-
Notifications
You must be signed in to change notification settings - Fork 69
open source alg_ext #1219
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
open source alg_ext #1219
Conversation
for more information, see https://pre-commit.ci
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Pull request overview
This PR open sources the alg_ext (algorithm extension) module, which provides advanced quantization algorithms and extensions for AutoRound. The changes introduce enhanced quantization techniques including support for MXFP, NVFP, and GGUF formats with double quantization capabilities.
Key changes:
- Added new
alg_ext.pymodule with advanced quantization algorithms and wrapper classes - Refactored
base.pyto eliminate duplicated conditional logic by introducing a reusableautocast_ctxcontext manager
Reviewed changes
Copilot reviewed 2 out of 4 changed files in this pull request and generated 6 comments.
| File | Description |
|---|---|
| auto_round/alg_ext.py | New file containing advanced quantization functions, custom wrapper classes (WrapperLinearV2, DQWrapperLinear), and GGUF-specific quantization algorithms |
| auto_round/compressors/base.py | Refactored to use a shared autocast context manager, reducing code duplication in _quantize_layer and _get_loss methods |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Co-authored-by: Copilot <[email protected]>
…to open_alg_ext
Co-authored-by: Copilot <[email protected]>
for more information, see https://pre-commit.ci
|
@chensuyue there is no need to compile alg_ext in the future |
xin3he
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It is recommended to add a detailed documentation (README file) about this algorithm.
Signed-off-by: n1ck-guo <[email protected]>
No description provided.