segformer-b0-finetuned-100by100PNG-50epochs-attempt2-100epochs-backgroundclass

This model is a fine-tuned version of nvidia/mit-b0 on the JCAI2000/100By100BranchPNG dataset. It achieves the following results on the evaluation set:

• Loss: 0.1491
• Mean Iou: 0.4520
• Mean Accuracy: 0.9041
• Overall Accuracy: 0.9041
• Accuracy Background: nan
• Accuracy Branch: 0.9041
• Iou Background: 0.0
• Iou Branch: 0.9041

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: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Branch	Iou Background	Iou Branch
0.5951	1.05	20	0.6067	0.4897	0.9793	0.9793	nan	0.9793	0.0	0.9793
0.4531	2.11	40	0.4117	0.4717	0.9435	0.9435	nan	0.9435	0.0	0.9435
0.4396	3.16	60	0.3235	0.4750	0.9499	0.9499	nan	0.9499	0.0	0.9499
0.2327	4.21	80	0.2603	0.4703	0.9405	0.9405	nan	0.9405	0.0	0.9405
0.1971	5.26	100	0.2336	0.4786	0.9572	0.9572	nan	0.9572	0.0	0.9572
0.1768	6.32	120	0.2441	0.4776	0.9552	0.9552	nan	0.9552	0.0	0.9552
0.2005	7.37	140	0.1811	0.4597	0.9194	0.9194	nan	0.9194	0.0	0.9194
0.1818	8.42	160	0.2437	0.4805	0.9610	0.9610	nan	0.9610	0.0	0.9610
0.2163	9.47	180	0.1803	0.4774	0.9548	0.9548	nan	0.9548	0.0	0.9548
0.1436	10.53	200	0.1897	0.4807	0.9615	0.9615	nan	0.9615	0.0	0.9615
0.0957	11.58	220	0.1682	0.4578	0.9157	0.9157	nan	0.9157	0.0	0.9157
0.1614	12.63	240	0.1603	0.4422	0.8844	0.8844	nan	0.8844	0.0	0.8844
0.0952	13.68	260	0.1732	0.4738	0.9477	0.9477	nan	0.9477	0.0	0.9477
0.1243	14.74	280	0.1432	0.4569	0.9139	0.9139	nan	0.9139	0.0	0.9139
0.0943	15.79	300	0.1539	0.4725	0.9451	0.9451	nan	0.9451	0.0	0.9451
0.0833	16.84	320	0.1176	0.4570	0.9140	0.9140	nan	0.9140	0.0	0.9140
0.1817	17.89	340	0.1270	0.4623	0.9246	0.9246	nan	0.9246	0.0	0.9246
0.0939	18.95	360	0.1561	0.4715	0.9431	0.9431	nan	0.9431	0.0	0.9431
0.0849	20.0	380	0.1496	0.4682	0.9363	0.9363	nan	0.9363	0.0	0.9363
0.1155	21.05	400	0.1204	0.4547	0.9094	0.9094	nan	0.9094	0.0	0.9094
0.0507	22.11	420	0.1323	0.4667	0.9335	0.9335	nan	0.9335	0.0	0.9335
0.0631	23.16	440	0.1219	0.4593	0.9185	0.9185	nan	0.9185	0.0	0.9185
0.0509	24.21	460	0.1178	0.4673	0.9346	0.9346	nan	0.9346	0.0	0.9346
0.0594	25.26	480	0.1193	0.4568	0.9136	0.9136	nan	0.9136	0.0	0.9136
0.0633	26.32	500	0.1321	0.4717	0.9434	0.9434	nan	0.9434	0.0	0.9434
0.0739	27.37	520	0.1361	0.4587	0.9174	0.9174	nan	0.9174	0.0	0.9174
0.0894	28.42	540	0.1286	0.4658	0.9317	0.9317	nan	0.9317	0.0	0.9317
0.0528	29.47	560	0.1296	0.4644	0.9288	0.9288	nan	0.9288	0.0	0.9288
0.0683	30.53	580	0.1434	0.4705	0.9410	0.9410	nan	0.9410	0.0	0.9410
0.0343	31.58	600	0.1154	0.4598	0.9196	0.9196	nan	0.9196	0.0	0.9196
0.0436	32.63	620	0.1417	0.4527	0.9053	0.9053	nan	0.9053	0.0	0.9053
0.0369	33.68	640	0.1185	0.4365	0.8730	0.8730	nan	0.8730	0.0	0.8730
0.0537	34.74	660	0.1369	0.4660	0.9319	0.9319	nan	0.9319	0.0	0.9319
0.0642	35.79	680	0.1351	0.4514	0.9027	0.9027	nan	0.9027	0.0	0.9027
0.0597	36.84	700	0.1441	0.4590	0.9180	0.9180	nan	0.9180	0.0	0.9180
0.0382	37.89	720	0.1413	0.4568	0.9136	0.9136	nan	0.9136	0.0	0.9136
0.0488	38.95	740	0.1369	0.4626	0.9252	0.9252	nan	0.9252	0.0	0.9252
0.0652	40.0	760	0.1477	0.4653	0.9306	0.9306	nan	0.9306	0.0	0.9306
0.0376	41.05	780	0.1320	0.4579	0.9158	0.9158	nan	0.9158	0.0	0.9158
0.0387	42.11	800	0.1298	0.4536	0.9071	0.9071	nan	0.9071	0.0	0.9071
0.0791	43.16	820	0.1431	0.4498	0.8997	0.8997	nan	0.8997	0.0	0.8997
0.0304	44.21	840	0.1368	0.4426	0.8852	0.8852	nan	0.8852	0.0	0.8852
0.0301	45.26	860	0.1523	0.4681	0.9363	0.9363	nan	0.9363	0.0	0.9363
0.0743	46.32	880	0.1396	0.4505	0.9009	0.9009	nan	0.9009	0.0	0.9009
0.1028	47.37	900	0.1354	0.4463	0.8926	0.8926	nan	0.8926	0.0	0.8926
0.0399	48.42	920	0.1497	0.4568	0.9136	0.9136	nan	0.9136	0.0	0.9136
0.0282	49.47	940	0.1489	0.4672	0.9343	0.9343	nan	0.9343	0.0	0.9343
0.0266	50.53	960	0.1574	0.4564	0.9128	0.9128	nan	0.9128	0.0	0.9128
0.0328	51.58	980	0.1540	0.4536	0.9072	0.9072	nan	0.9072	0.0	0.9072
0.0273	52.63	1000	0.1624	0.4572	0.9144	0.9144	nan	0.9144	0.0	0.9144
0.0311	53.68	1020	0.1459	0.4386	0.8771	0.8771	nan	0.8771	0.0	0.8771
0.0481	54.74	1040	0.1607	0.4597	0.9194	0.9194	nan	0.9194	0.0	0.9194
0.0384	55.79	1060	0.1718	0.4596	0.9192	0.9192	nan	0.9192	0.0	0.9192
0.0299	56.84	1080	0.1708	0.4589	0.9178	0.9178	nan	0.9178	0.0	0.9178
0.0315	57.89	1100	0.1458	0.4539	0.9078	0.9078	nan	0.9078	0.0	0.9078
0.2086	58.95	1120	0.1428	0.4590	0.9181	0.9181	nan	0.9181	0.0	0.9181
0.0355	60.0	1140	0.1575	0.4478	0.8957	0.8957	nan	0.8957	0.0	0.8957
0.0236	61.05	1160	0.1610	0.4471	0.8941	0.8941	nan	0.8941	0.0	0.8941
0.0775	62.11	1180	0.1688	0.4478	0.8955	0.8955	nan	0.8955	0.0	0.8955
0.026	63.16	1200	0.1513	0.4558	0.9117	0.9117	nan	0.9117	0.0	0.9117
0.03	64.21	1220	0.1583	0.4630	0.9260	0.9260	nan	0.9260	0.0	0.9260
0.0255	65.26	1240	0.1595	0.4565	0.9131	0.9131	nan	0.9131	0.0	0.9131
0.079	66.32	1260	0.1485	0.4503	0.9005	0.9005	nan	0.9005	0.0	0.9005
0.0366	67.37	1280	0.1658	0.4561	0.9123	0.9123	nan	0.9123	0.0	0.9123
0.0286	68.42	1300	0.1890	0.4667	0.9334	0.9334	nan	0.9334	0.0	0.9334
0.0303	69.47	1320	0.1469	0.4526	0.9052	0.9052	nan	0.9052	0.0	0.9052
0.0215	70.53	1340	0.1559	0.4548	0.9095	0.9095	nan	0.9095	0.0	0.9095
0.028	71.58	1360	0.1616	0.4598	0.9195	0.9195	nan	0.9195	0.0	0.9195
0.0228	72.63	1380	0.1445	0.4521	0.9041	0.9041	nan	0.9041	0.0	0.9041
0.0216	73.68	1400	0.1526	0.4542	0.9085	0.9085	nan	0.9085	0.0	0.9085
0.0202	74.74	1420	0.1525	0.4643	0.9285	0.9285	nan	0.9285	0.0	0.9285
0.0297	75.79	1440	0.1471	0.4590	0.9180	0.9180	nan	0.9180	0.0	0.9180
0.0237	76.84	1460	0.1603	0.4604	0.9208	0.9208	nan	0.9208	0.0	0.9208
0.0601	77.89	1480	0.1526	0.4581	0.9161	0.9161	nan	0.9161	0.0	0.9161
0.0299	78.95	1500	0.1625	0.4579	0.9159	0.9159	nan	0.9159	0.0	0.9159
0.0316	80.0	1520	0.1702	0.4593	0.9185	0.9185	nan	0.9185	0.0	0.9185
0.0274	81.05	1540	0.1741	0.4607	0.9214	0.9214	nan	0.9214	0.0	0.9214
0.0274	82.11	1560	0.1609	0.4594	0.9188	0.9188	nan	0.9188	0.0	0.9188
0.0345	83.16	1580	0.1652	0.4581	0.9163	0.9163	nan	0.9163	0.0	0.9163
0.018	84.21	1600	0.1645	0.4588	0.9176	0.9176	nan	0.9176	0.0	0.9176
0.0352	85.26	1620	0.1579	0.4588	0.9176	0.9176	nan	0.9176	0.0	0.9176
0.0202	86.32	1640	0.1741	0.4620	0.9239	0.9239	nan	0.9239	0.0	0.9239
0.0315	87.37	1660	0.1587	0.4543	0.9086	0.9086	nan	0.9086	0.0	0.9086
0.0208	88.42	1680	0.1610	0.4579	0.9158	0.9158	nan	0.9158	0.0	0.9158
0.0174	89.47	1700	0.1685	0.4596	0.9193	0.9193	nan	0.9193	0.0	0.9193
0.0278	90.53	1720	0.1698	0.4586	0.9173	0.9173	nan	0.9173	0.0	0.9173
0.0259	91.58	1740	0.1674	0.4593	0.9186	0.9186	nan	0.9186	0.0	0.9186
0.0642	92.63	1760	0.1586	0.4572	0.9144	0.9144	nan	0.9144	0.0	0.9144
0.0543	93.68	1780	0.1636	0.4591	0.9182	0.9182	nan	0.9182	0.0	0.9182
0.0264	94.74	1800	0.1572	0.4586	0.9172	0.9172	nan	0.9172	0.0	0.9172
0.0239	95.79	1820	0.1687	0.4596	0.9191	0.9191	nan	0.9191	0.0	0.9191
0.0238	96.84	1840	0.1595	0.4569	0.9137	0.9137	nan	0.9137	0.0	0.9137
0.0181	97.89	1860	0.1552	0.4552	0.9103	0.9103	nan	0.9103	0.0	0.9103
0.0354	98.95	1880	0.1645	0.4573	0.9146	0.9146	nan	0.9146	0.0	0.9146
0.0897	100.0	1900	0.1491	0.4520	0.9041	0.9041	nan	0.9041	0.0	0.9041

Framework versions

• Transformers 4.33.0
• Pytorch 2.0.1+cu117
• Datasets 2.14.4
• Tokenizers 0.13.3