Ollama - Can't get plain text output

[GitHubGeek]

Here's my doom emacs config.el section:

(use-package! ellama
  :config
  (setopt ellama-language "English")
  (require 'llm-ollama)
  (setopt ellama-major-mode 'fundamental-mode)
  (setopt ellama-provider
          (make-llm-ollama
           :chat-model "phind-codellama:34b-v2"
           :embedding-model "phind-codellama:34b-v2")))

With the line (setopt ellama-major-mode 'fundamental-mode) I'm still getting markdown (or formatted) text when running ellama-code-complete in a typescript buffer

p.s. related to #129

[s-kostyaev]

ellama-code-complete in a typescript buffer

Which model do you use? It's model outputs markdown.

[s-kostyaev]

@GitHubGeek if you want to complete code you need to use base code model, not instruction tuned.

[s-kostyaev]

Not a bug