Raphael

@cyrilXBT AI

Giving away 1 month of Claude AI Pro to 3 people. Follow + Repost to enter. Comment "AI" Winners announced Friday.

@BleapApp 🤗

giveaway time! win a Claude Max 5x subscription to enter > follow @BleapApp > rt and like this post winner will be selected in 24 hours (btw, you get 20% cashback when using a Bleap card for your claude subscription anyway)

@fahdmirza Могу ли я использовать это для vlm ? Есть ли инструкция как самому сделать это?

💥 We ran Qwen3.5 9B at INT4 with ParoQuant — and it barely lost anything 🌀 ♠ A brand new ICLR 2026 quantization method that finally solves reasoning accuracy at 4-bit 🔬 🔹 4x smaller than FP16 — runs locally on consumer hardware 🔹 AIME-24 score drops only 2.3 points vs FP16 — AWQ drops 13 🔹 Smarter than AWQ, faster than QTIP — best of both worlds 🔹 Single fused CUDA kernel — less than 10% overhead 🔹 Works with vLLM and Open WebUI out of the box 🔹 Tested live with a reasoning prompt to see if the quality holds 🔹 The quantization method that reasoning models actually needed 🔥 Watch the full video below 👇

I realized I can gift 3 people a week of Claude Code!! 🎉 RT and comment "Code" to enter into this giveaway. Winner's will be announced in 48 hours.

@VadimStrizheus Code

@JuwonAJ @Loganbricks1 @D5Render May I get your Revit file and your visualizations for my training?

Here we go again ✍️ A quick sketch of a Living room space with an open kitchen. Modelled by @Loganbricks1 in revit Visualized by yours sincerely, ME 🙂‍↔️ 💎Workflow - Revit + D5render. @D5Render Enjoy ✨

@wavespeed_ai Hi bro, can you think of something for the cameras in the scene (interior design, there is more than one object)

🚀 New on WaveSpeedAI: Qwen-Image-Edit Multiple Angles Get any camera angle you need with one image. Front, side, back, bird's-eye, even dramatic low angles. It's especially handy for: 🛍️ product shots 🧩 3D assets 🎭 character design 👕 virtual try-on 🎨 creative experiments One image, many perspectives. 📸✨

@cloudxdev Bro, I like what you did, but Tree.js updates quickly, I need the most updated skill threejs-loaders.md, and you have an old one, but thanks anyway☹️

🧊 Threejs Skills for Claude Code to create 3D Web Design Elements 📂 Bookmark it - Give Claude Code base level of Three.js knowledge. - 10 skill files covering scene setup, shaders, animations, post-processing. - Claude Code will have the knowledge of how to steer Threejs without bloating the context github.com/CloudAI-X/thre…

@MedOumoumad Thank you bro). To be honest, I was expecting—and am still expecting—the opposite: “material extraction.”

@__Rafail__ Here man, before I forget again : huggingface.co/oumoumad/Qwen-… Don't have time to prepare a proper release with image examples etc..

Thanks for the amazing response to the Material Transfer LoRA! While I’m still working on an improved version, here’s another LoRA you can try out, it does the opposite: “Extract Materials.” huggingface.co/oumoumad/Qwen-…

@MedOumoumad Love u 😘

@MedOumoumad @ltx_model 😭😭 bro..

New IC-LoRA to use with @ltx_model I trained this one with Lightness maps. So you can use it to drive luminance and light sources in your generated videos. I find it especially fun to use for Audio Reactive workflows. Grab it here : huggingface.co/oumoumad/ltx-2…

@onirenaud It's opencourse? git?

After 4 days, I think my Gaussian Splatting implementation in WebGPU is ready. 🚀 Next up: focus on performance and an online visualizer in Three.js Blocks by the end of the week. Also will ship GS3D "lit" mode (SH + PBR) and some GS3D+Physics examples! 1/2

@siyuhuh How can I get this project? github?

Check out the IOT floor plan demo with 2d layout editing. vibe coding with Gemini 3. Custom shader material applied. floorplan-demo.whothree.space #three #3DCG #IoT #floor #Ikea #threejs #webgl #Gemini #r3f

@MedOumoumad huggingface.co/gvecchio/Stabl…

@__Rafail__ I finished training it actually ! but didn't post yet because I wondered if I should retrain it on the 2511 version, I think I'll just post it as is for now, soon as I get a chance, lately I've been busy testing things on open source video models like LTX-2. But stay tuned.

Φeat Physically-Grounded Feature Representation

@wildmindai It would be great to have a model that performs the same task on 3D point clouds.

Orient Anything: a unified, symmetry-aware orientation estimator; based on VGGT + DINOv2 encoder; tops GPT-4o & Gemini-2.5 in symmetry recognition (65.2%). orient-anythingv2.github.io