on which folder should i keep my GPT4ALL model files?

[edwinjosechittilappilly]

I tried to run with GPT4ALL model ggml-gpt4all-j-v1.3-groovy.bin its giving error. May I know where should I kee the .bin files of the model?

[meherhendi]

When using docker to run Danswer, set this env variables in .env file in deployment/docker_compose/.env:

INTERNAL_MODEL_VERSION=gpt4all-chat-completion INTERNAL_MODEL_VERSION=ggml-gpt4all-j-v1.3-groovy.bin QA_TIMEOUT=120 (since the CPU is usually slower than GPU and needs more time to run)

Usually, the gpt4all library will detect the model and start downloading it, For some reason, the download progress does not appear in the docker logs

the downloaded model is not persisted to a volume, to avoid redownloading it, if you have the model on your host machine you can mount it under /root/.cache/gpt4all/ each time by adding this line under "volumes" in docker-compose.dev.yml: /path/to/model/ggml-gpt4all-j-v1.3-groovy.bin:/root/.cache/gpt4all/ggml-gpt4all-j-v1.3-groovy.bin