Skip to content

GGUF fix for unquantized types when using unquantize kernels #12498

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.

Sign up for GitHub

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
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

GGUF fix for unquantized types when using unquantize kernels #12498

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion src/diffusers/quantizers/gguf/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,8 @@
def _fused_mul_mat_gguf(x: torch.Tensor, qweight: torch.Tensor, qweight_type: int) -> torch.Tensor:
# there is no need to call any kernel for fp16/bf16
if qweight_type in UNQUANTIZED_TYPES:
return x @ qweight.T
weight = dequantize_gguf_tensor(qweight)
return x @ weight.T
Comment on lines +82 to +83
Copy link
Contributor

@Isotr0py Isotr0py Oct 17, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems dequantize_gguf_tensor missed implementation for FP16 and FP32 qweight:

diffusers/src/diffusers/quantizers/gguf/utils.py

Lines 487 to 501 in dbe4136

dequantize_functions = {
gguf.GGMLQuantizationType.IQ4_NL: dequantize_blocks_IQ4_NL,
gguf.GGMLQuantizationType.IQ4_XS: dequantize_blocks_IQ4_XS,
gguf.GGMLQuantizationType.BF16: dequantize_blocks_BF16,
gguf.GGMLQuantizationType.Q8_0: dequantize_blocks_Q8_0,
gguf.GGMLQuantizationType.Q5_1: dequantize_blocks_Q5_1,
gguf.GGMLQuantizationType.Q5_0: dequantize_blocks_Q5_0,
gguf.GGMLQuantizationType.Q4_1: dequantize_blocks_Q4_1,
gguf.GGMLQuantizationType.Q4_0: dequantize_blocks_Q4_0,
gguf.GGMLQuantizationType.Q6_K: dequantize_blocks_Q6_K,
gguf.GGMLQuantizationType.Q5_K: dequantize_blocks_Q5_K,
gguf.GGMLQuantizationType.Q4_K: dequantize_blocks_Q4_K,
gguf.GGMLQuantizationType.Q3_K: dequantize_blocks_Q3_K,
gguf.GGMLQuantizationType.Q2_K: dequantize_blocks_Q2_K,
}

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

it was bf16 in my use case.

fp16 and fp32 would fail in any case, whether native or dequant kernels are used.
this PR therefore currently only fixes the bf16 case for kernel dequant - for native bf16 already works


# TODO(Isotr0py): GGUF's MMQ and MMVQ implementation are designed for
# contiguous batching and inefficient with diffusers' batching,
Expand Down