Adeer Khan

Flux Gym 🚀Dead simple web UI for training FLUX LoRA with LOW VRAM (12GB/16GB/20GB) support. 🌐Github: github.com/cocktailpeanut… Amazing!

EarthCrafter: Scalable 3D Earth Generation via Dual-Sparse Latent Diffusion 🚀 Paper: arxiv.org/abs/2507.16535 📷 Project: attention-warp.github.io

They introduce a large-scale synthetic dataset, SpatialGen Dataset, featuring 12,328 structured annotated scenes with 57,440 rooms, and 4.7M photorealistic 2D renderings.

SpatialGen: Layout-guided 3D Indoor Scene Generation 👋Project: manycoreresearch.github.io/SpatialGen/ 🚀Paper: arxiv.org/pdf/2509.14981

🎙️ Meet Qwen3-ASR — the all-in-one speech recognition model! ✅ High-accuracy EN/CN + 9 more languages: ar, de, en, es, fr, it, ja, ko, pt, ru, zh ✅ Auto language detection ✅ Songs? Raps? Voice with BGM? No problem. <8% WER ✅ Works in noise, low quality, far-field ✅ Custom context? Just paste ANY text — names, jargon, even gibberish 🧠 ✅ One model. Zero hassle.Great for edtech, media, customer service & more. API：bailian.console.alibabacloud.com/?tab=doc#/doc/… ModelScope Demo: modelscope.cn/studios/Qwen/Q… Hugging Face Demo: huggingface.co/spaces/Qwen/Qw… Blog：qwen.ai/blog?id=41e4c0…

🚨Qwen released Qwen3-ASR speech recognition model but no open-source.

PhD Opportunities in Energy Transport and Conversion - University of Maryland We are seeking highly motivated PhD students who are passionate about transforming the global energy landscape and tackling grand challenges in science and engineering.

Generated isometric café

GeoSplat: A Deep Dive into Geometry-Constrained Gaussian Splatting Abstract: A few recent works explored incorporating geometric priors to regularize the optimization of Gaussian splatting, further improving its performance. However, those early studies mainly focused on the use of low-order geometric priors (e.g., normal vector), which are unreliably estimated by noise-sensitive methods like local principal component analysis. To address their limitations, we first present GeoSplat, a general geometry-constrained optimization framework that exploits both first-order and second-order geometric quantities to improve the entire training pipeline of Gaussian splatting, including Gaussian initialization, gradient update, and densification. As an example, we initialize the scales of 3D Gaussian primitives in terms of principal curvatures, leading to a better coverage of the object surface than random initialization. Secondly, based on certain geometric structures (e.g., local manifold), we introduce efficient and noise-robust estimation methods that provide dynamic geometric priors for our framework. We conduct extensive experiments on multiple datasets for novel view synthesis, showing that our framework, GeoSplat, significantly improves the performance of Gaussian splatting and outperforms previous baselines.

@rrmdp @AravSrinivas Reminds me of the good old days of Yahoo

@AravSrinivas oh YES! Please please please never ever turn it into this

Minimalism is back.

Cursor is now free for students. Enjoy!

Ideogram v3 again: > in swirling free-flowing words "Enjoy a slice of fofr's perfectly melted ai" written in beautiful stretched lettering with interesting interconnections and glyphs and ligatures, fills the space, repeating pattern of monstera leaves, all the letters effortlessly flow into each other

📢 New Comfy Drop /w @SyntaxDiffusion + @yoland_yan 🎤 Host: @PurzBeats ⌚ May 6th - 5:30pm EST / 2:30pm PST 📍 YouTube, X, and Twitch Change of plans - something's coming.

Been using Gemini 2.5 Pro in Cursor, and have to say that it has been better than Claude

LLaMA-Omni2 is out on Hugging Face LLM-based Real-time Spoken Chatbot with Autoregressive Streaming Speech Synthesis

@airkatakana Switch to Cursor

with claude on the $20 tier im coding for 30 mins at a time, reaching usage limits, and getting kicked off for 5 hours. no one works in 30 min sprints every 5 hours, this is completely unusable

After Flux, we get this!

Huge

BoltAI — All AI models in one app.

Love to see how beautifully run the QwQ-32B effectively via unsloth! Hope to see how fine-tune the QwQ-32B via @UnslothAI soon! 👍🤗