In a significant progression towards democratizing it, Google introduced GemmaA family of open source IA language models to allow developers and researchers from around the world. Gemma represents a step forward in Google's commitment to make AI accessible to everyone, according to the traces of Tensorflow, Bert and T5.
Gemma is inspired by his brother, GeminiWho feeds the sales AI sales AI google. The model was built from the same research and technologies as Gemini, but emphasizing opening and accessibility.
At the heart of Gemma is its versatility, with models available in two sizes: Gemma 2B and Gemma 7B, each equipped with pre-formulated variants and regulated by instruction. These models retain flexibility to operate directly on local devices, including office computers and laptops.
Key characteristics:
- Framework: Gemma supports inference and supervised fine adjustment (SFT) in the main frames such as Jax, Pytorch and Tensorflow via native Keras 3.0.
- Ease of use: Colard and Kaggle loans to use, as well as integration with popular tools such as Hugging Face, Mixtext, Nvidia Nemo and Tensorrt-LLM, facilitate the start-up of developers.
- Deployment options: Pre-formulated gemma models and set by instruction can work on the laptop, the workstation or Google Cloud. The deployment options include Vertex AI and Google Kubernetes Engine (GKE).
- Optimized performance: Gemma works exceptionally well on various AI material platforms, including NVIDIA and Google Cloud TPUS GPUs.
- Developer attraction: Google hopes to attract developers to its VERTEX AI cloud platform via Gemma.
The company aims to match Meta's open weight models, emphasizing transparency and confidentiality. Thus, Gemma provides an alternative more accessible to closed models.
While Gemma marks the most important version of Google in the field of Open language models from AI since the advent of Chatgpt, it also highlights the company's unshakable commitment to responsible AI. The AI generative toolbox responsible for Google, introduced alongside Gemma, allows developers the tools and advice necessary to guarantee the safe and ethical deployment of AI applications.
Gemma is already available worldwide! Whether you are a experienced developer or a curious researcher, explore the capacities of Gemma and contribute to the responsible development of AI.
