[ad_1]
Meta has Announce The latest addition to the Llama family of generative AI models: Llama 3.3 70B.
In a mail On
“By taking advantage of the latest advances in post-training techniques…this model improves baseline performance at a much lower cost,” Dahla wrote.
Al-Dahl published a chart showing the Llama 3.3 70B outperforming Google’s Gemini 1.5 Pro, OpenAI’s GPT-4o, and Amazon’s newly released Nova Pro in a number of industry benchmarks, including MMLU, which evaluates a model’s ability to understand language. Via email, a Meta spokesperson said the model should provide improvements in areas such as mathematics, general knowledge, following instructions and using the app.
Introducing the Llama 3.3 – a new 70B model that offers the performance of our 405B model but is easier and more cost effective to operate. By leveraging the latest advances in post-training techniques including online preference optimization, this model improves fundamental performance in… pic.twitter.com/6oQ7b3Yuzc
– Ahmed Al Dahle (@Ahmad_Al_Dahle) December 6, 2024
Llama 3.3 70B, which is available for download From the Hugging Face AI development platform and other sources, including the official Llama Websiteis Meta’s latest play to dominate the AI field with “open” models that can be used and marketed for a range of applications.
Meta terms restriction How some developers can use llama models; Platforms with more than 700 million monthly users must request a special license. But for many, it is unimportant that llama models are not “open” in the strict sense of the word. As an example, the Llama app has achieved more than 650 million downloads, according to Meta.
Meta has benefited from Llama internally as well. Meta AI, the company’s AI assistant, which is powered entirely by Llama Models, now has nearly 600 million monthly active users, per Meta CEO Mark Zuckerberg. Zuckerberg claims that Meta AI is on track to become the most widely used AI assistant in the world.
For Mita, the lama’s outgoing nature was both a blessing and a curse. In November, a report claimed that Chinese military researchers used a llama model to develop a defense chatbot. Meta responded by making its llama models available to American defense contractors.
Meta also expressed concerns about its ability to comply with the Artificial Intelligence Act, an EU law that sets a regulatory framework for AI, and described the implementation of the law as “too unpredictable” for its open release strategy. A relevant issue for the company is the provisions in the General Data Protection Regulation (GDPR), the European Union’s privacy law, relating to AI training. Meta trains AI models on public data of Instagram and Facebook users who have not opted out, data that in Europe is subject to GDPR safeguards.
EU regulators earlier this year asked Meta to halt European user data training while they assessed the company’s compliance with the GDPR. Meta acquiesced, while at the same time supporting Open letter Calling for a “modern interpretation” of the GDPR does not “reject progress”.
Meta, not immune to the technical challenges faced by other AI labs, is ramping up its computing infrastructure to train and serve future generations of llamas. The company announced Wednesday that it will build a $10 billion AI data center in Louisiana, the largest AI data center Meta has ever built.
Zuckerberg said at Meta’s fourth-quarter earnings call in August that to train the next major set of Llama models, Llama 4, the company will need 10 times more compute than is needed to train Llama 3. Metal has purchased a pool of more than 100,000 Nvidia GPUs to develop models, and compete with competing resources such as xAI.
Training generative AI models is an expensive business. Meta’s capital expenditures rose nearly 33% to $8.5 billion in the second quarter of 2024, up from $6.4 billion a year earlier, driven by investments in servers, data centers and network infrastructure.
[ad_2]