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

Tfrecord Dump Error #866

Closed
ghost opened this issue Nov 26, 2019 · 2 comments · Fixed by #894
Closed

Tfrecord Dump Error #866

ghost opened this issue Nov 26, 2019 · 2 comments · Fixed by #894
Assignees
Labels
enhancement New feature or request
Milestone

Comments

@ghost
Copy link

ghost commented Nov 26, 2019

If you dump annotation for TFRecord Zip 1.0 then there will be a lot of problem with the dumped file. (/annotation/tfrecord.py)

  1. First problem is with label_map.pbtxt. The problem is id = 0 is reserved for background. So, it needs to start from 1 NOT 0. This can easily be fixed.

  2. Tfrecord contains encoded images so that training can be faster. But if you look at the current file, it does not encode images into tfrecord. If you use this tfrecord in TF Detection API, it will throw an error.

@nmanovic
Copy link
Contributor

@san999p ,

Thanks for the report. The second problem will be fixed in our new dataset framework (datumaro). Dump annotations is used only to dump annotations (without images). We recently introduced export as dataset feature. Now it doesn't support TFRecord format but it will do. The feature will export images + annotations. Please stay tuned.

@zhiltsov-max
Copy link
Contributor

@san999p, please check if this PR satisfies your needs - you can test it by exporting a dataset.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants