deep-diver/semantic-segmentation-ml-pipeline

Templating Gradio app with placeholders

deep-diver opened this issue · 0 comments

deep-diver commented

Write a Gradio app in this folder based on this notebook.

To-dos

  • replace labels.txt with sidewalks dataset's labels
  • clean-up requirements.txt
  • re-write app.py. Include placeholders where to look for the model from 🤗 Model Hub. There should be two placeholders, one for the repo ID of the 🤗 Model Hub, and the other for the revision, and they are represented as $MODEL_REPO_ID@$MODEL_VERSION
    • for instance, if we want to look for the model from chansung/test-unet on main branch, $MODEL_REPO_ID@$MODEL_VERSION will be replaced with the chansung/test-unet@main in runtime.