segformer-b0-finetuned-segments-sidewalk-3

This model is a fine-tuned version of nvidia/mit-b0 on the jhaberbe/lipid-droplets-v3 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1127
• Mean Iou: 0.4462
• Mean Accuracy: 0.8924
• Overall Accuracy: 0.8924
• Accuracy Unlabeled: nan
• Accuracy Lipid: 0.8924
• Iou Unlabeled: 0.0
• Iou Lipid: 0.8924

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: 6e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 500

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabeled	Accuracy Lipid	Iou Unlabeled	Iou Lipid
0.4354	4.0	20	0.6338	0.3881	0.7762	0.7762	nan	0.7762	0.0	0.7762
0.3649	8.0	40	0.5519	0.4247	0.8494	0.8494	nan	0.8494	0.0	0.8494
0.249	12.0	60	0.3059	0.3220	0.6440	0.6440	nan	0.6440	0.0	0.6440
0.1959	16.0	80	0.3187	0.3624	0.7247	0.7247	nan	0.7247	0.0	0.7247
0.1541	20.0	100	0.2542	0.3600	0.7201	0.7201	nan	0.7201	0.0	0.7201
0.245	24.0	120	0.1705	0.0943	0.1886	0.1886	nan	0.1886	0.0	0.1886
0.1661	28.0	140	0.0909	0.2357	0.4713	0.4713	nan	0.4713	0.0	0.4713
0.0906	32.0	160	0.2466	0.3304	0.6608	0.6608	nan	0.6608	0.0	0.6608
0.0747	36.0	180	0.2400	0.3773	0.7546	0.7546	nan	0.7546	0.0	0.7546
0.077	40.0	200	0.3043	0.4416	0.8832	0.8832	nan	0.8832	0.0	0.8832
0.077	44.0	220	0.1589	0.2127	0.4254	0.4254	nan	0.4254	0.0	0.4254
0.0595	48.0	240	0.2388	0.4113	0.8226	0.8226	nan	0.8226	0.0	0.8226
0.0678	52.0	260	0.1919	0.3874	0.7747	0.7747	nan	0.7747	0.0	0.7747
0.0584	56.0	280	0.2784	0.4315	0.8631	0.8631	nan	0.8631	0.0	0.8631
0.0752	60.0	300	0.1283	0.3702	0.7404	0.7404	nan	0.7404	0.0	0.7404
0.058	64.0	320	0.0754	0.3292	0.6585	0.6585	nan	0.6585	0.0	0.6585
0.071	68.0	340	0.1524	0.3384	0.6768	0.6768	nan	0.6768	0.0	0.6768
0.1157	72.0	360	0.0839	0.2022	0.4043	0.4043	nan	0.4043	0.0	0.4043
0.0681	76.0	380	0.1706	0.3494	0.6989	0.6989	nan	0.6989	0.0	0.6989
0.0391	80.0	400	0.1079	0.3062	0.6123	0.6123	nan	0.6123	0.0	0.6123
0.0493	84.0	420	0.2278	0.4427	0.8854	0.8854	nan	0.8854	0.0	0.8854
0.0544	88.0	440	0.1832	0.3955	0.7911	0.7911	nan	0.7911	0.0	0.7911
0.029	92.0	460	0.1327	0.3851	0.7703	0.7703	nan	0.7703	0.0	0.7703
0.0926	96.0	480	0.1737	0.4200	0.8400	0.8400	nan	0.8400	0.0	0.8400
0.0437	100.0	500	0.1248	0.3420	0.6841	0.6841	nan	0.6841	0.0	0.6841
0.0474	104.0	520	0.1365	0.3171	0.6343	0.6343	nan	0.6343	0.0	0.6343
0.0267	108.0	540	0.1942	0.4441	0.8881	0.8881	nan	0.8881	0.0	0.8881
0.0495	112.0	560	0.1312	0.3710	0.7420	0.7420	nan	0.7420	0.0	0.7420
0.0257	116.0	580	0.1224	0.3549	0.7097	0.7097	nan	0.7097	0.0	0.7097
0.0276	120.0	600	0.1836	0.4088	0.8176	0.8176	nan	0.8176	0.0	0.8176
0.0278	124.0	620	0.1652	0.4134	0.8269	0.8269	nan	0.8269	0.0	0.8269
0.04	128.0	640	0.1380	0.4057	0.8115	0.8115	nan	0.8115	0.0	0.8115
0.0373	132.0	660	0.1566	0.4473	0.8947	0.8947	nan	0.8947	0.0	0.8947
0.024	136.0	680	0.0994	0.3670	0.7339	0.7339	nan	0.7339	0.0	0.7339
0.0301	140.0	700	0.1427	0.3591	0.7182	0.7182	nan	0.7182	0.0	0.7182
0.0372	144.0	720	0.1781	0.4245	0.8489	0.8489	nan	0.8489	0.0	0.8489
0.0364	148.0	740	0.1370	0.3996	0.7992	0.7992	nan	0.7992	0.0	0.7992
0.0497	152.0	760	0.1406	0.4023	0.8046	0.8046	nan	0.8046	0.0	0.8046
0.0442	156.0	780	0.1583	0.4168	0.8337	0.8337	nan	0.8337	0.0	0.8337
0.0358	160.0	800	0.1217	0.3739	0.7478	0.7478	nan	0.7478	0.0	0.7478
0.0408	164.0	820	0.0968	0.2903	0.5805	0.5805	nan	0.5805	0.0	0.5805
0.0225	168.0	840	0.1196	0.4256	0.8512	0.8512	nan	0.8512	0.0	0.8512
0.0399	172.0	860	0.1074	0.2849	0.5697	0.5697	nan	0.5697	0.0	0.5697
0.0436	176.0	880	0.0858	0.3241	0.6481	0.6481	nan	0.6481	0.0	0.6481
0.0351	180.0	900	0.1624	0.4129	0.8258	0.8258	nan	0.8258	0.0	0.8258
0.0291	184.0	920	0.1507	0.4153	0.8307	0.8307	nan	0.8307	0.0	0.8307
0.0417	188.0	940	0.1322	0.3823	0.7645	0.7645	nan	0.7645	0.0	0.7645
0.0277	192.0	960	0.1121	0.3679	0.7358	0.7358	nan	0.7358	0.0	0.7358
0.0289	196.0	980	0.1493	0.4154	0.8307	0.8307	nan	0.8307	0.0	0.8307
0.0366	200.0	1000	0.1342	0.4056	0.8111	0.8111	nan	0.8111	0.0	0.8111
0.0337	204.0	1020	0.1494	0.4110	0.8219	0.8219	nan	0.8219	0.0	0.8219
0.0287	208.0	1040	0.1065	0.4281	0.8561	0.8561	nan	0.8561	0.0	0.8561
0.0286	212.0	1060	0.1629	0.4310	0.8621	0.8621	nan	0.8621	0.0	0.8621
0.0319	216.0	1080	0.1547	0.3900	0.7799	0.7799	nan	0.7799	0.0	0.7799
0.038	220.0	1100	0.0830	0.4073	0.8146	0.8146	nan	0.8146	0.0	0.8146
0.0207	224.0	1120	0.1571	0.4229	0.8458	0.8458	nan	0.8458	0.0	0.8458
0.0406	228.0	1140	0.1477	0.4302	0.8604	0.8604	nan	0.8604	0.0	0.8604
0.0323	232.0	1160	0.1267	0.4136	0.8272	0.8272	nan	0.8272	0.0	0.8272
0.0269	236.0	1180	0.1262	0.4243	0.8485	0.8485	nan	0.8485	0.0	0.8485
0.0484	240.0	1200	0.1179	0.4161	0.8322	0.8322	nan	0.8322	0.0	0.8322
0.0366	244.0	1220	0.0954	0.4016	0.8031	0.8031	nan	0.8031	0.0	0.8031
0.0384	248.0	1240	0.0980	0.4147	0.8294	0.8294	nan	0.8294	0.0	0.8294
0.0205	252.0	1260	0.0946	0.3758	0.7516	0.7516	nan	0.7516	0.0	0.7516
0.0243	256.0	1280	0.1198	0.4322	0.8644	0.8644	nan	0.8644	0.0	0.8644
0.033	260.0	1300	0.1271	0.4357	0.8714	0.8714	nan	0.8714	0.0	0.8714
0.0219	264.0	1320	0.1041	0.4368	0.8736	0.8736	nan	0.8736	0.0	0.8736
0.0304	268.0	1340	0.1113	0.4122	0.8244	0.8244	nan	0.8244	0.0	0.8244
0.0241	272.0	1360	0.0802	0.4152	0.8303	0.8303	nan	0.8303	0.0	0.8303
0.0209	276.0	1380	0.1255	0.4476	0.8952	0.8952	nan	0.8952	0.0	0.8952
0.015	280.0	1400	0.1500	0.4440	0.8879	0.8879	nan	0.8879	0.0	0.8879
0.0209	284.0	1420	0.1275	0.4471	0.8941	0.8941	nan	0.8941	0.0	0.8941
0.0423	288.0	1440	0.1406	0.4135	0.8271	0.8271	nan	0.8271	0.0	0.8271
0.0179	292.0	1460	0.0999	0.4272	0.8544	0.8544	nan	0.8544	0.0	0.8544
0.028	296.0	1480	0.1374	0.4471	0.8942	0.8942	nan	0.8942	0.0	0.8942
0.0524	300.0	1500	0.1253	0.4339	0.8679	0.8679	nan	0.8679	0.0	0.8679
0.0182	304.0	1520	0.1077	0.4222	0.8444	0.8444	nan	0.8444	0.0	0.8444
0.0141	308.0	1540	0.1295	0.4478	0.8956	0.8956	nan	0.8956	0.0	0.8956
0.0255	312.0	1560	0.1309	0.4364	0.8728	0.8728	nan	0.8728	0.0	0.8728
0.0375	316.0	1580	0.0917	0.4369	0.8737	0.8737	nan	0.8737	0.0	0.8737
0.0312	320.0	1600	0.0967	0.3975	0.7951	0.7951	nan	0.7951	0.0	0.7951
0.0312	324.0	1620	0.1041	0.4184	0.8368	0.8368	nan	0.8368	0.0	0.8368
0.0294	328.0	1640	0.1041	0.4279	0.8558	0.8558	nan	0.8558	0.0	0.8558
0.0277	332.0	1660	0.1285	0.4322	0.8644	0.8644	nan	0.8644	0.0	0.8644
0.022	336.0	1680	0.0897	0.3872	0.7744	0.7744	nan	0.7744	0.0	0.7744
0.0185	340.0	1700	0.1148	0.4293	0.8586	0.8586	nan	0.8586	0.0	0.8586
0.0197	344.0	1720	0.1161	0.4448	0.8896	0.8896	nan	0.8896	0.0	0.8896
0.0243	348.0	1740	0.0981	0.4256	0.8511	0.8511	nan	0.8511	0.0	0.8511
0.0252	352.0	1760	0.0848	0.3893	0.7787	0.7787	nan	0.7787	0.0	0.7787
0.0273	356.0	1780	0.0852	0.4175	0.8350	0.8350	nan	0.8350	0.0	0.8350
0.0299	360.0	1800	0.0912	0.4016	0.8033	0.8033	nan	0.8033	0.0	0.8033
0.0223	364.0	1820	0.0877	0.4222	0.8444	0.8444	nan	0.8444	0.0	0.8444
0.0246	368.0	1840	0.1020	0.4455	0.8910	0.8910	nan	0.8910	0.0	0.8910
0.0354	372.0	1860	0.1018	0.4262	0.8524	0.8524	nan	0.8524	0.0	0.8524
0.0228	376.0	1880	0.1030	0.4422	0.8843	0.8843	nan	0.8843	0.0	0.8843
0.0208	380.0	1900	0.1135	0.4403	0.8805	0.8805	nan	0.8805	0.0	0.8805
0.0187	384.0	1920	0.1085	0.4384	0.8768	0.8768	nan	0.8768	0.0	0.8768
0.0211	388.0	1940	0.0929	0.4376	0.8752	0.8752	nan	0.8752	0.0	0.8752
0.0312	392.0	1960	0.0926	0.4450	0.8899	0.8899	nan	0.8899	0.0	0.8899
0.0211	396.0	1980	0.0946	0.4424	0.8847	0.8847	nan	0.8847	0.0	0.8847
0.0106	400.0	2000	0.1098	0.4487	0.8974	0.8974	nan	0.8974	0.0	0.8974
0.0244	404.0	2020	0.0960	0.4394	0.8788	0.8788	nan	0.8788	0.0	0.8788
0.0156	408.0	2040	0.0916	0.4257	0.8514	0.8514	nan	0.8514	0.0	0.8514
0.0186	412.0	2060	0.1090	0.4476	0.8953	0.8953	nan	0.8953	0.0	0.8953
0.0219	416.0	2080	0.1093	0.4297	0.8594	0.8594	nan	0.8594	0.0	0.8594
0.012	420.0	2100	0.0891	0.4347	0.8693	0.8693	nan	0.8693	0.0	0.8693
0.0367	424.0	2120	0.1247	0.4566	0.9132	0.9132	nan	0.9132	0.0	0.9132
0.0123	428.0	2140	0.0705	0.4162	0.8323	0.8323	nan	0.8323	0.0	0.8323
0.0246	432.0	2160	0.1167	0.4468	0.8936	0.8936	nan	0.8936	0.0	0.8936
0.0144	436.0	2180	0.1139	0.4523	0.9047	0.9047	nan	0.9047	0.0	0.9047
0.0278	440.0	2200	0.1015	0.4330	0.8660	0.8660	nan	0.8660	0.0	0.8660
0.0157	444.0	2220	0.0921	0.4196	0.8392	0.8392	nan	0.8392	0.0	0.8392
0.0207	448.0	2240	0.1109	0.4462	0.8924	0.8924	nan	0.8924	0.0	0.8924
0.0191	452.0	2260	0.1060	0.4421	0.8843	0.8843	nan	0.8843	0.0	0.8843
0.0214	456.0	2280	0.1137	0.4357	0.8714	0.8714	nan	0.8714	0.0	0.8714
0.0228	460.0	2300	0.0943	0.4340	0.8679	0.8679	nan	0.8679	0.0	0.8679
0.0213	464.0	2320	0.1003	0.4379	0.8759	0.8759	nan	0.8759	0.0	0.8759
0.0248	468.0	2340	0.1063	0.4525	0.9050	0.9050	nan	0.9050	0.0	0.9050
0.0186	472.0	2360	0.0915	0.4278	0.8556	0.8556	nan	0.8556	0.0	0.8556
0.0212	476.0	2380	0.0948	0.4375	0.8751	0.8751	nan	0.8751	0.0	0.8751
0.0201	480.0	2400	0.0961	0.4326	0.8652	0.8652	nan	0.8652	0.0	0.8652
0.0168	484.0	2420	0.1050	0.4372	0.8745	0.8745	nan	0.8745	0.0	0.8745
0.0242	488.0	2440	0.1120	0.4481	0.8963	0.8963	nan	0.8963	0.0	0.8963
0.017	492.0	2460	0.1035	0.4235	0.8470	0.8470	nan	0.8470	0.0	0.8470
0.0267	496.0	2480	0.1076	0.4465	0.8931	0.8931	nan	0.8931	0.0	0.8931
0.0148	500.0	2500	0.1127	0.4462	0.8924	0.8924	nan	0.8924	0.0	0.8924

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2