(*2*)Generative AI is evolving at a quick tempo.
Llama 2 is a generative AI mannequin however there are just a few key variations between Meta’s Llama 2 and its opponents—together with OpenAI’s GPT-4 (ChatGPT) and Google PaLM 2 (Bard). Check the variations right here.
2023 is actually shaping as much as be the 12 months of AI, with firms racing to launch their very own generative AI fashions. Meta has adopted swimsuit with an upgraded model of its LLaMa LLM mannequin, referred to as Llama 2. Meta has launched the mannequin as open supply—that means that it’s obtainable free of charge for each business and analysis functions.
Meta and Microsoft have partnered to distribute Llama 2—which is now obtainable within the Azure AI mannequin catalog. This will permit builders utilizing Microsoft Azure to make use of and profit from Llama 2.
Now, whereas Llama 2 is a generative AI mannequin, or extra particularly, a Large Language Model (LLM), there are just a few key variations between Meta’s Llama 2 and its opponents—together with OpenAI’s GPT-4 (ChatGPT) and Google PaLM 2 (Bard).
Key Differences Between Meta Llama 2, GPT-4 and PaLM 2:
Llama 2 Is Less Powerful Than GPT-4 and PaLM 2, Meta Admits
Percy Liang, director of Stanford’s Center for Research on Foundation Models, factors out that Meta’s analysis paper acknowledges that Llama 2 is much less highly effective than GPT-4 and that there’s a efficiency hole between the 2 fashions. In phrases of benchmarks, Llama 2 mannequin additionally performs barely under its rivals—GPT-4 and PaLM 2.
Trained Using Fewer ‘Tokens’ Compared to Competition
For those that aren’t aware of the time period, ‘tokens’ on this context consult with the textual content {that a} generative AI mannequin was skilled on. For instance, ‘bat’ and ‘cool’ are each tokens. The extra tokens an AI mannequin is skilled on, the higher it is going to be. According to TechCrunch, Llama 2 was skilled on two trillion tokens, which is as greater than the 1.4 trillion tokens that the unique Llama was skilled on. For comparability, Google’s PaLM 2, its fast rival, was skilled on 3.6 trillion tokens, in keeping with a CNBC report from May.
Language Support
Meta’s Llama 2 helps fewer languages than Google’s PaLM 2 and OpenAI GPT-4, with 20 languages in comparison with PaLM2’s 100, and GPT-4’s 26. Here, Google appears to be forward. Google Bard, which used the PaLM 2 mannequin now even helps 9 Indian languages: Hindi, Tamil, Telugu, Bengali, Kannada, Malayalam, Marathi, Gujarati, and Urdu. And, total, Google is now supporting over 40 languages for Bard together with Bahasa Indonesia.
