In this video, we review Guanaco, the new 65B parameter model that achieves 99% of the performance of ChatGPT. It is truly incredible. Since it is a large model, we use a cloud GPU to power it. This model can code, has logic and reasoning, can do creative writing, and so much more. Guacano was trained in under 24 hours on a single GPU, using a new technology called QloRA, which is mind-blowing. How does it do on the LLM rubric? Let's find out!
Enjoy ๐
Follow me on Twitter ๐ง –
Subscribe to my Substack ๐๏ธ –
Become a Patron ๐ฅ –
Join the Discord ๐ฌ –
Links:
Runpod –
Runpod Tutorial –
Runpod The Bloke Template –
HuggingFace –
Guanaco Model –
TextGen WebUI –