Multimodality
#5
by
Dampfinchen
- opened
Hello!
Fantastic model, great job on it. Great logic, nice writing style, fantastic creative writing, top coder. It's the best local model for sure right now.
However, just like the DeepSeek model it lacks multimodality or even omnimodality. For many, these are important use cases, video/image/audio/text in, audio/text out would be enough.
Without sacrifcing text generation performance, it would be a killer model. (Even though I can't run it, I really want downscaled versions of this model for consumer systems)
Have a good day, your work is appreciated.
There will be a multimodal model of course!
Great model - great work by your team - eagerly awaiting the multimodal version.