쉐도잉 연습: SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer, [ICLR 2026, Oral] - 영상으로 영어 말하기 배우기

연습 모드 선택

각 모드는 전체 화면 연습실로 열립니다.

125 문장

문장이 너무 짧거나 길면 Edit를 눌러 조정하세요.

Thank you, Luke, for the amazing opening.

So good morning and thanks for having me.

My name is Jun Song.

I'm from the University of Hong Kong and NVIDIA.

It's really my great honor to present this work, SANA Video, for our SANA team.

So currently, generative AI is evolving rapidly with model sizes growing larger and larger.

At At the same time, AVIDIA GPUs are becoming increasingly powerful.

We can foresee that the future visual generative models will demand even more computational resources.

So actually, generating images and videos are significantly more computationally demanding than processing just text.

For example, the inference for Flux or CogVideoX exceed the Lama 3,

and an order of magnitude higher than the VIT models.

So they are restricted to cloud deployment and remain impractical for consumer GPUs or age devices.

So actually, efficiency is a forever topic.

So far, we have explored several directions for efficient video generation.

First, in Sanaa, we tried to reduce input tokens by using our self -developed deep compression autoencoder, DCAE.

We also try to use linear tension to replace Softmax 1, enhance the architectural efficiency of the model.

And we also already explored inference step distillation in Sanaa's spring project.

The fewer the inference steps, the faster the speed.

So the last part is what we are going to talk about today, the hybrid AR plus diffusion for long video generation.

Actually, we can generate video block by block causally and diffuse within the block.

So video generation requires much more computation than images.

SANA can generate one image within just one second.

However, generating a five -second video with one model actually takes nearly 30 minutes.

So to boost the video generation efficiency, we focus on three key technical areas, the high ratio VAEs and leaning attention mechanisms,

especially optimized for video tasks.

Crucially, the block KV cache based on leaning attention enables the generation of much longer videos.

So to get a video generation model, actually, we have two ways to go to try to choose.

First, we can just train from scratch.

But only when we have hundreds or thousands of GPUs,

like H100 GPUs, and it will take you almost several months to train one model.

And it's, of course, not affordable.

Or actually, we can start from a semantic aligned text -to -image model.

Here we start from our SANA 1 .5, which is a 1 .6 billion parameter text -to -image model trained on our DCAE.

Actually, since our video model is based on image model, so the linear tension DIT actually is almost the same as our SANA,

but we're introducing two modules for temporal modeling, and we will talk about them later.

Finally, on the left at the top, we implement block causal linear tension module to achieve constant kv cache memory.

So for any length video generation, So first, let's talk about linear attention in video generation model.

Actually, in video tasks, linear attention provides even more significant acceleration than in images,

as the video length and resolution are much larger.

Our results show that for long videos, SANA video is 1 .5 times faster than the one model.

And additionally, at high resolution, our linear attention is four times faster than the flash attention.

And these improvements make a long form, high resolution video generation much more practical.

However, we find that the linear attention map is far different from the softmax attention.

As you can see in this figure, the linear attention is super dense, while the softmax attention is rather sparse.

And this demonstrates that the local contacts captured by the linear tension is actually insufficient.

To enhance the temporal modeling, we introduced two key improvements.

First is the 3D rope for the linear tension and temporal convolutions.

And both are highlighted in the orange box.

While 3D rope doesn't solve the tension map sparsity problem entirely, it do help improve attention low quality.

Combined with our spatial temporal mix FFN, it significantly boosts the overall performance.

Here we have the training loss.

The green line stands for the 3D rope and the additional temporal convolution.

And to further support long video generation, we leverage the unique properties of linear attention to implement an efficient block KV cache.

So let's quickly review how this works.

In linear attention, we call it the attention map, or the state.

Actually, it has a fixed shape of D times D, regardless of the input sequence length.

So when processing the first block here, we obtain the state, S1.

And for the second block, the state is updated additably,

meaning that we only need to add the current block's contribution to the previous cache.

And generally for any block I,

actually the current state is updated as adding the current block contribution the KV here to the previous cache to ask.

So now let's see how we enable long video generation with block of causal linear attention.

For the first block here, the green one, we compute the state as one, shown in the green box.

After we compute first block the state, we cache it, it's shown in the red box.

Then we calculate the KV for the next block and add it to the previous one.

Here we take the cached state from block one in the red box and updates it into the block two, which is in the green color.

And the subsequent process follows the same logic.

The new state is cached, and once the next block competition is finished, they are summed together. And so on.

So, token in Valenian attention will become larger, more and more as the video length become larger and larger.

While our causal linear tension maintains a constant KV cache potentially make minute length video generation possible.

So here in conclusion, we can have two advantages of block linear tension.

First, we have constant memory usage regardless of the input sequence length and second we have the global history context.

So one last question is how to achieve causal FFN with temporal conf.

So let's take two block situation as example.

During causal training the block one cannot foresee block two so we pad a zero token here between the two blocks.

However block two must see block one's token so that it's correct the causal modeling.

So here we catch the last token from block one as the first token in block two.

So finally the two blocks are separated by the zero padding token

which is subsequently removed to to avoid affecting the final computation.

And the block two used the last token from block one making the the causal modeling available.

So here's some results, compared to the current state -of -the -art video models, Sonar Video performs competitively on VBench,

and it takes only 18 seconds to generate a 1K resolution video on a single H100 GPU.

Here we show some examples.

On the left we have text to video examples, And on the right, we show the image to video examples.

And using SANA video as a base model, it can be applied to tasks like physical AI and real -world models.

Actually, by fine -tuning SANA video on a small data site with just a few safety steps,

we enable it to generate videos for robotic manipulation and gameplay play.

We also combine with the post -training tech in long lab, which is a three long test long method.

And this method is also accurate poster and is also from our group.

We also implement the DMD step distillation to further accelerate our model.

And finally, Sonar Video you can also generate minute -length video within just 36 seconds,

achieving 27 FPS real -time generation.

So, finally, there is one more thing.

Based on linear attention, we have much more potential applications to discover.

We believe the efficiency of linear attention

and deep compression auto encoder will facilitate streaming video generation

and also the development of long form like minute -length or even hour -length, real -world models.

And we are working on this project and we'll see if we can keep pushing the efficiency boundary.

And all of our projects, including the code and checkpoints,

100

are all open -source to encourage further exploration in efficient generative AI system.

101

So that's all for my talk today.

102

And we will have a poster session right after this aural, maybe in the afternoon.

103

Welcome everyone to discuss more.

104

Thanks so much.

105

Thank you.

106

Time for a quick question, if any.

107

Yeah, please go to the mic.

108

Thank you for your speech, and I have a question.

109

I think it's something similar to the auto -aggressive video diffusion models like cause of it

110

and something similar to the block causal linear attention you introduced.

111

So I wanted to ask what are the key differences and if we are thinking about infinite long video generation,

112

which one do you think would be better choice?

113

Thank you.

114

Thank you for your question.

115

Actually the cause V something like that, they actually perform on soft max extension.

116

So as I talked before, the soft max extension has an open square computation complexity.

117

So if you want to generate mini lens or even hour lens video, it's not possible to do it on soft max extension.

118

So if you want to generate mini lens video, so you have to keep the computation within like a sliding window or something like that.

119

So the main difference here is our linear tension has the global context.

120

It has memory from the first second to the final second.

121

So if you want to do some project or task like world model,

122

it has to be maintaining overall all the cache or the memory.

123

So I think that's the main advantages of getting attention.

124

Yeah, thank you.

125

Thank you.

앱 다운로드

유창하게 말하는 데 필요한 모든 것

AI 발음 평가문장마다 점수 확인

IPA 발음 연습모든 소리 마스터

단어 학습나만의 단어장 만들기

단어 게임놀면서 외우기

QR 스캔

Download on theApp Store Get it onGoogle Play

✨ 추천 영상

33 문장

Enhancing AV Simulation with Generative AI - DRIVE Labs Ep. 36

Daily ConversationUniversity StudentsWorking Adults

321 문장

Mixture of Experts Explained Visually: How Trillion-Parameter Models Actually Work

Daily ConversationUniversity Students

747 문장

이 레슨의 어휘와 말하기 포인트

이 말하기 레슨은 영상 “SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer,”을(를) 바탕으로 합니다. 가장 자주 반복되는 단어는 다음과 같습니다: block, attention, model, linear, generation. 이 영상에는 섀도잉할 문장 125개와 단어 1634개가 있습니다. 말하는 구간의 길이는 12:46입니다. 화자는 분당 약 128단어의 일정한 속도로 말해서 섀도잉하기에 편한 속도입니다. 영어에서 가장 많이 쓰이는 3,000단어에 속하는 단어가 80%뿐이라 어휘가 어려운 편입니다.

이 영상의 핵심 어휘

영상에 나오는 익혀 둘 만한 단어 15개를 발음, 뜻과 함께 정리했습니다.

단어	발음	뜻
cache 명사	/kæʃ/	캐시
tension 명사	/ˈtɛnʃən/	긴장, 불안
token 명사	/ˈtoʊkən/	토큰
efficient 형용사	/ɪˈfɪʃənt/	효율적인
introduce 동사	/ˌɪntɹəˈdus/	소개(紹介)하다
lens 명사	/lɛnz/	렌즈
rope 명사	/ɹoʊp/	밧줄, 로프
sequence 명사	/ˈsiː.kwəns/	순서
contribution 명사	/ˌkɑntɹɪˈbjuʃən/	기여, 공헌
explore 동사	/ɪkˈsploɹ/	탐험하다
poster 명사	/ˈpoʊstɚ/	포스터
lean 동사	/ˈliːn/	기울다
combine 동사	/kəmˈbaɪn/	조합하다, 결합하다
automobile 명사	/ˈɔː.tə.məˌbil/	자동차, 차
module 명사	/ˈmɑd͡ʒul/	모듈

이 영상의 문법

화자가 가장 많이 쓰는 문형을 영상 속 실제 표현과 함께 정리했습니다.

문형	영상 속 표현
수동태 be + 과거분사 — 누가 하는지보다 무슨 일이 일어나는지에 초점	are restricted · is based · are highlighted
관계절 who / which + 절 — 사람이나 사물에 대한 추가 정보	two, which is · lab, which is

주의할 발음

화자는 doesn't, I'm, we'll 같은 축약형과 약화된 형태를 4번 사용합니다. 들리는 대로 짧게 발음하세요.

“sh”와 “zh” 소리: cache /kæʃ/, tension /ˈtɛnʃən/, efficiency /ɪˈfɪʃ.ən.si/, extension /ɪkˈstɛnʃən/, efficient /ɪˈfɪʃənt/
긴 단어 — 강세 위치에 주의: efficiency /ɪˈfɪʃ.ən.si/, contribution /ˌkɑntɹɪˈbjuʃən/, automobile /ˈɔː.tə.məˌbil/, potentially /pəˈtɛnʃ(ə)li/, mechanism /ˈmɛk.ə.nɪ.zəm/

한국어 화자가 어려워하는 소리:

/f/ — ㅍ(/p/)으로 바꾸지 말고 윗니를 아랫입술에 대기: efficiency /ɪˈfɪʃ.ən.si/, efficient /ɪˈfɪʃənt/, flash /flæʃ/, fewer /ˈfjuː.ɚ/, affordable /əˈfɔɹ.də.bəl/
/v/ — ㅂ(/b/)과 구별하기: solve /sɒlv/, discover /dɪˈskʌvə/, aggressive /əˈɡɹɛs.ɪv/, leverage /ˈlɛv.(ə.)ɹɪd͡ʒ/, evolve /ɪˈvɑlv/
/z/ — ㅈ이 아니라 성대를 울리는 /s/: lens /lɛnz/, mechanism /ˈmɛk.ə.nɪ.zəm/, causal /ˈkɔːzəl/, diffuse /dɪˈfjuz/, optimize /ˈɑptɪmaɪz/

이 영상으로 연습하는 방법

먼저 말하지 않고 영상을 끝까지 듣고 모르는 단어를 적어 둡니다.
보통 속도로 한 문장씩 섀도잉하고, 화자의 리듬과 맞을 때까지 반복합니다.
자신의 목소리를 녹음해 원본과 비교하고, cache, tension, token 같은 단어에 특히 주의합니다.

쉐도잉이란? 영어 실력을 빠르게 키우는 과학적 방법

쉐도잉(Shadowing)은 원래 전문 통역사 훈련을 위해 개발된 언어 학습 기법으로, 다언어 학자인 Dr. Alexander Arguelles에 의해 대중화된 방법입니다. 핵심 원리는 간단하지만 매우 강력합니다: 원어민의 영어를 들으면서 1~2초의 짧은 지연으로 즉시 소리 내어 따라 말하는 것——마치 '그림자(shadow)'처럼 화자를 따라가는 것입니다. 문법 공부나 수동적인 청취와 달리, 쉐도잉은 뇌와 입 근육이 동시에 실시간으로 영어를 처리하고 재현하도록 훈련합니다. 연구에 따르면 이 방법은 발음 정확도, 억양, 리듬, 연음, 청취력, 말하기 유창성을 크게 향상시킵니다. IELTS 스피킹 준비와 자연스러운 영어 소통을 원하는 분들에게 특히 효과적입니다.

섀도잉 방법: 단계별 전체 가이드 읽기 →

쉐도잉 연습: SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer, [ICLR 2026, Oral] - 영상으로 영어 말하기 배우기

연습 모드 선택

앱 다운로드

Enhancing AV Simulation with Generative AI - DRIVE Labs Ep. 36

Mixture of Experts Explained Visually: How Trillion-Parameter Models Actually Work

MIT Introduction to Deep Learning | 6.S191

Why Memory Prices Are Crazy High

Controlled Multi-modal Image Generation for Plant Growth Modeling

Trust the Process: How to Design Teams That Actually Work

Neural Radiance Fields in 2 Minutes

How I Learned Procedural Generation

The future unfolds! How to optimize Android apps for adaptive layouts

LangChain in 5 Minutes (Explained Clearly)

What AI can (and can't) currently do in a real lab | Jennifer Doudna

NVIDIA GTC 2025 - Keynote Highlights by NVIDIA CEO Jensen Huang

이 레슨의 어휘와 말하기 포인트

이 영상의 핵심 어휘

이 영상의 문법

주의할 발음

이 영상으로 연습하는 방법

쉐도잉이란? 영어 실력을 빠르게 키우는 과학적 방법