swin-tiny-patch4-window7-224-finetuned-st-ucimhar-stacked-tiny

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: 1.2422
• Accuracy: 0.7833

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: 12
• eval_batch_size: 12
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 48
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 200

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.7491	0.9977	107	1.6778	0.3781
1.3073	1.9953	214	1.1312	0.5299
1.011	2.9930	321	0.9034	0.5961
0.922	4.0	429	0.7951	0.6179
0.8234	4.9977	536	0.6725	0.6709
0.7737	5.9953	643	0.7821	0.6319
0.7946	6.9930	750	0.6644	0.6854
0.8128	8.0	858	0.7123	0.6582
0.715	8.9977	965	0.6771	0.6759
0.7034	9.9953	1072	0.6267	0.7090
0.6079	10.9930	1179	0.6672	0.6895
0.6178	12.0	1287	0.6762	0.6750
0.6561	12.9977	1394	0.5786	0.7085
0.6171	13.9953	1501	0.5703	0.7244
0.6261	14.9930	1608	0.6156	0.7058
0.646	16.0	1716	0.6111	0.7203
0.5834	16.9977	1823	0.5608	0.7407
0.5469	17.9953	1930	0.5682	0.7425
0.5964	18.9930	2037	0.5712	0.7452
0.6095	20.0	2145	0.5856	0.7226
0.5992	20.9977	2252	0.6288	0.7085
0.5572	21.9953	2359	0.6966	0.7049
0.4914	22.9930	2466	0.6251	0.7131
0.5694	24.0	2574	0.5781	0.7407
0.5108	24.9977	2681	0.5544	0.7493
0.4507	25.9953	2788	0.5737	0.7421
0.5308	26.9930	2895	0.5404	0.7529
0.4754	28.0	3003	0.5940	0.7439
0.4562	28.9977	3110	0.5465	0.7507
0.4617	29.9953	3217	0.5298	0.7665
0.4347	30.9930	3324	0.5796	0.7566
0.466	32.0	3432	0.5378	0.7602
0.4335	32.9977	3539	0.5227	0.7724
0.4335	33.9953	3646	0.6071	0.7471
0.492	34.9930	3753	0.5336	0.7715
0.3798	36.0	3861	0.5833	0.7679
0.4314	36.9977	3968	0.5538	0.7747
0.4269	37.9953	4075	0.5880	0.7616
0.3804	38.9930	4182	0.6006	0.7665
0.4089	40.0	4290	0.5728	0.7747
0.3446	40.9977	4397	0.5992	0.7747
0.3786	41.9953	4504	0.5686	0.7706
0.3944	42.9930	4611	0.6180	0.7634
0.3295	44.0	4719	0.5682	0.7743
0.3104	44.9977	4826	0.5924	0.7702
0.3465	45.9953	4933	0.6169	0.7765
0.3611	46.9930	5040	0.5923	0.7829
0.2893	48.0	5148	0.6474	0.7697
0.3345	48.9977	5255	0.7161	0.7575
0.2894	49.9953	5362	0.6239	0.7661
0.2979	50.9930	5469	0.6255	0.7765
0.2933	52.0	5577	0.6235	0.7729
0.3072	52.9977	5684	0.6996	0.7638
0.3179	53.9953	5791	0.6933	0.7792
0.2318	54.9930	5898	0.6738	0.7811
0.2083	56.0	6006	0.6706	0.7792
0.1927	56.9977	6113	0.7160	0.7806
0.2316	57.9953	6220	0.7098	0.7706
0.2582	58.9930	6327	0.7736	0.7656
0.257	60.0	6435	0.6893	0.7879
0.2449	60.9977	6542	0.7491	0.7820
0.2335	61.9953	6649	0.7232	0.7770
0.2251	62.9930	6756	0.7697	0.7733
0.2055	64.0	6864	0.7920	0.7679
0.1984	64.9977	6971	0.7432	0.7788
0.2104	65.9953	7078	0.7694	0.7652
0.2279	66.9930	7185	0.7625	0.7788
0.2356	68.0	7293	0.7823	0.7747
0.2097	68.9977	7400	0.8695	0.7711
0.2393	69.9953	7507	0.7937	0.7770
0.1599	70.9930	7614	0.8057	0.7792
0.1836	72.0	7722	0.7616	0.7883
0.1823	72.9977	7829	0.8693	0.7688
0.1938	73.9953	7936	0.8137	0.7647
0.1821	74.9930	8043	0.8643	0.7688
0.2125	76.0	8151	0.8180	0.7824
0.1872	76.9977	8258	0.9031	0.7684
0.219	77.9953	8365	0.8406	0.7738
0.1639	78.9930	8472	0.8814	0.7720
0.1532	80.0	8580	0.8852	0.7792
0.2049	80.9977	8687	0.8322	0.7788
0.1379	81.9953	8794	0.9790	0.7711
0.145	82.9930	8901	0.9681	0.7620
0.201	84.0	9009	0.9446	0.7801
0.1433	84.9977	9116	0.8740	0.7711
0.127	85.9953	9223	0.8781	0.7833
0.1646	86.9930	9330	0.8880	0.7815
0.1671	88.0	9438	0.9304	0.7733
0.123	88.9977	9545	0.9443	0.7738
0.1261	89.9953	9652	0.9818	0.7638
0.1143	90.9930	9759	0.9140	0.7738
0.1621	92.0	9867	0.8911	0.7733
0.1607	92.9977	9974	0.8875	0.7801
0.1501	93.9953	10081	0.9756	0.7652
0.1313	94.9930	10188	0.9266	0.7724
0.1304	96.0	10296	0.9165	0.7811
0.1178	96.9977	10403	0.8847	0.7806
0.1022	97.9953	10510	0.9896	0.7747
0.1535	98.9930	10617	0.9597	0.7634
0.1235	100.0	10725	1.0785	0.7811
0.1537	100.9977	10832	0.9804	0.7665
0.1034	101.9953	10939	1.0229	0.7747
0.1151	102.9930	11046	0.9565	0.7792
0.1207	104.0	11154	0.9972	0.7752
0.1004	104.9977	11261	1.0217	0.7693
0.1129	105.9953	11368	1.0180	0.7743
0.1461	106.9930	11475	1.0894	0.7634
0.1058	108.0	11583	1.0378	0.7801
0.1182	108.9977	11690	0.9832	0.7756
0.1234	109.9953	11797	0.9694	0.7915
0.0876	110.9930	11904	1.0163	0.7783
0.1114	112.0	12012	1.0190	0.7720
0.1102	112.9977	12119	1.0097	0.7788
0.1341	113.9953	12226	1.0256	0.7797
0.0925	114.9930	12333	1.0942	0.7797
0.099	116.0	12441	1.1353	0.7715
0.0949	116.9977	12548	1.0752	0.7765
0.0999	117.9953	12655	1.0974	0.7765
0.0942	118.9930	12762	1.0812	0.7761
0.1149	120.0	12870	1.0109	0.7824
0.1101	120.9977	12977	1.0238	0.7879
0.0971	121.9953	13084	1.0740	0.7724
0.1146	122.9930	13191	1.0517	0.7788
0.1378	124.0	13299	0.9993	0.7765
0.1491	124.9977	13406	1.0226	0.7797
0.0993	125.9953	13513	1.0196	0.7806
0.1103	126.9930	13620	1.0618	0.7829
0.0628	128.0	13728	1.1314	0.7752
0.125	128.9977	13835	1.0911	0.7806
0.1051	129.9953	13942	1.1129	0.7729
0.07	130.9930	14049	1.1152	0.7774
0.1128	132.0	14157	1.1385	0.7815
0.1186	132.9977	14264	1.0660	0.7915
0.0828	133.9953	14371	1.0861	0.7788
0.081	134.9930	14478	1.0989	0.7783
0.084	136.0	14586	1.0952	0.7770
0.0958	136.9977	14693	1.0558	0.7747
0.0943	137.9953	14800	1.0902	0.7833
0.055	138.9930	14907	1.1308	0.7797
0.0972	140.0	15015	1.0727	0.7842
0.0819	140.9977	15122	1.1066	0.7851
0.0885	141.9953	15229	1.1115	0.7752
0.0769	142.9930	15336	1.0922	0.7788
0.0668	144.0	15444	1.1498	0.7788
0.0836	144.9977	15551	1.1783	0.7729
0.1068	145.9953	15658	1.1379	0.7783
0.0656	146.9930	15765	1.1223	0.7806
0.0815	148.0	15873	1.1083	0.7783
0.06	148.9977	15980	1.1309	0.7783
0.0812	149.9953	16087	1.1336	0.7774
0.0771	150.9930	16194	1.1676	0.7833
0.0867	152.0	16302	1.1745	0.7824
0.075	152.9977	16409	1.1625	0.7779
0.0833	153.9953	16516	1.1656	0.7743
0.0788	154.9930	16623	1.1623	0.7783
0.0647	156.0	16731	1.1728	0.7788
0.0915	156.9977	16838	1.1869	0.7765
0.0814	157.9953	16945	1.1643	0.7824
0.0644	158.9930	17052	1.1955	0.7820
0.0899	160.0	17160	1.2017	0.7811
0.0746	160.9977	17267	1.1638	0.7801
0.0744	161.9953	17374	1.1992	0.7774
0.0868	162.9930	17481	1.2164	0.7824
0.0351	164.0	17589	1.2245	0.7738
0.0586	164.9977	17696	1.1803	0.7829
0.0558	165.9953	17803	1.2584	0.7756
0.0982	166.9930	17910	1.2349	0.7765
0.0711	168.0	18018	1.3022	0.7756
0.0416	168.9977	18125	1.1992	0.7783
0.053	169.9953	18232	1.2331	0.7752
0.0458	170.9930	18339	1.1994	0.7833
0.0698	172.0	18447	1.2182	0.7779
0.0823	172.9977	18554	1.2031	0.7788
0.0705	173.9953	18661	1.1887	0.7874
0.0489	174.9930	18768	1.2278	0.7733
0.0513	176.0	18876	1.2315	0.7824
0.0587	176.9977	18983	1.2252	0.7806
0.0291	177.9953	19090	1.2223	0.7788
0.0549	178.9930	19197	1.2333	0.7797
0.0433	180.0	19305	1.2576	0.7811
0.0859	180.9977	19412	1.2297	0.7820
0.1015	181.9953	19519	1.2171	0.7801
0.0396	182.9930	19626	1.2269	0.7783
0.0892	184.0	19734	1.2359	0.7824
0.0561	184.9977	19841	1.2501	0.7820
0.0391	185.9953	19948	1.2284	0.7869
0.0673	186.9930	20055	1.2562	0.7783
0.0464	188.0	20163	1.2398	0.7815
0.0798	188.9977	20270	1.2727	0.7792
0.0543	189.9953	20377	1.2414	0.7838
0.0421	190.9930	20484	1.2468	0.7811
0.0563	192.0	20592	1.2546	0.7833
0.0638	192.9977	20699	1.2530	0.7824
0.0571	193.9953	20806	1.2487	0.7824
0.05	194.9930	20913	1.2502	0.7797
0.0743	196.0	21021	1.2507	0.7797
0.0282	196.9977	21128	1.2481	0.7820
0.0629	197.9953	21235	1.2416	0.7829
0.0285	198.9930	21342	1.2423	0.7842
0.0475	199.5338	21400	1.2422	0.7833

Framework versions

• Transformers 4.44.0
• Pytorch 1.12.1+cu113
• Datasets 2.21.0
• Tokenizers 0.19.1