Local Embedding Model

Support
1

Description

I get the following error :

get vector input from modules provider: remote client vectorize: failed with status: 429 error: Rate limit reached. Please log in or use a HF access token

Collection Information

Here is my class definition :

type or paste code hereself.class_obj = {
            "class": "TextChunk",
            "vectorizer": "text2vec-huggingface",
            "vectorIndexType": "hnsw",
            "vectorIndexConfig": {
                "type": "hnsw",
                "params": {"efConstruction": 128, "m": 16},
            },
            "moduleConfig": {
                "text2vec-huggingface": {
                    "model": "sentence-transformers/all-MiniLM-L6-v2",
                }
            },
        }

Question

I was wondering how the vectorizer was working cause the model that I use is load locally on my machine. But from what I understand the class definition goes trought the Hugging Face API to vectorize the data.
Is there any possibility to use the model locally ? If yes how to do it

2

If I understand your requirements you should look into the text2vec-transformer documentation. The text2vec-huggingface will use HF APIs not the local model hence the need for their API key.

Maybe this other thread will help you: Storing the data to weaviate - #3 by garcia.e

Take care

1 Like