Skip to content
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.

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

[deeplab] Add Cityscapes and ADE20K to the notebook demo #6991

Open
sdll opened this issue Jun 10, 2019 · 1 comment
Open

[deeplab] Add Cityscapes and ADE20K to the notebook demo #6991

sdll opened this issue Jun 10, 2019 · 1 comment

Comments

@sdll
Copy link

@sdll sdll commented Jun 10, 2019

System information

  • What is the top-level directory of the model you are using: deeplab
  • Have I written custom code (as opposed to using a stock example script provided in TensorFlow): n/a
  • OS Platform and Distribution (e.g., Linux Ubuntu 16.04): n/a
  • TensorFlow installed from (source or binary): n/a
  • TensorFlow version (use command below): n/a
  • Bazel version (if compiling from source): n/a
  • CUDA/cuDNN version: n/a
  • GPU model and memory: n/a
  • Exact command to reproduce: n/a

Feature Request

The current demo only includes an example for using the DeepLab model trained on the Pascal dataset.

Judging from the issues, this is a common issue, and since all of the required infra seems to be there, adding support for Cityscapes and ADE20K will serve as a good first PR for the community newcomers and help eliminate the confusion around different labelling schemes for those who see the demo first. Personally, this bit me while porting DeepLab to tfjs .

In particular, adding Cityscapes should be easy in terms of the size of additional code necessary for the visualization. I have translated the colormap in the utils to the corresponding labels:

cityscapes_labels = [
      'road',
      'sidewalk',
      'building',
      'wall',
      'fence',
      'pole',
      'traffic light',
      'traffic sign',
      'vegetation',
      'terrain',
      'sky',
      'person',
      'rider',
      'car',
      'truck',
      'bus',
      'train',
      'motorcycle',
      'bicycle',
]

Do I understand correctly that the frozen model trained on Cityscapes produces labels according to the list above?

UPDATE 11/06: I can confirm that it does. See the segmentation demo below:

cityscapes-example

@sdll sdll changed the title [deeplab demo] Adding Cityscapes and ADE20K to the notebook demo [deeplab] Adding Cityscapes and ADE20K to the notebook demo Jun 10, 2019
@sdll sdll changed the title [deeplab] Adding Cityscapes and ADE20K to the notebook demo [deeplab] Add Cityscapes and ADE20K to the notebook demo Jun 10, 2019
@tensorflowbutler

This comment has been minimized.

Copy link
Member

@tensorflowbutler tensorflowbutler commented Jun 12, 2019

Thank you for your post. We noticed you have not filled out the following field in the issue template. Could you update them if they are relevant in your case, or leave them as N/A? Thanks.
What is the top-level directory of the model you are using
Have I written custom code
OS Platform and Distribution
TensorFlow installed from
TensorFlow version
Bazel version
CUDA/cuDNN version
GPU model and memory
Exact command to reproduce

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Linked pull requests

Successfully merging a pull request may close this issue.

None yet
3 participants
You can’t perform that action at this time.