cannot change the Chat model Provider to ollama

ItzCrazyKns / Perplexica

Perplexica is an AI-powered search engine. It is an Open source alternative to Perplexity AI

MIT License

13.81k stars 1.31k forks source link

cannot change the Chat model Provider to ollama #117

Closed adithyaric closed 4 months ago

adithyaric commented 4 months ago

I want to change the Chat model Provider to my ollama on wsl but its not automatically changed

i have tried this

OLLAMA: Your Ollama API URL. You should enter it as http://host.docker.internal:PORT_NUMBER. If you installed Ollama on port 11434, use http://host.docker.internal:11434. For other ports, adjust accordingly. You need to fill this if you wish to use Ollama's models instead of OpenAI's.

and this https://github.com/ItzCrazyKns/Perplexica/?tab=readme-ov-file#ollama-connection-errors but result are same. my ollama run at http://localhost:11434

Screenshot

ItzCrazyKns commented 4 months ago

Seems like its not able to connect to Ollama. Why are you using localhost:11434 instead of host.docker.internal:11434 or private_ip:11434?

adithyaric commented 4 months ago

Seems like its not able to connect to Ollama. Why are you using localhost:11434 instead of host.docker.internal:11434 or private_ip:11434?

yes i think that the connection problem with ollama. My Ollama server is running at localhost:11434, and I also tried hosting it at host.docker.internal:11434, but it seems like it's not working.

TomWest61 commented 4 months ago

i have the same proble under win 11: I renamed sample.config.toml to config.toml then: "docker compose up -d" and start at: http://localhost:3000/. Message: "internal server error" and no models are shown. (ollama is runnung. docker is running. Testet with "ollama list" and "docker --version")

DriKel commented 4 months ago

Seems like its not able to connect to Ollama. Why are you using localhost:11434 instead of host.docker.internal:11434 or private_ip:11434?

yes i think that the connection problem with ollama. My Ollama server is running at localhost:11434, and I also tried hosting it at host.docker.internal:11434, but it seems like it's not working.

using lm-studio loaded model hosted on localhost, but custom_openai base url in settings is set to host.docker.internal. works for me

DriKel commented 4 months ago

your error is not about ollama or backend, but searxng not being able to run checks by connecting to soundcloud, find its lines (soundcloud) in ./searxng/settings.yml and remove

IAmStoxe commented 4 months ago

I'm having a similar issue unrelated to soundcloud.

I have removed the network specification even from the docker-compose. Other docker containers connect without issue to the Ollama instance.

I am not seeing any errors when tailing the logs of the searchnx and background services.

Opened PR #137 that seemingly fixes this.