PulseAugur
中
实时 20:02:59
Português(PT) Tencent lança Hy-Memory 2.0: stack de memória persistente para atores autônomos com 3 camadas, MoE 30B-A3B, 72,6% no LoCoMo e app MemHub sob licença aberta Apac

Kyutai 和腾讯发布新AI内存和机器人模型

作者 PulseAugur 编辑部 · [2 个来源] ·

Kyutai 发布了 REKI 2.4B VLA，一个拥有32个内存寄存器的机器人模型，在LIBERO-Plus基准测试中达到90.4%，在NVIDIA H100硬件上的延迟为21.8毫秒。同时，腾讯发布了 Hy-Memory 2.0，一个面向自主智能体的持久化内存栈。该内存栈包含一个3层MoE 30B-A3B模型和MemHub应用，在Apache 2.0开源许可下，在LoCoMo基准测试中表现为72.6%。 AI

影响 这些发布展示了AI内存系统和机器人模型能力的进步，可能提高自主智能体的性能和机器人控制。

排序理由 多个研究实验室发布新模型及相关基准测试。

在 Mastodon — sigmoid.social 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →

Kyutai 和腾讯发布新AI内存和机器人模型

本文如何被排名

Signal score
6 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
多个研究实验室发布新模型及相关基准测试。
Source corroboration
2 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
model release, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [2]

  1. Mastodon — sigmoid.social TIER_1 Português(PT) · [email protected] ·

    Kyutai推出REKI 2.4B VLA：具有32个寄存器内存的机器人模型（+0.1毫秒），在LIBERO-Plus上达到90.4%，在H100上延迟21.8毫秒。查看代码。# ai # m

    Kyutai lança REKI 2.4B VLA: modelo robótico com memória em 32 registradores (+0,1 ms), 90,4% no LIBERO-Plus e latência de 21,8 ms em H100. Veja código. # ai # machinelearning # python # programming # software # coding # development # engineering # inclusive # community Kyutai REK…

    链接 dev.to/…/kyutai-reki-24b-vla-32-registrad…
  2. Mastodon — sigmoid.social TIER_1 Português(PT) · [email protected] ·

    腾讯推出Hy-Memory 2.0：面向自主代理的持久化内存栈，具有3层、MoE 30B-A3B，在Apac开源许可下LoCoMo和MemHub应用上达到72.6%

    Tencent lança Hy-Memory 2.0: stack de memória persistente para atores autônomos com 3 camadas, MoE 30B-A3B, 72,6% no LoCoMo e app MemHub sob licença aberta Apache 2.0. # ai # machinelearning # python # programming # software # coding # development # engineering # inclusive # comm…

    链接 dev.to/…/tencent-hunyuan-revela-hy-memory…

相关实体

相关话题