qengineering/pytorch-raspberry-pi-64-os

PyTorch installation wheels for Raspberry Pi 64 OS

PyTorch wheels for Raspberry Pi 4

Find your operating system and Pytorch version in the table below. Follow the instructions in the provided guide.
There is no Raspberry Pi 32-bit wheel available due to unsupported libraries.
The Jetson Nano wheels support CUDA 10.2, cuDNN 8.0 and NEON. They can also be used on the (AGX) Xavier.

---

Bookworm.

If you have a Bookworm OS on your Rpi4 or Rpi5, you can use the wheel provided by PyTorch.
The wheels here do not support Bookworm with its Python version 3.11.
Please see our site: install pytorch on raspberry pi 5.

---

Wheel: the installation wheel torch-version-cpxx-cpxx-linux_aarch64.whl (xx is the used python version)
Vision: the accompanying torchvision.
LibTorch: the C++ API for those who like to program. (The aarch64 version of libtorch-cxx11-abi-shared-with-deps-1.10.1+cpu.zip)
Guide: link to the installation tutorial.

Roadmap.

Operating system	PyTorch 2.0.0	PyTorch 1.13.0	PyTorch 1.12.0	PyTorch 1.11.0	PyTorch 1.10.0	PyTorch 1.9.0	PyTorch 1.8.0	PyTorch 1.7.0
Raspberry Pi 64-bit Bullseye (Python 3.9)	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide
Raspberry Pi 64-bit Buster (Python 3.7)		Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide
Raspberry Pi Ubuntu 18.04 (Python 3.6)							Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide
Raspberry Pi Ubuntu 20.04 (Python 3.8)					Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide
Jetson Nano JetPack 4.6 (Python 3.6)					Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision Guide	Wheel Vision Guide
Jetson Nano Ubuntu 20.04 (Python 3.8)		Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide	Wheel Vision LibTorch Guide

We compiled the Pytorch 1.x versions with the clang compiler to prevent issues with the ARM NEON registers and the GNU compiler. For instance #61110 and #65673.




Pytorch 2.0 was build with the GNU compiler.

❗ C++ programmers please note

You should also use the clang compiler if you want to compile Pytorch 1.x versions C++ code yourself.
The GNU GCC compiler will give you 'no expression errors'.

# set clang compiler at the command line
$ export CC=clang
$ export CXX=clang++

Don't worry if you plan to use Python. It only applies to C++ users.

---

Find PyTorch and TorchVision with other frameworks and deep-learning examples on our SD-image

---

pytorch-raspberry-pi-64-os's Issues

are these installation wheels cpuonly version?

Hey there,

I was wondering if these wheels are cpuonly installations of pytorch? or are they cuda versions?

Thanks.

slow forward processing in Pi4

Hello, I hope you don't mind as my question is a bit unrelated about this repo.

I'm trying to use PyTorch to do some ML training tasks on Raspberry pi4. Specifically, I'm trying to do an inference task using MobileNet model. I realized that the forward processing step is done very slow in Pi4 as it doesn't have any GPU. I was wondering if you possibly have any suggestions to increase the performance?

Cheers,

Torch wheels for bookworm version

It seems that the bullseye version wheels cannot fit the newest bookworm version, is there any wheels for bookworm version? Or how can I complie the bookworm version wheels on my own? For those who in urgent need, is switching the OS to older bullseye version a better choice? (sad)
ERROR: torch-2.0.0a0+gite9ebda2-cp39-cp39-linux_aarch64.whl is not a supported wheel on this platform.

Wheel does not work on Raspbian Bullseye

How can I fix this?

Torchaudio & pocketfft 0.10

I think finally Torch & Torch audio can be compiled with alternatives for cpu than intel MKL

I had a go with https://github.com/KumaTea/pytorch-aarch64 but the examples I ran from speechbrain just seem to return null without error.

I was wondering if your planning on providing torchaudio also?

Stuart

RPI 5 install not supported

I tried to install torch-1.11.0 on RPI 5 but it failed:

pip install torch-1.11.0a0+gitbc2c6ed-cp39-cp39-linux_aarch64.whl --no-index
ERROR: torch-1.11.0a0+gitbc2c6ed-cp39-cp39-linux_aarch64.whl is not a supported wheel on this platform.

I followed the instruction for RPI 4 OS 64 bit

model for 32 bit ?

do you have the model for 32 bits ? could you share ?