Fix GGMLLayer reconstruction for GGUF quantized weights#6
Draft
maxious wants to merge 3 commits intoFearL0rd:mainfrom
Draft
Fix GGMLLayer reconstruction for GGUF quantized weights#6maxious wants to merge 3 commits intoFearL0rd:mainfrom
maxious wants to merge 3 commits intoFearL0rd:mainfrom
Conversation
added 2 commits
April 7, 2026 01:47
GGMLLayer (from gguf.ops or gguf.pig) stores quantized weights as plain tensor attributes rather than nn.Parameter. When clone_module_simple() tries to reconstruct it via the generic handler, it wraps the quantized weight in nn.Parameter(), which fails with 'Only Tensors of floating point and complex dtype can require gradients'. This special case handles GGMLLayer by: - Creating new instance via __new__ + nn.Module.__init__ - Deep-copying non-weight/bias attributes - Moving weight/bias tensors to target device via .to() instead of wrapping This stops the warning flood: 'Warning: Failed to reconstruct... Only Tensors of floating point and complex dtype can require gradients'
Handles GGMLOps.Linear from ComfyUI-GGUF in addition to gguf.ops and gguf.pig
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
GGMLLayer (from gguf.ops or gguf.pig) stores quantized weights as plain tensor attributes rather than nn.Parameter. When clone_module_simple() tries to reconstruct it via the generic handler, it wraps the quantized weight in nn.Parameter(), which fails with 'Only Tensors of floating point and complex dtype can require gradients'.
This special case handles GGMLLayer by:
This stops the warning flood: 'Warning: Failed to reconstruct... Only Tensors of floating point and complex dtype can require gradients'