Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

Gemma 2 is now available to researchers and developers

Jul 02, 2024 - blog.google
Google DeepMind has officially released Gemma 2, a high-performing AI model, to researchers and developers globally. Available in 9 billion (9B) and 27 billion (27B) parameter sizes, Gemma 2 offers improved performance and efficiency compared to its predecessor, with significant safety advancements. The model is designed to run efficiently on a single Google Cloud TPU host, NVIDIA A100 80GB Tensor Core GPU, or NVIDIA H100 Tensor Core GPU, reducing deployment costs. It is also optimized to run at high speed across a range of hardware, making it accessible for various AI deployments.

Gemma 2 is designed to integrate easily into workflows and is compatible with major AI frameworks like Hugging Face Transformers, JAX, PyTorch, and TensorFlow. It is available under a commercially-friendly license, allowing developers and researchers to share and commercialize their innovations. Google Cloud customers will be able to deploy and manage Gemma 2 on Vertex AI starting next month. The company is also committed to responsible AI development and has provided resources for developers and researchers to build and deploy AI responsibly.

Key takeaways:

  • Gemma 2, a state-of-the-art open model built by Google DeepMind, is now officially released to researchers and developers globally. It offers best-in-class performance and easily integrates with other AI tools.
  • Gemma 2 is built on a redesigned architecture, offering outsized performance, unmatched efficiency and cost savings, and fast inference across different hardware.
  • It is designed to be open and accessible, with broad framework compatibility and effortless deployment. It is available under a commercially-friendly license, and is compatible with major AI frameworks.
  • Google DeepMind is committed to responsible AI development, and has followed robust internal safety processes in the development of Gemma 2. They are also working on open sourcing their text watermarking technology, SynthID, for Gemma models.
View Full Article

Comments (0)

Be the first to comment!