GOOGLE
GEMMA – SCI & TECH
News:
Gemma: Introducing new
state-of-the-art open models
What's
in the news?
●
Google
has unveiled Gemma, its latest open artificial intelligence (AI)
offering, aiming to provide accessible AI models while adhering to strict
principles of responsibility.
●
This move comes hot on the heels of
OpenAI’s Sora making headlines recently, showcasing the rapid evolution within
the AI landscape.
Gemma:
●
The name “Gemma” is derived from the Latin
word for “precious stone”.
●
It is a set of lightweight models inspired
by Google’s Gemini models, developed by Google DeepMind and other teams across
the company.
Performance
and Responsible Design:
●
Gemma models, despite their lightweight
design, offer top-tier performance
compared to larger models.
●
They can run directly on developer devices
and surpass larger models on key benchmarks while ensuring safe and responsible
outputs.
Key
Features of Gemma:
1.
Two Sizes:
●
Gemma 2B and Gemma 7B come with
pre-trained and instruction-tuned variants.
2.
Responsible AI Toolkit:
●
Google offers tools supporting safer AI
applications with Gemma.
3.
Framework Support:
●
Inference and supervised fine-tuning (SFT)
toolchains are provided for major frameworks like JAX, PyTorch, and TensorFlow.
4.
Easy Deployment:
●
Gemma models can run on various platforms
and integrate seamlessly with tools like Colab and Kaggle.
5.
Commercial Usage:
●
Gemma permits responsible commercial usage
and distribution for all organizations.