14:23 · Apr 17, 2025 · Thu https://x.com/imxiaohu/status/1912262728301047895 X (formerly Twitter) 小互 (@imxiaohu) on X 重大好消息Veo 2 登陆Google AI Studio了意味着大家可以免费白嫖了🤤在Google AI Studio中你可以测试提示,调整宽高比和持续时间等参数,然后立即查看生成的视频,供开发者调试测试!Google还给出了Prompt 编写建议为了获得高质量视频,提示词应该强调以下几个维度:
14:16 · Apr 17, 2025 · Thu https://github.com/lllyasviel/FramePack GitHub GitHub - lllyasviel/FramePack: Lets make video diffusion practical! Lets make video diffusion practical! Contribute to lllyasviel/FramePack development by creating an account on GitHub.
13:17 · Apr 17, 2025 · Thu https://x.com/dotey/status/1912558003783414240?t=2n-C-jeyKEXvIDNhXBKeUg&s=35
14:56 · Apr 16, 2025 · Wed https://github.com/droidrun/droidrun GitHub GitHub - droidrun/mobilerun: Automate your mobile devices with natural language commands - an LLM agnostic mobile Agent 🤖 Automate your mobile devices with natural language commands - an LLM agnostic mobile Agent 🤖 - droidrun/mobilerun
12:42 · Apr 16, 2025 · Wed https://x.com/dotey/status/1912181719605612904?t=guEEuJhTEI0Bqsjhl5ACig&s=35
09:02 · Apr 16, 2025 · Wed https://x.com/op7418/status/1912324183406833929?t=WPkNr0b9FOHevD57h8Ha9Q&s=35 X (formerly Twitter) 歸藏(guizang.ai) (@op7418) on X 谷歌的 Veo 2 终于在他们自己的平台上线了Gemini Advanced 会员可以免费通过 Gemini APP 使用文生视频而且每个视频是 8 秒的长度
09:01 · Apr 16, 2025 · Wed https://x.com/vista8/status/1912411516990812263?t=6uI9NX6xk2MAgXDBz-ho1Q&s=35
08:59 · Apr 16, 2025 · Wed https://x.com/YILS2048/status/1911975701756166317?t=z11pCrygLoncSoa4FY4wPA&s=35 X (formerly Twitter) YILS (@YILS2048) on X 看了宝玉老师的帖子,把 65 页的 Google 官方提示工程白皮书 PDF 文件转为Markdown文件再进行翻译,效果确实会非常好!顺便介绍推荐下MinerU,一个开源高质量数据提取工具,将PDF转换成Markdown和JSON格式。且官方提供了带美观界面的一键客户端,如此一来人人可用。1/3https://t.co/h0AYcJw438
08:58 · Apr 16, 2025 · Wed https://x.com/dotey/status/1912276883196039520?t=Dm0C1ApvqKDZb5SeZ-hRhw&s=35
08:57 · Apr 16, 2025 · Wed https://x.com/karminski3/status/1912287572065415582?t=4kgi73oZCfIWD3KOiLAkJA&s=35 X (formerly Twitter) karminski-牙医 (@karminski3) on X 微软研究院整了个活,发布了个原生 1-bit 的大语言模型 —— bitnet-b1.58-2B-4T有啥意义吗?有的,这个模型虽然将权重量化到超低精度(实际是1.58位,权重只有{-1, 0, +1}三个值),但它在性能上几乎能与其它2B参数规模的全精度模型相媲美。与传统模型相比,这个1-bit模型带来了惊人的效率提升:-
11:25 · Apr 15, 2025 · Tue https://s3.cn-north-1.amazonaws.com.cn/sides-share/%E6%B7%B1%E5%85%A5%E6%8E%A2%E8%AE%A8+Amazon+ElastiCache+.pdf
09:52 · Apr 15, 2025 · Tue https://x.com/HiTw93/status/1911921912399421851?t=tdK8eRDyQsU9Ae0hHdYuag&s=35 X (formerly Twitter) Tw93 (@HiTw93) on X 这个开源的小说转语音智能解决方案 EasyVoice,展示效果做得挺有趣,支持处理大型文本文件,轻松将超长小说转换为语音,这里比较难处理的是如何转成很流畅的对话、多音字处理、分割等地方。https://t.co/JKW7I7MCUe
23:43 · Apr 14, 2025 · Mon https://x.com/Aurimas_Gr/status/1911403414094758029?t=OSEfMcJs_6ZHYW2RliHDng&s=35 X (formerly Twitter) Aurimas Griciūnas (@Aurimas_Gr) on X 𝗠𝗖𝗣 and 𝗔𝟮𝗔: Friends or Foes? In my latest Newsletter episode I talk about both protocols. Could A2A eat up MCP in the long term?I have been asked multiple times why I think the two protocols could become competitive in the future. I tried to outline my…