image
Tshaka Armstrong is a nerd. Co-Founder of the non-profit digital literacy organization, Digital Shepherds, he’s also been a broadcast technology reporter, wr…
Tshaka Armstrong is a nerd. Co-Founder of the non-profit digital literacy organization, Digital Shepherds, he’s also been a broadcast technology reporter, wr…
Tshaka Armstrong is a nerd. Co-Founder of the non-profit digital literacy organization, Digital Shepherds, he’s also been a broadcast technology reporter, wr…
Tshaka Armstrong is a nerd. Co-Founder of the non-profit digital literacy organization, Digital Shepherds, he’s also been a broadcast technology reporter, wr…
Tshaka Armstrong is a nerd. Co-Founder of the non-profit digital literacy organization, Digital Shepherds, he’s also been a broadcast technology reporter, wr…
Tshaka Armstrong is a nerd. Co-Founder of the non-profit digital literacy organization, Digital Shepherds, he’s also been a broadcast technology reporter, wr…
B站上有不少详细测评，我看了一些，自己也跑了一些测试。 总体水平在第一梯队，但是比较靠后的位置，稍微落后顶尖闭源image模型一点。 实力是其次，重要的是minimax终于补齐了全模态的最后一块拼图。 文，图，视频，音频，语音，音乐。 闫俊杰前两天放出来的采访中说到，要把视频模型和语言模型合并。 把理解和生成统一。 大…
据昆仑芯消息，11月10日，MiniMax正式开源MiniMax H3-Image模型。昆仑芯同步完成该模型的首发日适配与调优，成为首批完成适配的国产算力厂商之一。
11月11日消息，MiniMax发布图像生成模型MiniMax H3-Image，并同步开放模型权重。该模型由MiniMax H3衍生，支持文生图、图像参考和图像编辑。在H3预训练赋予的广泛图像能力基础上，H3-Image进一步强化了后训练链路，涉及数据、强化学习、步数压缩、PE、部署优化等方面，以提升模型速度、画面质…
大家好啊，前段时间做了个 MBTI 梗化测试 app，想加个"结果分享"功能——用户测完 16 人格，截图给朋友看。听起来简单，真做下去发现要串三段 API： componentSnapshot 拿 PixelMap → ImagePacker 打 PNG → startAbilityByType('share') 调…
