search

karthink / gptel

A simple LLM client for Emacs
GNU General Public License v3.0
1.04k stars 111 forks source link

issues with ollama #268

Open lbraglia opened 3 months ago

lbraglia commented 3 months ago

Hi i've somewhat copypasted the config with minor modification to use dolphin-mistral

(setq                                                                                                                                
 gptel-model "dolphin-mistral:latest"                                                                                                
 gptel-backend (gptel-make-ollama "Ollama"                                                                                           
                 :host "localhost:11434"                                                                                             
                 :stream t                                                                                                           
                 :models '("dolphin-mistral:latest")))

ollama is up and running but gptel starts chatgpt 3.5 and furthermore ollama/dolphin-mistral is not available in the M-x gptel-menu, GPT-model section

am i missing something or this is a bug?

karthink commented 3 months ago
  1. Are you on the latest commit? If not please update gptel
  2. Could you run this: (map-keys gptel--known-backends)and paste the output here?
karthink commented 3 months ago

  1. Are you on the latest commit? If not please update gptel

    1. Could you run this: (map-keys gptel--known-backends)and paste the output here?

@lbraglia Did you have a chance to try this?

lbraglia commented 2 months ago

not yet, i'll try it asap. BTW the version i've used is the one available in MELPA.

Thanks for follow up

karthink commented 2 months ago

Any updates?

lbraglia commented 2 months ago

ok, ive updated now to 

  gptel                          20240425.2058  installed             Interact with ChatGPT or other LLMs

and with

(map-keys gptel--known-backends)

i obtain ("ChatGPT")

while from gptel-menu, selecting -m GPT Model i have

Click on a completion to select it.
In this buffer, type RET to select the completion near point.

8 possible completions:
ChatGPT:gpt-3.5-turbo   ChatGPT:gpt-3.5-turbo-16k   ChatGPT:gpt-4
ChatGPT:gpt-4-0125-preview  ChatGPT:gpt-4-1106-preview  ChatGPT:gpt-4-32k
ChatGPT:gpt-4-turbo     ChatGPT:gpt-4-turbo-preview
karthink commented 2 months ago

and with

(map-keys gptel--known-backends)

i obtain ("ChatGPT")

The Ollama backend is not being defined then. Does your setq form throw any errors? Can you try evaluating it manually?