Question: Can current open source efforts replicate AutoGlm's performance

[Chord403]

This is a question asking, surprised by the great performance of AutoGLM at the launch, and with the greatest respect for your choice to open source this work. As a rookie, I really want to contribute to this project. So wondering if AutoGLM's performance can be replicated based on existing open source work? Such as data or training resources? If not, what else do I need to prepare? What are your plans for the future?