Qwen2.7b.OpenDevin

brought to you by skratos115 (HF) / Kingatlas115 (GH) in collaboration with the official Opendevin Team ~xingyaoww

Qwen2-7B-Instruct with OpenDevin Tool Calling

Overview

This project involves the fine-tuning of the Qwen2-7B-Instruct model using the opendevin-code-act dataset with the help of Unsloth. The primary goal is to develop a more powerful LLM capable of effectively using the CodeAct framework for tool calling. This is still in early development and should not be used in production. We are working on building a bigger dataset for tool paths/ trajectories.

We Need Your Help

Using the feedback integration will help us build better trajectories and performance for model training.
Read more here: Twitter and LinkedIn

Model Details

• Model Name: Qwen2-7B-Instruct
  
• Dataset: opendevin-code-act
  
• Training Platform: Unsloth
  

Provided f16 and q4_k_m. Q5_k_m, and Q8_0 gguf files can be found on my Hugging Face profile.

I used the qwen2.7b.OpenDevin.q4_k_m.gguf for my testing and got it to write me a simple script. More testing to come.

Running the Model

You can run this model using vLLM or ollama. The following instructions are for using ollama.

ollama run {name of model here]

Prerequisites

• Docker
  
• Hugging Face transformers library (version >= 4.37.0 is recommended)
  

Running with Ollama

1. Install Docker: Ensure you have Docker installed on your machine.

2. Pull the Latest Hugging Face Transformers:

   pip install transformers>=4.37.0

1. Set Up Your Workspace:
   

   WORKSPACE_BASE=$(pwd)/workspace

1. Run the Docker Command:
   

   docker run -it \
      --pull=always \
      -e SANDBOX_USER_ID=$(id -u) \
      -e PERSIST_SANDBOX="true" \
      -e LLM_API_KEY="ollama" \
      -e LLM_BASE_URL="http://[yourIPhere or 0.0.0.0]:11434" \
      -e SSH_PASSWORD="make something up here" \
      -e WORKSPACE_MOUNT_PATH=$WORKSPACE_BASE \
      -v $WORKSPACE_BASE:/opt/workspace_base \
      -v /var/run/docker.sock:/var/run/docker.sock \
      -p 3000:3000 \
      --add-host host.docker.internal:host-gateway \
      --name opendevin-app-$(date +%Y%m%d%H%M%S) \
      ghcr.io/opendevin/opendevin:main

Replace [yourIPhere or 0.0.0.0] with your actual IP address or use 0.0.0.0 for localhost.
this uses the newest opendev image you can change the last line of code from ghcr.io/opendevin/opendevin:main to ghcr.io/opendevin/opendevin:0.6.2. this was tested on main not on the current stable build but feel free to play.

Early Development

This project is in its early stages, and we are continuously working to improve the model and its capabilities. Contributions and feedback are welcome.

Support my work

Right now all of my work has been funded personally, if you like my work and can help support growth in the AI community consider joining or donating to my Patreon.
Patreon Link

License

This project is licensed under the MIT License.