Running Ollama on Windows host file upload error

kbarto323 commented 4 months ago

If I run Ollama on a docker container it appears to work, but when I run Ollama on the Windows host to leverage the GPU and upload a document, I get privategpt | requests.exceptions.ConnectionError: HTTPConnectionPool(host='localhost', port=11434): Max retries exceeded with url: /api/embeddings (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x74b5469ca110>: Failed to establish a new connection: [Errno 111] Connection refused')) I have changed the OLLAMA_API_BASE to the Windows host IP so it appears it isn't being updated everywhere it needs to be.

3x3cut0r commented 4 months ago

i missed that privategpt added a new option "embedding_api_base" to the settings-ollama.yaml. i added OLLAMA_EMBEDDING_API_BASE to setup the api_base for the embedding. you may now need to set both OLLAMA_API_BASE and OLLAMA_EMBEDDING_API_BASE to change the api base. wait for the gitaction for my new version and try it out.

kbarto323 commented 4 months ago

Excellent, thank you!

On Tue, May 14, 2024, 5:50 AM 3x3cut0r @.***> wrote:

i missed that privategpt added a new option "embedding_api_base" to the settings-ollama.yaml https://github.com/zylon-ai/private-gpt/blob/main/settings-ollama.yaml. i added OLLAMA_EMBEDDING_API_BASE to setup the api_base for the embedding. you may now need to set both OLLAMA_API_BASE and OLLAMA_EMBEDDING_API_BASE to change the api base. wait for the gitaction for my new version and try it out.

— Reply to this email directly, view it on GitHub https://github.com/3x3cut0r/docker/issues/12#issuecomment-2110002816, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABM2JCGZVCALYGIORYUMF23ZCH3AVAVCNFSM6AAAAABHUNO2QGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMJQGAYDEOBRGY . You are receiving this because you authored the thread.Message ID: @.***>

3x3cut0r / docker

Running Ollama on Windows host file upload error #12