Facebook (Meta) introduces LLaMA, a Large Language Model for Open Science
OpenAI’s ChatGPT, Google’s Bard, and Microsoft’s Prometheus are no longer the only large language model-powered AIs out there. Meta, formerly known as Facebook, has introduced their own AI, named LLaMA, which stands for Large Language Model Meta AI, with an animal-themed acronym.
LLaMA is a large language model released by Meta for open science. It’s a foundational model available in several sizes, designed to democratize access to large language models for researchers. LLaMA is trained on 1.4 trillion tokens, can be fine-tuned for different tasks, and is released under a non-commercial license. Meta emphasizes the importance of responsible AI guidelines for large language models.
LLaMA is currently not accessible to the general public, but individuals can request access here
REFERENCES
Zuckerberg Introduces Meta’s Answer to ChatGPT, LLaMA: https://gizmodo.com/facebook-chatgpt-google-ai-chatbot-google-bard-1850155514
LLaMA: Open and Efficient Foundation Language Models: Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, Aurelien Rodriguez, Armand Joulin, Edouard Grave, Guillaume Lample: https://arxiv.org/abs/2302.13971
LLaMA Ai Facebook:https://ai.facebook.com/blog/large-language-model-llama-meta-ai/
LLaMA Access request https://docs.google.com/forms/d/e/1FAIpQLSfqNECQnMkycAp2jP4Z9TFX0cGR4uf7b_fBxjY_OjhJILlKGA/viewform
Wiki Hyphen Website | Updates 5th March 2023 | Link: https://www-wiki.com/Facebook-Ai-Chatbot-llama