wizard-vicuna instead of codellama?

[rosmur]

Hello, I'd like to use wizard-vicuna instead of codellama. Is this feasible? (I have limited disk space and cant afford a separate coding optimized model..)

I tried changing the model and i get the error "Sorry I dont understand. Please try again" and a spinning indicator

I've changed to

[rjmacarthy]

Please include more details about which provider you are using, e.g Ollama and the exact model you are trying to use. Also please include details for the chat.hbs template file.

Thanks.

[rosmur]

Its Ollama wizard-vicuna-uncensored: https://ollama.com/library/wizard-vicuna-uncensored.

Sorry I dont know how to get the chat.hbs template

[rjmacarthy]

It should work by default now using Ollama as twinny now supports OpenAI specification.