Skip to content

[DNL] executorch export faster-rcnn #8788

New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

#

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

Jump to bottom
Draft
wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

[DNL] executorch export faster-rcnn #8788

wants to merge 1 commit into from

Conversation

yipjustin
Copy link

No description provided.

@pytorch-bot PyTorch Bot
Copy link

pytorch-bot bot commented Dec 6, 2024 •
edited
Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/8788

Note: Links to docs will display an error until the docs builds have been completed.

❌ 16 New Failures

As of commit 916ba03 with merge base 6279faa (image):

NEW FAILURES - The following jobs have failed:

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@yipjustin
Copy link
Author 

    return guard_scalar(self.sym_num)

For more information, run with TORCH_LOGS="dynamic"
For extended logs when we create symbols, also add TORCHDYNAMO_EXTENDED_DEBUG_CREATE_SYMBOL="u1"
If you suspect the guard was triggered from C++, add TORCHDYNAMO_EXTENDED_DEBUG_CPP=1
For more debugging help, see https://docs.google.com/document/d/1HSuTTVvYH1pTew89Rtpeu84Ht3nQEFTYhAX3Ypa_xJs/edit?usp=sharing

User Stack (most recent call last):
  (snipped, see stack below for prefix)
  File "/home/yipjustin/et-arm/vision/torchvision/models/detection/generalized_rcnn.py", line 104, in forward
    proposals, proposal_losses = self.rpn(images, features, targets)
  File "/home/yipjustin/miniconda3/envs/vision/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1747, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/yipjustin/et-arm/vision/torchvision/models/detection/rpn.py", line 373, in forward
    boxes, scores = self.filter_proposals(proposals, objectness, images.image_sizes, num_anchors_per_level)
  File "/home/yipjustin/et-arm/vision/torchvision/models/detection/rpn.py", line 289, in filter_proposals
    keep = box_ops.batched_nms(boxes, scores, lvl, self.nms_thresh)
  File "/home/yipjustin/et-arm/vision/torchvision/ops/boxes.py", line 72, in batched_nms
    return _batched_nms_vanilla(boxes, scores, idxs, iou_threshold)
  File "/home/yipjustin/miniconda3/envs/vision/lib/python3.12/site-packages/torch/_dynamo/polyfills/__init__.py", line 133, in getattr_and_trace
    return fn(*args[2:], **kwargs)
  File "/home/yipjustin/et-arm/vision/torchvision/ops/boxes.py", line 109, in _batched_nms_vanilla
    for class_id in idxs:

For C++ stack trace, run with TORCHDYNAMO_EXTENDED_DEBUG_CPP=1
For more information about this error, see: https://pytorch.org/docs/main/generated/exportdb/index.html#constrain-as-size-example

from user code:
   File "/home/yipjustin/et-arm/vision/torchvision/models/detection/generalized_rcnn.py", line 104, in forward
    proposals, proposal_losses = self.rpn(images, features, targets)
  File "/home/yipjustin/miniconda3/envs/vision/lib/python3.12/site-packages/torch/nn/modules/module.py", line 1747, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/yipjustin/et-arm/vision/torchvision/models/detection/rpn.py", line 373, in forward
    boxes, scores = self.filter_proposals(proposals, objectness, images.image_sizes, num_anchors_per_level)
  File "/home/yipjustin/et-arm/vision/torchvision/models/detection/rpn.py", line 289, in filter_proposals
    keep = box_ops.batched_nms(boxes, scores, lvl, self.nms_thresh)
  File "/home/yipjustin/et-arm/vision/torchvision/ops/boxes.py", line 72, in batched_nms
    return _batched_nms_vanilla(boxes, scores, idxs, iou_threshold)
  File "/home/yipjustin/miniconda3/envs/vision/lib/python3.12/site-packages/torch/_dynamo/polyfills/__init__.py", line 133, in getattr_and_trace
    return fn(*args[2:], **kwargs)
  File "/home/yipjustin/et-arm/vision/torchvision/ops/boxes.py", line 109, in _batched_nms_vanilla
    for class_id in idxs:

Set TORCH_LOGS="+dynamo" and TORCHDYNAMO_VERBOSE=1 for more information

digantdesai
digantdesai reviewed Dec 7, 2024
View reviewed changes
torchvision/ops/boxes.py
return _batched_nms_vanilla(boxes, scores, idxs, iou_threshold)
else:
return _batched_nms_coordinate_trick(boxes, scores, idxs, iou_threshold)
return _batched_nms_vanilla(boxes, scores, idxs, iou_threshold)
Copy link

Choose a reason for hiding this comment

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

Can we do what Angela suggested here?

# for free to join this conversation on GitHub. Already have an account? # to comment
Reviewers

@digantdesai digantdesai digantdesai left review comments

Assignees
No one assigned
Labels
cla signed
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@yipjustin @digantdesai @facebook-github-bot