[General Quant Linear] Register quant params of general quant linear for friendly post process. #226
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Outdated suggestions cannot be applied.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.
Hi all, I have implemented support for loading auto-gptq pre-quantized checkpoints into MLC-LLM (based on the release 0.1.0 ). However this flow has encountered some issues with the
GeneralQuantLinear
in the release later than 0.2.0 has some issues of the load flow.As from the compiler side, we load params with three different ways: named_buffers(), named_paramters(), state_dicts(), however, because the qweight, qzeros is not registered in the GeneralQantLinear, which means we cannot retrieve them using any of these methods.
This issue prevents us from fully utilizing the pre-quantized auto-gptq checkpoints in MLC-LLM. I believe addressing this issue will significantly enhance the functionality and flexibility of our model loading or post processing.
Thanks, and please CC @PanQiWei