首页 News 正文

On November 14th, Alibaba's Intelligent Information Business Group announced the launch of a fully stack self-developed quark large model with hundreds of billions of parameters. It is a multi-modal large model based on the Transformer architecture and independently developed. The quark side stated that the overall capability of its self-developed large model exceeds GPT-3.5, with some scenarios superior to GPT-4.
您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

六月清晨搅 注册会员
  • 粉丝

    0

  • 关注

    0

  • 主题

    30