segformer-b0-scene-parse-150-lr-3-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: 0.1448
• Mean Iou: 0.5020
• Mean Accuracy: 0.5211
• Overall Accuracy: 0.9636
• Per Category Iou: [0.04038452943608308, 0.9635414972513529]
• Per Category Accuracy: [0.04908134789959329, 0.993061727806312]

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: 0.001
• 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	0.1656	0.4844	0.5	0.9688	[0.0, 0.9687870873345269]	[0.0, 1.0]
No log	2.0	224	0.1537	0.4844	0.5	0.9688	[0.0, 0.9687870873345269]	[0.0, 1.0]
No log	3.0	336	0.1432	0.4844	0.5000	0.9688	[0.0, 0.9687868224249946]	[0.0, 0.9999997265554673]
No log	4.0	448	0.1475	0.4850	0.5005	0.9686	[0.0013222566963458553, 0.9685858800631549]	[0.0013324868788234735, 0.9997507279640349]
0.2536	5.0	560	0.1711	0.4845	0.5001	0.9687	[0.0003201696729864885, 0.9687338453238151]	[0.0003208153122262885, 0.999935029579042]
0.2536	6.0	672	0.1638	0.4859	0.5015	0.9680	[0.0039027635518112448, 0.9679476439113655]	[0.004022922169186793, 0.9990080526133531]
0.2536	7.0	784	0.1410	0.4869	0.5025	0.9685	[0.005323932455427736, 0.9684674246442314]	[0.00540633211344301, 0.9995013465502566]
0.2536	8.0	896	0.1433	0.4844	0.5000	0.9688	[0.0, 0.9687869283888075]	[0.0, 0.9999998359332805]
0.2055	9.0	1008	0.1506	0.4852	0.5008	0.9687	[0.0016878008192091351, 0.9687229148795367]	[0.0016940406433959573, 0.999880887561577]
0.2055	10.0	1120	0.1453	0.4935	0.5096	0.9671	[0.020029693488101304, 0.9670700472858372]	[0.021548943855622924, 0.9975562262116929]
0.2055	11.0	1232	0.1517	0.4845	0.5001	0.9688	[0.00011372817946646207, 0.9687905263352534]	[0.00011372817946646207, 1.0]
0.2055	12.0	1344	0.1431	0.4857	0.5013	0.9687	[0.0027137294106124215, 0.9687079701798799]	[0.002727778871680665, 0.9998331988350826]
0.2055	13.0	1456	0.1414	0.4933	0.5092	0.9672	[0.019346379945589794, 0.9672062565458419]	[0.020713805582525918, 0.9977227539320777]
0.1952	14.0	1568	0.3025	0.4616	0.5605	0.8715	[0.052657461092701, 0.870574568072721]	[0.2288363740061515, 0.8922046651387315]
0.1952	15.0	1680	0.1681	0.5006	0.5284	0.9534	[0.04789077424916502, 0.9533115168989387]	[0.07506229588337938, 0.9817203423700553]
0.1952	16.0	1792	0.1410	0.4898	0.5053	0.9683	[0.011200534555931552, 0.9683138222802735]	[0.011495033303684793, 0.9991528688378454]
0.1952	17.0	1904	0.1923	0.4976	0.5436	0.9374	[0.05802705926695029, 0.9371260247893431]	[0.12360895159592887, 0.9635867588094531]
0.184	18.0	2016	0.1869	0.5041	0.5434	0.9464	[0.062043684016905756, 0.9461805831562268]	[0.11365179486831295, 0.9732005216884172]
0.184	19.0	2128	0.1451	0.4945	0.5108	0.9667	[0.0224270262537552, 0.9666655973811165]	[0.02448211242454899, 0.9970476740698675]
0.184	20.0	2240	0.1495	0.5034	0.5236	0.9627	[0.044245675510015896, 0.9625936882687055]	[0.0553839259646526, 0.9918894164059282]
0.184	21.0	2352	0.1666	0.5074	0.5361	0.9560	[0.05892929526242253, 0.9558424300678736]	[0.0883447287837535, 0.9839176332566303]
0.184	22.0	2464	0.1359	0.4952	0.5117	0.9661	[0.024373231478689413, 0.9660300331460354]	[0.02716575797285461, 0.9963086081870168]
0.172	23.0	2576	0.1373	0.4947	0.5109	0.9667	[0.022619186612181357, 0.9667099473494948]	[0.024663738024592447, 0.9970877610383542]
0.172	24.0	2688	0.1447	0.5034	0.5235	0.9628	[0.044081368434508966, 0.9626945273298088]	[0.055020674764565694, 0.992004591243081]
0.172	25.0	2800	0.1408	0.5024	0.5209	0.9645	[0.04033427475747473, 0.9644272946188437]	[0.04782864048994779, 0.9940135882244722]
0.172	26.0	2912	0.1487	0.5049	0.5279	0.9602	[0.049801967806652817, 0.9600934920373924]	[0.06687047209076527, 0.9889578175713707]
0.1592	27.0	3024	0.1466	0.5061	0.5282	0.9614	[0.05082699206223086, 0.9613120480339439]	[0.0662356312083704, 0.9902326706714988]
0.1592	28.0	3136	0.1395	0.5022	0.5210	0.9641	[0.04037355060044024, 0.9640515837293752]	[0.04838200446765027, 0.9936091637606804]
0.1592	29.0	3248	0.1477	0.5054	0.5287	0.9601	[0.05084271645010736, 0.9600046044343804]	[0.06849152300704096, 0.9888161186145507]
0.1592	30.0	3360	0.1448	0.5020	0.5211	0.9636	[0.04038452943608308, 0.9635414972513529]	[0.04908134789959329, 0.993061727806312]

Framework versions

• Transformers 4.37.0
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0