Bionic-GPT is designed to be installed in your data center and then serve Generative AI to potentially hundreds of people via their browser.

    However, before that can happen you need to be able to trial the software.

    Using Kubernetes on Docker desktop or K3s we're able to package a version of the software that show cases all it can do. Many people are using Bionic-GPT in this way as they like the user interface and the easy install. We hope you like it too.

    Bionic-GPT can run on modest hardware (16GB ram) and then scale into more powerful hardware when required. For example 70B running across multiple GPU cards.

    We're open source. If you have any suggestions, recommendations are just have any questions you can either contact us or raise an issue on our github.

    Alt text

    LLama 3 8B

    For installations on modest hardware we run a quantized model with 8 billion parameters.

    However typically larger models give better results and if you have access to a model with for example 70 billion parameters you'll get some great results.

    We still believe this is a great way to start to look at the ways LLM's can help out in your company. We've tried to minimise the time it takes to go from idea to practical proof of concept.