Skip to content

Fixes patches on attention, call contiguous in torch.cond branches #288

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 6, 2025
Merged

Fixes patches on attention, call contiguous in torch.cond branches #288

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
17ea1c0
Fixes patches on attention
xadupre Nov 6, 2025
150d399
doc
xadupre Nov 6, 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
1 change: 1 addition & 0 deletions CHANGELOGS.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,7 @@ Change Logs
0.8.1
+++++

* :pr:`288`: add .contiguous in torch.cond branch (attention patch for sdpa implementation)
* :pr:`286`: adds variable to track random nodes in models

0.8.0
Expand Down
4 changes: 2 additions & 2 deletions onnx_diagnostic/torch_export_patches/patches/patch_transformers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1452,7 +1452,7 @@ def patched_sdpa_attention_forward(
scale=scaling,
is_causal=True,
**sdpa_kwargs,
),
).contiguous(),
lambda query, key, value: torch.nn.functional.scaled_dot_product_attention(
query,
key,
Expand All @@ -1461,7 +1461,7 @@ def patched_sdpa_attention_forward(
scale=scaling,
is_causal=False,
**sdpa_kwargs,
),
).contiguous(),
[query, key, value],
)
attn_output = attn_output.transpose(1, 2).contiguous()
Expand Down
Loading