据科技媒体The Information报道，Meta公司计划在今年7月发布其人工智能模型Llama 3。该模型旨在与OpenAI的GPT-4相媲美，后者是一款强大的多模态模型，能够处理长文本并支持图像输入。然而，Meta的一位员工透露，由于研究团队尚未开始对Llama 3进行微调，因此尚不确定该模型是否也将具备多模态功能。此外，Llama 3的参数规模预计将达到1400亿，远超去年发布的Llama 2的700亿参数。值得注意的是，Meta还在努力解决Llama 2中的一个问题，即模型无法处理有争议的话题。
英文翻译内容：
Title: Meta to Release Llama 3 in July, Modeled After GPT-4
Keywords: AI model, Llama 3, GPT-4
News content:
According to a report by tech media The Information, Meta plans to release its AI model Llama 3 in July. The model is designed to rival OpenAI’s GPT-4, a powerful multimodal model capable of handling long text and image input. However, an employee at Meta has revealed that researchers have not yet started fine-tuning Llama 3, so it is uncertain if the model will also be multimodal. Furthermore, Llama 3 is expected to reach a parameter scale of 14 billion, significantly surpassing Llama 2’s 7 billion parameters from last year. Notably, Meta is also working to address an issue found in Llama 2 where the model could not handle controversial topics.
【来源】https://wallstreetcn.com/articles/3709312
Views: 1