Zhengyi Wang

PhD student @ Tsinghua University.

[Zhengyi Wang @coolboywzy]

Project page: https://t.co/btgjUx70lr

[Zhengyi Wang @coolboywzy]

🚀 Introducing Nano3D — a training-free framework for precise, coherent 3D object editing without masks!

By integrating FlowEdit into TRELLIS and introducing Voxel/Slat-Merge, Nano3D preserves structure & consistency while delivering superior 3D quality.

[Cheng Lu @clu_cheng]

This is a very solid and promising research that scales consistency models to 10B+ video diffusion models. The combination of sCM and Variational Score Distillation is a very promising direction for few-step generation!

[Zhengyi Wang @coolboywzy]

Meet RDT2, our latest foundation model that zero-shot deploys on any robot arms with unseen scenes, objects & instructions.🔥

Fully open-sourced: https://t.co/0SFXmMhBFM
Project page: https://t.co/78IsnrbKzD

[Songming Liu @songming_liu]

😠💢😵‍💫Tired of endless data collection & fine-tuning every time you try out VLA?

Meet RDT2, the first foundation model that zero-shot deploys on any robot arms with unseen scenes, objects & instructions.

No collection. No tuning. Just plug and play🚀

Witness a clear sign of embodied superintelligence

- 7B one-step diffusion → 23 Hz inference⚡
- Re-designed UMI @chichengcc @SongShuran and manufactured 100 portable devices
- Trained on 10K-hour UMI data on 100 real houses
- Zero-shot: pick, place, press, wipe… open-vocabulary
- Demos: block 30 m/s arrows in 500 ms🛡️; first to play ping-pong with an end-to-end model 🏓; extinguish burning incense by shaking quickly🥢

Fully open source at https://t.co/juzBEY4H45
Project page: https://t.co/z8ce3sAvFs

Thanks to awesome collaborators @bang_guo96535 @D0g4M74794 @EthanNg51931527

[Elon Musk @elonmusk]

Cool

[Zhengyi Wang @coolboywzy]

So excited to share Qwen-Image—a 20B MMDiT model! 🚀

It’s been amazing to watch its accurate text rendering start emerging and getting better and better during training. Also, it starts to show preliminary abilities in understanding 3D space and handling spatial transformation.

[Qwen @Alibaba_Qwen]

🚀 Meet Qwen-Image — a 20B MMDiT model for next-gen text-to-image generation. Especially strong at creating stunning graphic posters with native text. Now open-source.

🔍 Key Highlights:
🔹 SOTA text rendering — rivals GPT-4o in English, best-in-class for Chinese
🔹 In-pixel text generation — no overlays, fully integrated
🔹 Bilingual support, diverse fonts, complex layouts

🎨 Also excels at general image generation — from photorealistic to anime, impressionist to minimalist. A true creative powerhouse.

Blog:https://t.co/FZ415nwcAp
Hugging Face：https://t.co/XSLwjJ6lYa
ModelScope：https://t.co/mcZ0dHeD64
Github：https://t.co/A9yvJZ6TJc
Technical report：https://t.co/tld83OrOUP
Demo: https://t.co/MgEAr4uqoR

[Zhengyi Wang @coolboywzy]

DeepMesh V2 drops soon! Upgraded autoregressive 3D mesh generator.🔥🔥🔥

[Zhengyi Wang @coolboywzy]

Dive into the paper 👉 https://t.co/2ZKI1P0lBV

[Zhengyi Wang @coolboywzy]

🚀 Introducing ShapeLLM-Omni, a 3D-native multimodal large language model finetuned from Qwen2.5-VL-7B. It builds on voxel-based 3D VQVAE and a 2.56M-dialogue 3D-Alpaca dataset, enabling 4 tasks: text/image-to-3D, 3D comprehension and editing. Code, model, data open-sourced!

[Zhengyi Wang @coolboywzy]

@cocolitron Thank you, Ben!

[Zhengyi Wang @coolboywzy]

Illustration video for DeepMesh, our latest model for 3D mesh generation! 🔥
DeepMesh generates high-quality meshes from raw point clouds. It can also refine existing meshes, improving their structure and quality.
Open-sourced: https://t.co/RD63HJEWyq

[Zhengyi Wang @coolboywzy]

@ronnie_brownie @_akhaliq We are open-sourced and Meshtron is close-sourced.

[Zhengyi Wang @coolboywzy]

Thank @_akhaliq for sharing our work! We're thrilled to announce DeepMesh, our latest auto-regressive artist-mesh generative model. The model weights and inference code are fully open-sourced!🎉
Code: https://t.co/ee9yBsI3jo…
Project page: https://t.co/NUr5XtDDmA

[AK @_akhaliq]

DeepMesh is out on Hugging Face

Auto-Regressive Artist-mesh Creation with Reinforcement Learning

Conditioned on point clouds and images, DeepMesh generates meshes with intricate details and precise topology, outperforming state-of-the-art methods in both precision and quality.

[Zhengyi Wang @coolboywzy]

@dylan_ebert_ @_akhaliq Thank you, Dylan!

[Zhengyi Wang @coolboywzy]

@yuliangxiu Congratulations!

[Zhengyi Wang @coolboywzy]

@Koven_Yu Nice work!

[Zhengyi Wang @coolboywzy]

@fangfu0830 looking forward to it😍

[Zhengyi Wang @coolboywzy]

🚀 So excited to see LLaMa-Mesh integrated into Blender with #meshgen! 🎉 Now you can generate 3D meshes locally with AI in Blender. Open-source and available now! 🙌 #AI #3D #Blender #LLaMaMesh #OpenSource

[dylan @dylan_ebert_]

Generate meshes with AI locally in Blender

📢 New open-source release

meshgen, a local blender integration of LLaMa-Mesh, is open source and available now 🤗

[dylan @dylan_ebert_]

LLaMa-Mesh running locally in blender

official @huggingface release soon 🤗

[Zhengyi Wang @coolboywzy]

Code https://t.co/CzROqj2NRn

[Zhengyi Wang @coolboywzy]

🚀 Introducing LLaMA-Mesh! 🎉
We fine-tuned LLaMA on 3D Mesh data, enabling LLMs to natively generate 3D meshes via chatting while retaining their original language capabilities. ✨
Model weights and inference code are fully open-sourced.
🌐 Project page: https://t.co/dSMlf6ieCs

[Zhengyi Wang @coolboywzy]

Paper link https://t.co/TcDADyR2Gs