Loading video...

Video Failed to Load

Go Home

🚀 Introducing Sparse VideoGen2 (SVG2) — Pareto-frontier video generation acceleration with semantic-aware sparse attention! 🏆Spotlight paper accepted by #NeurIPS2025 ✅ Training-free & plug-and-play ✅ Up to 2.5× faster on HunyuanVideo, 1.9× faster on Wan 2.1 ✅ SOTA quality preserved (no artifacts, no flickering, no distortion) 📚Paper: 💻Code: 🌐Website: ⚡Attention...

HaochengXiUCB's profile picture

Haocheng Xi

1,621 subscribers

43,470 views • 10 months ago •via X (Twitter)

#NeurIPS2025
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

NVIDIA released Sol-Attn A training-free sparse attention method accelerating video generation inference by up to 2.1× while preserving visual quality, unifying dynamic routing and approximate correction in a single online-softmax pass.
0:10

NVIDIA released Sol-Attn A training-free sparse attention method accelerating video generation inference by up to 2.1× while preserving visual quality, unifying dynamic routing and approximate correction in a single online-softmax pass.

DailyPapers

22,868 views • 4 days ago

🚀 Sol Video Inference Engine is here! An agent-native, training-free full-stack accelerator for video diffusion. It auto-tunes cache + sparse attn + token pruning + quant + kernel fusion for any model/hardware/config. >2× end-to-end speedup on 64B Cosmos3-Super, 22B LTX-2.3 and 2B SANA-Video — near-lossless VBench quality, minimal human effort. Practical acceleration for real video gen deployment. 📄 Paper: 🌐 Project: 💻 Code: Proud of the team! 🎉
0:10

🚀 Sol Video Inference Engine is here! An agent-native, training-free full-stack accelerator for video diffusion. It auto-tunes cache + sparse attn + token pruning + quant + kernel fusion for any model/hardware/config. >2× end-to-end speedup on 64B Cosmos3-Super, 22B LTX-2.3 and 2B SANA-Video — near-lossless VBench quality, minimal human effort. Practical acceleration for real video gen deployment. 📄 Paper: 🌐 Project: 💻 Code: Proud of the team! 🎉

Enze Xie

35,993 views • 1 month ago

Introducing DeepConf: Deep Think with Confidence 🚀 First method to achieve 99.9% on AIME 2025 with open-source models! Using GPT-OSS-120B even without tools, we reached this almost-perfect accuracy while saving up to 85% generated tokens. It also delivers many strong advantages for parallel thinking: 🔥 Performance boost: ~10% accuracy across models & datasets ⚡ Ultra-efficient: Up to 85% fewer tokens generated 🔧 Plug & play: Works with ANY existing model - zero training needed (no hyperparameter tuning as well!) ⭐ Easy to deploy: Just ~50 lines of code in vLLM (see PR below) 📚 Paper: 🌐 Project: joint work with: Yichao Fu , xuewei_wang, Yuandong Tian (see details in the comments below)
0:16

Introducing DeepConf: Deep Think with Confidence 🚀 First method to achieve 99.9% on AIME 2025 with open-source models! Using GPT-OSS-120B even without tools, we reached this almost-perfect accuracy while saving up to 85% generated tokens. It also delivers many strong advantages for parallel thinking: 🔥 Performance boost: ~10% accuracy across models & datasets ⚡ Ultra-efficient: Up to 85% fewer tokens generated 🔧 Plug & play: Works with ANY existing model - zero training needed (no hyperparameter tuning as well!) ⭐ Easy to deploy: Just ~50 lines of code in vLLM (see PR below) 📚 Paper: 🌐 Project: joint work with: Yichao Fu , xuewei_wang, Yuandong Tian (see details in the comments below)

Jiawei Zhao

464,518 views • 11 months ago

pip install spectralquant ✂️ Up to 6.62x KV cache compression for LLMs and transformers. Same model. Faster outputs. Smaller KV cache. Try now (2 mins): - KV cache integration via Hugging Face's DynamicCache - Three presets: 5.95x (paper), 6.55x (validated), 6.68x (edge) - Mistral 7B / Qwen 2.5 7B / Llama 3.1 8B verified - Pure PyTorch + future CUDA kernel support - Auto-calibration from a bundled corpus 📰 Paper: 💻 Code, quickstart, and benchmarks: #LLM #Inference #PyTorch #OpenSource #MachineLearning #LLM #KVCache #Inference
0:15

pip install spectralquant ✂️ Up to 6.62x KV cache compression for LLMs and transformers. Same model. Faster outputs. Smaller KV cache. Try now (2 mins): - KV cache integration via Hugging Face's DynamicCache - Three presets: 5.95x (paper), 6.55x (validated), 6.68x (edge) - Mistral 7B / Qwen 2.5 7B / Llama 3.1 8B verified - Pure PyTorch + future CUDA kernel support - Auto-calibration from a bundled corpus 📰 Paper: 💻 Code, quickstart, and benchmarks: #LLM #Inference #PyTorch #OpenSource #MachineLearning #LLM #KVCache #Inference

ani

17,269 views • 2 months ago

🚀 Introducing CodeWarp: Jump to Definition with MicroGPT! 🚀 No more endless scrolling—CodeWarp lets you instantly navigate to function, variable, and class definitions with ease. Code faster, smarter, and more efficiently! ⚡
0:14

🚀 Introducing CodeWarp: Jump to Definition with MicroGPT! 🚀 No more endless scrolling—CodeWarp lets you instantly navigate to function, variable, and class definitions with ease. Code faster, smarter, and more efficiently! ⚡

MicroGPT

213,368 views • 1 year ago

🎥 Video generation is hitting the memory wall. As videos get longer, the KV cache quietly explodes — and long-horizon consistency starts to break. We built Quant VideoGen: a training-free KV cache compression method for auto-regressive video diffusion. Instead of storing every KV in high precision, QVG exploits video’s spatiotemporal redundancy with semantic-aware smoothing + progressive residual quantization. 🚀 Up to 7× KV memory reduction ⚡ <4% overhead ✅ Strong long-video quality 🕹️ Deploy HYWorldPlay on your own RTX 5090 locally KV compression is becoming a core scaling primitive — not just for LLMs, but for video generation too. Paper: Code: (1/5)
0:10

🎥 Video generation is hitting the memory wall. As videos get longer, the KV cache quietly explodes — and long-horizon consistency starts to break. We built Quant VideoGen: a training-free KV cache compression method for auto-regressive video diffusion. Instead of storing every KV in high precision, QVG exploits video’s spatiotemporal redundancy with semantic-aware smoothing + progressive residual quantization. 🚀 Up to 7× KV memory reduction ⚡ <4% overhead ✅ Strong long-video quality 🕹️ Deploy HYWorldPlay on your own RTX 5090 locally KV compression is becoming a core scaling primitive — not just for LLMs, but for video generation too. Paper: Code: (1/5)

Haocheng Xi

65,008 views • 3 months ago

🚀 LongCat-Video Now Open-Source: Text/Image-to-Video + Video Continuation in One Model 🏆 Text/Image-to-Video Performance Hits Open-Source SOTA 🎬 Minutes-Long High-Quality Videos: No Color Drift/Quality Loss (Industry-Standout) ⚙ 13.6B Params | Strong Open-Source DiT-Based Unified Multitask Video Base Model ⚡ C2F Pipeline + Block Sparse Attention: 720p/30fps Video in Minutes 🤗 Open-Source Links: GitHub： Hugging Face： Project Page：
0:06

🚀 LongCat-Video Now Open-Source: Text/Image-to-Video + Video Continuation in One Model 🏆 Text/Image-to-Video Performance Hits Open-Source SOTA 🎬 Minutes-Long High-Quality Videos: No Color Drift/Quality Loss (Industry-Standout) ⚙ 13.6B Params | Strong Open-Source DiT-Based Unified Multitask Video Base Model ⚡ C2F Pipeline + Block Sparse Attention: 720p/30fps Video in Minutes 🤗 Open-Source Links: GitHub： Hugging Face： Project Page：

Meituan LongCat

43,802 views • 9 months ago

Attention fellow YouTube creators! Launching TubeToolkit in beta 🚀 Built in ~72 hours with Claude Code. It turns your video transcript into the full supporting content package (titles, timestamps, description, hooks, thumbnail ideas) so you can publish faster and stay on-brand. ✅ Free credits to start ✅ No credit card required 💳 Paid tiers available if you want to scale Sign up today: If you’re testing + giving useful feedback and you run out of credits, DM/reply — let me know.
0:14

Attention fellow YouTube creators! Launching TubeToolkit in beta 🚀 Built in ~72 hours with Claude Code. It turns your video transcript into the full supporting content package (titles, timestamps, description, hooks, thumbnail ideas) so you can publish faster and stay on-brand. ✅ Free credits to start ✅ No credit card required 💳 Paid tiers available if you want to scale Sign up today: If you’re testing + giving useful feedback and you run out of credits, DM/reply — let me know.

Branden Flasch

17,253 views • 6 months ago

We have released the code and weights for our #CVPR2023 paper "Avatars Grow Legs: Generating Smooth Human Motion from Sparse Tracking Inputs with Diffusion Model"! code: abs: project: The demo is below:
0:14

We have released the code and weights for our #CVPR2023 paper "Avatars Grow Legs: Generating Smooth Human Motion from Sparse Tracking Inputs with Diffusion Model"! code: abs: project: The demo is below:

Artsiom Sanakoyeu

35,712 views • 3 years ago

New freemint collection coming right up! 🛸 Locus Fleet: Special Proto Craft 🛸 ✅ No FCFS, No Public, Whitelist Only! ✅ 350 confirmed collabs and more on the way! ✅ Many opportunities to secure your whitelist spot! Mint Details: 📦 Amount: 4,500 WL 🪙 Price: FREE 🌐 Chain: Polygon 📅 Date: September 26, 2024 🚀 Launchpad: Space Mavericks Website
0:14

New freemint collection coming right up! 🛸 Locus Fleet: Special Proto Craft 🛸 ✅ No FCFS, No Public, Whitelist Only! ✅ 350 confirmed collabs and more on the way! ✅ Many opportunities to secure your whitelist spot! Mint Details: 📦 Amount: 4,500 WL 🪙 Price: FREE 🌐 Chain: Polygon 📅 Date: September 26, 2024 🚀 Launchpad: Space Mavericks Website

Gravity Blasters

42,226 views • 1 year ago

⚡ Want to help your athlete get faster? Single-Leg Pogo Hops improve: ✅ Speed ✅ Balance ✅ Explosiveness ✅ Coordination The goal: spend less time on the ground and more time moving. 3 sets of 15-20 hops per leg before training. Train smart. Get faster. ⚡ #SpeedTraining
0:10

⚡ Want to help your athlete get faster? Single-Leg Pogo Hops improve: ✅ Speed ✅ Balance ✅ Explosiveness ✅ Coordination The goal: spend less time on the ground and more time moving. 3 sets of 15-20 hops per leg before training. Train smart. Get faster. ⚡ #SpeedTraining

USG

42,122 views • 1 month ago

⚡ Video diffusion is slow. PDD changes that. 🚀 Introducing Parallel Decoding Distillation, our latest work on accelerating video diffusion models. Paper: Page: By Neta Shaul, Julius Berner, and Chao Liu. Below PDD on LTX-2.3 🔊
0:10

⚡ Video diffusion is slow. PDD changes that. 🚀 Introducing Parallel Decoding Distillation, our latest work on accelerating video diffusion models. Paper: Page: By Neta Shaul, Julius Berner, and Chao Liu. Below PDD on LTX-2.3 🔊

Arash Vahdat

19,965 views • 8 days ago

Can $300 RTX3060 be faster than $10,000 A100 for LLM decoding?🧐Yes Introduce EAGLE-2: - 4.26x🚀vanilla (13B) - 1.4x🚀EAGLE (13B) - No extra training than EAGLE - Maintain distribution Paper: Code: Demo:
0:13

Can $300 RTX3060 be faster than $10,000 A100 for LLM decoding?🧐Yes Introduce EAGLE-2: - 4.26x🚀vanilla (13B) - 1.4x🚀EAGLE (13B) - No extra training than EAGLE - Maintain distribution Paper: Code: Demo:

Hongyang Zhang

35,696 views • 2 years ago

Can robots learn without training❓ [𝗜𝘁'𝘀 𝗼𝗽𝗲𝗻 𝘀𝗼𝘂𝗿𝗰𝗲𝗱 ⬇ ] Teaching robots to do complex tasks WITHOUT spending hours training them. Sounds cool, right? That's exactly what DIAL-MPC does! The first training-free method for whole-body torque control using full-order dynamics: ✅ Instantly checks if a robot's moves are right or wrong ✅ Adapts quickly to new tasks without needing extra training ✅ Could work hand-in-hand with other robot learning methods Robots are getting smarter AND faster without the need for long training sessions. Website: Paper: Code: Saw this first Haoru Xue ✈️ CVPR 🙏
0:10

Can robots learn without training❓ [𝗜𝘁'𝘀 𝗼𝗽𝗲𝗻 𝘀𝗼𝘂𝗿𝗰𝗲𝗱 ⬇ ] Teaching robots to do complex tasks WITHOUT spending hours training them. Sounds cool, right? That's exactly what DIAL-MPC does! The first training-free method for whole-body torque control using full-order dynamics: ✅ Instantly checks if a robot's moves are right or wrong ✅ Adapts quickly to new tasks without needing extra training ✅ Could work hand-in-hand with other robot learning methods Robots are getting smarter AND faster without the need for long training sessions. Website: Paper: Code: Saw this first Haoru Xue ✈️ CVPR 🙏

Ilir Aliu

71,502 views • 1 year ago

OSS devs shouldn’t have to beg for compute. Exabits is the launchpad for open-source AI to go orbital. Training on Exabits means: ✅ No throttling ✅ No permission slips ✅ No infra bottlenecks Just plug in and scale.
0:19

OSS devs shouldn’t have to beg for compute. Exabits is the launchpad for open-source AI to go orbital. Training on Exabits means: ✅ No throttling ✅ No permission slips ✅ No infra bottlenecks Just plug in and scale.

Exabits

20,265 views • 10 months ago

✅ Rewind on mobile ✅ 2x faster ✅ Wont melt your phone Kick app 3.0 now live 🚀
0:17

✅ Rewind on mobile ✅ 2x faster ✅ Wont melt your phone Kick app 3.0 now live 🚀

KICK

510,007 views • 1 year ago

(1/n) 🚀 With FastVideo, you can now generate a 5-second video in 5 seconds on a single H200 GPU! Introducing FastWan series, a family of fast video generation models trained via a new recipe we term as “sparse distillation”, to speed up video denoising time by 70X! 🖥️ Live demo: (Thanks to @gmicloud for the support!) 🔗 Blog: 🔓 We fully open-source our models, code, and data with Apache-2.0 licenses
0:13

(1/n) 🚀 With FastVideo, you can now generate a 5-second video in 5 seconds on a single H200 GPU! Introducing FastWan series, a family of fast video generation models trained via a new recipe we term as “sparse distillation”, to speed up video denoising time by 70X! 🖥️ Live demo: (Thanks to @gmicloud for the support!) 🔗 Blog: 🔓 We fully open-source our models, code, and data with Apache-2.0 licenses

Hao AI Lab

78,660 views • 1 year ago

🚀 #NVlabs #InstantSplat now supports high-quality sparse-view surface reconstruction in just seconds! 📸 Build your scene with just 3 images. It's effortless, fast, and ready for you to explore. 💻 Code is now available
0:17

🚀 #NVlabs #InstantSplat now supports high-quality sparse-view surface reconstruction in just seconds! 📸 Build your scene with just 3 images. It's effortless, fast, and ready for you to explore. 💻 Code is now available

Zhiwen(Aaron) Fan

22,939 views • 1 year ago

Introduce EAGLE, a new method for fast LLM decoding based on compression: - 3x🚀than vanilla - 2x🚀 than Lookahead (on its benchmark) - 1.6x🚀 than Medusa (on its benchmark) - provably maintains text distribution - trainable (in 1~2 days) and testable on RTX 3090s Playground: Blog: Code: ⚒️First Principle: Compression! Yi Ma We find that the sequence of second-top-layer features is compressible, making the prediction of subsequent feature vectors from previous ones easy by a small model. 🙏Acknowledge: This project is greatly inspired by the Medusa team (Tianle Cai @yli3521 Zhengyang Geng Hongwu Peng Tri Dao), the Lookahead team (Hao Zhang LMSYS Org), and others. Joint work with Yuhui Li and Chao Zhang
0:15

Introduce EAGLE, a new method for fast LLM decoding based on compression: - 3x🚀than vanilla - 2x🚀 than Lookahead (on its benchmark) - 1.6x🚀 than Medusa (on its benchmark) - provably maintains text distribution - trainable (in 1~2 days) and testable on RTX 3090s Playground: Blog: Code: ⚒️First Principle: Compression! Yi Ma We find that the sequence of second-top-layer features is compressible, making the prediction of subsequent feature vectors from previous ones easy by a small model. 🙏Acknowledge: This project is greatly inspired by the Medusa team (Tianle Cai @yli3521 Zhengyang Geng Hongwu Peng Tri Dao), the Lookahead team (Hao Zhang LMSYS Org), and others. Joint work with Yuhui Li and Chao Zhang

Hongyang Zhang

118,857 views • 2 years ago