yolox-burn continous issues without cuda

[Raj2032]

I am having errors like this:

lib/python3.14/site-packages/torch/cuda/__init__.py", line 417, in _lazy_init
    raise AssertionError("Torch not compiled with CUDA enabled")

AssertionError: Torch not compiled with CUDA enabled

when running:

CUDA_VISIBLE_DEVICES="" python tools/train.py -f ../yolox_custom.py -b 

I want to train model with only cpu, is there a way I can do so?

[laggui]

This is not related to burn at all 😅

But it looks like your pytorch installation was not compiled with CUDA. If you want to use CUDA, make sure you download the correct wheel (see https://pytorch.org/get-started/locally/).

If you don't want to use CUDA, make sure you're not using any cuda features in your training script.

Though training on the CPU will be really slow I think.

[Raj2032]

This is not related to burn at all 😅

Hi mate yes that is true I will keep that in mind :)

I don't want to at all use cuda I just want to use cpu only, so I installed pytorch only cpu version, however I don't know why that error msg keeps popping up.

[laggui]

The yolox repo code leverages cuda for training, with a couple of .cuda() calls. I haven't looked at the whole code but perhaps CPU training is not supported.

Your cpu-only pytorch version therefore returns an assertion error on cuda initialization because it's not available.