{"id":118312,"date":"2026-09-17T01:49:43","date_gmt":"2026-09-17T01:49:43","guid":{"rendered":"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/"},"modified":"2026-09-17T01:49:43","modified_gmt":"2026-09-17T01:49:43","slug":"nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers","status":"publish","type":"post","link":"https:\/\/youzum.net\/zh\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/","title":{"rendered":"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers"},"content":{"rendered":"<p class=\"wp-block-paragraph\">Nunchux AI has released <a href=\"https:\/\/arxiv.org\/abs\/2609.15810\">VC-Attention<\/a>, a training-free low-bit attention kernel built for video Diffusion Transformers (DiTs). It targets 2 problems at once: value quantization error and a slow softmax stage.<\/p>\n<h2 class=\"wp-block-heading\"><strong>Why Attention is the Video Bottleneck<\/strong><\/h2>\n<p class=\"wp-block-paragraph\">Video DiTs flatten a clip into 1 sequence of spatiotemporal tokens and run full self-attention at every layer. A 5-second 720p Wan2.2-14B clip spans about 70K tokens. On the RTX 5090, attention takes more than 64% of generation time. The research team states that attention is about two thirds of every MiniMax-H3 denoising step on a single B200.<\/p>\n<p class=\"wp-block-paragraph\">Low-bit Tensor Cores speed up the 2 matrix products, QK and PV. 2 obstacles remain. First, prior methods like <a href=\"https:\/\/arxiv.org\/abs\/2411.10958\">SageAttention2<\/a> smooth queries and keys. After QK smoothing and rotation, the value term accounts for 82% of output error on Wan2.2. Second, the softmax between the products still runs in FP32. On B200 and H200, that exponential and its FP8 cast become the longest pipeline stage.<\/p>\n<h2 class=\"wp-block-heading\"><strong>V-Smooth: Fixing Value Outliers<\/strong><\/h2>\n<p class=\"wp-block-paragraph\">Value outliers sit in a few tokens, and their channels shift across heads, layers, and steps. A Hadamard rotation preserves token norms, so it does not remove them. Rotating V changes value error by just 0.2%.<\/p>\n<p class=\"wp-block-paragraph\"><strong>V-Smooth takes a different route:<\/strong><\/p>\n<ul class=\"wp-block-list\">\n<li><strong>Group:<\/strong> An online k-means clusters value tokens per batch and head. Keys and values are permuted together, so non-causal attention output is unchanged.<\/li>\n<li><strong>Demean:<\/strong> Each 128-token hardware block subtracts its mean. Only the residual is quantized, using per-channel E4M3 at 8 bits or NVFP4 at 4 bits.<\/li>\n<li><strong>Restore:<\/strong> The mean is added back using the row sum online softmax already keeps. No second pass or extra buffer is needed.<\/li>\n<\/ul>\n<p class=\"wp-block-paragraph\">Averaged over 100 Wan2.2 heads, the block mean removes 8% of block energy in sequence order. It removes 12% under DeltaQuant\u2019s static cube and 36% after sorting. Each mean costs 0.125 bit per value element.<\/p>\n<p class=\"wp-block-paragraph\">Grouping runs only on the first 25% of denoising steps. The permutation is reused across 4 adjacent steps. Averaged over the full schedule, grouping costs 3 to 4% of attention time.<\/p>\n<h2 class=\"wp-block-heading\"><strong>ExpCast-FP8: Removing the Softmax Bottleneck<\/strong><\/h2>\n<p class=\"wp-block-paragraph\">An E4M3 byte is already close to a logarithm of the value it stores. Read as an integer, it equals roughly 8 log2(v) + 56. So ExpCast-FP8 writes the byte directly from the log-domain score with 1 fused multiply-add. The constant \u03b2 = -0.35 centers the leftover error, and no constant is fitted per model.<\/p>\n<p class=\"wp-block-paragraph\">The direct path writes the same byte as the FP32 exponent-then-cast path on 79.6% of each doubling. Elsewhere it lands 1 code away. The paper proves a per-row total variation bound under 3.64%, plus any underflow tail. Across 204.8K Wan2.2 attention rows, the measured average is 1.6%. ExpCast-FP8 applies only to the 8-bit kernel, since NVFP4 has no single affine log-to-code map.<\/p>\n<p class=\"wp-block-paragraph\">Hand-written CuTe\/CUDA fusion of the preprocessing chain cuts 1 V-Smooth call from 42.2 ms to 4.8 ms on B200.<\/p>\n<h2 class=\"wp-block-heading\"><strong>Explainer: How VC-Attention Works<\/strong><\/h2>\n<p><!-- VC-Attention interactive explainer | Marktechpost | paste into a Custom HTML block --><\/p>\n<div><\/div>\n<p class=\"wp-block-paragraph\">\n<h2 class=\"wp-block-heading\"><strong>Benchmarks<\/strong><\/h2>\n<\/p><p class=\"wp-block-paragraph\">Tests cover 4 open-weight video DiTs: Wan2.2-T2V-A14B, LongCat-Video, HunyuanVideo-1.5, and MiniMax-H3. Fidelity is scored against BF16 <a href=\"https:\/\/proceedings.mlsys.org\/paper_files\/paper\/2026\/hash\/ae8b0b5838ba510daff1198474e7b984-Abstract-Conference.html\">FlashAttention-4<\/a> outputs over 100 prompts.<\/p>\n<figure class=\"wp-block-table\">\n<table class=\"has-fixed-layout\">\n<thead>\n<tr>\n<th>GPU (Wan2.2)<\/th>\n<th>Precision<\/th>\n<th>Attention speedup<\/th>\n<th>End-to-end speedup<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>B200<\/td>\n<td>8-bit<\/td>\n<td>1.59\u00d7<\/td>\n<td>1.19\u00d7<\/td>\n<\/tr>\n<tr>\n<td>H200<\/td>\n<td>8-bit<\/td>\n<td>1.46\u00d7<\/td>\n<td>1.13\u00d7<\/td>\n<\/tr>\n<tr>\n<td>RTX PRO 6000<\/td>\n<td>4-bit<\/td>\n<td>2.27\u00d7<\/td>\n<td>1.36\u00d7<\/td>\n<\/tr>\n<tr>\n<td>RTX 5090<\/td>\n<td>4-bit<\/td>\n<td>3.58\u00d7<\/td>\n<td>1.70\u00d7<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/figure>\n<p class=\"wp-block-paragraph\">On B200, VC-Attention is 6.02\u00d7 faster than SageAttention2, which ships no Blackwell kernel. On H200, the gap is 1.16\u00d7. On workstation cards, 4-bit V-Smooth matches <a href=\"https:\/\/github.com\/thu-ml\/sageattention\">SageAttention3<\/a> on the RTX PRO 6000. It stays within 5% on the RTX 5090, so fidelity separates them.<\/p>\n<p class=\"wp-block-paragraph\"><strong>Fidelity results:<\/strong><\/p>\n<ul class=\"wp-block-list\">\n<li>At 8 bits, V-Smooth adds 2.3 dB PSNR over SageAttention2 on Wan2.2 and 2.8 dB on HunyuanVideo-1.5.<\/li>\n<li>Adding ExpCast-FP8 gives back 0.7 to 2.1 dB but still beats SageAttention2 on all 4 models.<\/li>\n<li>At 4 bits, V-Smooth beats SageAttention3 by 2.9 dB on Wan2.2 and 3.6 dB on LongCat-Video.<\/li>\n<li>Run training-free, <a href=\"https:\/\/arxiv.org\/abs\/2603.00040\">Attn-QAT<\/a> falls 3.4 to 6.7 dB below SageAttention2.<\/li>\n<\/ul>\n<p class=\"wp-block-paragraph\">On MiniMax-H3 at 1344\u00d7768, attention runs 1.60\u00d7 faster than BF16 FlashAttention-4 on B200. PSNR is 20.2 dB versus 19.9 dB for SageAttention2. On B300, the paper reports 1.47\u00d7 versus 1.31\u00d7 for a naive FP8 kernel. The blog chart lists 1.51\u00d7 for B300.<\/p>\n<p class=\"wp-block-paragraph\">Nunchux Attention, the company&#8217;s proprietary extension, reaches 1.91\u00d7 on B200 and 1.83\u00d7 on B300 for MiniMax-H3 attention.<\/p>\n<p class=\"wp-block-paragraph\">The method changes only per-interaction cost. So it can compose with sparse attention like <a href=\"https:\/\/arxiv.org\/abs\/2502.01776\">Sparse VideoGen<\/a> and <a href=\"https:\/\/arxiv.org\/abs\/2506.19852\">Radial Attention<\/a>, and distillation and multi-GPU execution. Nunchux says free MiniMax-H3 access is coming through its <a href=\"https:\/\/www.nunchux.ai\/waitlist\">Modelverse waitlist<\/a>.<\/p>\n<h2 class=\"wp-block-heading\"><strong>Key Takeaways<\/strong><\/h2>\n<ul class=\"wp-block-list\">\n<li>VC-Attention is training-free low-bit attention for video DiTs from Nunchux AI.<\/li>\n<li>V-Smooth clusters value tokens, then quantizes only residuals after block-mean subtraction.<\/li>\n<li>ExpCast-FP8 replaces the FP32 exponential and cast with 1 multiply-add.<\/li>\n<li>Wan2.2 attention runs 1.59\u00d7 faster on B200 and 3.58\u00d7 on RTX 5090.<\/li>\n<li>No public kernel release yet; Nunchux runs a proprietary extension in its stack.<\/li>\n<\/ul>\n<p class=\"wp-block-paragraph\">\n<hr class=\"wp-block-separator has-alpha-channel-opacity\" \/>\n<\/p><p class=\"wp-block-paragraph\">\n<\/p><p class=\"wp-block-paragraph\">Check out the\u00a0<strong><a href=\"https:\/\/arxiv.org\/pdf\/2609.15810\" target=\"_blank\" rel=\"noreferrer noopener\">Paper<\/a>\u00a0<\/strong>and\u00a0<a href=\"https:\/\/github.com\/jmiao24\/Paper2Agent\"><strong>Technical details<\/strong><\/a>. All credit goes to the researcher of this project. Also,\u00a0feel free to follow us on\u00a0<strong><a href=\"https:\/\/x.com\/intent\/follow?screen_name=marktechpost\" target=\"_blank\" rel=\"noopener\"><mark>Twitter<\/mark><\/a><\/strong>\u00a0and don\u2019t forget to join our\u00a0<strong><a href=\"https:\/\/www.reddit.com\/r\/machinelearningnews\/\" target=\"_blank\" rel=\"noopener\">150k+ML SubReddit<\/a><\/strong>\u00a0and Subscribe to\u00a0<strong><a href=\"https:\/\/magic.beehiiv.com\/v1\/f5e63dd4-5653-4f09-83e2-321a8b1ba526?email=%7B%7Bemail%7D%7D\" target=\"_blank\" rel=\"noopener\">our Newsletter<\/a><\/strong>. Wait! are you on telegram?\u00a0<strong><a href=\"https:\/\/t.me\/machinelearningresearchnews\" target=\"_blank\" rel=\"noopener\">now you can join us on telegram as well.<\/a><\/strong><\/p>\n<p class=\"wp-block-paragraph\">Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.?\u00a0<strong><a href=\"https:\/\/forms.gle\/wbash1wF6efRj8G58\" target=\"_blank\" rel=\"noopener\"><mark>Connect with us<\/mark><\/a><\/strong><\/p>\n<p>The post <a href=\"https:\/\/www.marktechpost.com\/2026\/09\/16\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/\">Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers<\/a> appeared first on <a href=\"https:\/\/www.marktechpost.com\/\">MarkTechPost<\/a>.<\/p>","protected":false},"excerpt":{"rendered":"<p>Nunchux AI has released VC-Attention, a training-free low-bit attention kernel built for video Diffusion Transformers (DiTs). It targets 2 problems at once: value quantization error and a slow softmax stage. Why Attention is the Video Bottleneck Video DiTs flatten a clip into 1 sequence of spatiotemporal tokens and run full self-attention at every layer. A 5-second 720p Wan2.2-14B clip spans about 70K tokens. On the RTX 5090, attention takes more than 64% of generation time. The research team states that attention is about two thirds of every MiniMax-H3 denoising step on a single B200. Low-bit Tensor Cores speed up the 2 matrix products, QK and PV. 2 obstacles remain. First, prior methods like SageAttention2 smooth queries and keys. After QK smoothing and rotation, the value term accounts for 82% of output error on Wan2.2. Second, the softmax between the products still runs in FP32. On B200 and H200, that exponential and its FP8 cast become the longest pipeline stage. V-Smooth: Fixing Value Outliers Value outliers sit in a few tokens, and their channels shift across heads, layers, and steps. A Hadamard rotation preserves token norms, so it does not remove them. Rotating V changes value error by just 0.2%. V-Smooth takes a different route: Group: An online k-means clusters value tokens per batch and head. Keys and values are permuted together, so non-causal attention output is unchanged. Demean: Each 128-token hardware block subtracts its mean. Only the residual is quantized, using per-channel E4M3 at 8 bits or NVFP4 at 4 bits. Restore: The mean is added back using the row sum online softmax already keeps. No second pass or extra buffer is needed. Averaged over 100 Wan2.2 heads, the block mean removes 8% of block energy in sequence order. It removes 12% under DeltaQuant\u2019s static cube and 36% after sorting. Each mean costs 0.125 bit per value element. Grouping runs only on the first 25% of denoising steps. The permutation is reused across 4 adjacent steps. Averaged over the full schedule, grouping costs 3 to 4% of attention time. ExpCast-FP8: Removing the Softmax Bottleneck An E4M3 byte is already close to a logarithm of the value it stores. Read as an integer, it equals roughly 8 log2(v) + 56. So ExpCast-FP8 writes the byte directly from the log-domain score with 1 fused multiply-add. The constant \u03b2 = -0.35 centers the leftover error, and no constant is fitted per model. The direct path writes the same byte as the FP32 exponent-then-cast path on 79.6% of each doubling. Elsewhere it lands 1 code away. The paper proves a per-row total variation bound under 3.64%, plus any underflow tail. Across 204.8K Wan2.2 attention rows, the measured average is 1.6%. ExpCast-FP8 applies only to the 8-bit kernel, since NVFP4 has no single affine log-to-code map. Hand-written CuTe\/CUDA fusion of the preprocessing chain cuts 1 V-Smooth call from 42.2 ms to 4.8 ms on B200. Explainer: How VC-Attention Works Benchmarks Tests cover 4 open-weight video DiTs: Wan2.2-T2V-A14B, LongCat-Video, HunyuanVideo-1.5, and MiniMax-H3. Fidelity is scored against BF16 FlashAttention-4 outputs over 100 prompts. GPU (Wan2.2) Precision Attention speedup End-to-end speedup B200 8-bit 1.59\u00d7 1.19\u00d7 H200 8-bit 1.46\u00d7 1.13\u00d7 RTX PRO 6000 4-bit 2.27\u00d7 1.36\u00d7 RTX 5090 4-bit 3.58\u00d7 1.70\u00d7 On B200, VC-Attention is 6.02\u00d7 faster than SageAttention2, which ships no Blackwell kernel. On H200, the gap is 1.16\u00d7. On workstation cards, 4-bit V-Smooth matches SageAttention3 on the RTX PRO 6000. It stays within 5% on the RTX 5090, so fidelity separates them. Fidelity results: At 8 bits, V-Smooth adds 2.3 dB PSNR over SageAttention2 on Wan2.2 and 2.8 dB on HunyuanVideo-1.5. Adding ExpCast-FP8 gives back 0.7 to 2.1 dB but still beats SageAttention2 on all 4 models. At 4 bits, V-Smooth beats SageAttention3 by 2.9 dB on Wan2.2 and 3.6 dB on LongCat-Video. Run training-free, Attn-QAT falls 3.4 to 6.7 dB below SageAttention2. On MiniMax-H3 at 1344\u00d7768, attention runs 1.60\u00d7 faster than BF16 FlashAttention-4 on B200. PSNR is 20.2 dB versus 19.9 dB for SageAttention2. On B300, the paper reports 1.47\u00d7 versus 1.31\u00d7 for a naive FP8 kernel. The blog chart lists 1.51\u00d7 for B300. Nunchux Attention, the company&#8217;s proprietary extension, reaches 1.91\u00d7 on B200 and 1.83\u00d7 on B300 for MiniMax-H3 attention. The method changes only per-interaction cost. So it can compose with sparse attention like Sparse VideoGen and Radial Attention, and distillation and multi-GPU execution. Nunchux says free MiniMax-H3 access is coming through its Modelverse waitlist. Key Takeaways VC-Attention is training-free low-bit attention for video DiTs from Nunchux AI. V-Smooth clusters value tokens, then quantizes only residuals after block-mean subtraction. ExpCast-FP8 replaces the FP32 exponential and cast with 1 multiply-add. Wan2.2 attention runs 1.59\u00d7 faster on B200 and 3.58\u00d7 on RTX 5090. No public kernel release yet; Nunchux runs a proprietary extension in its stack. Check out the\u00a0Paper\u00a0and\u00a0Technical details. All credit goes to the researcher of this project. Also,\u00a0feel free to follow us on\u00a0Twitter\u00a0and don\u2019t forget to join our\u00a0150k+ML SubReddit\u00a0and Subscribe to\u00a0our Newsletter. Wait! are you on telegram?\u00a0now you can join us on telegram as well. Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.?\u00a0Connect with us The post Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers appeared first on MarkTechPost.<\/p>","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"pmpro_default_level":"","site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"_pvb_checkbox_block_on_post":false,"footnotes":""},"categories":[52,5,7,1],"tags":[],"class_list":["post-118312","post","type-post","status-publish","format-standard","hentry","category-ai-club","category-committee","category-news","category-uncategorized","pmpro-has-access"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v25.3 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers - YouZum<\/title>\n<meta name=\"description\" content=\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/youzum.net\/zh\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers - YouZum\" \/>\n<meta property=\"og:description\" content=\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\" \/>\n<meta property=\"og:url\" content=\"https:\/\/youzum.net\/zh\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/\" \/>\n<meta property=\"og:site_name\" content=\"YouZum\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/DroneAssociationTH\/\" \/>\n<meta property=\"article:published_time\" content=\"2026-09-17T01:49:43+00:00\" \/>\n<meta name=\"author\" content=\"admin NU\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin NU\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"4 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/\"},\"author\":{\"name\":\"admin NU\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c\"},\"headline\":\"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers\",\"datePublished\":\"2026-09-17T01:49:43+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/\"},\"wordCount\":839,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\"},\"articleSection\":[\"AI\",\"Committee\",\"News\",\"Uncategorized\"],\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/\",\"url\":\"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/\",\"name\":\"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers - YouZum\",\"isPartOf\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#website\"},\"datePublished\":\"2026-09-17T01:49:43+00:00\",\"description\":\"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19\",\"breadcrumb\":{\"@id\":\"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/youzum.net\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/yousum.gpucore.co\/#website\",\"url\":\"https:\/\/yousum.gpucore.co\/\",\"name\":\"YouSum\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/yousum.gpucore.co\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/yousum.gpucore.co\/#organization\",\"name\":\"Drone Association Thailand\",\"url\":\"https:\/\/yousum.gpucore.co\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png\",\"contentUrl\":\"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png\",\"width\":300,\"height\":300,\"caption\":\"Drone Association Thailand\"},\"image\":{\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/DroneAssociationTH\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c\",\"name\":\"admin NU\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/yousum.gpucore.co\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png\",\"contentUrl\":\"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png\",\"caption\":\"admin NU\"},\"url\":\"https:\/\/youzum.net\/zh\/members\/adminnu\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers - YouZum","description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/youzum.net\/zh\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/","og_locale":"zh_CN","og_type":"article","og_title":"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers - YouZum","og_description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","og_url":"https:\/\/youzum.net\/zh\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/","og_site_name":"YouZum","article_publisher":"https:\/\/www.facebook.com\/DroneAssociationTH\/","article_published_time":"2026-09-17T01:49:43+00:00","author":"admin NU","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin NU","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"4 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/#article","isPartOf":{"@id":"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/"},"author":{"name":"admin NU","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c"},"headline":"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers","datePublished":"2026-09-17T01:49:43+00:00","mainEntityOfPage":{"@id":"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/"},"wordCount":839,"commentCount":0,"publisher":{"@id":"https:\/\/yousum.gpucore.co\/#organization"},"articleSection":["AI","Committee","News","Uncategorized"],"inLanguage":"zh-Hans","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/","url":"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/","name":"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers - YouZum","isPartOf":{"@id":"https:\/\/yousum.gpucore.co\/#website"},"datePublished":"2026-09-17T01:49:43+00:00","description":"\u0e01\u0e34\u0e08\u0e01\u0e23\u0e23\u0e21\u0e40\u0e01\u0e35\u0e48\u0e22\u0e27\u0e01\u0e31\u0e1a\u0e42\u0e14\u0e23\u0e19","breadcrumb":{"@id":"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/youzum.net\/nunchux-ai-introduces-vc-attention-a-training-free-low-bit-attention-kernel-that-speeds-up-video-diffusion-transformers\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/youzum.net\/"},{"@type":"ListItem","position":2,"name":"Nunchux AI Introduces VC-Attention: A Training-Free Low-Bit Attention Kernel That Speeds Up Video Diffusion Transformers"}]},{"@type":"WebSite","@id":"https:\/\/yousum.gpucore.co\/#website","url":"https:\/\/yousum.gpucore.co\/","name":"YouSum","description":"","publisher":{"@id":"https:\/\/yousum.gpucore.co\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/yousum.gpucore.co\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"zh-Hans"},{"@type":"Organization","@id":"https:\/\/yousum.gpucore.co\/#organization","name":"Drone Association Thailand","url":"https:\/\/yousum.gpucore.co\/","logo":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/","url":"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png","contentUrl":"https:\/\/youzum.net\/wp-content\/uploads\/2024\/11\/tranparent-logo.png","width":300,"height":300,"caption":"Drone Association Thailand"},"image":{"@id":"https:\/\/yousum.gpucore.co\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/DroneAssociationTH\/"]},{"@type":"Person","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/97fa48242daf3908e4d9a5f26f4a059c","name":"admin NU","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/yousum.gpucore.co\/#\/schema\/person\/image\/","url":"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png","contentUrl":"https:\/\/youzum.net\/wp-content\/uploads\/avatars\/2\/1746849356-bpfull.png","caption":"admin NU"},"url":"https:\/\/youzum.net\/zh\/members\/adminnu\/"}]}},"rttpg_featured_image_url":null,"rttpg_author":{"display_name":"admin NU","author_link":"https:\/\/youzum.net\/zh\/members\/adminnu\/"},"rttpg_comment":0,"rttpg_category":"<a href=\"https:\/\/youzum.net\/zh\/category\/ai-club\/\" rel=\"category tag\">AI<\/a> <a href=\"https:\/\/youzum.net\/zh\/category\/committee\/\" rel=\"category tag\">Committee<\/a> <a href=\"https:\/\/youzum.net\/zh\/category\/news\/\" rel=\"category tag\">News<\/a> <a href=\"https:\/\/youzum.net\/zh\/category\/uncategorized\/\" rel=\"category tag\">Uncategorized<\/a>","rttpg_excerpt":"Nunchux AI has released VC-Attention, a training-free low-bit attention kernel built for video Diffusion Transformers (DiTs). It targets 2 problems at once: value quantization error and a slow softmax stage. Why Attention is the Video Bottleneck Video DiTs flatten a clip into 1 sequence of spatiotemporal tokens and run full self-attention at every layer. A&hellip;","_links":{"self":[{"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/posts\/118312","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/comments?post=118312"}],"version-history":[{"count":0,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/posts\/118312\/revisions"}],"wp:attachment":[{"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/media?parent=118312"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/categories?post=118312"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/youzum.net\/zh\/wp-json\/wp\/v2\/tags?post=118312"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}