Closed djav1985 closed 2 months ago
Yes, the current one can handle a larger repo. I tested it on MinIO Docs with around 700+ files. It may take some time to create the embeddings.
For the embedding model, I tested Nomic via Ollama and text-ada embedding from OpenAI. I do not recommend using all-minilm for larger repo
What about the two newer text embedding models the small and large one from open AI?
It will work fine, I mean, in case you want to use an embedding model running locally
I see in change log you added new GitHub loader. What is the exact change? is it more capable of larger repos? if so any rough deal what it can handle?