segformer-b0-scene-parse-150-lr-5-e-30

This model is a fine-tuned version of DiTo97/binarization-segformer-b3 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1499
• Mean Iou: 0.4846
• Mean Accuracy: 0.5002
• Overall Accuracy: 0.9687
• Per Category Iou: [0.0004776099180272339, 0.9687318368688158]
• Per Category Accuracy: [0.0004786768150677956, 0.9999280293990062]

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: 1e-05
• train_batch_size: 1
• eval_batch_size: 1
• 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
No log	1.0	112	2.2080	0.0208	0.4867	0.0409	[0.030359692962016637, 0.011234027937699289]	[0.9620657111236344, 0.011247758028276789]
No log	2.0	224	1.4759	0.0198	0.4977	0.0391	[0.03105738195300228, 0.008579759290056214]	[0.9867430286323423, 0.008583423879932913]
No log	3.0	336	0.9545	0.2044	0.5106	0.3895	[0.03167545114091724, 0.3770365875250204]	[0.63983643511383, 0.38141170005716085]
No log	4.0	448	0.9261	0.1630	0.5110	0.3102	[0.03177509935992916, 0.2941818944688777]	[0.7253108004426911, 0.29678542982527334]
1.4226	5.0	560	0.5223	0.4799	0.5020	0.9428	[0.017040433028331244, 0.9427249005417263]	[0.031779387701061236, 0.9721328848551324]
1.4226	6.0	672	0.2859	0.4844	0.5000	0.9661	[0.0026066960283743228, 0.9661016634825252]	[0.0028381121801182773, 0.997139715499558]
1.4226	7.0	784	0.3327	0.4845	0.5001	0.9666	[0.0023741570079603156, 0.9666249471662205]	[0.0025444558361227857, 0.9976889561878419]
1.4226	8.0	896	0.3413	0.4861	0.5020	0.9655	[0.006810437487136094, 0.965477402345672]	[0.0075824444429356125, 0.996347820132999]
0.4243	9.0	1008	0.2237	0.4845	0.5001	0.9670	[0.002003100646207948, 0.9669488069696599]	[0.002125189264059858, 0.998036285433216]
0.4243	10.0	1120	0.2312	0.4845	0.5001	0.9673	[0.0017493474271832114, 0.9672550380074268]	[0.001838322662122066, 0.9983613016047257]
0.4243	11.0	1232	0.2255	0.4846	0.5001	0.9683	[0.0008165323590605036, 0.9682956004963261]	[0.0008300459665537307, 0.9994667831613294]
0.4243	12.0	1344	0.2229	0.4845	0.5001	0.9681	[0.0009280118466771697, 0.9681143324063521]	[0.0009488664525634671, 0.9992759735664452]
0.4243	13.0	1456	0.1762	0.4844	0.5000	0.9688	[0.0, 0.9687853389316134]	[0.0, 0.9999981952660845]
0.2658	14.0	1568	0.1799	0.4844	0.5000	0.9688	[0.0, 0.9687686496310763]	[0.0, 0.9999809682605274]
0.2658	15.0	1680	0.1895	0.4844	0.5000	0.9688	[0.0, 0.9687842263115777]	[0.0, 0.9999970467990473]
0.2658	16.0	1792	0.1751	0.4846	0.5002	0.9687	[0.0005055791248239351, 0.9686817929933391]	[0.0005075332188130173, 0.9998754733598305]
0.2658	17.0	1904	0.1570	0.4844	0.5000	0.9688	[0.0, 0.9687863985697428]	[0.0, 0.9999992890442151]
0.2281	18.0	2016	0.2391	0.4892	0.5049	0.9676	[0.010804445997179198, 0.9675856416536037]	[0.01133886923635771, 0.9984063652637066]
0.2281	19.0	2128	0.1763	0.4846	0.5002	0.9687	[0.00047390236595756206, 0.9687105941224816]	[0.00047528194403894595, 0.9999062085253005]
0.2281	20.0	2240	0.1614	0.4844	0.5000	0.9688	[8.31166363884634e-05, 0.96876793276621]	[8.317434020681555e-05, 0.9999776322372291]
0.2281	21.0	2352	0.1576	0.4844	0.5000	0.9688	[0.0, 0.9687801996866862]	[0.0, 0.999992890442151]
0.2281	22.0	2464	0.1509	0.4848	0.5004	0.9687	[0.0008725737882892481, 0.9686954205214706]	[0.0008758767254432005, 0.9998780437384374]
0.2197	23.0	2576	0.1575	0.4844	0.5000	0.9688	[0.00013731233980227023, 0.9687503427281783]	[0.00013749227666840936, 0.9999577801641586]
0.2197	24.0	2688	0.1522	0.4847	0.5002	0.9687	[0.0006022476015827608, 0.96869965961662]	[0.0006042870431352313, 0.999890895631472]
0.2197	25.0	2800	0.1532	0.4844	0.5000	0.9688	[0.0, 0.9687816831800673]	[0.0, 0.9999944217315339]
0.2197	26.0	2912	0.1589	0.4844	0.5000	0.9688	[7.462281558667101e-05, 0.9687624309087838]	[7.46871626346915e-05, 0.9999722180354826]
0.2129	27.0	3024	0.1612	0.4845	0.5001	0.9687	[0.0003791835028091742, 0.9687128062602468]	[0.00038022555523115677, 0.9999114586603274]
0.2129	28.0	3136	0.1526	0.4844	0.5000	0.9688	[4.750029687685548e-05, 0.9687701927590581]	[4.7528194403894597e-05, 0.9999810776383405]
0.2129	29.0	3248	0.1501	0.4844	0.5000	0.9688	[2.375603233535373e-05, 0.9687772095405226]	[2.3764097201947298e-05, 0.999989062218694]
0.2129	30.0	3360	0.1499	0.4846	0.5002	0.9687	[0.0004776099180272339, 0.9687318368688158]	[0.0004786768150677956, 0.9999280293990062]

Framework versions

• Transformers 4.37.1
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1