TheoremExplainAgent
Basic Information
This repository contains the full codebase, data pointers and scripts for TheoremExplainAgent, an AI system developed to generate long-form Manim videos that visually explain mathematical theorems. It implements the methods described in the ACL/ArXiv paper and provides generation and evaluation pipelines, example commands, environment setup instructions and a paired dataset called TheoremExplainBench. The code supports multimodal generation including voiceover, Manim scene rendering and optional retrieval-augmented generation (RAG). The repo also hosts utilities for configuring models via a .env, downloading pretrained TTS models used in the project, and producing batch or single-topic video outputs. The project is intended for research use and includes instructions to reproduce baseline videos used in the paper.
Links
Categorization
App Details
-
$$$$ -
$$$$