swinv2-tiny-patch4-window8-256-DMAE-da3-colab

This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.3992
• Accuracy: 0.3696

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: 0.001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 120

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.3523	0.9778	22	1.4024	0.3261
1.3805	2.0	45	1.3775	0.2609
1.3221	2.9778	67	1.4419	0.3043
1.297	4.0	90	1.3582	0.3261
1.353	4.9778	112	1.3406	0.3478
1.2627	6.0	135	1.3824	0.1522
1.3006	6.9778	157	1.4008	0.1522
1.2438	8.0	180	1.3769	0.3261
1.222	8.9778	202	1.4212	0.3043
1.2221	10.0	225	1.4223	0.2391
1.2262	10.9778	247	1.4154	0.2609
1.2381	12.0	270	1.3327	0.2391
1.227	12.9778	292	1.2887	0.2826
1.2158	14.0	315	1.3465	0.2609
1.2174	14.9778	337	1.3476	0.3043
1.1767	16.0	360	1.4024	0.1957
1.2067	16.9778	382	1.3664	0.1739
1.2303	18.0	405	1.4260	0.2826
1.222	18.9778	427	1.4807	0.1739
1.2026	20.0	450	1.3851	0.1739
1.2185	20.9778	472	1.3214	0.2609
1.2773	22.0	495	1.4404	0.1957
1.227	22.9778	517	1.4535	0.2391
1.2032	24.0	540	1.3967	0.3043
1.2223	24.9778	562	1.4090	0.3261
1.2527	26.0	585	1.4858	0.2609
1.2203	26.9778	607	1.4366	0.1739
1.1993	28.0	630	1.4056	0.2609
1.2014	28.9778	652	1.3755	0.3043
1.2027	30.0	675	1.4579	0.2609
1.1961	30.9778	697	1.4524	0.2609
1.1939	32.0	720	1.4488	0.2391
1.1889	32.9778	742	1.4568	0.1522
1.1871	34.0	765	1.3814	0.3261
1.1778	34.9778	787	1.4403	0.1304
1.2404	36.0	810	1.4437	0.1957
1.197	36.9778	832	1.4765	0.2174
1.2161	38.0	855	1.3720	0.2391
1.221	38.9778	877	1.3750	0.3478
1.229	40.0	900	1.3405	0.2391
1.2046	40.9778	922	1.4231	0.2609
1.2077	42.0	945	1.4384	0.2391
1.1865	42.9778	967	1.4346	0.2609
1.1882	44.0	990	1.3679	0.2826
1.2528	44.9778	1012	1.3451	0.2174
1.1836	46.0	1035	1.4913	0.2391
1.2009	46.9778	1057	1.4841	0.3261
1.203	48.0	1080	1.4326	0.3043
1.1679	48.9778	1102	1.3935	0.3043
1.179	50.0	1125	1.4185	0.1957
1.1687	50.9778	1147	1.3686	0.2826
1.1779	52.0	1170	1.4319	0.1957
1.1566	52.9778	1192	1.3801	0.1957
1.192	54.0	1215	1.3746	0.2174
1.1803	54.9778	1237	1.4017	0.1957
1.194	56.0	1260	1.4288	0.1957
1.1486	56.9778	1282	1.3920	0.3043
1.1429	58.0	1305	1.4616	0.2391
1.1655	58.9778	1327	1.4119	0.2174
1.1697	60.0	1350	1.3812	0.2609
1.1898	60.9778	1372	1.4009	0.2391
1.1882	62.0	1395	1.4221	0.2391
1.134	62.9778	1417	1.6190	0.2609
1.1748	64.0	1440	1.4336	0.2391
1.1439	64.9778	1462	1.3744	0.1957
1.1585	66.0	1485	1.3992	0.3696
1.1344	66.9778	1507	1.3952	0.2391
1.1374	68.0	1530	1.3666	0.2174
1.1252	68.9778	1552	1.3705	0.2826
1.1339	70.0	1575	1.3983	0.2826
1.1344	70.9778	1597	1.3792	0.3043
1.1343	72.0	1620	1.4467	0.2826
1.1555	72.9778	1642	1.4823	0.2174
1.1329	74.0	1665	1.5136	0.1522
1.1513	74.9778	1687	1.4791	0.2391
1.1278	76.0	1710	1.4527	0.2609
1.0956	76.9778	1732	1.4840	0.2391
1.1131	78.0	1755	1.4900	0.2174
1.1376	78.9778	1777	1.5395	0.2174
1.0883	80.0	1800	1.5038	0.1957
1.1017	80.9778	1822	1.5392	0.1957
1.1608	82.0	1845	1.4875	0.2174
1.1308	82.9778	1867	1.5080	0.1957
1.1382	84.0	1890	1.4835	0.1739
1.1195	84.9778	1912	1.4076	0.1957
1.1149	86.0	1935	1.4840	0.1739
1.1344	86.9778	1957	1.4733	0.1957
1.1268	88.0	1980	1.4446	0.2391
1.1267	88.9778	2002	1.4360	0.2174
1.1034	90.0	2025	1.4329	0.1522
1.1113	90.9778	2047	1.4670	0.1739
1.0957	92.0	2070	1.4802	0.2391
1.1227	92.9778	2092	1.4715	0.1739
1.1083	94.0	2115	1.4813	0.1957
1.0583	94.9778	2137	1.5203	0.1957
1.093	96.0	2160	1.5394	0.1739
1.0809	96.9778	2182	1.4620	0.1739
1.0888	98.0	2205	1.4407	0.1739
1.1292	98.9778	2227	1.4578	0.1957
1.0754	100.0	2250	1.5031	0.1739
1.0817	100.9778	2272	1.4461	0.2174
1.0671	102.0	2295	1.4723	0.2391
1.0815	102.9778	2317	1.4989	0.1957
1.0967	104.0	2340	1.4654	0.2174
1.091	104.9778	2362	1.4559	0.2174
1.0895	106.0	2385	1.4221	0.2826
1.0847	106.9778	2407	1.4293	0.2826
1.102	108.0	2430	1.4582	0.2391
1.0404	108.9778	2452	1.4656	0.2174
1.0488	110.0	2475	1.4890	0.2174
1.0966	110.9778	2497	1.4632	0.2174
1.0901	112.0	2520	1.4495	0.2174
1.1008	112.9778	2542	1.4333	0.2174
1.0884	114.0	2565	1.4406	0.2174
1.0889	114.9778	2587	1.4474	0.2174
1.0729	116.0	2610	1.4561	0.2174
1.0671	116.9778	2632	1.4538	0.2174
1.0937	117.3333	2640	1.4532	0.2174

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3