Spaces:

chansung

/

co-write-with-llama2

like 161

Sleeping

This project is designed to showcase the usage of two Generative AI toolbox (LLM and TTI)
I am using Hugging Face Inference API for the LLM part (LLaMA2 chat 70b), but I had to use
T4 instance for TTI purpose. I have tried out different quantized models, but the output qualities
weren't good enough, so I had to switch to use StabilityAI's Stable Diffusion 2.1.

• one more consideration
  : SD 2.1 is good, but SDXL is much better. However, in order to use SDXL, T4 instance is not enough.
  I need A10 instance instead.

I hope I could get community grant for A10 instance (at least T4 instance).
Thank you so much.