LogoThread Easy
  • 探索
  • 線程創作
LogoThread Easy

Twitter 線程的一站式夥伴

© 2025 Thread Easy All Rights Reserved.

探索

Newest first — browse tweet threads

Keep on to blur preview images; turn off to show them clearly

Deep Gaussian from Motion: Exploring 3D Geometric Foundation Models for Gaussian Splatting

Contributions:
• Pose-free Foundation Model Adaptation: Unlike VGGT/MegaSAM, which rely on pre-computed poses (potentially affected by inaccuracies), our pipeline operates without pose annotations. This is achieved by refining Gaussian geometries dynamically to align photometric appearance with ray-consistent novel view synthesis.

• Progressive and Modular Framework Design: The progressive design enables iterative scalability, addressing GPU bottlenecks evident in VGGT-like pipelines. Modularity ensures robustness against scene diversity, allowing refinement of components independent of memory constraints imposed by dense image sets.

• Scene-specific Gaussian Prediction: Our method dynamically predicts Gaussian geometries for each input scene, adapting to its unique photometric and geometric characteristics for high-quality synthesis—a flexibility less evident in feed-forward methods like VGGT/MegaSAM, which process inputs less adaptively.

Deep Gaussian from Motion: Exploring 3D Geometric Foundation Models for Gaussian Splatting Contributions: • Pose-free Foundation Model Adaptation: Unlike VGGT/MegaSAM, which rely on pre-computed poses (potentially affected by inaccuracies), our pipeline operates without pose annotations. This is achieved by refining Gaussian geometries dynamically to align photometric appearance with ray-consistent novel view synthesis. • Progressive and Modular Framework Design: The progressive design enables iterative scalability, addressing GPU bottlenecks evident in VGGT-like pipelines. Modularity ensures robustness against scene diversity, allowing refinement of components independent of memory constraints imposed by dense image sets. • Scene-specific Gaussian Prediction: Our method dynamically predicts Gaussian geometries for each input scene, adapting to its unique photometric and geometric characteristics for high-quality synthesis—a flexibility less evident in feed-forward methods like VGGT/MegaSAM, which process inputs less adaptively.

Paper (pdf): https://t.co/YLwEAEIDqS

avatar for MrNeRF
MrNeRF
Wed Nov 05 07:55:32
RT @NiohBerg: NYC will go in the same direction as London. Things will get progressively worse, dirtier and more dangerous. Everything will…

RT @NiohBerg: NYC will go in the same direction as London. Things will get progressively worse, dirtier and more dangerous. Everything will…

AI Optimist. Empiricist, not 'rationalist'. Anti world government.

avatar for renji
renji
Wed Nov 05 07:54:14
https://t.co/cI60bXuVAz

https://t.co/cI60bXuVAz

Believing is seeing

avatar for Yangyi
Yangyi
Wed Nov 05 07:53:35
一起来看看11labs如何在AI时代保证高效敏捷的市场策略👇🏻

三个月原则
AI产品在研发过程中会出现一种“前沿研究”与“产品落地”之间难以调和的矛盾

11Labs通过三个月原则,如果一个体验需要模型优化三个月以上,那么产品团队就先行开发“过渡”方案来满足市场需求,避免因等待长期研究而错失良机

双轨制管理
Pre-PMF,给团队6个月时间去证明价值,快速试错验证市场,鼓励探索和冒险,时间到了如果无法证明价值就迅速砍掉

Post-PMF,追求稳定可靠的长期增长,进行大量测试和评估,准备好才部署上线,面相企业用户时明确提供Alpha/Beta版本,让用户自行选择

人才与组织
全球化远程招募,随规模增大加设线下Hub集体办公

并且主动寻找「非传统背景」人才,例如曾雇佣了一位在呼叫中心工作、但业余时间开发了顶尖开源文本转语音模型的工程师,并让他成为核心研究员。

扁平化、小团队、高所有权,有20个5-10人小队,团队拥有极高的自主权去推进和发布产品。

一年前取消了所有内部头衔,旨在确保决策权和影响力来源于个人能力和贡献,而非资历或职位。

刻意控制信息透明度。为避免团队因信息过载而分心,会有选择性地限制某些Slack频道的访问权限,以强制团队保持专注。

市场策略
找到多方共赢模式,从创作者(PLG)到企业客户
不仅和用户,客户共赢,还需要设置好激励机制和员工共赢(毕竟企业销售周期长,和快速迭代的文化有冲突)

有意思的是,11labs会向销售承诺,如果为了公司长期利益而放弃一个“坏”订单,公司依然会支付销售佣金,以此鼓励团队做出正确决策。

一起来看看11labs如何在AI时代保证高效敏捷的市场策略👇🏻 三个月原则 AI产品在研发过程中会出现一种“前沿研究”与“产品落地”之间难以调和的矛盾 11Labs通过三个月原则,如果一个体验需要模型优化三个月以上,那么产品团队就先行开发“过渡”方案来满足市场需求,避免因等待长期研究而错失良机 双轨制管理 Pre-PMF,给团队6个月时间去证明价值,快速试错验证市场,鼓励探索和冒险,时间到了如果无法证明价值就迅速砍掉 Post-PMF,追求稳定可靠的长期增长,进行大量测试和评估,准备好才部署上线,面相企业用户时明确提供Alpha/Beta版本,让用户自行选择 人才与组织 全球化远程招募,随规模增大加设线下Hub集体办公 并且主动寻找「非传统背景」人才,例如曾雇佣了一位在呼叫中心工作、但业余时间开发了顶尖开源文本转语音模型的工程师,并让他成为核心研究员。 扁平化、小团队、高所有权,有20个5-10人小队,团队拥有极高的自主权去推进和发布产品。 一年前取消了所有内部头衔,旨在确保决策权和影响力来源于个人能力和贡献,而非资历或职位。 刻意控制信息透明度。为避免团队因信息过载而分心,会有选择性地限制某些Slack频道的访问权限,以强制团队保持专注。 市场策略 找到多方共赢模式,从创作者(PLG)到企业客户 不仅和用户,客户共赢,还需要设置好激励机制和员工共赢(毕竟企业销售周期长,和快速迭代的文化有冲突) 有意思的是,11labs会向销售承诺,如果为了公司长期利益而放弃一个“坏”订单,公司依然会支付销售佣金,以此鼓励团队做出正确决策。

https://t.co/cI60bXuVAz

avatar for Yangyi
Yangyi
Wed Nov 05 07:53:34
RT @VittoStack: Building an AI security auditor?

You can now join the Solodit API closed beta to get access to:

- 49,000+ vulnerabilities…

RT @VittoStack: Building an AI security auditor? You can now join the Solodit API closed beta to get access to: - 49,000+ vulnerabilities…

CPO @CyfrinAudits | Ex @Alchemy | Created @cyfrinupdraft and @AlchemyLearn | Robotics | Making web3 mainstream

avatar for Vitto Rivabella
Vitto Rivabella
Wed Nov 05 07:52:31
This is that youtube history unc
you know, it's nice that everyone feels compelled to yap about everything these days, even topics obscure to them. As a result, we have a powerful system of selecting for true, general epistemic rigor. I value it highly. 
This guy is sloppy.

This is that youtube history unc you know, it's nice that everyone feels compelled to yap about everything these days, even topics obscure to them. As a result, we have a powerful system of selecting for true, general epistemic rigor. I value it highly. This guy is sloppy.

We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1

avatar for Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Wed Nov 05 07:52:27
  • Previous
  • 1
  • More pages
  • 918
  • 919
  • 920
  • More pages
  • 2131
  • Next