Turn Python Scripts into Beautiful ML Tools
summary
Claim: a new open-source app framework called streamlet turns normal Python scripts into interactive ML tools. Concrete demos: pip install streamlet, a 300-line self-driving data browser with ~30 streamlet calls that runs a YOLO v3 model in real time and lets users filter and visualize pickups on a deck GL map. Key primitives shown are st.cache for fast reruns, widget variables like st.slider, and spatial views via st.map. Runs locally, supports saving to S3 via boto3 and a team deployment product is in development.
AI-written summary · about
chapters (9)
- 0:00Streamlet framework intro
- 1:14ML app development problems
- 8:59What is Streamlet
- 12:13Installing Streamlet demos
- 26:12Streamlet’s three principles
- 31:36Building a data app
- 46:19ML debugging use case
- 55:03Self-driving car demo
- 1:01:13Q&A and future
related talks
42mStreamlit: A Faster Way to Build and Share Data Apps
21mWeb apps for data science using streamlit
18mStreamlit: The Fastest Way to build Data Apps
01h 58mHow to create beautiful interactive GUIs and web apps
28mPython Apps in the Browser made simple by PyCafe
18mRapid Prototyping in AI Project using Colab and Streamlit
09mBuild Interactive Web Apps with Python: Streamlit and Supabase
28mBeyond Prototyping: Building Production-Level Apps with Streamlit