swin-tiny-patch4-window7-224-ve-U11-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: 0.7088
• Accuracy: 0.7826

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	1.3860	0.1304
1.3859	2.0	13	1.3832	0.2609
1.3859	2.92	19	1.3773	0.2609
1.3791	4.0	26	1.3569	0.2174
1.3347	4.92	32	1.3177	0.2609
1.3347	6.0	39	1.2093	0.3913
1.2088	6.92	45	1.1083	0.4348
1.0456	8.0	52	1.0340	0.4565
1.0456	8.92	58	1.0120	0.5
0.9278	10.0	65	0.9282	0.5652
0.847	10.92	71	0.9934	0.5217
0.847	12.0	78	1.0171	0.4783
0.7142	12.92	84	0.8889	0.5870
0.5959	14.0	91	0.9392	0.5870
0.5959	14.92	97	0.9018	0.6304
0.5344	16.0	104	0.8327	0.6739
0.4438	16.92	110	0.7308	0.7391
0.4438	18.0	117	0.6834	0.7174
0.4419	18.92	123	0.7909	0.6304
0.3989	20.0	130	0.9103	0.6739
0.3989	20.92	136	0.7534	0.7391
0.3534	22.0	143	0.8043	0.7391
0.3534	22.92	149	0.7648	0.7174
0.3265	24.0	156	0.7088	0.7826
0.2808	24.92	162	0.8845	0.6957
0.2808	26.0	169	0.7756	0.7609
0.2753	26.92	175	0.9944	0.6087
0.2837	28.0	182	0.8091	0.7174
0.2837	28.92	188	0.9966	0.6739
0.2667	30.0	195	0.7711	0.7826
0.2325	30.92	201	0.8946	0.6957
0.2325	32.0	208	0.9079	0.6739
0.2096	32.92	214	1.0338	0.6522
0.1733	34.0	221	0.8191	0.7391
0.1733	34.92	227	1.0068	0.6957
0.1975	36.0	234	0.8644	0.7174
0.1844	36.92	240	0.8682	0.6739
0.1844	38.0	247	0.7915	0.7609
0.1701	38.92	253	0.7554	0.7609
0.1696	40.0	260	0.8762	0.7174
0.1696	40.92	266	1.0173	0.6739
0.1556	42.0	273	0.9080	0.7174
0.1556	42.92	279	1.2456	0.6739
0.153	44.0	286	0.9820	0.7391
0.1343	44.92	292	0.9908	0.7174
0.1343	46.0	299	0.9435	0.7391
0.1513	46.92	305	0.8842	0.7826
0.1402	48.0	312	1.0207	0.6739
0.1402	48.92	318	0.9915	0.7174
0.1648	50.0	325	1.1576	0.6739
0.1047	50.92	331	1.2283	0.6739
0.1047	52.0	338	1.0869	0.6957
0.1223	52.92	344	1.1203	0.7174
0.1223	54.0	351	0.9685	0.7174
0.1223	54.92	357	1.1926	0.7174
0.1236	56.0	364	1.0088	0.7174
0.1115	56.92	370	0.9149	0.7391
0.1115	58.0	377	0.8820	0.7391
0.1173	58.92	383	0.9653	0.7391
0.102	60.0	390	1.0046	0.7174
0.102	60.92	396	1.0585	0.6957
0.1206	62.0	403	1.0490	0.6957
0.1206	62.92	409	0.9683	0.7609
0.1124	64.0	416	0.9627	0.7609
0.0927	64.92	422	0.9771	0.7609
0.0927	66.0	429	1.0002	0.7174
0.0906	66.92	435	0.9607	0.7391
0.084	68.0	442	0.9414	0.7391
0.084	68.92	448	0.9863	0.7174
0.0866	70.0	455	0.9930	0.7174
0.0944	70.92	461	0.9981	0.7174
0.0944	72.0	468	1.0039	0.7174
0.1064	72.92	474	0.9987	0.7174
0.1074	73.85	480	0.9964	0.7174

Framework versions

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