.0948	0.7174
0.1046	60.0	390	1.0563	0.7391
0.1046	60.92	396	1.1155	0.6957
0.1129	62.0	403	1.0922	0.6957
0.1129	62.92	409	1.0364	0.6957
0.1031	64.0	416	1.0675	0.7174
0.0808	64.92	422	1.1133	0.6957
0.0808	66.0	429	1.2029	0.7174
0.0783	66.92	435	1.1453	0.7174
0.09	68.0	442	1.0925	0.6957
0.09	68.92	448	1.0999	0.7174
0.0796	70.0	455	1.0971	0.7391
0.0828	70.92	461	1.0923	0.7391
0.0828	72.0	468	1.1061	0.7391
0.0923	72.92	474	1.1173	0.7391
0.092	73.85	480	1.1208	0.7391

Framework versions

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