Loading video...

Video Failed to Load

Go Home

bro casually walks and explains 5 GPU performance optimization methods for LLMs. one of the most simple and intuitive explanations for beginners.

Hesamation's profile picture

ℏεsam

83,463 subscribers

942,602 views • 6 months ago •via X (Twitter)

GamingEducationScience & Technology
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

bro casually explains RL tuning for LLMs and the three critical components: training, inference, and environments. basically any RLVR algorithm such as GRPO comes down to this super simple concept.
1:42

bro casually explains RL tuning for LLMs and the three critical components: training, inference, and environments. basically any RLVR algorithm such as GRPO comes down to this super simple concept.

ℏεsam

102,344 views • 5 months ago

Announcing EllesmereUI: WoW's Most Intuitive UI Suite COMPLETELY FREE and Performance Optimized, EllesmereUI combines the most intuitive UI editor in WoW with extreme performance optimization and endless customization. Bring your UI to life like never before. Install:
1:25

Announcing EllesmereUI: WoW's Most Intuitive UI Suite COMPLETELY FREE and Performance Optimized, EllesmereUI combines the most intuitive UI editor in WoW with extreme performance optimization and endless customization. Bring your UI to life like never before. Install:

Ellesmere

419,716 views • 4 months ago

Evidence of the local sun and moon—sunsets and seasons explained, too. Simple explanations for repeated observations.
3:05

Evidence of the local sun and moon—sunsets and seasons explained, too. Simple explanations for repeated observations.

truthache

11,881 views • 7 months ago

Never forget when he was with the Red Sox for about 5 minutes and casually had one of the most epic moments of their historic season
0:45

Never forget when he was with the Red Sox for about 5 minutes and casually had one of the most epic moments of their historic season

Isaac

470,822 views • 2 months ago

This is one of the most clear and concise explanations about what is going with MicroStrategy and Bitcoin for those that don’t understand.
2:50

This is one of the most clear and concise explanations about what is going with MicroStrategy and Bitcoin for those that don’t understand.

The ₿itcoin Therapist

317,914 views • 1 year ago

Can LLMs invent better ways to train LLMs? At Sakana AI, we’re pioneering AI-driven methods to automate AI research and discovery. We’re excited to release DiscoPOP: a new SOTA preference optimization algorithm that was discovered and written by an LLM! Our method leverages LLMs to propose and implement new preference optimization algorithms. We then train models with those algorithms and evaluate their performance, providing feedback to the LLM. By repeating this process for multiple generations in an evolutionary loop, the LLM discovers many highly-performant and novel preference optimization objectives! Paper: GitHub: Model: We proudly collaborated with the University of Oxford (Foerster Lab for AI Research (now part of BOLD)) and Cambridge University (Mihaela van der Schaar) on this groundbreaking project. Looking ahead, we envision a future where AI-driven research reduces the need for extensive human intervention and computational resources. This will accelerate scientific discoveries and innovation, pushing the boundaries of what AI can achieve.
0:22

Can LLMs invent better ways to train LLMs? At Sakana AI, we’re pioneering AI-driven methods to automate AI research and discovery. We’re excited to release DiscoPOP: a new SOTA preference optimization algorithm that was discovered and written by an LLM! Our method leverages LLMs to propose and implement new preference optimization algorithms. We then train models with those algorithms and evaluate their performance, providing feedback to the LLM. By repeating this process for multiple generations in an evolutionary loop, the LLM discovers many highly-performant and novel preference optimization objectives! Paper: GitHub: Model: We proudly collaborated with the University of Oxford (Foerster Lab for AI Research (now part of BOLD)) and Cambridge University (Mihaela van der Schaar) on this groundbreaking project. Looking ahead, we envision a future where AI-driven research reduces the need for extensive human intervention and computational resources. This will accelerate scientific discoveries and innovation, pushing the boundaries of what AI can achieve.

Sakana AI

555,922 views • 2 years ago

Today we announce TRINDs, a dataset and benchmarking pipeline that uses synthetic personas to train and optimize performance of LLMs for tropical and infectious diseases, which are out-of-distribution for most models. Learn more →
1:21

Today we announce TRINDs, a dataset and benchmarking pipeline that uses synthetic personas to train and optimize performance of LLMs for tropical and infectious diseases, which are out-of-distribution for most models. Learn more →

Google AI

39,237 views • 1 year ago

Bro just casually showed one of the most beautiful things you’ll ever see
0:59

Bro just casually showed one of the most beautiful things you’ll ever see

One Last Shitpost

12,950,796 views • 9 months ago

you’re in a “who’s the most simple?” battle, then your opponent is phuwin😭 bro just casually walking ALONE and quietly waiting for the train after the SF event. like bro forgot he’s famous. but tbh, that’s exactly what makes him so admirable and lovable🫶🏻
0:31

you’re in a “who’s the most simple?” battle, then your opponent is phuwin😭 bro just casually walking ALONE and quietly waiting for the train after the SF event. like bro forgot he’s famous. but tbh, that’s exactly what makes him so admirable and lovable🫶🏻

phuwin.th

194,413 views • 7 months ago

enjoy beyoncé’s full performance of be alive for the oscars (one of her greatest and most underrated performances ever)
5:24

enjoy beyoncé’s full performance of be alive for the oscars (one of her greatest and most underrated performances ever)

ᶻ 𝗓 𐰁

43,377 views • 9 months ago

Ultra is the most powerful execution engine to trade onchain. Jupiter aims to provide the best performance and most seamless swap experience for users. Wei Liang explains the motivation, improvements and the growth of Ultra:
8:01

Ultra is the most powerful execution engine to trade onchain. Jupiter aims to provide the best performance and most seamless swap experience for users. Wei Liang explains the motivation, improvements and the growth of Ultra:

Jupiter

19,920 views • 5 months ago

H100s, Blackwell & the GPU Revolution GPU supply is tight and demand is surging. We aggregate thousands of H100s, A100s, L40S, MI300X & more into one on-demand pool. One click and you’re training LLMs or vision models—no sticker shock. Skip the shortage. Build on NodeAI today. ⚡️ #AI #GPU #NODEAI
0:46

H100s, Blackwell & the GPU Revolution GPU supply is tight and demand is surging. We aggregate thousands of H100s, A100s, L40S, MI300X & more into one on-demand pool. One click and you’re training LLMs or vision models—no sticker shock. Skip the shortage. Build on NodeAI today. ⚡️ #AI #GPU #NODEAI

Node AI | $GPU

12,966 views • 1 year ago

Introducing Limit Order V2 – the most advanced limit order system on Solana. Intuitive. Flexible. Private. Built for both beginners and pros. Limit V2 gives traders full control and precision – set your exact price targets, automate entries and exits, and protect positions privately. 👇
0:58

Introducing Limit Order V2 – the most advanced limit order system on Solana. Intuitive. Flexible. Private. Built for both beginners and pros. Limit V2 gives traders full control and precision – set your exact price targets, automate entries and exits, and protect positions privately. 👇

Jupiter

231,129 views • 8 months ago

The fact Aaliyah murdered this live performance and casually sat down for an interview...
5:53

The fact Aaliyah murdered this live performance and casually sat down for an interview...

Aaliyah Hub

20,216 views • 9 days ago

🔥 Today we're shipping something special at Platformatic: Next-generation flamegraph visualization for Node.js! We've built a complete profiling ecosystem that makes performance optimization intuitive and actionable. Let me show you what we've cooked up 🧵
0:27

🔥 Today we're shipping something special at Platformatic: Next-generation flamegraph visualization for Node.js! We've built a complete profiling ecosystem that makes performance optimization intuitive and actionable. Let me show you what we've cooked up 🧵

Matteo Collina

13,862 views • 10 months ago

New to OSINT and not sure where to start? The OSINT Investigator’s Toolkit is a searchable directory of free and paid OSINT tools Perfect for beginners and for discovering the core staples Simple, practical, and very handy
0:36

New to OSINT and not sure where to start? The OSINT Investigator’s Toolkit is a searchable directory of free and paid OSINT tools Perfect for beginners and for discovering the core staples Simple, practical, and very handy

The OSINT Newsletter

29,011 views • 5 months ago

Property hooks, one of the new features in PHP 8.4, can remove the need for traditional getter and setter methods entirely. Got 5 minutes for me to show you the essentials?
5:15

Property hooks, one of the new features in PHP 8.4, can remove the need for traditional getter and setter methods entirely. Got 5 minutes for me to show you the essentials?

Laracasts

36,044 views • 1 year ago