Does AnyGPT support multi images input? - Githubissues

OpenMOSS / AnyGPT

Code for "AnyGPT: Unified Multimodal LLM with Discrete Sequence Modeling"

779 stars 61 forks source link

Does AnyGPT support multi images input? #43

Closed QAQdev closed 1 month ago

QAQdev commented 1 month ago

Thanks for your great work!

I am trying to apply AnyGPT on videos, since I need to clip video into frames as multi-images input for AnyGPT, I want to ask that:

Does AnyGPT support multi-images input?
What is the context length for AnyGPT? -How many tokens will each image have?

Based on these, I could calculate a maximum input frames for AnyGPT!

Very looking forward to your reply! @JunZhan2000 Thanks in advance!