Skip to content

fix: get_padded_shape_tensors can now handle dynamic pads #3123

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

Jump to bottom
Merged
merged 2 commits into from
Sep 3, 2024
Merged

fix: get_padded_shape_tensors can now handle dynamic pads #3123

merged 2 commits into from
Sep 3, 2024

Conversation

jiwoong-choi
Copy link

@jiwoong-choi jiwoong-choi commented Aug 28, 2024 •
edited
Loading

Description

Exported programs with dynamic shapes often include aten.constant_pad_nd with the argument pad containing mixture of int and TRTTensor values. This makes pad_before + pad_after fails in get_padded_shape_tensors.

Type of change

Please delete options that are not relevant and/or add your own.

  • New feature (non-breaking change which adds functionality)

Checklist:

  • My code follows the style guidelines of this project (You can use the linters)
  • I have performed a self-review of my own code
  • I have commented my code, particularly in hard-to-understand areas and hacks
  • I have made corresponding changes to the documentation
  • I have added tests to verify my fix or my feature
  • New and existing unit tests pass locally with my changes
  • I have added the relevant labels to my PR in so that relevant reviewers are notified

fix: get_padded_shape_tensors can now handle dynamic pads
06afd81 
Exported programs with dynamic shapes often include aten.constant_pad_nd with `pad` containing TRTTensor values
@github-actions github-actions bot added component: conversion Issues re: Conversion stage component: converters Issues re: Specific op converters component: api [Python] Issues re: Python API component: dynamo Issues relating to the `torch.compile` or `torch._dynamo.export` paths labels Aug 28, 2024
@github-actions github-actions bot requested a review from peri044 August 28, 2024 12:41
@jiwoong-choi jiwoong-choi marked this pull request as ready for review August 28, 2024 12:42
peri044
peri044 reviewed Aug 29, 2024
View reviewed changes
Copy link
Collaborator

@peri044 peri044 left a comment

Choose a reason for hiding this comment

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

Added a minor comment. Changes LGTM

py/torch_tensorrt/dynamo/conversion/impl/pad.py Outdated
)
padded_shape_tensor = impl.cat.cat(
ctx, target, source_ir, f"{name}_cat_dim_{i}", slices, 0
ctx, target, source_ir, f"{name}_cat_dim_{i}", slices, 0, cast_dtype=np.int32
Copy link
Collaborator

Choose a reason for hiding this comment

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

can you use cast_dtype=padded_shape_tensor.dtype instead of assuming int32 ?

@jiwoong-choi jiwoong-choi requested a review from peri044 August 30, 2024 10:46
@peri044 peri044 merged commit ae7e6c8 into pytorch:main Sep 3, 2024
69 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@peri044 peri044 Awaiting requested review from peri044

Assignees
No one assigned
Labels
cla signed component: api [Python] Issues re: Python API component: conversion Issues re: Conversion stage component: converters Issues re: Specific op converters component: dynamo Issues relating to the `torch.compile` or `torch._dynamo.export` paths
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@jiwoong-choi @peri044 @facebook-github-bot