kyegomez/Gemini
Open-source PyTorch implementation of Google's multi-modal foundation model Gemini supporting text, image, audio, and video inputs.
Not currently ranked — collecting fresh signals.
star history
The repository implements Google’s Gemini model as an open-source project. It uses a transformer architecture that processes multiple modalities directly through special decoders for text or image generation. The model accepts text, audio, images, and video as input tokens processed by a transformer with conditional decoding for generation. Key features include Multi Grouped Query Attention, Flash Attention, RoPE, ALiBi, and KV cache support.
Frequently asked
- What is kyegomez/Gemini?
- Open-source PyTorch implementation of Google's multi-modal foundation model Gemini supporting text, image, audio, and video inputs.
- Is Gemini open source?
- Yes — kyegomez/Gemini is open source, released under the MIT license.
- What language is Gemini written in?
- kyegomez/Gemini is primarily written in Python.
- How popular is Gemini?
- kyegomez/Gemini has 467 stars on GitHub.
- Where can I find Gemini?
- kyegomez/Gemini is on GitHub at https://github.com/kyegomez/Gemini.