segformer_rust

This model is a fine-tuned version of nvidia/mit-b3 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1989
• Mean Iou: 0.4909
• Mean Accuracy: 0.5770
• Overall Accuracy: 0.9351
• Per Category Iou: [0.9330033577422647, 0.43802844257518975, 0.10174998526948427]
• Per Category Accuracy: [0.964764208993714, 0.619516063444351, 0.14676017920088064]

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Per Category Iou	Per Category Accuracy
0.3245	1.0	288	0.2058	0.3737	0.4006	0.9292	[0.928254594835728, 0.19278478772636354, 0.0]	[0.9934403780274128, 0.20834926267819698, 0.0]
0.2376	2.0	576	0.2186	0.4440	0.5394	0.9205	[0.9174093139660287, 0.3952780826522951, 0.019450366169241784]	[0.9466914641620554, 0.6519016208812702, 0.019680530318053235]
0.2267	3.0	864	0.1989	0.4456	0.5055	0.9349	[0.932781347474167, 0.4035973409203434, 0.0003156374054605271]	[0.9714340031497313, 0.5447040039563373, 0.0003156655078774732]
0.2227	4.0	1152	0.1916	0.4707	0.5302	0.9370	[0.9346588332598218, 0.4138838160094419, 0.06341712806343888]	[0.9731864476801606, 0.54656085524886, 0.07078191965098726]
0.2144	5.0	1440	0.1896	0.4891	0.5724	0.9328	[0.9302444676641086, 0.43202079299936413, 0.10517087813054486]	[0.9610582832573524, 0.6347191782872807, 0.12137743477257919]
0.2003	6.0	1728	0.2024	0.4630	0.5614	0.9260	[0.9229095306993824, 0.428931900333899, 0.037186726485148515]	[0.9487448882036036, 0.6966673460182111, 0.03891184433643468]
0.1941	7.0	2016	0.2073	0.4860	0.5888	0.9219	[0.918639874525932, 0.4060024340936176, 0.13340049908214777]	[0.9461779703553824, 0.669512151538538, 0.15057649425124547]
0.1867	8.0	2304	0.1807	0.4634	0.5301	0.9372	[0.9349180068188713, 0.43659922555690217, 0.018734234161859927]	[0.9689566528873623, 0.602518494621371, 0.018875178573596604]
0.1854	9.0	2592	0.1921	0.4725	0.5473	0.9345	[0.9321974693686552, 0.43270280485825646, 0.052579663560228966]	[0.9642507962436797, 0.6216561173996741, 0.05587279489431276]
0.18	10.0	2880	0.1847	0.4771	0.5591	0.9335	[0.9311550048182651, 0.4407180368387474, 0.05928079306817498]	[0.9605858689552134, 0.6528219329451048, 0.06379276154708474]
0.1838	11.0	3168	0.1887	0.4652	0.5400	0.9344	[0.9321143595621934, 0.43344353554824266, 0.02996913887506222]	[0.9638921692510604, 0.6257803286194064, 0.030457674516485428]
0.1852	12.0	3456	0.1873	0.4783	0.5588	0.9338	[0.931356267339074, 0.43971228563091297, 0.06380926339455255]	[0.9613249758102673, 0.6471117095414763, 0.06799758799176032]
0.1682	13.0	3744	0.1778	0.4578	0.5035	0.9411	[0.939396147353813, 0.410371405648735, 0.023504333548772998]	[0.981291057821883, 0.5056175933375897, 0.023695148059264176]
0.1793	14.0	4032	0.1728	0.4720	0.5296	0.9402	[0.938163188412165, 0.436672912065328, 0.04117603263361246]	[0.974742378495531, 0.5715719389819965, 0.042525809702262676]
0.178	15.0	4320	0.1825	0.4794	0.5492	0.9369	[0.9348434866894885, 0.42443186583904435, 0.07881529998132747]	[0.9714614165049668, 0.5650121075899666, 0.11103331889905584]
0.1651	16.0	4608	0.1912	0.4640	0.5322	0.9362	[0.9340976310450665, 0.4306683229566104, 0.02708604107096439]	[0.9679043107574059, 0.6011619809118315, 0.02741028826736059]
0.1761	17.0	4896	0.1757	0.4724	0.5310	0.9389	[0.9368263731599995, 0.42525112196356074, 0.055183904634890216]	[0.9744989654092739, 0.5559049976770041, 0.06246130062283233]
0.1633	18.0	5184	0.1917	0.4620	0.5384	0.9354	[0.9330768754070563, 0.44480333780621734, 0.00820601828493007]	[0.9636645135753821, 0.6433031553142516, 0.008227538173267988]
0.1641	19.0	5472	0.1875	0.5028	0.5812	0.9354	[0.9331706683726091, 0.4415413111697262, 0.133770201971674]	[0.9634674000411962, 0.6392245347005364, 0.1408556154060956]
0.1629	20.0	5760	0.1793	0.4675	0.5268	0.9393	[0.9374054652680682, 0.4330992564780264, 0.03193276637035433]	[0.973550456833595, 0.5746419640278527, 0.032230257753028166]
0.1562	21.0	6048	0.2019	0.4716	0.5546	0.9325	[0.9300698286146659, 0.43671042563474527, 0.04809297394799092]	[0.9592354491418064, 0.6554619381223432, 0.049203349291978456]
0.1626	22.0	6336	0.1970	0.4814	0.5547	0.9367	[0.9344970508696149, 0.4450380854494638, 0.06452712765123607]	[0.9658097423660128, 0.6317005392634606, 0.066633751117982]
0.1568	23.0	6624	0.1956	0.4917	0.5775	0.9355	[0.9330854195306507, 0.4390415550387406, 0.10283427594052297]	[0.9658042305145758, 0.6119827953416446, 0.15474894474639514]
0.1595	24.0	6912	0.1817	0.4769	0.5411	0.9383	[0.9362300601150796, 0.4346106438489147, 0.05977008091938032]	[0.9712496965644727, 0.586809439959973, 0.06537513608016285]
0.1539	25.0	7200	0.1893	0.4825	0.5650	0.9358	[0.933537238237644, 0.4394361715176897, 0.07460055177704666]	[0.9667388784050233, 0.6055558721569863, 0.12278174158326487]
0.1521	26.0	7488	0.1933	0.4841	0.5721	0.9355	[0.9332085743268924, 0.44365656488299987, 0.07547064782962047]	[0.9657158949896331, 0.6132794230856533, 0.13717285114752506]
0.1468	27.0	7776	0.2049	0.4861	0.5735	0.9327	[0.9304477363857275, 0.43790721804897553, 0.08980648951011608]	[0.9595730176196706, 0.6520673697794168, 0.10889650623034679]
0.1534	28.0	8064	0.1950	0.4976	0.5800	0.9355	[0.9333228433994155, 0.4429051769820105, 0.11658642467846855]	[0.9637939610275141, 0.636643256405625, 0.13955248343767832]
0.1478	29.0	8352	0.2049	0.4913	0.5841	0.9344	[0.9322490633597582, 0.4375642417642195, 0.10407288263883356]	[0.9639606458995019, 0.6190808277013159, 0.169160289279109]
0.1414	30.0	8640	0.1989	0.4909	0.5770	0.9351	[0.9330033577422647, 0.43802844257518975, 0.10174998526948427]	[0.964764208993714, 0.619516063444351, 0.14676017920088064]

Framework versions

• Transformers 4.33.1
• Pytorch 2.0.1
• Datasets 2.14.5
• Tokenizers 0.13.3