Skip to content

Avoid autocast at onnx export if fp32 model is desired #304

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 3 commits into from
Sep 10, 2025
Merged

Avoid autocast at onnx export if fp32 model is desired #304

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
b5f912e
Avoid autocast at onnx export if fp32 model is desired
i-riyad Sep 9, 2025
ce490b3
Merge branch 'main' into rislam/autocast-avoid-for-fp32
i-riyad Sep 9, 2025
ad8efc2
Merge branch 'main' into rislam/autocast-avoid-for-fp32
i-riyad Sep 10, 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
4 changes: 3 additions & 1 deletion modelopt/torch/_deploy/utils/torch_onnx.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -393,7 +393,9 @@ def get_onnx_bytes_and_metadata(
# during inference.
input_none_names = list(set(tree_spec_input.names) - set(input_names))

use_torch_autocast = not (is_fp4_quantized(model) or is_mxfp8_quantized(model))
use_torch_autocast = not (
is_fp4_quantized(model) or is_mxfp8_quantized(model) or weights_dtype == "fp32"
)
autocast = torch.autocast("cuda") if use_torch_autocast else nullcontext()

# Get output once (we export in inference mode - so also using inference mode here!)
Expand Down
Loading