AIDA Pulse: Gems, Think, Sora

Stay in the rhythm of AI, Data and Analytics with our new weekly newsletter.

Today you will read about Gemma, IBM Think and Sora

Google introduced Gemma

Few hours ago Google introduced Gemma - a family of lightweight, open models built from the same research and technology used to create the Gemini models.

Along with the model weights, tools are also being released to support developer innovation, foster collaboration, and guide responsible use of Gemma models.

Google says that Gemma is available worldwide, starting today. Here are the key details to know:

• There is model weights in two sizes: Gemma 2B and Gemma 7B. Each size is released with pre-trained and instruction-tuned variants.
• A new Responsible Generative AI Toolkit provides guidance and essential tools for creating safer AI applications with Gemma.
• Google is providing toolchains for inference and supervised fine-tuning (SFT) across all major frameworks: JAX, PyTorch, and TensorFlow through native Keras 3.0.
• Ready-to-use Colab and Kaggle notebooks, alongside integration with popular tools such as Hugging Face, MaxText, NVIDIA NeMo and TensorRT-LLM, make it easy to get started with Gemma.
• Pre-trained and instruction-tuned Gemma models can run on your laptop, workstation, or Google Cloud with easy deployment on Vertex AI and Google Kubernetes Engine (GKE).
• Optimization across multiple AI hardware platforms ensures industry-leading performance, including NVIDIA GPUs and Google Cloud TPUs.
• Terms of use permit responsible commercial usage and distribution for all organizations, regardless of size.

IBM Think on Tour 2024

Visiting seven cities around the world, Think on Tour is helping local leaders scale AI for their businesses.

#IBM is going global to bring #Think local. First event will take place on 20–23 May in Boston @ Boston Convention & Exhibition Center. The other events are as follow:

• Think Japan—July
• Think Madrid—July
• Think Singapore—August
• Think Mumbai—September
• Think Toronto—September
• Think Munich—October
• Think London—October

You can learn more about IBM events on their website.

OpenAI has presented Sora

A text-to-video generator that can create realistic videos based on text instructions. The purpose of Sora is to enhance AI's ability to not only recreate but also understand the physical world and its movements. The model can produce videos up to one minute in length with high visual quality and precise implementation of user specifications. Sample videos include a walk through Tokyo and creative depictions such as a young man sitting on a cloud and reading.

Initially, Sora access is limited to a select group of experts and creative professionals to test and develop the model in various contexts. OpenAI intends to introduce safeguards, such as generated content detection technologies and C2PA metadata protections, in future products. Sora is based on diffusion model technology and Transformer architecture and is considered by the company to be a significant step towards general artificial intelligence.

The development of Sora could have a significant impact on the film industry, particularly in the areas of special effects and stock videos, as it has the potential to replace a significant amount of manually created or filmed content.