feat(spinquant): refactor SpinQuant with CPU offload, parallel fuse, meta device support, and vLLM export#269
Merged
Conversation
…meta device support, and vLLM export
- Move rotation fuse computation to CPU to avoid GPU OOM on large models
- Add multi-threaded parallel fuse for R1/R2/R4 with ThreadPoolExecutor
- Support meta device weights via accelerate hook-based materialization
- Add block-diagonal Hadamard (had_dim > 0) for R4 rotation
- Auto-generate transform_config for vLLM deployment when R4 is enabled
- Save transform_config in quantization_config during model export
- Fix fuse_ln_linear to compute on CPU and write back to original device
- Add .detach() in pack_weight_to_int8 to prevent gradient errors
- Add comprehensive SpinQuant documentation (spinquant.md)
gavingavin99
force-pushed
the
dev_rotation0320
branch
from
0000af1 to
bdf4fe8
Compare
added 3 commits
March 24, 2026 12:19
yghstill
approved these changes
Mar 24, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
feat(spinquant): refactor SpinQuant with CPU offload, parallel fuse, meta device support, and vLLM export