Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

NVIDIA silently released the Llama 3.1 70B Instruct Nemotron model which is topping leaderboards. You can try it here: I sat down with one of the researchers to learn how it performed so well and why they built this 🤙 Anything else I should ask him?

Nader Khalil🍊

13,173 subscribers

48,344 views • 2 years ago •via X (Twitter)

News & Politics Science & Technology

Anya Rossi• Live Now

Private livecam show

11 Comments

Isaac2 years ago

As a Nvidia shareholder, I would like to vote for Nader for CMO

Tony Kam2 years ago

Nader for Nvidia CMO!

dar2 years ago

The stache strikes again

Nader Khalil🍊2 years ago

Put this on my tombstone 😂

Ozlem Peksoy Bishop2 years ago

not so quite anymore 🤓

Erika⚡️⚡️2 years ago

When the biggest technologists in the industry go direct. Long NVIDIA!!!

Dave Font2 years ago

Can you do this with Jensen next!?

Oleksii Kuchaiev2 years ago

thanks for the great conversation! A key part of model’s recipe is hiring great interns (@AlexanderBukha1 ) who implement REINFORCE and actually train the model :)

Shadi Khalil2 years ago

🤙

Xander Dunn2 years ago

By the way, your audio here is stereo right-only. With headphones, I hear it only in my right ear.

Nader Khalil🍊2 years ago

Ahh wow thanks for the heads up. I thought I fixed this. Will fix it for the next video

Related Videos

Our Llama-3.1-Nemotron-70B-Instruct model is a leading model on the 🏆 Arena Hard benchmark (85) from Arena. Arena Hard uses a data pipeline to build high-quality benchmarks from live data in Chatbot Arena, and is known for its predictive ability of Chatbot Arena Elo score as well as separability between helpful and less helpful models. Use our customized model Llama-3.1-Nemotron-70B to improve the helpfulness of LLM generated responses in your applications. 📥 Try on our API catalog: 📥 On GitHub: 📥 Or on Hugging Face:

Our Llama-3.1-Nemotron-70B-Instruct model is a leading model on the 🏆 Arena Hard benchmark (85) from Arena. Arena Hard uses a data pipeline to build high-quality benchmarks from live data in Chatbot Arena, and is known for its predictive ability of Chatbot Arena Elo score as well as separability between helpful and less helpful models. Use our customized model Llama-3.1-Nemotron-70B to improve the helpfulness of LLM generated responses in your applications. 📥 Try on our API catalog: 📥 On GitHub: 📥 Or on Hugging Face:

NVIDIA AI Developer

140,901 views • 2 years ago

“There's two reasons why we're building Nemotron. The first is because it helps us build the GPUs and systems for the future.” The second reason is because trust is the foundation of impactful AI–and it starts with understanding. That’s why Nemotron was built open source at its core—so developers, researchers, and enterprises can build together. Hear from Bryan Catanzaro, VP of Applied Deep Learning Research at NVIDIA, on why making AI open benefits everyone. 👉

“There's two reasons why we're building Nemotron. The first is because it helps us build the GPUs and systems for the future.” The second reason is because trust is the foundation of impactful AI–and it starts with understanding. That’s why Nemotron was built open source at its core—so developers, researchers, and enterprises can build together. Hear from Bryan Catanzaro, VP of Applied Deep Learning Research at NVIDIA, on why making AI open benefits everyone. 👉

NVIDIA

64,911 views • 1 year ago

This is the fastest I've seen Llama 3.3 running anywhere! Llama 3.3 70B running at 652 t/s is lightning fast. And if you want Llama 3.1, here are the speeds I was able to get: • Llama 3.1 8B: 1006 t/s • Llama 3.1 70B: 709 t/s • Llama 3.1 405B: 206 t/s (You can access all of these models for free! See the link below.) This speed is incredible, but the interesting part is what's happening behind the scenes: These models aren't running on a GPU! In this video, I'm using the SambaNova cloud to access these models. They built a custom chip (SN40L) optimized for AI workflows. A single SN40L chip can hold hundreds of models (trillions of parameters) in memory! The speed alone is a huge deal, but the big advantage is for agentic workflows running multiple specialized models. A GPU can only host a single model and switch (unload and load) to a different model if necessary. An SN40L, on the other hand, can host every model at once, making it much faster. Here is the video where you can see how fast these chips are:

This is the fastest I've seen Llama 3.3 running anywhere! Llama 3.3 70B running at 652 t/s is lightning fast. And if you want Llama 3.1, here are the speeds I was able to get: • Llama 3.1 8B: 1006 t/s • Llama 3.1 70B: 709 t/s • Llama 3.1 405B: 206 t/s (You can access all of these models for free! See the link below.) This speed is incredible, but the interesting part is what's happening behind the scenes: These models aren't running on a GPU! In this video, I'm using the SambaNova cloud to access these models. They built a custom chip (SN40L) optimized for AI workflows. A single SN40L chip can hold hundreds of models (trillions of parameters) in memory! The speed alone is a huge deal, but the big advantage is for agentic workflows running multiple specialized models. A GPU can only host a single model and switch (unload and load) to a different model if necessary. An SN40L, on the other hand, can host every model at once, making it much faster. Here is the video where you can see how fast these chips are:

Santiago

97,561 views • 1 year ago

You can now try Llama 3.1 405B for free (link below)! This is the largest open-source model out there, and for the first time, an open model is competitive with closed models. This time around, Meta did something new: Llama 3.1 has a license that allows developers to use it to enhance other models. For the first time, you can distill Llama 3.1 405B's capabilities into a smaller, more practical model for your use case. First, here is the link where you can play with Llama 3.1 for free: The model is hosted in Tune Studio, an end-to-end platform for developing applications using Large Language Models. They are sponsoring this post. Take a look at the attached video. It will show you how you can fine-tune a simple model using Llama 3.1 without leaving the platform: 1. You can create an empty dataset 2. Use the playground to generate and record interactions with Llama 3.1 3. Modify the dataset directly using the playground 4. Export the data and fine-tune a smaller model Fast and easy! As long as you have a web browser, you can start experimenting with fine-tuning and Llama 3.1. That's all it takes!

You can now try Llama 3.1 405B for free (link below)! This is the largest open-source model out there, and for the first time, an open model is competitive with closed models. This time around, Meta did something new: Llama 3.1 has a license that allows developers to use it to enhance other models. For the first time, you can distill Llama 3.1 405B's capabilities into a smaller, more practical model for your use case. First, here is the link where you can play with Llama 3.1 for free: The model is hosted in Tune Studio, an end-to-end platform for developing applications using Large Language Models. They are sponsoring this post. Take a look at the attached video. It will show you how you can fine-tune a simple model using Llama 3.1 without leaving the platform: 1. You can create an empty dataset 2. Use the playground to generate and record interactions with Llama 3.1 3. Modify the dataset directly using the playground 4. Export the data and fine-tune a smaller model Fast and easy! As long as you have a web browser, you can start experimenting with fine-tuning and Llama 3.1. That's all it takes!

Santiago

55,609 views • 2 years ago

What is OpenClaw Enterprise, and why does it matter? McBain 🦞 sat down with Kevin S Lin to talk about: - What OpenClaw Enterprise is - Why he built it - What it does that other agents don't do - How you can get involved + Learn why Red Hat and NVIDIA are contributing🦞

What is OpenClaw Enterprise, and why does it matter? McBain 🦞 sat down with Kevin S Lin to talk about: - What OpenClaw Enterprise is - Why he built it - What it does that other agents don't do - How you can get involved + Learn why Red Hat and NVIDIA are contributing🦞

OpenClaw🦞

33,275 views • 15 days ago

so i'm at the central state museum of Kazakhstan, and they have this dombra exhibit. i notice that one of them is modern-looking and out of a case, so i ask one of the docents if i could try playing it. i sit down & start strumming away, basically playing it like a balalaika one of the other guests, a middle-aged Kazakh guy there with his family, is watching me, and i can recognize the antsy look of someone watching someone else butchering their instrument, so i hand it to him he pops off his shoes, sits down, and starts *wailing* on it after two songs, the docent is like "wait, sit here" and slides this table over to him, which has this little string attached. he slips it around his pinky, and… 🎶🐐🎶

so i'm at the central state museum of Kazakhstan, and they have this dombra exhibit. i notice that one of them is modern-looking and out of a case, so i ask one of the docents if i could try playing it. i sit down & start strumming away, basically playing it like a balalaika one of the other guests, a middle-aged Kazakh guy there with his family, is watching me, and i can recognize the antsy look of someone watching someone else butchering their instrument, so i hand it to him he pops off his shoes, sits down, and starts wailing on it after two songs, the docent is like "wait, sit here" and slides this table over to him, which has this little string attached. he slips it around his pinky, and… 🎶🐐🎶

iwsfutcmd

37,376 views • 1 year ago

Ilya and David Chalmers, the same vibe.. "Anything which I can learn, anything which any one of you can learn, the AI could do as well. The reason is that all of us have a brain, and the brain is a biological computer. So why can't the digital computer, do the same things?"

Ilya and David Chalmers, the same vibe.. "Anything which I can learn, anything which any one of you can learn, the AI could do as well. The reason is that all of us have a brain, and the brain is a biological computer. So why can't the digital computer, do the same things?"

Rohan Paul

119,333 views • 1 year ago

Head of Nemotron Bryan Catanzaro says NVIDIA is sitting on internal "teacher" models that beat everything it has publicly shipped and he's trying to release these as well "This technology of using teacher models and multi-domain on-policy distillation has been really key for us in these ways." "Now, what does that mean? Well, it definitely means that inside NVIDIA we have a collection of teacher models that are individually better at different domains than any of the models that we've released." "And actually, I'm trying to release those teacher models as part of the Nemotron project as well -- because as you know, Nemotron is not just a model." "Nemotron is an open technology effort. It includes data sets, it includes techniques, all of our ideas about how to build a model, which we publish." "Of course it includes the models themselves, but I think it should also include collections of teacher models." "I'd imagine this could be a really great way of federating open technologies for AI across the ecosystem." _______ Wrote more on a prior Bryan interview here:

Head of Nemotron Bryan Catanzaro says NVIDIA is sitting on internal "teacher" models that beat everything it has publicly shipped and he's trying to release these as well "This technology of using teacher models and multi-domain on-policy distillation has been really key for us in these ways." "Now, what does that mean? Well, it definitely means that inside NVIDIA we have a collection of teacher models that are individually better at different domains than any of the models that we've released." "And actually, I'm trying to release those teacher models as part of the Nemotron project as well -- because as you know, Nemotron is not just a model." "Nemotron is an open technology effort. It includes data sets, it includes techniques, all of our ideas about how to build a model, which we publish." "Of course it includes the models themselves, but I think it should also include collections of teacher models." "I'd imagine this could be a really great way of federating open technologies for AI across the ecosystem." _______ Wrote more on a prior Bryan interview here:

Fireside Alpha

29,125 views • 2 months ago

"Elon Musk is a genius by all accounts. His commitment and his work, like I asked him one time about how does he learn. And the amount of research, the amount of study that he does is unheard of, but he'll always say the most important thing is imagination. You can learn anything that you want to learn. You can study all these things that you have in a book. If you don't have the imagination to take it to another level, it doesn't mean anything. I sat down and spoke with Elon for a couple hours. It was really, really a joy." — Kobe Bryant

"Elon Musk is a genius by all accounts. His commitment and his work, like I asked him one time about how does he learn. And the amount of research, the amount of study that he does is unheard of, but he'll always say the most important thing is imagination. You can learn anything that you want to learn. You can study all these things that you have in a book. If you don't have the imagination to take it to another level, it doesn't mean anything. I sat down and spoke with Elon for a couple hours. It was really, really a joy." — Kobe Bryant

DogeDesigner

242,605 views • 4 months ago

There has never been a better time to be a curious mind. AI is the most patient tutor you'll ever have so you can learn anything. This morning I woke up wanting to understand how NVIDIA NVLink works, so I built a 3D explainer, from the rack down to the GPU.

There has never been a better time to be a curious mind. AI is the most patient tutor you'll ever have so you can learn anything. This morning I woke up wanting to understand how NVIDIA NVLink works, so I built a 3D explainer, from the rack down to the GPU.

Nicolas Bustamante

45,496 views • 5 days ago

“Holiness is pure motives”. This break down of holiness is one of the reasons I beg people to try and study the etymology of a word so that you can truly understand weight and depths not the vibes and emotions built around it

“Holiness is pure motives”. This break down of holiness is one of the reasons I beg people to try and study the etymology of a word so that you can truly understand weight and depths not the vibes and emotions built around it

Dr. God Abeg ooo

33,516 views • 8 months ago

One of the things I always ask people in political organizing is, “What is your why?” I sat down with Zayed Kadir, the chair of the High School Democrats of America, to hear what inspires him to do this work.

One of the things I always ask people in political organizing is, “What is your why?” I sat down with Zayed Kadir, the chair of the High School Democrats of America, to hear what inspires him to do this work.

Ken Martin

15,911 views • 1 year ago

I would like to personally thank the Carolina Panthers for using this bouncing kickoff in the preseason Here is why I think it can work and why more teams should try it out:

I would like to personally thank the Carolina Panthers for using this bouncing kickoff in the preseason Here is why I think it can work and why more teams should try it out:

Shawn Syed

374,685 views • 1 year ago

Andrew Garfield says a woman at Glastonbury asked him to take a photo of her instead of with him “I remember being at Glastonbury one year. You don’t quite realize how it’s affecting how you see the world until you get the opposite experience or you remember what it’s like to not be recognizable in some way” “And there were these two lovely girls who - it was their first Glastonbury, and I was there with my group of friends, and we just kind of sat, you know, you plop down wherever you plop down” “And then these two girls, you can feel it. You’re like, ‘Oh, I think I’m getting a little protective.’ But then we were just chatting and chatting and chatting” “And then at the end of the conversation, one of them was like, ‘Can I-’ And I thought, ‘Oh, here it comes.’ But she was like, ‘Can I ask you to take a photo of us?’” “And I was like, ‘Yeah, I would love to do that.’ It made me so happy. I was like, ‘Yeah! Now I feel like I’m just a part of your experience, and you’re a part of mine, and it’s not anything else’” “They were so cool. We had such a nice chat and they just kind of were treating me like anyone else. It was just the best. It’s the best”

Andrew Garfield says a woman at Glastonbury asked him to take a photo of her instead of with him “I remember being at Glastonbury one year. You don’t quite realize how it’s affecting how you see the world until you get the opposite experience or you remember what it’s like to not be recognizable in some way” “And there were these two lovely girls who - it was their first Glastonbury, and I was there with my group of friends, and we just kind of sat, you know, you plop down wherever you plop down” “And then these two girls, you can feel it. You’re like, ‘Oh, I think I’m getting a little protective.’ But then we were just chatting and chatting and chatting” “And then at the end of the conversation, one of them was like, ‘Can I-’ And I thought, ‘Oh, here it comes.’ But she was like, ‘Can I ask you to take a photo of us?’” “And I was like, ‘Yeah, I would love to do that.’ It made me so happy. I was like, ‘Yeah! Now I feel like I’m just a part of your experience, and you’re a part of mine, and it’s not anything else’” “They were so cool. We had such a nice chat and they just kind of were treating me like anyone else. It was just the best. It’s the best”

Casper

415,562 views • 2 months ago

A thread Mr Dad is explaining how the biogas digester is built and how it works as well... I hope I managed to ask all the questions, and you got the answers as well.

A thread Mr Dad is explaining how the biogas digester is built and how it works as well... I hope I managed to ask all the questions, and you got the answers as well.

Cing'iNkosi

10,769 views • 3 years ago

Dominion AI (Dominion AI) is a new, Christian-built alternative to mainstream AI models like ChatGPT and Claude. I had a conversation with the founders to learn about why they created it, what it can do, and who it's for. The full interview is now on X. Try Dominion here:

Dominion AI (Dominion AI) is a new, Christian-built alternative to mainstream AI models like ChatGPT and Claude. I had a conversation with the founders to learn about why they created it, what it can do, and who it's for. The full interview is now on X. Try Dominion here:

ZUBY:

18,223 views • 5 months ago

I redid the NYT blue wave map they put out, and extended it to previous years as well. You can really understand how big this wave was compared to everything else in recent history. 3.1 GB of data in 23 seconds.

I redid the NYT blue wave map they put out, and extended it to previous years as well. You can really understand how big this wave was compared to everything else in recent history. 3.1 GB of data in 23 seconds.

Louis Riel

72,323 views • 12 hours ago

I just sat down with Tom Bilyeu on his podcast. Here are the 8 things I told him about how inflation is silently hurting the average American, America's $40T debt problem and Japan's carry trade: 1) If you JUST have a salary, it will lead to poverty

I just sat down with Tom Bilyeu on his podcast. Here are the 8 things I told him about how inflation is silently hurting the average American, America's $40T debt problem and Japan's carry trade: 1) If you JUST have a salary, it will lead to poverty

Felix Nikolas Prehn 🐶

223,097 views • 1 month ago

With today’s launch of our Llama 3.1 collection of models we’re making history with the largest and most capable open source AI model ever released. 128K context length, multilingual support, and new safety tools. Download 405B and our improved 8B & 70B here.

With today’s launch of our Llama 3.1 collection of models we’re making history with the largest and most capable open source AI model ever released. 128K context length, multilingual support, and new safety tools. Download 405B and our improved 8B & 70B here.

Ahmad Al-Dahle

867,212 views • 2 years ago

I turned Llama 3.1 into my new phone assistant. It can answer anything, and look at how fast it does it using Groq’s API 🤯

I turned Llama 3.1 into my new phone assistant. It can answer anything, and look at how fast it does it using Groq’s API 🤯

Alvaro Cintas

356,398 views • 2 years ago