Skip to content

Latest commit

 

History

History
26 lines (18 loc) · 1.08 KB

File metadata and controls

26 lines (18 loc) · 1.08 KB

Training & Testing

We trained MAFF-Net for a total of 60 epochs on a single NVIDIA RTX 4090 GPU with a batch size of 4.

The final trained checkpoints for both VoD and TJ4DRadSet are available for download:

Place all checkpoint files under the following directory: MAFF-Net/checkpoints/.

Test and evaluate the models

  • Test with a model:
python tools/test.py --cfg_file tools/cfgs/MAFF-Net/MAFF-Net_vod.yaml --batch_size 1 --ckpt checkpoints/vod2025010810360.pth
python tools/test.py --cfg_file tools/cfgs/MAFF-Net/MAFF-Net_TJ4D.yaml --batch_size 1 --ckpt checkpoints/tj4d2025060814210.pth

Train a model

  • Train with a single GPU:
python tools/train.py --cfg_file tools/cfgs/MAFF-Net/MAFF-Net_vod.yaml
python tools/train.py --cfg_file tools/cfgs/MAFF-Net/MAFF-Net_TJ4D.yaml