Experience with Streamlit Sharing

Community Cloud
1

I finally got an invite from Streamlit for their fantastic cloud-based share platform. For fun, I hosted the web app that was focused on making streamlit easier to use for image classification prototypes as well as quick development.

Here is the deploy

[ This is an open-source simplified image classification web template that I built for making prototyping deep learning models easier, faster, and more manageable. I have also included guides for Firebase and hosting on multiple cloud services. This is a WORK IN PROGRESS repo]

Features I am hoping to see in the coming times:

  • GPU Support

  • Creating a Pipeline for the apps so that one can decide which version to keep for beta testing and which one to push for production

  • Workflow Support similar to netlify to get a preview for each PR

  • Ability to host from private repos

  • DNS Setting changes [really required. I am currently masking the app link with my own domain instead of setting a CNAME record]

  • Ability to control the visibility of apps

  • Increase the no of apps allowed per account. Perhaps some discount with the GitHub Developer Pack for Students

  • Ability to form exclusive teams on the platform rather than on the GitHub repo

  • Integration of an analytics feature

4 Likes
2

Hi @smaranjitghose, welcome to the Streamlit community!

Thanks for your feedback, glad you are finding Streamlit sharing useful! I’ll be sure to pass on this feedback to our product team, as I’m sure they are interested in hearing real-world use-cases!

Best,
Randy

1 Like
3

Hey Team,
Is there any way I can cite Streamlit in my research paper?

Regards,
Smaranjit Ghose

4

@smaranjitghose I’m curious how you did the domain masking, can you elaborate on that?

1 Like
5

The version of the template above that I use (Not streamlit sharing) for my personal projects is on Heroku mostly ( can be used for AWS, GCP too). So adding a CNAME record and modifying the DNS records in my personal domain’s control panel does the work for me.

This is an elaborate documentation of the same for Heroku.

The link to my github repo for the template is: GitHub - smaranjitghose/img_ai_app_boilerplate: An image classification app boilerplate to serve your deep learning models asap!

1 Like