Fix quant_state None on AMD GPUs by caching quant_state_dict at load time#3
Open
0xDELUXA wants to merge 1 commit intomengqin:mainfrom
Open
Fix quant_state None on AMD GPUs by caching quant_state_dict at load time#30xDELUXA wants to merge 1 commit intomengqin:mainfrom
0xDELUXA wants to merge 1 commit intomengqin:mainfrom
Conversation
0xDELUXA
force-pushed
the
fix/quant-state-none-on-amd
branch
from
77d32df to
2af2e9d
Compare
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
AMD support was added to bitsandbytes in: bitsandbytes-foundation/bitsandbytes#1846.
Params4bit.from_prequantizedcorrectly loads weights on AMD butquant_stateends upNoneby inference time, causing:Fix: Cache the raw
quant_state_dictasself._bnb_quant_state_dictat load time, and reconstruct theQuantStateobject on demand inforward()ifquant_stateis missing.Impact: No behavioral change on Nvidia - the fallback only triggers when
quant_state is None.Tested on AMD Radeon RX 9060 XT with Windows ROCm, bitsandbytes, and
z_image_turbo_nf4_v2.safetensors.