Add support for llama.cpp's --tensor-split parameter #460
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
added 2 commits
July 9, 2023 23:00
|
Hi @abetlen , Thank you for your time and consideration. |
|
@shouyiwang thank you for the contribution, lgtm |
|
@abetlen Thank you so much!! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
The current llama-cpp-python does not include support for --tensor-split parameter. When running a large model across two GPUs, it currently loads the model by default in a half-by-half manner. However, this approach presents certain issues. For example, when a user has two GPUs with different VRAM sizes, it can lead to OOM. Implementing the --tensor-split parameter will address this problem by empowering users to define the proportion of the model distributed across multiple GPUs.
I'm uncertain if importing ctypes into llama.py is the most appropriate approach. However, I'm currently unsure of an alternative solution in llama_cpp.py. I would greatly appreciate any advice or suggestions regarding this matter.
Tested thouroughly with text-generation-webui. I'll sumbit a PR there after this PR get merged. Thx!