Skip to content

GGUF: torch.compile cannot trace sets #12556

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
+7 −7
Open

GGUF: torch.compile cannot trace sets #12556

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
14 changes: 7 additions & 7 deletions src/diffusers/quantizers/gguf/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,23 +41,23 @@
else:
ops = None

UNQUANTIZED_TYPES = {gguf.GGMLQuantizationType.F32, gguf.GGMLQuantizationType.F16, gguf.GGMLQuantizationType.BF16}
STANDARD_QUANT_TYPES = {
UNQUANTIZED_TYPES = [gguf.GGMLQuantizationType.F32, gguf.GGMLQuantizationType.F16, gguf.GGMLQuantizationType.BF16]
STANDARD_QUANT_TYPES = [
Copy link
Collaborator

Choose a reason for hiding this comment

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

we use set operations on them like here

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

Line 74 in df8dd77

DEQUANT_TYPES = STANDARD_QUANT_TYPES | KQUANT_TYPES | IMATRIX_QUANT_TYPES

would this work?

Copy link
Author

@dxqb dxqb Oct 30, 2025 •
edited
Loading

Choose a reason for hiding this comment

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

you are right, this would not work and another solution is required

gguf.GGMLQuantizationType.Q4_0,
gguf.GGMLQuantizationType.Q4_1,
gguf.GGMLQuantizationType.Q5_0,
gguf.GGMLQuantizationType.Q5_1,
gguf.GGMLQuantizationType.Q8_0,
gguf.GGMLQuantizationType.Q8_1,
}
KQUANT_TYPES = {
]
KQUANT_TYPES = [
gguf.GGMLQuantizationType.Q2_K,
gguf.GGMLQuantizationType.Q3_K,
gguf.GGMLQuantizationType.Q4_K,
gguf.GGMLQuantizationType.Q5_K,
gguf.GGMLQuantizationType.Q6_K,
}
IMATRIX_QUANT_TYPES = {
]
IMATRIX_QUANT_TYPES = [
gguf.GGMLQuantizationType.IQ1_M,
gguf.GGMLQuantizationType.IQ1_S,
gguf.GGMLQuantizationType.IQ2_XXS,
Expand All @@ -67,7 +67,7 @@
gguf.GGMLQuantizationType.IQ3_S,
gguf.GGMLQuantizationType.IQ4_XS,
gguf.GGMLQuantizationType.IQ4_NL,
}
]
# TODO(Isotr0py): Currently, we don't have MMQ kernel for I-Matrix quantization.
# Consolidate DEQUANT_TYPES, MMVQ_QUANT_TYPES and MMQ_QUANT_TYPES after we add
# MMQ kernel for I-Matrix quantization.
Expand Down
Loading