This is really just a question: does the resulting image access the native GPUs on M1 Macs? In other words, can I use this image to train models on an M1 Mac?
================================================================================
PyTorch and TensorFlow for AArch64
================================================================================
ubuntu@6c43f6619444:~$ ipython
bash: ipython: command not found
ubuntu@6c43f6619444:~$ python
Python 3.8.10 (default, Jun 22 2022, 20:18:18)
[GCC 9.4.0] on linux
Type "help", "copyright", "credits" or "license"for more information.
>>> import platform
>>> platform.platform()
'Linux-5.15.49-linuxkit-aarch64-with-glibc2.29'
>>> import torch
>>> torch.backends.mps.is_available()
False
I am using M2 OSX 13.2. Python 3.11.1
How can I build docker image that enable MPS + PyTorch
To build on 64GB M1 Max MacBook Pro
Make sure to allocate 48GB of memory and 8 cpu to Docker.
$ ./build.sh --build-type full --jobs 8 --bazel_memory_limit 24576