Welcome to the forefront of artificial intelligence innovation, where Meta’s cutting-edge advancements are redefining the landscape. In a groundbreaking move, Meta AI has undergone a significant upgrade with the integration of Llama 3 technology, ushering in a new era of enhanced capabilities and seamless integration.
But the evolution doesn’t stop there. With the launch of Meta Llama 3 AI models boasting 8B and 70B parameters, Meta emerges as a frontrunner, poised to surpass even the formidable Google Gemini 1.5 Pro.
Join us as we delve into the details of this monumental leap forward in AI technology.
Meta AI Gets Upgraded With Llama 3 to Add New Features, Better Integration
Meta AI is getting a massive upgrade.
The Meta announce two new Llama 3 artificial intelligence (AI) models, the Llama 3 8B and 70B, which are said to offer improve capabilities compare to the predecessor.
Meta also upgrade its native AI assistant with Llama 3 models.
Meta AI can now respond more efficiently, generate images faster, and even animate images, as per Meta.
Accessibility of the chatbot is also being improve by integrating it into different interfaces.
It is also being expand to more countries.
In post, Meta said that Meta AI is now power by Llama and can be use for free as long as a user has an account on any of its platforms.
Meta announce the chatbot in Meta Connect 2023 and soon began adding it to Facebook, Instagram and WhatsApp in the US.
The AI assistant was recently expand to India and now Meta has reveal that it is expanding it to more regions.
The countries where Meta AI is being roll out include Australia, Canada, Ghana, Jamaica, Malawi, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Uganda, Zambia, and Zimbabwe.
New Features of Meta AI
As the AI assistant is getting better in terms of the efficiency of responses and image generation quality as a result of Llama 3 integration, Meta is also adding new capabilities as well as wider integration of the chatbot.
Meta AI will now be able to generate images in real-time.
Means when you begin typing your prompt, you can see the AI generating an image.
As you keep typing and describing the image better, the image also keeps changing based on the prompt.
As this does make the generation speed faster, it also allows users to preview the image and make real-time changes for better results.
Other new feature is image animation.
Meta is also offering image editing capabilities.
If you do not like the generated image, you can ask the AI to make changes to it or iterate on it in a new style.
You will also be able to animate the image and turn it into GIF.
Meta is also making the chatbot available at more user touch-points.
You will now be able to find the AI assistant in the Facebook feed under posts.
A small Meta AI logo with a couple of suggest queries will pop up and the user can ask questions regarding the topics shown in the video.
It is also being integrate into search across all platforms.
This will allow users to go to the search bar on Facebook, Instagram, WhatsApp, or Messenger and type a question, and the AI will answer it.
Real-time events such as flight booking prices and stock market updates can also be queried.
But, Meta’s AI uses both Google and Bing to show results, as users cannot control which search engine it shows.
Lastly, Meta has also launch a new meta.ai website where users can jump to have a conversation with the chatbot or ask it to solve a math problem or generate content.
This new platform is likely for those who want to use the AI assistant but do not want to open their social media account for it, especially in a professional setting.
You will be able to save conversations here for future reference as well.
Instagram Said to Be Testing AI-Powered Chatbots for Influencers That Can Interact With Followers
— 2YoDoINDIA News Network (@2yodoindia) April 17, 2024
For more news visit https://t.co/98KV4yIruC#2YoDoINDIA #Instagram #ArtificialIntelligence #AI #Meta pic.twitter.com/Y8IoyLNUfc
Meta Llama 3 AI Models With 8B and 70B Parameters Launched, Said to Outperform Google’s Gemini 1.5 Pro
Meta introduce the next generation of its artificial intelligence (AI) models, Llama 3 8B and 70B.
Shortened for Large Language Model Meta AI, Llama 3 comes with improve capabilities over its predecessor.
Meta also adopt new training methods to optimise the efficiency of the models.
As, with Llama 2, the largest model was 70B, but this time the company said its large models will contain more than 400 billion parameters.
As per a report last week reveal that Meta will unveil its smaller AI models in April and its larger models later in the summer.
Availability of Meta Llama 3
Those interested in trying out the new AI models are in luck as Meta is taking a community-first approach with the Llama 3.
The new foundation models will be open source just like previous models.
Meta said in its blog post :
“Llama 3 models will soon be available on AWS, Databricks, Google Cloud, Hugging Face, Kaggle, IBM WatsonX, Microsoft Azure, NVIDIA NIM, and Snowflake, and with support from hardware platforms offered by AMD, AWS, Dell, Intel, NVIDIA, and Qualcomm.”
The list includes all major cloud, hosting, and hardware platforms, which should make it easier for enthusiasts to get their hands on the AI models.
Meta has also integrate Llama 3 with its own Meta AI that can be access via Facebook Messenger, Instagram, and WhatsApp in support countries.
Performance and Architecture of Meta Llama 3
Meta share benchmark scores of Llama 3 for both its pre-trained and instruct models.
For reference, pre-trained is the general conversational AI whereas the instruct models are aimed at completing specific tasks.
The pre-trained model of Llama 3 70B outscore Google’s Gemini 1.0 Pro in the MMLU (79.5 vs 71.8), BIG-Bench Hard (81.3 vs 75.0), and DROP (79.7 vs 74.1) benchmarks and the 70B Instruct model outscore the Gemini 1.5 Pro model in MMLU, HumanEval, and GSM-8K benchmarks, based on data share by Meta.
Meta has opt for a decoder-only transformer architecture for the new AI models but has made many improvements over the predecessor.
Llama 3 now uses a tokeniser with a vocabulary of 128K tokens, and Meta has adopt group query attention (GQA) to improve inference efficiency.
GQA helps in improving the attention of the AI so it does not move outside of its designated context when answering queries.
Meta has pre-trained the models with more than 15T tokens, which it claims to have sourced from publicly available data.
