ROCm support #342
Open
ROCm support #342
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…l doesn't support 3.12.x and newer).
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Issue #141 proposed to modify poetry file to include ROCm support (for amdgpu acceleration), so I did. And it does work, however:
PYYAML support max Python 3.11, so the requires-python has to be changed (so it won't even attempt to build in Python 3.12 and newer);
memoryview is too large while installing torch on my system (poetry bug?), so I had to fallback to installing it manually:
pip install torch==2.7.1 torchvision==0.22.1 torchaudio==2.7.1 --index-url https://download.pytorch.org/whl/rocm6.3and only then runpoetry install --extras rocmto install the rest;Under glibc 2.41, libraries that request an executable stack (
libctranslate2in this case) cannot be loaded by default, so I had to also addGLIBC_TUNABLES=glibc.rtld.execstack=2into env for a workaround.My system in question is Debian Unstable. After all of that, it does work and does utilises GPU (RX 7800 XT, which works despite being officially unsupported by ROCm). For acceleration, it may be a bit slower that NVIDIA, but it's definetely faster than CPU.