Skip to content

[None][autodeploy] fix weight extraction for graph based quantized checkpoints #9109

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

Merged
merged 2 commits into from
Nov 13, 2025
Merged

[None][autodeploy] fix weight extraction for graph based quantized checkpoints #9109

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
dbf7e5a
fix weight extraction for graph based quantized checkpoints
Fridah-nv Nov 13, 2025
ae5711c
minor
Fridah-nv Nov 13, 2025
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
13 changes: 7 additions & 6 deletions tensorrt_llm/_torch/auto_deploy/utils/node_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -139,14 +139,15 @@ def find_get_attr_node(weight_node: Node) -> Node:
n for n in node.args if isinstance(n, Node) and find_get_attr_node(n) is not None
]
# can be two weights (if bias weight is present)
assert len(weight_nodes) >= 1, (
weight_node = None
if weight_nodes:
weight_node = weight_nodes[0]
# for modelopt quantized graph, there will be a quantize_op
_, weight_params, _ = get_quantization_params_from_linear_node(node)
weight_node = weight_params.input_node if weight_params else weight_node
assert weight_node is not None, (
"Expected exactly at least one weight node in the parametrized node"
)
weight_node = weight_nodes[0]
# for modelopt quantized graph, there will be a quantize_op
_, weight_params, _ = get_quantization_params_from_linear_node(node)
weight_node = weight_params.input_node if weight_params else weight_node

return find_get_attr_node(weight_node)


Expand Down