qdurllm

Search your favorite websites and chat with them, on your desktop🌐

Docs in active development!👷‍♀️

They will be soon available on: https://astrabert.github.io/qdurllm/

In the meantime, refer to the Quickstart guide in this README!

Quickstart

1. Prerequisites

• conda package manager
• docker and docker compose.

2. Installation

Important

This is only for the pre-release of v1.0.0, i.e. v1.0.0-rc.0

1. Clone the january-2025 branch of this GitHub repo:

git clone -b january-2025 --single-branch https://github.com/AstraBert/qdurllm.git
cd qdurllm/

2. Create the conda environment:

conda env create -f environment.yml

3. Pull qdrant from Docker Hub:

docker pull qdrant/qdrant

3. Launching

1. Launch qdrant vector database services with docker compose (from within the qdurllm folder):

docker compose up

2. Activate the qdurllm conda environment you just created:

conda activate qdurllm

3. Go inside the app directory and launch the Gradio application:

cd app/
python3 app.py

You should see the app running on http://localhost:7860 once all the models are downloaded from HuggingFace Hub.

Relies on

• Qwen2.5-1.5B-Instruct, with Apache 2.0 license
• nomic-ai/modernbert-embed-base, with Apache 2.0 license
• prithivida/Splade_PP_en_v1, with Apache 2.0 license

Give feedback!

Comment on the discussion thread created for this release with your feedback or create issues :)