据报道，社交媒体巨头Meta计划在今年7月份发布其最新的语言模型Llama 3。该模型据称将拥有超过1400亿个参数，接近于OpenAI的GPT-4的水平。这意味着Llama 3有望成为市场上最强大的语言模型之一，能够处理更长的文本，甚至可能支持图像输入。然而，Meta尚未决定Llama 3是否将是一个多模态模型，因为研究人员尚未开始对模型进行微调。此外，Meta还在努力解决Llama 2的一个关键问题，即模型无法处理有争议的话题。
英文标题：Meta to Release Llama 3 in July, Targeting GPT-4-Level Capabilities
英文关键词：Meta AI, Llama 3, Language Models
英文新闻内容：
Meta, the social media giant, is reportedly planning to release its latest language model, Llama 3, in July. The model is expected to have over 14 billion parameters, approaching the capabilities of OpenAI’s GPT-4. This would make Llama 3 one of the most powerful language models on the market, capable of handling longer texts and possibly image inputs. However, Meta has not yet decided whether Llama 3 will be a multimodal model, as researchers have yet to start fine-tuning the model. Additionally, Meta is working to address a key issue with Llama 2, the model’s inability to handle controversial topics.
【来源】https://wallstreetcn.com/articles/3709312
Views: 1