T
traeai
RSS登录
返回首页
AI Engineer

Building Generative Image & Video models at Scale - Sander Dieleman (Veo and Nano Banana)

8.5Score
用这条生成生成视频方案
Building Generative Image & Video models at Scale - Sander Dieleman (Veo and Nano Banana)
AI 深度提炼
  • 探讨了生成模型在扩展性上的关键挑战与解决方案。
  • 分享了训练超大规模生成模型的硬件与软件优化策略。
  • 提供了对生成式 AI 在实际应用中的未来展望。
#AI#生成模型#深度学习
打开原文

Building Generative Image & Video models at Scale - Sander Dieleman (Veo and Nano Banana) - YouTube

Back ![Image 1](http://www.youtube.com/ "YouTube Home")

Skip navigation

Search

Search with your voice

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)

Sign in

![Image 2](http://www.youtube.com/ "YouTube Home")

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)

Building Generative Image & Video models at Scale - Sander Dieleman (Veo and Nano Banana)

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)

Tap to unmute

2x

![Image 3](http://www.youtube.com/watch?v=xOP1PM8fwnk)

Building Generative Image & Video models at Scale - Sander Dieleman (Veo and Nano Banana)

AI Engineer 1,795 views 14 hours ago

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)

Search

Copy link

Info

Shopping

Image 4

![Image 5](http://www.youtube.com/watch?v=xOP1PM8fwnk)

If playback doesn't begin shortly, try restarting your device.

•

You're signed out

Videos you watch may be added to the TV's watch history and influence TV recommendations. To avoid this, cancel and sign in to YouTube on your computer.

Cancel Confirm

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)

Share

[](http://www.youtube.com/watch?v=xOP1PM8fwnk "Share link")- [x] Include playlist

An error occurred while retrieving sharing information. Please try again later.

Image 6

0:00

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)[](https://www.youtube.com/watch?v=wjk0ulMAkbc "Next (SHIFT+n)")

0:00 / 0:00

Live

•Watch full video

•

•

20:37 How Google DeepMind is researching the next Frontier of AI for Gemini —Raia Hadsell, VP of Research AI Engineer 67K views • 3 days ago Live Playlist ()Mix (50+)29:17 Taste & Craft: A Conversation with Tuomas Artman, CTO Linear & Gergely Orosz, @The Pragmatic Engineer AI Engineer 2.4K views • 19 hours ago Live Playlist ()Mix (50+)1:01:41 This is not the AI we were promised | The Royal Society The Royal Society 157K views • Streamed 2 months ago Live Playlist ()Mix (50+)12:17 Shrimp Forever | Trump: I Would Have Won Vietnam | Jet Fuel Running Low The Late Show with Stephen Colbert 641K views • 6 hours ago Live Playlist ()Mix (50+)1:45:00 Yann LeCun: Special Lecture on AI and World Models Al-Khwarizmi Applied Mathematics Webinar 14K views • 7 days ago Live Playlist ()Mix (50+)18:46 The Future of MCP — David Soria Parra, Anthropic AI Engineer 91K views • 2 days ago Live Playlist ()Mix (50+)1:08:22 Ken Goldberg: Good Old-Fashioned Engineering Can Close the 100,000 Year “Data Gap” in Robotics MIT Robotics 19K views • 2 weeks ago Live Playlist ()Mix (50+)14:11 Trump Extends Iran Ceasefire Again, Reads from Bible & Tucker Carlson Apologizes for Supporting Him Jimmy Kimmel Live 1M views • 6 hours ago Live Playlist ()Mix (50+)26:46 Everything We Got Wrong About Research-Plan-Implement - Dexter Horthy MLOps.community 90K views • 4 weeks ago Live Playlist ()Mix (50+)15:26 Gemma, DeepMind's Family of Open Models —Omar Sanseviero, Google DeepMind AI Engineer 15K views • 1 day ago Live Playlist ()Mix (50+)39:58 Ex-Google Exec: How to Position Yourself Now Before the Next AI Phase (2026–2027) | Mo Gawdat Silicon Valley Girl and Mo Gawdat 573K views • 3 weeks ago Live Playlist ()Mix (50+)46:21 Harness Engineering: How to Build Software When Humans Steer, Agents Execute —Ryan Lopopolo, OpenAI AI Engineer 49K views • 5 days ago Live Playlist ()Mix (50+)

Sign in to confirm you’re not a bot This helps protect our community. Learn more

Sign in

Building Generative Image & Video models at Scale - Sander Dieleman (Veo and Nano Banana)

![Image 7](http://www.youtube.com/@aiDotEngineer)

AI Engineer

AI Engineer

411K subscribers

Subscribe

Subscribed

70

Share

Save

Download

Download

1,795 views 14 hours ago

1,795 views • Apr 21, 2026

https://sander.ai/2025/04/15/latents.... Speaker info:…...more

...more

How this was made

Auto-dubbed

Audio tracks for some languages were automatically generated. Learn more

Transcript

Follow along using the transcript.

Show transcript

![Image 8 ### AI Engineer 411K subscribers](http://www.youtube.com/@aiDotEngineer)

VideosAbout

Show less

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)

Building Generative Image & Video models at Scale - Sander Dieleman (Veo and Nano Banana)

1,795 views

Apr 21, 2026

70

Share

Save

Download

Download

0 Comments

Sort comments

Sort by

Top Show featured commentsNewest Show recent comments, including potential spam

Image 9: Default profile photo

Add a comment...

Comments

Top Show featured commentsNewest Show recent comments, including potential spam

In this video

Transcript

Description

Building Generative Image & Video models at Scale - Sander Dieleman (Veo and Nano Banana)

70 Likes

1,795 Views

14h Ago

https://sander.ai/2025/04/15/latents.... Speaker info:

  • https://sander.ai/
  • https://github.com/benanne
  • ![Image 10/sanderdieleman](https://www.youtube.com/redirect?event=video_description&redir_token=QUFFLUhqbXRCbjl4eVhscVdFcGY3RktTU2dWSUFqSlNPQXxBQ3Jtc0ttSjV4Z083cGxBRjlET3hXb3JBeV9SbGVERVRxZnd2dmJ3Q19CanYwWE9rbm1LbDgtblYwRWN2c0NBYV91MFlObjBlMzBMMG1YTXc2TG4yakh4YktKVGNhamN5WXY5MXRxVGo1YXdfSWRZd1dGRV9Vaw&q=https%3A%2F%2Fwww.linkedin.com%2Fin%2Fsanderdieleman&v=xOP1PM8fwnk)
  • https://x.com/sedielem

Timestamps 0:00 Introduction 2:55 Data Curation 4:02 Representation 9:39 Modeling: Diffusion Mechanism 20:01 Network Architecture 22:25 Training at Scale 23:33 Sampling & Guidance 28:03 Distillation 30:03 Control Signals…...more

...more Show less

How this was made

Auto-dubbed

Audio tracks for some languages were automatically generated. Learn more

Transcript

Follow along using the transcript.

Show transcript

![Image 11 ### AI Engineer 411K subscribers](http://www.youtube.com/@aiDotEngineer)

VideosAbout

Transcript

NaN / NaN

![Image 12 20:37](http://www.youtube.com/watch?v=zZsTVBXcbow&pp=ugUEEgJlbg%3D%3D)

Image 13

[How Google DeepMind is researching the next Frontier of AI for Gemini —Raia Hadsell, VP of Research](http://www.youtube.com/watch?v=zZsTVBXcbow&pp=ugUEEgJlbg%3D%3D)

AI Engineer

67K views • 3 days ago

New

![Image 14 29:17](http://www.youtube.com/watch?v=wjk0ulMAkbc)

Image 15

[Taste & Craft: A Conversation with Tuomas Artman, CTO Linear & Gergely Orosz, @The Pragmatic Engineer](http://www.youtube.com/watch?v=wjk0ulMAkbc)

AI Engineer

2.4K views • 19 hours ago

New

![Image 16 1:01:41](http://www.youtube.com/watch?v=CyyL0yDhr7I)

Image 17

[This is not the AI we were promised | The Royal Society](http://www.youtube.com/watch?v=CyyL0yDhr7I)

The Royal Society

157K views • Streamed 2 months ago

![Image 18 12:17](http://www.youtube.com/watch?v=b4LiQ-qhOcM)

Image 19

[Shrimp Forever | Trump: I Would Have Won Vietnam | Jet Fuel Running Low](http://www.youtube.com/watch?v=b4LiQ-qhOcM)

The Late Show with Stephen Colbert

641K views • 6 hours ago

New

![Image 20 1:45:00](http://www.youtube.com/watch?v=vJKC31YpA8c)

Image 21

[Yann LeCun: Special Lecture on AI and World Models](http://www.youtube.com/watch?v=vJKC31YpA8c)

Al-Khwarizmi Applied Mathematics Webinar

14K views • 7 days ago

![Image 22 18:46](http://www.youtube.com/watch?v=v3Fr2JR47KA&pp=ugUEEgJlbg%3D%3D)

Image 23

[The Future of MCP — David Soria Parra, Anthropic](http://www.youtube.com/watch?v=v3Fr2JR47KA&pp=ugUEEgJlbg%3D%3D)

AI Engineer

91K views • 2 days ago

New

![Image 24 1:08:22](http://www.youtube.com/watch?v=EtK5es0jVM4&pp=0gcJCcMKAYcqIYzv)

Image 25

[Ken Goldberg: Good Old-Fashioned Engineering Can Close the 100,000 Year “Data Gap” in Robotics](http://www.youtube.com/watch?v=EtK5es0jVM4&pp=0gcJCcMKAYcqIYzv)

MIT Robotics

19K views • 2 weeks ago

![Image 26 14:11](http://www.youtube.com/watch?v=TS6WWmzcCpA&pp=ugUHEgVlbi1VUw%3D%3D)

Image 27

[Trump Extends Iran Ceasefire Again, Reads from Bible & Tucker Carlson Apologizes for Supporting Him](http://www.youtube.com/watch?v=TS6WWmzcCpA&pp=ugUHEgVlbi1VUw%3D%3D)

Jimmy Kimmel Live

1M views • 6 hours ago

New

![Image 28 26:46](http://www.youtube.com/watch?v=YwZR6tc7qYg&pp=ugUEEgJlbg%3D%3D)

Image 29

[Everything We Got Wrong About Research-Plan-Implement - Dexter Horthy](http://www.youtube.com/watch?v=YwZR6tc7qYg&pp=ugUEEgJlbg%3D%3D)

MLOps.community

90K views • 4 weeks ago

![Image 30 15:26](http://www.youtube.com/watch?v=_gVFUEdhCyI&pp=ugUEEgJlbg%3D%3D)

Image 31

[Gemma, DeepMind's Family of Open Models —Omar Sanseviero, Google DeepMind](http://www.youtube.com/watch?v=_gVFUEdhCyI&pp=ugUEEgJlbg%3D%3D)

AI Engineer

15K views • 1 day ago

New

![Image 32 39:58](http://www.youtube.com/watch?v=E0Q96IKXx6Q)

Image 33
Image 34

[Ex-Google Exec: How to Position Yourself Now Before the Next AI Phase (2026–2027) | Mo Gawdat](http://www.youtube.com/watch?v=E0Q96IKXx6Q)

Silicon Valley Girl and Mo Gawdat

573K views • 3 weeks ago

![Image 35 46:21](http://www.youtube.com/watch?v=am_oeAoUhew&pp=ugUEEgJlbg%3D%3D)

Image 36

[Harness Engineering: How to Build Software When Humans Steer, Agents Execute —Ryan Lopopolo, OpenAI](http://www.youtube.com/watch?v=am_oeAoUhew&pp=ugUEEgJlbg%3D%3D)

AI Engineer

49K views • 5 days ago

New

![Image 37 25:18](http://www.youtube.com/watch?v=twLoOT3H9SY&pp=ugUHEgVlbi1VUw%3D%3D)

Image 38

[The Decline of Trump Presidency and Revisioning of the United States | Explainer](http://www.youtube.com/watch?v=twLoOT3H9SY&pp=ugUHEgVlbi1VUw%3D%3D)

Heather Cox Richardson

17K views • 5 hours ago

New

![Image 39 18:52](http://www.youtube.com/watch?v=XKup1pj-34M&pp=ugUEEgJlbg%3D%3D)

Image 40

[The New Application Layer - Malte Ubl, CTO Vercel](http://www.youtube.com/watch?v=XKup1pj-34M&pp=ugUEEgJlbg%3D%3D)

AI Engineer

11K views • 1 day ago

New

![Image 41 32:23](http://www.youtube.com/watch?v=SSya123u9Yk&pp=ugUHEgVlbi1VUw%3D%3D)

Image 42

[Demis Hassabis: Why AGI is Bigger than the Industrial Revolution & Where Are The Bottlenecks in AI](http://www.youtube.com/watch?v=SSya123u9Yk&pp=ugUHEgVlbi1VUw%3D%3D)

20VC with Harry Stebbings

264K views • 2 weeks ago

![Image 43 18:38](http://www.youtube.com/watch?v=_Zcw_sVF6hU&pp=ugUEEgJlbg%3D%3D)

Image 44

[The Friction is Your Judgment — Armin Ronacher & Cristina Poncela Cubeiro, Earendil](http://www.youtube.com/watch?v=_Zcw_sVF6hU&pp=ugUEEgJlbg%3D%3D)

AI Engineer

17K views • 3 days ago

New

![Image 45 1:02:33](http://www.youtube.com/watch?v=N5JDzS9MQYI)

Image 46
Image 47
Image 48

[Anthropic's CEO: ‘We Don’t Know if the Models Are Conscious’ | Interesting Times with Ross Douthat](http://www.youtube.com/watch?v=N5JDzS9MQYI)

Interesting Times with Ross Douthat and 2 more

436K views • 2 months ago

![Image 49 1:07:09](http://www.youtube.com/watch?v=21EYKqUsPfg)

Image 50

[Richard Sutton – Father of RL thinks LLMs are a dead end](http://www.youtube.com/watch?v=21EYKqUsPfg)

Dwarkesh Patel

712K views • 6 months ago

![Image 51 1:57:03](http://www.youtube.com/watch?v=mYSRn6PC1mc&pp=ugUEEgJlbg%3D%3D)

Image 52

[Full Workshop: Build Your Own Deep Research Agents - Louis-François Bouchard, Paul Iusztin, Samridhi](http://www.youtube.com/watch?v=mYSRn6PC1mc&pp=ugUEEgJlbg%3D%3D)

AI Engineer

9.9K views • 1 day ago

New

![Image 53 1:22:56](http://www.youtube.com/watch?v=8SF_h3xF3cE)

Image 54

[Practical Deep Learning for Coders: Lesson 1](http://www.youtube.com/watch?v=8SF_h3xF3cE)

Jeremy Howard

541K views • 3 years ago

Show more

[](http://www.youtube.com/watch?v=xOP1PM8fwnk)