Open JayCroghan opened 1 month ago
I have OpenWebUI running fine with Ollama on another machine. I am trying to use the same endpoint with PrivateGPT but it gives me 502.
When attempting to query just the Ollama LLM:
HTTP Request: POST http://ADDRESS:11434/api/chat "HTTP/1.1 502 Bad Gateway"
But this works fine from the bash prompt on the same machine that PrivateGPT is running:
curl http://ADDRESS:11434/api/chat -d '{ "model": "llama3-chatqa", "prompt": "What is water made of?" }'
embedding_api_base is set to the same as api_base which are both set to http://ADDRESS:11434
This page: https://docs.privategpt.dev/installation/getting-started/installation
Local, Ollama-powered setup - RECOMMENDED - this does not work out of the box. Ontop of the 502 error, now it gives me tons of other errors in the background. I have changed nothing except the API URL.
Can you confirm us your ollama version? I'm using 0.1.48 without any issue
ollama
Has you pulled last changes? Has you reinstalled dependencies and extras?
I have OpenWebUI running fine with Ollama on another machine. I am trying to use the same endpoint with PrivateGPT but it gives me 502.
When attempting to query just the Ollama LLM:
HTTP Request: POST http://ADDRESS:11434/api/chat "HTTP/1.1 502 Bad Gateway"
But this works fine from the bash prompt on the same machine that PrivateGPT is running:
curl http://ADDRESS:11434/api/chat -d '{ "model": "llama3-chatqa", "prompt": "What is water made of?" }'
embedding_api_base is set to the same as api_base which are both set to http://ADDRESS:11434
This page: https://docs.privategpt.dev/installation/getting-started/installation
Local, Ollama-powered setup - RECOMMENDED - this does not work out of the box. Ontop of the 502 error, now it gives me tons of other errors in the background. I have changed nothing except the API URL.