dusty-nv / jetson-containers

Machine Learning Containers for NVIDIA Jetson and JetPack-L4T
MIT License
2.09k stars 435 forks source link

How to rebuild PyTorch in l4t-pytorch Jetpack4.6.1 container #344

Closed yinfan98 closed 8 months ago

yinfan98 commented 8 months ago

@dusty-nv Hi, I need Python 3.8 in Jetpack4.6.1 l4t-pytorch docker.

only see Python 3.6 - torch-1.10.0-cp36-cp36m-linux_aarch64.whl at PyTorch for Jetson

How can I rebuild the PyTorch ? Thanks

dusty-nv commented 8 months ago

Hi @yinfan98, you can follow a similar procedure as in this dockerfile that builds pytorch wheels:

https://github.com/dusty-nv/jetson-containers/blob/master/packages/pytorch/Dockerfile.builder

however, install/use python3.8 instead of normal python3

also you may need to apply some patches to the source for older versions of PyTorch, you can find them under the Build from Source section of this thread: https://forums.developer.nvidia.com/t/pytorch-for-jetson/72048

yinfan98 commented 8 months ago

Thanks!