Distill Any Depth Distillation Creates a Stronger Monocular Depth Estimator a SOTA depth estimation model trained with knowledge distillation algorithm, which efficiently distills knowledge from multiple teacher models to achieve superior performance

thanks AK for sharing our work

13. Distill Any Depth: Distillation Creates a Stronger Monocular Depth Estimator 🔑 Keywords: Monocular Depth Estimation, Cross-Context Distillation, depth normalization, pseudo-labels, multi-teacher framework 💡 Category: Computer Vision 🌟 Research Objective: The paper aims

🚀 The AGI Lab at Westlake University is currently recruiting international PhD students and summer research interns to work on cutting-edge research in Generative AI and Multimodal AI. Feel free to email me for more details!

thanks Gradio for sharing! More models will be released soon.

新しいSOTAの単眼深度推定「Distill-Any-Depth」 使ってみた！ アニメ調でフォトリアルでも普通に良い！！

Caricature to depth! 🤯 huggingface.co/spaces/xingyan…

Distill-Any-Depth: A New SOTA Monocular Depth Estimation Model Achieves Unprecedented Performance agientry.com/blog/380

🚀 Introducing AppAgent X – the next-generation of AppAgent, a self-evolving GUI agent that gets smarter and more efficient over time! 🔓 Open-sourced with a Gradio interface! 📂 Check out our code & models at: appagentx.github.io Paper: huggingface.co/papers/2503.02…

昨日のより軽い奥行き推定（Distill-Any-Depthのbase）モデル。でもTDPTと一緒に動かすとまだ重いな

Recently, the AGI Lab at Westlake University in China introduced #AppAgentX, a GUI agent with self-evolution capabilities. By continuously learning and optimizing its behavior during task execution, it enhances operational efficiency. Experimental results show that AppAgentX

西湖大学发布AppAgentX：会自我进化的AI智能体，手机操作更高效！Westlake University Launches AppAgen... youtube.com/shorts/K3YgtBh… 来自 YouTube

🚀 Excited to share Ultra3D — our latest research on high-resolution 3D asset generation! Achieves SOTA quality with ultra-detailed geometry & texture. Check it out on HuggingFace Papers 🧠👇 huggingface.co/papers/2507.17…

Image-to-3D has come a long way. Tried the new Ultra3D framework from the team at Hitem3D. Great output in ~15 seconds, with decent-quality mesh and texture. The speed + fidelity makes 3D finally feel like something you can reach for early in the process.

We are excited to introduce Ultra3D, a novel 3D generation model that produces extremely detailed 3D models. Project Page: buaacyw.github.io/ultra3d/ Arxiv: arxiv.org/abs/2507.17745

thanks AK