AimenGPT is an open-source AI chatbot that allows you to host your own private ChatGPT alternative without sending data to external servers. Supports document uploads and Code Llama programming models up to 34B parameters.
It runs locally using Docker and provides robust AI capabilities leveraging Llama 2, ChromaDB, and Langchain.
You can have natural conversations, get code assistance, and more while keeping your information completely secure.
Supported Models
- Nous Hermes Llama 2 7B Chat (GGML q4_0)
- Nous Hermes Llama 2 13B Chat (GGML q4_0)
- Nous Hermes Llama 2 70B Chat (GGML q4_0)
- Code Llama 7B Chat (GGUF Q4_K_M)
- Code Llama 13B Chat (GGUF Q4_K_M)
- Phind Code Llama 34B Chat (GGUF Q4_K_M)
How to use it:
1. Install Docker on any x86, arm64, or M1/M2 Mac system.
2. Clone ChromaDB and AimenGPT repositories from GitHub.
git clone https://github.com/chroma-core/chroma.git git clone https://github.com/aietal/aimengpt.git
3. Navigate to the AimenGPT directory and uses the following command to run AimenGP:
cd aimengpt
./run.sh --model 70b
4. Access AimenGPT via your browser at http://localhost:3000.
5. For M1/M2 Mac users: ensure Docker and Xcode are installed on your device.
6. Follow the cloning process mentioned in step 2.
7. Run AimenGPT with the Mac-specific command:
./run-mac.sh --model 70b
8. Modify the model as needed:
- 7b
- 13b
- 70b
- code-7b
- code-13b
- code-34b
