Skip to content

[6106576] Restore qdq_utils symbols used by edgellm 0.6.1 #1408

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
ajrasane wants to merge 2 commits into
base: main
Choose a base branch
from
+493 −1
Open

[6106576] Restore qdq_utils symbols used by edgellm 0.6.1 #1408

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
c61ac69
[6106576] Restore qdq_utils symbols used by edgellm 0.6.1
ajrasane May 6, 2026
c7e10f4
[6106576] Address PR review feedback on edgellm shim restoration
ajrasane May 8, 2026
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
8 changes: 8 additions & 0 deletions modelopt/onnx/export/nvfp4_exporter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,6 +39,10 @@ def _cast_fp4(array: np.ndarray) -> np.ndarray:

Note: The first dimension of the array must be divisible by 2
as two FP4 values are packed into a single byte.

Also reused by the deprecated ``modelopt.onnx.quantization.qdq_utils.fp4qdq_to_2dq``
compatibility shim. Do not rename or change the signature without updating that
shim (it is a load-bearing re-export for TensorRT-Edge-LLM 0.6.1).
"""
array_f32_t = torch.from_numpy(array)
array_f32_t_shape = array_f32_t.shape
Expand Down Expand Up @@ -76,6 +80,10 @@ def _replace_fp4qdq_with_2dq(
):
"""Replaces the given node in the ONNX graph with a subgraph consisting of two DequantizeLinear nodes.

Also reused by the deprecated ``modelopt.onnx.quantization.qdq_utils.fp4qdq_to_2dq``
compatibility shim. Do not rename or change the signature without updating that
shim (it is a load-bearing re-export for TensorRT-Edge-LLM 0.6.1).

Args:
graph: The ONNX graph containing the node to replace.
node: The node to be replaced.
Expand Down
Loading
Loading