🚀 Medial Secures Investment on Shark Tank India - Fueling the Future of Professional Social Networking. 🔥

Startup Showcase

Premium Content

Try our Valuation Calculator →

News on Medial

Google goes “open AI” with Gemma, a free, open-weights chatbot family

Arstechnica

Arstechnica · 1y ago

Google goes “open AI” with Gemma, a free, open-weights chatbot family

Google has introduced Gemma, a new family of free, open-weights AI language models. These models, similar to Google's more powerful Gemini models, can be run on desktop or laptop computers. Gemma comes in two sizes, 2 billion and 7 billion parameters, and includes pre-trained and instruction-tuned variants. This is Google's first significant open large language model release since OpenAI's ChatGPT. Gemma aims to match Meta's open-weights model strategy and attract more developers to Google's Vertex AI cloud platform. Google also released a "Responsible Generative AI Toolkit" to provide guidance and tools for developing safe AI applications.

Related News

Google launches two new open LLMs | TechCrunch

TechCrunch · 1y ago

Google launches two new open LLMs | TechCrunch

Google has launched a new family of lightweight open-weight models called Gemma, which are inspired by its Gemini models. The Gemma models, starting with Gemma 2B and Gemma 7B, are dense decoder-only models that can be used for commercial and research purposes. Although these models are not open source, developers can access ready-to-use notebooks and integrations with other platforms. Google emphasizes that the generation quality of these models has significantly improved over the past year, allowing for new ways of developing AI applications. Additionally, Google has released a responsible generative AI toolkit and a debugging tool for use with Gemma.

Google is giving away some of the AI that powers chatbots

Economic Times

Economic Times · 1y ago

Google is giving away some of the AI that powers chatbots

Google has released the computer code for its online chatbot, following in the footsteps of Meta. The move is seen as a shift in the industry, with companies becoming more open in sharing AI technology. Google's two AI language models, Gemma 2B and Gemma 7B, have been made available to help companies and developers build chatbots. While some companies have opted for secrecy due to concerns about misuse, others argue that open sourcing the code allows for problem identification and solution suggestions. Google stressed that the released models have been made as safe as possible.

Google I/O 2024: From Gemini 1.5 Pro, Project Astra to Trillium, ET brings you the low down

Economic Times

Economic Times · 1y ago

Google I/O 2024: From Gemini 1.5 Pro, Project Astra to Trillium, ET brings you the low down

Google has unveiled Gemini 1.5 Pro at its annual developer conference, Google I/O. The new version offers an expanded context window of two million tokens and is available for developers in private preview. This advanced version allows users to interact with Gemini multimodal foundation model directly through an app on Android and iOS. Additionally, Google introduced Trillium, its sixth-generation tensor processing unit (TPU), which delivers improved compute performance. The company also showcased Project Astra, an AI assistant that can remember and process visual information. Lastly, Gemma, a family of open language models, was used to train Navarasa 2.0 for processing content in 15 Indian languages.

Mistral AI makes its first large language model free for everyone

TechCrunch · 1y ago

Mistral AI makes its first large language model free for everyone

French AI startup, Mistral, has released its first language model, Mistral 7B, which it claims outperforms others in its size. The model is available for free download and use, without any restrictions, under the Apache 2.0 license. Mistral aims to support the open generative AI community and bring open models to state-of-the-art performance. While the model is free, Mistral also offers a paid commercial offering, providing white-box solutions with access to weights and code sources.

Nvidia’s “Chat With RTX” is a ChatGPT-style app that runs on your own GPU

Arstechnica

Arstechnica · 1y ago

Nvidia’s “Chat With RTX” is a ChatGPT-style app that runs on your own GPU

Nvidia has released Chat With RTX, a personalized AI chatbot that can be run locally on PCs equipped with Nvidia RTX graphics cards. It utilizes open-weights LLMs and can search through local files to answer questions. The application supports various file formats and can even incorporate information from YouTube videos without requiring an internet connection. However, the installation process and dependencies can be problematic, with occasional crashes during testing. Despite this, the local processing capability prioritizes user privacy by avoiding the need to send sensitive data to cloud-based services. The application is available for free download on the Nvidia website.

Google preps Bard integration to enhance Google Assistant

Livemint

Livemint · 1y ago

Google preps Bard integration to enhance Google Assistant

Google is set to revolutionize its virtual assistant, Google Assistant, with the integration of its AI chatbot, Bard. The integration is expected to leverage Google's Gemini family of Large Language Models to enhance the capabilities of the assistant. Users will have multiple input methods and the assistant will continue to offer hands-free assistance. The exact rollout details and features of Bard-integrated Google Assistant are yet to be officially announced by Google.

xAI open sources Grok

The Verge

The Verge · 1y ago

xAI open sources Grok

Elon Musk's xAI has open sourced its AI chatbot Grok, allowing researchers and developers to enhance and update it in the future. The release includes the base model weights and network architecture of the Grok-1 model, which boasts 314 billion parameters. The open release is under the Apache 2.0 license, enabling commercial use but without the data used for training or connections to X for real-time data. This move aligns with Musk's beliefs in open sourcing AI models, as he previously criticized OpenAI, a company he co-founded but is now suing.

Google releases new AI model capable of running on a single GPU

YourStory

YourStory · 4m ago

Google releases new AI model capable of running on a single GPU

Google has launched Gemma 3, a new AI model series designed to operate on a single GPU, enhancing portability and speed for developers. Built on the technology of Gemini 2.0, Gemma 3 offers models with up to 27 billion parameters, outperforming rivals in human preference evaluations. It supports over 35 languages and excels in handling complex data with a 128k-token context window. This release aligns with industry efforts to develop powerful models and improve AI safety practices.

Google’s Gemini AI now accessible to kids under 13 via Family Link

YourStory

YourStory · 2m ago

Google’s Gemini AI now accessible to kids under 13 via Family Link

Google plans to make its Gemini AI chatbot available to children under 13 through parent-managed Google accounts. Access is facilitated via Google's Family Link, allowing parents to oversee digital activities. The chatbot aids with tasks like homework, equipped with safeguards against inappropriate content. Although child data won't train AI models, the move has raised safety concerns among advocates. International bodies like UNICEF emphasize regulating AI’s impact on children.

Elon Musk's xAI makes Grok open source, billionaire takes another jibe a OpenAI

Livemint

Livemint · 1y ago

Elon Musk's xAI makes Grok open source, billionaire takes another jibe a OpenAI

Elon Musk's xAI has fulfilled its promise of making its generative AI chatbot, Grok, open source. The company released the base model weights and network architecture of Grok-1, a 314 billion parameter Mixture-of-Experts model. This release comes shortly after Musk took a jab at OpenAI, with whom he has been engaged in a legal battle. The release of Grok-1 is a significant milestone for xAI and opens up possibilities for further development and applications.

Trackers

Active Indian VC’s

OG Capital Email

With a hands-on approach, OG Capital aims to invest in over 20 promising...

Accel Partners Email

Early and growth-stage investments in disruptive technology companies with...

Early-stage venture capital firm investing in technology startups in India. Focus on...

Access All Trackers

Startup Showcase Winners

June 2025

Helping your parents when you are miles away

The Pit Stop Your Cravings Deserve

The next generation E-commerce platform

Enter Ongoing Startup Showcase

Top Users

Trending News on Medial

Download the medial app to read full posts, comements and news.