Use Gradio or Streamlit on SageMaker Studio Lab

This repository shows a quick demo for how to run Gradio or Streamlit applications on SageMaker Studio Lab. Following the same capability for Tensorboard on SageMaker Studio, you can now apply the same to work with your Streamlit/Gradio application, except the default port (8051) set by Streamlit is not open.

streamlit-app.mp4

Watch YouTube Explainer Video

Getting started

• https://streamlit.io/
• https://gradio.app/
• https://github.com/nicolasmetallo/deploy-streamlit-on-fargate-with-aws-cdk
• https://huggingface.co/spaces/keras-io/Monocular-Depth-Estimation
• https://docs.aws.amazon.com/sagemaker/latest/dg/studio-tensorboard.html

Requirements

• SageMaker Studio Lab account. See video for more information.
• Python 3.8+
• Keras
• Numpy
• Streamlit
• Gradio

How to run your apps

• 0_demo_notebook.ipynb Notebook that runs Monocular Depth Estimation in Keras.
• 1_launch_gradio_streamlit.ipynb Notebook with quick start to launch your apps.
• app_gradio.py Gradio application
• app_streamlit.py Streamlit application

Gradio

Use inline=False and server_port=6006.

gr.Interface(
    fn=infer,
    title="Monocular Depth Estimation",
    description = "Keras Implementation of Unet architecture with Densenet201 backbone for estimating the depth of image 📏",
    inputs=[gr.inputs.Image(label="image", type="numpy", shape=(640, 480))],
    outputs="image",
    article = "Author: <a href=\"https://huggingface.co/vumichien\">Vu Minh Chien</a>. Based on the Keras example from <a href=\"https://keras.io/examples/vision/depth_estimation/\">Victor Basu</a>. Repo: https://github.com/machinelearnear/use-gradio-streamlit-sagemaker-studiolab",
    examples=examples).launch(inline=False, server_port=6006, debug=True, cache_examples=True)

Then run with !python app_gradio.py either from the Terminal or from the 1_launch_gradio_streamlit.ipynb Notebook.

Streamlit

Use server.port 6006 and run from the Terminal or from the 1_launch_gradio_streamlit.ipynb Notebook.

!streamlit run app_streamlit.py --server.port 6006 # or 80/8080

References

See more implementations here https://paperswithcode.com/task/monocular-depth-estimation