yisol
Package compatibility issue on macOS ARM64 (M3 Pro)
Description
Unable to create the environment on macOS with ARM64 architecture (M3 Pro chip) due to package compatibility issues.
Body
When attempting to create the environment using the provided environment.yaml file on a macOS system with an ARM64 architecture (M3 Pro chip), the following error occurs:
> conda env create -f environment.yaml
Channels:
- pytorch
- nvidia
- defaults
Platform: osx-arm64
Collecting package metadata (repodata.json): done
Solving environment: failed
PackagesNotFoundError: The following packages are not available from current channels:
- torchvision==0.15.2=py310_cu118
- torchtriton==2.0.0=py310
- torchaudio==2.0.2=py310_cu118
- pytorch-cuda==11.8=h7e8668a_5
- pytorch==2.0.1=py3.10_cuda11.8_cudnn8.7.0_0
- python==3.10.0=h12debd9_5
- pip==23.3.1=py310h06a4308_0
Current channels:
- https://conda.anaconda.org/pytorch/osx-arm64
- https://conda.anaconda.org/nvidia/osx-arm64
- https://repo.anaconda.com/pkgs/main/osx-arm64
- https://repo.anaconda.com/pkgs/r/osx-arm64
To search for alternate channels that may provide the conda package you're
looking for, navigate to
https://anaconda.org
and use the search bar at the top of the page.
The error indicates that the required packages are not available for the osx-arm64 platform in the current channels.
Possible Solutions
- Check for ARM64-compatible versions of the required packages and update the environment.yaml file accordingly.
- Provide an alternative environment.yaml file specifically for macOS ARM64 systems, including the M3 Pro chip.
- Include instructions or workarounds for running the code on ARM64 systems, such as using Docker containers or virtual machines.
System Information
- Operating System: macOS Sonoma 14.1
- CPU Architecture: ARM64 (Apple M3 Pro chip)
Hi, i just was trying to run the conda environment and i was getting the same error on my M2. Did you modify the environment.yaml to get it working?
Yes, I did modify the environment.yaml to make the packages Mac compatible. However, the main issue that arises is that all the code is written for NVIDIA CUDA. So, you would have to modify the code base to make it compatible with MPS.
Which is too much of a hassle, so I abandoned attempting to run it locally.
That's what i concluded as well. I'll explore the possibility of running it with Docker.
I let you know if i get it working
you can run on cloud i have a full tutorial
IDM-VTON: The Most Amazing Virtual Clothing Try On Application - RunPod - Massed Compute - Kaggle
That's what i concluded as well. I'll explore the possibility of running it with Docker.
I let you know if i get it working
Hi there, did you achieve this ?
