rbehzadan/bge-m3-f16 repository overview

This Docker image provides a multi-architecture (AMD64 & ARM64) server for generating text embeddings using the BGE-M3⁠ model with llama.cpp⁠. It exposes a simple API endpoint (/embedding) to retrieve 1024-dimensional embeddings for any input text.

⁠Quick Start

Run the container:

docker run -d --name bge-m3-f16 -p 8080:8080 rbehzadan/bge-m3-f16

⁠API Usage

Send a request to get embeddings:

import requests
response = requests.post("http://localhost:8080/embedding", json={"content": "Hello, world!"})
embedding = response.json()[0]["embedding"][0]
print(embedding[:5])  # First 5 values

⁠Source Code

GitHub Repository: rbehzadan/bge-m3-f16⁠