Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Conda env fails to run #77

Open
DukeofCambridge opened this issue Nov 13, 2024 · 2 comments
Open

Conda env fails to run #77

DukeofCambridge opened this issue Nov 13, 2024 · 2 comments

Comments

@DukeofCambridge
Copy link

Hi, I can run successfully with virtual environment but failed with conda environment.
All the packages are the same (torch 2.4.1, torch-vision 0.19.1, xformer 0.0.28.post1, etc.) but I just cannot run successfully in a conda environment.
20241113 mochi failure

I also encountered problem mentioned in (fixed by downgrading torch from 2.5.x to 2.4.1):

@DukeofCambridge
Copy link
Author

The following are my commands to create the env:

git clone https://github.com/genmoai/models
cd models
conda create -n mochi python=3.10
conda activate mochi
pip install uv
uv pip install setuptools
uv pip install -e . --no-build-isolation

I also tried downloading flash-attn from https://github.com/Dao-AILab/flash-attention/releases/tag/v2.6.3.
After flash-attn is installed, I ran "uv pip install -e .[flash] --no-build-isolation". No problem here. But still fails to run the inference (both single-GPU and multi-GPU)

@vedantroy
Copy link
Contributor

It looks like it is not detecting flash attention. Can you run uv pip list | grep flash and make sure that flash attention is installed?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants