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

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: 0.2286
• Mean Iou: 0.8224
• Mean Accuracy: 1.0
• Overall Accuracy: 1.0
• Accuracy Branch: 1.0
• Iou Branch: 0.8224

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 Branch	Iou Branch
0.5864	1.05	20	0.6005	0.8224	1.0	1.0	1.0	0.8224
0.4461	2.11	40	0.4242	0.8224	1.0	1.0	1.0	0.8224
0.2556	3.16	60	0.3245	0.8224	1.0	1.0	1.0	0.8224
0.234	4.21	80	0.3176	0.8224	1.0	1.0	1.0	0.8224
0.1917	5.26	100	0.2751	0.8224	1.0	1.0	1.0	0.8224
0.2608	6.32	120	0.2997	0.8224	1.0	1.0	1.0	0.8224
0.2789	7.37	140	0.2508	0.8224	1.0	1.0	1.0	0.8224
0.2173	8.42	160	0.2684	0.8224	1.0	1.0	1.0	0.8224
0.1552	9.47	180	0.2374	0.8224	1.0	1.0	1.0	0.8224
0.1809	10.53	200	0.2596	0.8224	1.0	1.0	1.0	0.8224
0.1342	11.58	220	0.2375	0.8224	1.0	1.0	1.0	0.8224
0.1946	12.63	240	0.2211	0.8224	1.0	1.0	1.0	0.8224
0.1215	13.68	260	0.2135	0.8224	1.0	1.0	1.0	0.8224
0.1212	14.74	280	0.2470	0.8224	1.0	1.0	1.0	0.8224
0.1536	15.79	300	0.2224	0.8224	1.0	1.0	1.0	0.8224
0.1282	16.84	320	0.2466	0.8224	1.0	1.0	1.0	0.8224
0.11	17.89	340	0.2316	0.8224	1.0	1.0	1.0	0.8224
0.1228	18.95	360	0.2233	0.8224	1.0	1.0	1.0	0.8224
0.1243	20.0	380	0.1996	0.8224	1.0	1.0	1.0	0.8224
0.0893	21.05	400	0.2074	0.8224	1.0	1.0	1.0	0.8224
0.1012	22.11	420	0.1941	0.8224	1.0	1.0	1.0	0.8224
0.1587	23.16	440	0.2007	0.8224	1.0	1.0	1.0	0.8224
0.0913	24.21	460	0.2211	0.8224	1.0	1.0	1.0	0.8224
0.0949	25.26	480	0.2621	0.8224	1.0	1.0	1.0	0.8224
0.0863	26.32	500	0.2195	0.8224	1.0	1.0	1.0	0.8224
0.066	27.37	520	0.2221	0.8224	1.0	1.0	1.0	0.8224
0.0738	28.42	540	0.2126	0.8224	1.0	1.0	1.0	0.8224
0.0808	29.47	560	0.2068	0.8224	1.0	1.0	1.0	0.8224
0.062	30.53	580	0.2599	0.8224	1.0	1.0	1.0	0.8224
0.0787	31.58	600	0.2366	0.8224	1.0	1.0	1.0	0.8224
0.0535	32.63	620	0.2165	0.8224	1.0	1.0	1.0	0.8224
0.0681	33.68	640	0.2212	0.8224	1.0	1.0	1.0	0.8224
0.0574	34.74	660	0.2160	0.8224	1.0	1.0	1.0	0.8224
0.109	35.79	680	0.2281	0.8224	1.0	1.0	1.0	0.8224
0.0702	36.84	700	0.2403	0.8224	1.0	1.0	1.0	0.8224
0.0578	37.89	720	0.2141	0.8224	1.0	1.0	1.0	0.8224
0.0643	38.95	740	0.2101	0.8224	1.0	1.0	1.0	0.8224
0.0948	40.0	760	0.2118	0.8224	1.0	1.0	1.0	0.8224
0.0453	41.05	780	0.2048	0.8224	1.0	1.0	1.0	0.8224
0.0472	42.11	800	0.1924	0.8224	1.0	1.0	1.0	0.8224
0.0699	43.16	820	0.2197	0.8224	1.0	1.0	1.0	0.8224
0.0492	44.21	840	0.2172	0.8224	1.0	1.0	1.0	0.8224
0.0888	45.26	860	0.2196	0.8224	1.0	1.0	1.0	0.8224
0.0438	46.32	880	0.2196	0.8224	1.0	1.0	1.0	0.8224
0.0524	47.37	900	0.2232	0.8224	1.0	1.0	1.0	0.8224
0.0453	48.42	920	0.2184	0.8224	1.0	1.0	1.0	0.8224
0.1319	49.47	940	0.2080	0.8224	1.0	1.0	1.0	0.8224
0.0423	50.53	960	0.2180	0.8224	1.0	1.0	1.0	0.8224
0.0592	51.58	980	0.2251	0.8224	1.0	1.0	1.0	0.8224
0.0395	52.63	1000	0.2198	0.8224	1.0	1.0	1.0	0.8224
0.0451	53.68	1020	0.1953	0.8224	1.0	1.0	1.0	0.8224
0.0446	54.74	1040	0.2072	0.8224	1.0	1.0	1.0	0.8224
0.048	55.79	1060	0.2222	0.8224	1.0	1.0	1.0	0.8224
0.0623	56.84	1080	0.2264	0.8224	1.0	1.0	1.0	0.8224
0.0765	57.89	1100	0.2419	0.8224	1.0	1.0	1.0	0.8224
0.0661	58.95	1120	0.2251	0.8224	1.0	1.0	1.0	0.8224
0.0625	60.0	1140	0.2254	0.8224	1.0	1.0	1.0	0.8224
0.0373	61.05	1160	0.2209	0.8224	1.0	1.0	1.0	0.8224
0.0402	62.11	1180	0.2178	0.8224	1.0	1.0	1.0	0.8224
0.0339	63.16	1200	0.2076	0.8224	1.0	1.0	1.0	0.8224
0.0608	64.21	1220	0.2177	0.8224	1.0	1.0	1.0	0.8224
0.0301	65.26	1240	0.2048	0.8224	1.0	1.0	1.0	0.8224
0.0461	66.32	1260	0.2124	0.8224	1.0	1.0	1.0	0.8224
0.0395	67.37	1280	0.2188	0.8224	1.0	1.0	1.0	0.8224
0.1034	68.42	1300	0.2251	0.8224	1.0	1.0	1.0	0.8224
0.0368	69.47	1320	0.2169	0.8224	1.0	1.0	1.0	0.8224
0.0501	70.53	1340	0.2180	0.8224	1.0	1.0	1.0	0.8224
0.0417	71.58	1360	0.2255	0.8224	1.0	1.0	1.0	0.8224
0.0351	72.63	1380	0.2154	0.8224	1.0	1.0	1.0	0.8224
0.0423	73.68	1400	0.2216	0.8224	1.0	1.0	1.0	0.8224
0.032	74.74	1420	0.2211	0.8224	1.0	1.0	1.0	0.8224
0.0586	75.79	1440	0.2144	0.8224	1.0	1.0	1.0	0.8224
0.0315	76.84	1460	0.2166	0.8224	1.0	1.0	1.0	0.8224
0.039	77.89	1480	0.2258	0.8224	1.0	1.0	1.0	0.8224
0.0838	78.95	1500	0.2290	0.8224	1.0	1.0	1.0	0.8224
0.0543	80.0	1520	0.2414	0.8224	1.0	1.0	1.0	0.8224
0.0444	81.05	1540	0.2240	0.8224	1.0	1.0	1.0	0.8224
0.0501	82.11	1560	0.2253	0.8224	1.0	1.0	1.0	0.8224
0.0275	83.16	1580	0.2190	0.8224	1.0	1.0	1.0	0.8224
0.0356	84.21	1600	0.2188	0.8224	1.0	1.0	1.0	0.8224
0.0378	85.26	1620	0.2214	0.8224	1.0	1.0	1.0	0.8224
0.0306	86.32	1640	0.2195	0.8224	1.0	1.0	1.0	0.8224
0.0348	87.37	1660	0.2172	0.8224	1.0	1.0	1.0	0.8224
0.0363	88.42	1680	0.2180	0.8224	1.0	1.0	1.0	0.8224
0.0454	89.47	1700	0.2234	0.8224	1.0	1.0	1.0	0.8224
0.0367	90.53	1720	0.2224	0.8224	1.0	1.0	1.0	0.8224
0.0265	91.58	1740	0.2308	0.8224	1.0	1.0	1.0	0.8224
0.0404	92.63	1760	0.2269	0.8224	1.0	1.0	1.0	0.8224
0.0335	93.68	1780	0.2229	0.8224	1.0	1.0	1.0	0.8224
0.0292	94.74	1800	0.2269	0.8224	1.0	1.0	1.0	0.8224
0.041	95.79	1820	0.2277	0.8224	1.0	1.0	1.0	0.8224
0.0325	96.84	1840	0.2225	0.8224	1.0	1.0	1.0	0.8224
0.0371	97.89	1860	0.2250	0.8224	1.0	1.0	1.0	0.8224
0.0336	98.95	1880	0.2259	0.8224	1.0	1.0	1.0	0.8224
0.0445	100.0	1900	0.2286	0.8224	1.0	1.0	1.0	0.8224

Framework versions

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