roboflow
Questions about Multi-GPU training
Search before asking
- [x] I have searched the RF-DETR issues and found no similar bug report.
Bug
why i can not train rfdetr on multi-GPUs with the scripts in readme
python -m torch.distributed.launch --nproc_per_node=3 --use_env main.py
Traceback (most recent call last): File "/rf-detr/main.py", line 560, in <module> parser = argparse.ArgumentParser('LWDETR training and evaluation script', parents=[get_args_parser()]) NameError: name 'get_args_parser' is not defined parser = argparse.ArgumentParser('LWDETR training and evaluation script', parents=[get_args_parser()]) NameError: name 'get_args_parser' is not defined parser = argparse.ArgumentParser('LWDETR training and evaluation script', parents=[get_args_parser()]) NameError: name 'get_args_parser' is not defined
Environment
- OS: Ubuntu 20.04
- Python: 3.9.0
- PyTorch: 2.2.0
- GPU: A30
Minimal Reproducible Example
Additional
No response
Are you willing to submit a PR?
- [ ] Yes, I'd like to help by submitting a PR!
You should first create a main.py script that initializes your model and calls .train() as usual than run it in terminal. like this. hope it works
model = RFDETRBase()
model.train(dataset_dir=<DATASET_PATH>, epochs=10, batch_size=4, grad_accum_steps=4, lr=1e-4, output_dir=<OUTPUT_PATH>)```
