rf-detr-seg-satellite

This model is a fine-tuned version of Roboflow/rf-detr-seg-medium on the merve/satellite-building-segmentation dataset. It achieves the following results on the evaluation set:

• Loss: 57.3546
• Mean Iou: 0.3320

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: 0.0001
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 0.1
• num_epochs: 10.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou
No log	0.0047	2	64.5888	0.1572
No log	1.0213	11	63.9776	0.1583
No log	2.0071	14	63.8052	0.1588
No log	3.0024	15	63.9838	0.1598
No log	4.0189	23	62.0472	0.1897
No log	5.0024	24	61.8414	0.1994
No log	6.0118	29	60.5323	0.2599
No log	7.0308	42	58.2103	0.2854
No log	8.0024	43	58.0191	0.2838
No log	9.0166	50	57.3546	0.3320

Framework versions

• Transformers 5.10.0.dev0
• Pytorch 2.12.0+cu130
• Datasets 5.0.0
• Tokenizers 0.22.2