deep-diver / semantic-segmentation-ml-pipeline

Machine Learning Pipeline for Semantic Segmentation with TensorFlow Extended (TFX) and various GCP products
https://blog.tensorflow.org/2023/01/end-to-end-pipeline-for-segmentation-tfx-google-cloud-hugging-face.html
Apache License 2.0
93 stars 20 forks source link

feat: add gradio demo notebook #17

Closed sayakpaul closed 2 years ago

sayakpaul commented 2 years ago

Closes #16.

review-notebook-app[bot] commented 2 years ago

Check out this pull request on  ReviewNB

See visual diffs & provide feedback on Jupyter Notebooks.


Powered by ReviewNB

deep-diver commented 2 years ago

There are three things to comment:

  1. Since we are going to inject which model to use from 🤗 Model Hub, we need to reference a model hosted in 🤗 Model Hub. I have prepared a test repository just now: https://huggingface.co/chansung/test-unet/tree/main.

  2. In order to generalize a bit, I think it could be better to make labels.txt file while removing def sidewalk_palette() from the code.

  3. It would be nice to include a couple of default images to run inference

WDYT?

sayakpaul commented 2 years ago

In order to generalize a bit, I think it could be better to make labels.txt file while removing def sidewalk_palette() from the code.

Sure we can do that when writing code for the Hugging Face space. For a notebook that benefits from interactive usage, I think it would introduce unnecessary back and forth.

I will address rest of the comments tomorrow.

sayakpaul commented 2 years ago

@deep-diver up for another review.

deep-diver commented 2 years ago

LGTM! Thanks for the great work :)