segformer-finetuned-obb-1k-steps

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

• Loss: 0.0511
• Mean Iou: 0.2238
• Mean Accuracy: 0.4477
• Overall Accuracy: 0.4477
• Accuracy Backgound : nan
• Accuracy Rwy Obb: 0.4477
• Iou Backgound : 0.0
• Iou Rwy Obb: 0.4477

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: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Backgound	Accuracy Rwy Obb	Iou Backgound	Iou Rwy Obb
0.3927	1.0	173	0.1096	0.1590	0.3180	0.3180	nan	0.3180	0.0	0.3180
0.0969	2.0	346	0.0704	0.2112	0.4224	0.4224	nan	0.4224	0.0	0.4224
0.0651	3.0	519	0.0598	0.2186	0.4371	0.4371	nan	0.4371	0.0	0.4371
0.0576	4.0	692	0.0530	0.2250	0.4500	0.4500	nan	0.4500	0.0	0.4500
0.0531	5.0	865	0.0529	0.2212	0.4424	0.4424	nan	0.4424	0.0	0.4424
0.0467	5.7803	1000	0.0511	0.2238	0.4477	0.4477	nan	0.4477	0.0	0.4477

Framework versions

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