Chat gives no answer

[mamema]

just tried this project out, as i looks promising and should be able to index folder structures
what idid:
local install with:
docker compose -f docker-compose-ollama.yml --env-file .env up --build.

.env file:
LOCAL_FILES_PATH=/mnt/docker2/minima/documents
EMBEDDING_MODEL_ID=sentence-transformers/all-mpnet-base-v2
EMBEDDING_SIZE=768
OLLAMA_MODEL=llama3.2:latest # must be LLM model id from Ollama models page
RERANKER_MODEL=BAAI/bge-reranker-base # please, choose any BAAI reranker model

put two pdfs into the documents folder

see them indexed in the indexer container
want to chat and get no answer
nothing special in the logs.
How to proceed from here?

[dmayboroda]

Hey @mamema
Thank you for issue.
Did you setup a LOCAL_FILES_PATH as a folder inside a docker container ? That must be a folder on your machine (not inside a docker), this folder will be mounted automatically.

Can you share with me all logs that you have ? When chat loaded, there must be a log that tests the connection (you'll see an embedding in your logs).

Also, the model that you specified can be too big for your allocated Dockers RAM. Can you please try to use something like Llama 3.2 1b (smallest Llama model) or something like Qwen 0.5.

Thank you

[mamema]

in the indexer container i have a volume mapping of:
/mnt/docker2/minima/documents /usr/src/app/local_files

logs attached

_qdrant_logs.txt
_minima-chat-1_logs.txt
_ollama_logs.txt
_minima-llm-1_logs.txt

_minima-indexer-1_logs.txt

[mamema]

i have 192 GB RAM not restricted on the docker containers and 12 GB GPU with a NVIDIA 3060

[dmayboroda]

Taking a look at your problem
Thank you!