Building AI-powered apps with LlamaIndex and Streamlit

In this video, Andrei Fajardo from ‪@LlamaIndex‬ shows us how to build an Abstraction and Reasoning Corpus (ARC) Task Solver app with LlamaIndex Workflows and Streamlit. ARC Task Solver with Human Input Streamlit App built with LlamaIndex Workflows. ARC Tasks are challenging puzzles that come from the ARC-AGI benchmark, on which even the best AI systems struggle to perform well. In contrast, humans can solve these ARC Tasks at fairly high success rates (85%), so a viable approach could be to inject human reasoning into LLMs by first using humans in the loop and then subsequently fine-tuning on positive cases.
Building the demo app ...
🕹️ Demo app arc-solver-lla...
🐙 LlamaIndex recipe in Streamlit Cookbook github.com/str...
Connect with Andrei
👤 LinkedIn: / nerdai
𝕏 Twitter: x.com/_nerdai_
#streamlit #streamlitcookbook #llamaindex #llm #llms