🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

I investigated this in 2017 and even at the time it looked encouraging. Glad to see it make a comeback. github.com/unixpickle/sgd…

Very happy to see the TTT-series reaching yet another milestone! This time it serves as an inspiration for next-generation architecture post-Transformer, and by connecting TTT to Transformer, it can explain why (autoregressive) Transformers are so good at in-context learning!

🔥new language modelling layer in town - more expressive than RNN - more efficient (linear comp.) than attention! key perspective: LM layers are ML models trained to memorize tokens in a sequence: - Linear memorizer => RNN - Kernel mem. => attention - Neural mem. => our layer

Thinking about a PhD? Don’t miss the chance to work with Elliott / Shangzhe Wu! He’s not only a brilliant researcher but also an inspiring mentor and collaborator. Excited to see the amazing projects his new team will bring to life! 🌟

This is so cooool!

Meet Pika 2.0! Besides improved quality and motion, our new model can embed user-provided concepts into the generated videos, without any training! Combined with an unprecedented level of text-alignment, you can now create YOUR OWN personalized content with minimal effort 🎬

Impressive work on scaling test-time compute for diffusion models!

Introducing GSPN: A Leap Forward in Vision Attention Mechanisms Paper: arxiv.org/pdf/2501.12381 Project: whj363636.github.io/GSPN/ We present GSPN (Generalized Spatial Propagation Network), a novel attention mechanism developed at NVIDIA. Unlike pixel-to-pixel scans like mamba,

Introducing “Diffusion Autoencoders are Scalable Image Tokenizers” (DiTo). We show that with proper designs and scaling up, diffusion autoencoders (a single L2 loss) can outperform the GAN-LPIPS tokenizers (hybrid losses) used in current SOTA generative models. (1/4)

Me when the deadline is tomorrow but vibes come first. Need more memes rigged for maximum chill🤤 #chillguy

Meet our first general-purpose robot at Dexmate dexmate.ai/vega Adjustable height from 0.66m to 2.2m: compact enough for an SUV, tall enough to reach those impossible high shelves. Powerful dual arms (15lbs payload each) and omni-directional mobility for ultimate

Today, we're releasing a new paper – One-Minute Video Generation with Test-Time Training. We add TTT layers to a pre-trained Transformer and fine-tune it to generate one-minute Tom and Jerry cartoons with strong temporal consistency. Every video below is produced directly by

Excited to share our new paper, "One-Minute Video Generation with Test-Time Training (TTT)" in collaboration with NVIDIA. We augment a pre-trained Transformer with TTT-layers and finetune it to generate one-minute Tom and Jerry cartoons with strong temporal and spatial

Test-Time Training (TTT) is now on Video! And not just a 5-second video. We can generate a full 1-min video! TTT module is an RNN module that provides an explicit and efficient memory mechanism. It models the hidden state of an RNN with a machine learning model, which is updated

Kids in 2025: writing Tom & Jerry instead of watching them. TTT + diffusion transformer is wild!