segformer-finetuned-rwymarkings-3k-steps

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

• Loss: 0.0182
• Mean Iou: 0.0441
• Mean Accuracy: 0.0510
• Overall Accuracy: 0.0800
• Accuracy Backgound : nan
• Accuracy Tdz: 0.0908
• Accuracy Aim: 0.2203
• Accuracy Desig: 0.0
• Accuracy Rwythr: 0.0971
• Accuracy Thrbar: 0.0
• Accuracy Disp: 0.0
• Accuracy Chevron: 0.0
• Accuracy Arrow: 0.0
• Iou Backgound : 0.0
• Iou Tdz: 0.0818
• Iou Aim: 0.2189
• Iou Desig: 0.0
• Iou Rwythr: 0.0958
• Iou Thrbar: 0.0
• Iou Disp: 0.0
• Iou Chevron: 0.0
• Iou Arrow: 0.0

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: 1337
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: polynomial
• training_steps: 3000

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Backgound	Accuracy Tdz	Accuracy Aim	Accuracy Desig	Accuracy Rwythr	Accuracy Thrbar	Accuracy Disp	Accuracy Chevron	Accuracy Arrow	Iou Backgound	Iou Tdz	Iou Aim	Iou Desig	Iou Rwythr	Iou Thrbar	Iou Disp	Iou Chevron	Iou Arrow
1.6294	1.0	173	0.5448	0.0	0.0	0.0	nan	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.3371	2.0	346	0.1107	0.0	0.0	0.0	nan	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0724	3.0	519	0.0483	0.0	0.0	0.0	nan	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0508	4.0	692	0.0331	0.0	0.0	0.0	nan	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0369	5.0	865	0.0289	0.0002	0.0002	0.0004	nan	0.0	0.0019	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0019	0.0	0.0	0.0	0.0	0.0	0.0
0.0272	6.0	1038	0.0276	0.0106	0.0120	0.0195	nan	0.0107	0.0853	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0105	0.0845	0.0	0.0	0.0	0.0	0.0	0.0
0.0258	7.0	1211	0.0233	0.0066	0.0075	0.0122	nan	0.0118	0.0480	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0117	0.0480	0.0	0.0	0.0	0.0	0.0	0.0
0.0235	8.0	1384	0.0221	0.0150	0.0171	0.0277	nan	0.0233	0.1108	0.0	0.0024	0.0	0.0	0.0	0.0	0.0	0.0224	0.1107	0.0	0.0024	0.0	0.0	0.0	0.0
0.0213	9.0	1557	0.0209	0.0177	0.0200	0.0326	nan	0.0237	0.1351	0.0	0.0016	0.0	0.0	0.0	0.0	0.0	0.0231	0.1346	0.0	0.0016	0.0	0.0	0.0	0.0
0.0201	10.0	1730	0.0206	0.0277	0.0318	0.0512	nan	0.0595	0.1734	0.0	0.0211	0.0	0.0	0.0	0.0	0.0	0.0559	0.1726	0.0	0.0211	0.0	0.0	0.0	0.0
0.0203	11.0	1903	0.0198	0.0246	0.0281	0.0450	nan	0.0463	0.1512	0.0	0.0277	0.0	0.0	0.0	0.0	0.0	0.0432	0.1505	0.0	0.0277	0.0	0.0	0.0	0.0
0.0172	12.0	2076	0.0192	0.0377	0.0435	0.0690	nan	0.0744	0.2145	0.0	0.0592	0.0	0.0	0.0	0.0	0.0	0.0680	0.2119	0.0	0.0589	0.0	0.0	0.0	0.0
0.0168	13.0	2249	0.0189	0.0331	0.0381	0.0607	nan	0.0704	0.1884	0.0	0.0462	0.0	0.0	0.0	0.0	0.0	0.0645	0.1876	0.0	0.0461	0.0	0.0	0.0	0.0
0.0169	14.0	2422	0.0185	0.0383	0.0442	0.0701	nan	0.0786	0.2124	0.0	0.0628	0.0	0.0	0.0	0.0	0.0	0.0716	0.2112	0.0	0.0623	0.0	0.0	0.0	0.0
0.0172	15.0	2595	0.0184	0.0476	0.0551	0.0864	nan	0.0917	0.2463	0.0	0.1028	0.0	0.0	0.0	0.0	0.0	0.0830	0.2443	0.0	0.1013	0.0	0.0	0.0	0.0
0.0159	16.0	2768	0.0182	0.0523	0.0615	0.0964	nan	0.1202	0.2493	0.0	0.1225	0.0	0.0	0.0	0.0	0.0	0.1044	0.2468	0.0	0.1199	0.0	0.0	0.0	0.0
0.0163	17.0	2941	0.0181	0.0492	0.0571	0.0892	nan	0.0987	0.2414	0.0	0.1167	0.0	0.0	0.0	0.0	0.0	0.0885	0.2397	0.0	0.1146	0.0	0.0	0.0	0.0
0.0152	17.3410	3000	0.0182	0.0441	0.0510	0.0800	nan	0.0908	0.2203	0.0	0.0971	0.0	0.0	0.0	0.0	0.0	0.0818	0.2189	0.0	0.0958	0.0	0.0	0.0	0.0

Framework versions

• Transformers 4.43.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1