Загрузка видео...

Не удалось загрузить видео

Возникла проблема при загрузке этого видео. Это может быть связано с временными проблемами сети или видео может быть недоступно.

На главную

Today, we’re releasing Power Retention, a new architecture beyond Transformers. It enables LLMs to handle millions of tokens efficiently, unlocking long-context applications that were too costly before.

Manifest AI

1,126 subscribers

140,614 просмотров • 10 месяцев назад •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

HTML-in-canvas is now a first-class primitive in Remotion! It enables new types of effects that were impossible before.

HTML-in-canvas is now a first-class primitive in Remotion! It enables new types of effects that were impossible before.

Remotion

580,243 просмотров • 3 месяцев назад

I have made it my mission to give a voice to those who were ignored for too long. I promised that my government would deliver the Hillsborough Law. I’m proud that today we’re making good on that promise.

I have made it my mission to give a voice to those who were ignored for too long. I promised that my government would deliver the Hillsborough Law. I’m proud that today we’re making good on that promise.

Keir Starmer

379,236 просмотров • 28 дней назад

Announcing How Transformer LLMs Work, created with Jay Alammar and Maarten Grootendorst, co-authors of the beautifully illustrated book, “Hands-On Large Language Models.” This course offers a deep dive into the inner workings of the transformer architecture that powers large language models (LLMs). The transformer architecture revolutionized generative AI; in fact, the "GPT" in ChatGPT stands for "Generative Pre-Trained Transformer." Originally introduced in the Google Brain team's groundbreaking 2017 paper "Attention Is All You Need," by Vaswani and others, transformers were a highly scalable model for machine translation tasks. Variants of this architecture now power today’s LLMs such as those from OpenAI, Google, Meta, Cohere, Anthropic and DeepSeek. In this course, you’ll learn in detail how LLMs process text. You'll also work through code examples that illustrate that transformer's individual components. In details, you’ll learn: - How the representation of language has evolved, from Bag-of-Words to Word2Vec embeddings to the transformer architecture that captures a word's meanings taking into account the context of other words in the input. - How inputs are broken down into tokens before they are sent to the language model. - The details of a transformer's main stages: Tokenization and embedding, the stack of transformer blocks, and the language model head. - The inner workings of the transformer block, including attention, which calculates relevance scores, and the feedforward layer, which incorporates stored information learned in training. - How cached calculations make transformers faster. - Some of the most recent ideas in the latest models such as Mixture-of-Experts (MoE) which uses multiple sub-models and a router on each layer to improve the quality of LLMs. By the end of this course, you’ll have a deep understanding of how LLMs actually process text and be able to read through papers describing the latest models and understand the details. Gaining this intuition will improve your approach to building LLM applications. Please sign up here:

Announcing How Transformer LLMs Work, created with Jay Alammar and Maarten Grootendorst, co-authors of the beautifully illustrated book, “Hands-On Large Language Models.” This course offers a deep dive into the inner workings of the transformer architecture that powers large language models (LLMs). The transformer architecture revolutionized generative AI; in fact, the "GPT" in ChatGPT stands for "Generative Pre-Trained Transformer." Originally introduced in the Google Brain team's groundbreaking 2017 paper "Attention Is All You Need," by Vaswani and others, transformers were a highly scalable model for machine translation tasks. Variants of this architecture now power today’s LLMs such as those from OpenAI, Google, Meta, Cohere, Anthropic and DeepSeek. In this course, you’ll learn in detail how LLMs process text. You'll also work through code examples that illustrate that transformer's individual components. In details, you’ll learn: - How the representation of language has evolved, from Bag-of-Words to Word2Vec embeddings to the transformer architecture that captures a word's meanings taking into account the context of other words in the input. - How inputs are broken down into tokens before they are sent to the language model. - The details of a transformer's main stages: Tokenization and embedding, the stack of transformer blocks, and the language model head. - The inner workings of the transformer block, including attention, which calculates relevance scores, and the feedforward layer, which incorporates stored information learned in training. - How cached calculations make transformers faster. - Some of the most recent ideas in the latest models such as Mixture-of-Experts (MoE) which uses multiple sub-models and a router on each layer to improve the quality of LLMs. By the end of this course, you’ll have a deep understanding of how LLMs actually process text and be able to read through papers describing the latest models and understand the details. Gaining this intuition will improve your approach to building LLM applications. Please sign up here:

Andrew Ng

259,920 просмотров • 1 год назад

The CGI used in this Transformers: Revenge of the Fallen (2009) scene melted a computer because of the sheer detail — they had to store it on multiple machines. Devastator had 52,632 parts, 30+ GB of textures, and millions of polygons, too heavy for a single system to handle.

The CGI used in this Transformers: Revenge of the Fallen (2009) scene melted a computer because of the sheer detail — they had to store it on multiple machines. Devastator had 52,632 parts, 30+ GB of textures, and millions of polygons, too heavy for a single system to handle.

Movies Scenes 🎫

536,734 просмотров • 8 месяцев назад

$Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. Subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.$

Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. Subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.

Alexander Whedon

13,208,877 просмотров • 3 месяцев назад

New course on serving LLMs efficiently -- how do you serve models to many concurrent users at low latency and reasonable cost? This short course is built with Red Hat and taught by Cedric Clyburn. Efficient LLM serving requires efficient memory management. A 70B-parameter model takes ~140 GB just to load the weights. On top of that, every active request needs its own chunk of GPU memory, the KV cache, to store the token context it has built up so far. In this course, you'll learn to reduce a model's memory footprint with quantization and serve it using vLLM, which handles many concurrent requests efficiently through smart memory management. Skills you'll gain: - Quantize a model and measure the accuracy tradeoff - Serve a model with vLLM and watch it handle concurrent requests efficiently - Benchmark your deployment and make informed tradeoffs between speed, cost, and accuracy Join and learn to serve LLMs efficiently:

New course on serving LLMs efficiently -- how do you serve models to many concurrent users at low latency and reasonable cost? This short course is built with Red Hat and taught by Cedric Clyburn. Efficient LLM serving requires efficient memory management. A 70B-parameter model takes ~140 GB just to load the weights. On top of that, every active request needs its own chunk of GPU memory, the KV cache, to store the token context it has built up so far. In this course, you'll learn to reduce a model's memory footprint with quantization and serve it using vLLM, which handles many concurrent requests efficiently through smart memory management. Skills you'll gain: - Quantize a model and measure the accuracy tradeoff - Serve a model with vLLM and watch it handle concurrent requests efficiently - Benchmark your deployment and make informed tradeoffs between speed, cost, and accuracy Join and learn to serve LLMs efficiently:

Andrew Ng

130,549 просмотров • 2 месяцев назад

Sam Altman: "I bet there is another new architecture to find" Sam Altman believes we are on the verge of discovering a new underlying architecture that will be as big of a leap forward as Transformers were over LSTMs. He noted that we finally have AI models that are smart enough to help conduct this level of research (GPT 5.4 and above 👀) His direct advice to builders looking for the next major leap is to look for a "mega breakthrough" and use current models to help them find it.

Sam Altman: "I bet there is another new architecture to find" Sam Altman believes we are on the verge of discovering a new underlying architecture that will be as big of a leap forward as Transformers were over LSTMs. He noted that we finally have AI models that are smart enough to help conduct this level of research (GPT 5.4 and above 👀) His direct advice to builders looking for the next major leap is to look for a "mega breakthrough" and use current models to help them find it.

Chris

289,423 просмотров • 4 месяцев назад

Millions of verifiable data points, connected in a shared context and controlled by those who created them. 🕸️ As Phil Archer (@philarcher.bsky.social) of GS1 put it, that makes it possible to build applications on data users can trust, because they can see where it came from.

Millions of verifiable data points, connected in a shared context and controlled by those who created them. 🕸️ As Phil Archer (@philarcher.bsky.social) of GS1 put it, that makes it possible to build applications on data users can trust, because they can see where it came from.

OriginTrail

129,149 просмотров • 4 месяцев назад

wtf is MCP? Paul Klein IV, CEO of explains "Model Context Protocol" (MCP) in plain terms: "There's this new way that people are exposing tools to AI & the 'Model Context Protocol' is one of those ways that you can standardize & give your AI a set of tools to call. BrowserBase has a way to expose the browser to your agent via MCP or MCP server. More & more developers are trying to find ways to integrate more functionalities to their AI agents & BrowserBase's browsers are one of those primitive functionalities, & the MCP is just one way to integrate it into your software." Paul Klein IV From 🅱️'s GitHub: Model Context Protocol (MCP) is an open protocol that enables seamless integration between LLM applications and external data sources and tools. Whether you’re building an AI-powered IDE, enhancing a chat interface, or creating custom AI workflows, MCP provides a standardized way to connect LLMs with the context they need. This server provides cloud browser automation capabilities using Browserbase and Stagehand. This server enables LLMs to interact with web pages, take screenshots, and execute JavaScript in a cloud browser environment.

wtf is MCP? Paul Klein IV, CEO of explains "Model Context Protocol" (MCP) in plain terms: "There's this new way that people are exposing tools to AI & the 'Model Context Protocol' is one of those ways that you can standardize & give your AI a set of tools to call. BrowserBase has a way to expose the browser to your agent via MCP or MCP server. More & more developers are trying to find ways to integrate more functionalities to their AI agents & BrowserBase's browsers are one of those primitive functionalities, & the MCP is just one way to integrate it into your software." Paul Klein IV From 🅱️'s GitHub: Model Context Protocol (MCP) is an open protocol that enables seamless integration between LLM applications and external data sources and tools. Whether you’re building an AI-powered IDE, enhancing a chat interface, or creating custom AI workflows, MCP provides a standardized way to connect LLMs with the context they need. This server provides cloud browser automation capabilities using Browserbase and Stagehand. This server enables LLMs to interact with web pages, take screenshots, and execute JavaScript in a cloud browser environment.

Molly O’Shea

20,408 просмотров • 1 год назад

A new Golden Age of exploration is here and we’re building the team to lead it. NASA Force applications open today at 1pm ET. Join NASA through a 1–2 year appointment and work on the missions pushing America higher, farther, and faster than ever before.

A new Golden Age of exploration is here and we’re building the team to lead it. NASA Force applications open today at 1pm ET. Join NASA through a 1–2 year appointment and work on the missions pushing America higher, farther, and faster than ever before.

NASA Administrator Jared Isaacman

960,475 просмотров • 3 месяцев назад

New short course: Build Long-Context AI Apps with Jamba. Learn about state space models (SSMs), which have emerged as an alternative to transformers! Specifically, Jamba is a hybrid transformer-Mamba architecture that combines strengths of the transformer with ideas from SSMs. This course is built with AI21 Labs and taught by Chen Wang and Chen Almagor. The transformer architecture is computationally expensive when handling very long input contexts. But there's an alternative called Mamba, a selective state space model that can process very long contexts with a much lower computational cost. However, researchers found that the pure Mamba architecture underperforms in understanding the context, and gives lower-quality responses. To overcome this, AI21 developed the Jamba model, which combines Mamba's computational efficiency with the transformer's attention mechanism to help with the output quality. In this course, you’ll learn about how state space models, and Jamba, work. You’ll also learn how to prompt Jamba, use it to process long documents, and build long-context RAG apps. - Learn how Jamba combines transformer and state space model architectures to achieve high performance and quality - Use the AI21 SDK, with an example of prompting over a large 200k-token annual financial report of Nvidia - Use Jamba for tool-calling, with hands-on examples from calling simple arithmetic calculations to a function that returns quarterly company financial reports. - Learn how training for long context is done, and the metrics used for its evaluation - Create a RAG app using the AI21 Conversational RAG tool and build your own RAG pipeline that uses Jamba and LangChain. By the end of this course, you'll learn how to build applications that can handle context as long as an entire book. Please sign up here:

New short course: Build Long-Context AI Apps with Jamba. Learn about state space models (SSMs), which have emerged as an alternative to transformers! Specifically, Jamba is a hybrid transformer-Mamba architecture that combines strengths of the transformer with ideas from SSMs. This course is built with AI21 Labs and taught by Chen Wang and Chen Almagor. The transformer architecture is computationally expensive when handling very long input contexts. But there's an alternative called Mamba, a selective state space model that can process very long contexts with a much lower computational cost. However, researchers found that the pure Mamba architecture underperforms in understanding the context, and gives lower-quality responses. To overcome this, AI21 developed the Jamba model, which combines Mamba's computational efficiency with the transformer's attention mechanism to help with the output quality. In this course, you’ll learn about how state space models, and Jamba, work. You’ll also learn how to prompt Jamba, use it to process long documents, and build long-context RAG apps. - Learn how Jamba combines transformer and state space model architectures to achieve high performance and quality - Use the AI21 SDK, with an example of prompting over a large 200k-token annual financial report of Nvidia - Use Jamba for tool-calling, with hands-on examples from calling simple arithmetic calculations to a function that returns quarterly company financial reports. - Learn how training for long context is done, and the metrics used for its evaluation - Create a RAG app using the AI21 Conversational RAG tool and build your own RAG pipeline that uses Jamba and LangChain. By the end of this course, you'll learn how to build applications that can handle context as long as an entire book. Please sign up here:

Andrew Ng

77,792 просмотров • 1 год назад

This season proved it: New Mexico Football and The University of New Mexico are building something special. Now we’re launching the Real Ones Giving Campaign to fuel recruiting, retention, and the future of Lobo Football. Be a Real One and support the program today.

This season proved it: New Mexico Football and The University of New Mexico are building something special. Now we’re launching the Real Ones Giving Campaign to fuel recruiting, retention, and the future of Lobo Football. Be a Real One and support the program today.

New Mexico Lobos

27,753 просмотров • 7 месяцев назад

Gemini-1.5 Pro has its spotlight stolen today, and people are poking fun at Sora vs Google memes. Well, I think it's the biggest boost in LLM capability so far in 2024. v1.5's 10M token context (1) excels at retrieval; (2) generalizes zero-shot to extremely long instructions like full tutorials and codebases; and (3) works across modalities such as text, audio, and video. Here's a stunning example: v1.5 learns to translate from English to Kalamang purely in context, following a full linguistic manual at inference time. Kalamang is a language spoken by fewer than 200 speakers in western New Guinea. Gemini has never seen this language during training and is only provided with 500 pages of linguistic documentation, a dictionary, and ~400 parallel sentences in context. It basically acquires a sophisticated new skill in the neural activations, instead of gradient finetuning. I talked about the Myth of Context Length many times before: don't get too excited by claims of 1M or even 1B context tokens. LSTMs already achieved literally infinite context length 25 yrs ago! What truly matters is how well the model actually uses the context to solve real-world problems, and Gemini-1.5 has surpassed the SOTA with flying colors. The paper is also well-written with lots of solid quantitative analysis on in-context memorization and generalization. Paper: “Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context” Congrats to Jeff Dean Oriol Vinyals Sundar Pichai and team!

Gemini-1.5 Pro has its spotlight stolen today, and people are poking fun at Sora vs Google memes. Well, I think it's the biggest boost in LLM capability so far in 2024. v1.5's 10M token context (1) excels at retrieval; (2) generalizes zero-shot to extremely long instructions like full tutorials and codebases; and (3) works across modalities such as text, audio, and video. Here's a stunning example: v1.5 learns to translate from English to Kalamang purely in context, following a full linguistic manual at inference time. Kalamang is a language spoken by fewer than 200 speakers in western New Guinea. Gemini has never seen this language during training and is only provided with 500 pages of linguistic documentation, a dictionary, and ~400 parallel sentences in context. It basically acquires a sophisticated new skill in the neural activations, instead of gradient finetuning. I talked about the Myth of Context Length many times before: don't get too excited by claims of 1M or even 1B context tokens. LSTMs already achieved literally infinite context length 25 yrs ago! What truly matters is how well the model actually uses the context to solve real-world problems, and Gemini-1.5 has surpassed the SOTA with flying colors. The paper is also well-written with lots of solid quantitative analysis on in-context memorization and generalization. Paper: “Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context” Congrats to Jeff Dean Oriol Vinyals Sundar Pichai and team!

Jim Fan

278,517 просмотров • 2 лет назад

NEW - President Trump refers to the UFO files drops as “extraterrestrial” information “We’re releasing a lot of information having to do with extraterrestrial things and people are totally fascinated by it. It’s amazing, I wasn’t sure they would enjoy it. It’s literally trending number 1, could you believe it, a lot of people out there like it. We’re releasing all of that information.”

NEW - President Trump refers to the UFO files drops as “extraterrestrial” information “We’re releasing a lot of information having to do with extraterrestrial things and people are totally fascinated by it. It’s amazing, I wasn’t sure they would enjoy it. It’s literally trending number 1, could you believe it, a lot of people out there like it. We’re releasing all of that information.”

Red Panda Koala

159,101 просмотров • 2 месяцев назад

Sam Altman just said in his new interview, that a new AI architecture is coming that will be a massive upgrade, just like Transformers were over Long Short-Term Memory. And also now the current class of frontier models are powerful enough to have the brainpower needed to help us research these ideas. His advice is to use the current AI to help you find that next giant step forward. --- From 'TreeHacks' YT Channel (link in comment)

Sam Altman just said in his new interview, that a new AI architecture is coming that will be a massive upgrade, just like Transformers were over Long Short-Term Memory. And also now the current class of frontier models are powerful enough to have the brainpower needed to help us research these ideas. His advice is to use the current AI to help you find that next giant step forward. --- From 'TreeHacks' YT Channel (link in comment)

Rohan Paul

658,383 просмотров • 4 месяцев назад

Work from home and other shifts mean there is a large supply of office space that we can use for housing, but for too long, outdated city regulations stood in the way. We’re now unlocking tens of thousands of housing units to conversion, which can be much faster and cost-effective than new construction.

Work from home and other shifts mean there is a large supply of office space that we can use for housing, but for too long, outdated city regulations stood in the way. We’re now unlocking tens of thousands of housing units to conversion, which can be much faster and cost-effective than new construction.

Karen Bass

16,515 просмотров • 3 месяцев назад

“Lightning is great for payments. But platforms need more than payments.” 🌈 Today we're introducing Prism, a new payment architecture powering Geyser. It improves payment reliability, keeps us non-custodial, & enables native payment splits. We built it to solve challenges we faced at Geyser, and think it could be useful for other Bitcoin apps too. 🧵

“Lightning is great for payments. But platforms need more than payments.” 🌈 Today we're introducing Prism, a new payment architecture powering Geyser. It improves payment reliability, keeps us non-custodial, & enables native payment splits. We built it to solve challenges we faced at Geyser, and think it could be useful for other Bitcoin apps too. 🧵

Geyser ⚡️

16,644 просмотров • 4 месяцев назад

𝗢𝗻𝗲 𝗺𝗲𝗺𝗼𝗿𝘆 𝗰𝗮𝗻’𝘁 𝗿𝘂𝗹𝗲 𝘁𝗵𝗲𝗺 𝗮𝗹𝗹. We present 𝗟𝗼𝗚𝗲𝗥, a new 𝗵𝘆𝗯𝗿𝗶𝗱 𝗺𝗲𝗺𝗼𝗿𝘆 architecture for long-context geometric reconstruction. LoGeR enables stable reconstruction over up to 𝟭𝟬𝗸 𝗳𝗿𝗮𝗺𝗲𝘀 / 𝗸𝗶𝗹𝗼𝗺𝗲𝘁𝗲𝗿 𝘀𝗰𝗮𝗹𝗲, with 𝗹𝗶𝗻𝗲𝗮𝗿-𝘁𝗶𝗺𝗲 𝘀𝗰𝗮𝗹𝗶𝗻𝗴 in sequence length, 𝗳𝘂𝗹𝗹𝘆 𝗳𝗲𝗲𝗱𝗳𝗼𝗿𝘄𝗮𝗿𝗱 inference, and 𝗻𝗼 𝗽𝗼𝘀𝘁-𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻. Yet it matches or surpasses strong optimization-based pipelines. (1/5) Google DeepMind Berkeley AI Research

𝗢𝗻𝗲 𝗺𝗲𝗺𝗼𝗿𝘆 𝗰𝗮𝗻’𝘁 𝗿𝘂𝗹𝗲 𝘁𝗵𝗲𝗺 𝗮𝗹𝗹. We present 𝗟𝗼𝗚𝗲𝗥, a new 𝗵𝘆𝗯𝗿𝗶𝗱 𝗺𝗲𝗺𝗼𝗿𝘆 architecture for long-context geometric reconstruction. LoGeR enables stable reconstruction over up to 𝟭𝟬𝗸 𝗳𝗿𝗮𝗺𝗲𝘀 / 𝗸𝗶𝗹𝗼𝗺𝗲𝘁𝗲𝗿 𝘀𝗰𝗮𝗹𝗲, with 𝗹𝗶𝗻𝗲𝗮𝗿-𝘁𝗶𝗺𝗲 𝘀𝗰𝗮𝗹𝗶𝗻𝗴 in sequence length, 𝗳𝘂𝗹𝗹𝘆 𝗳𝗲𝗲𝗱𝗳𝗼𝗿𝘄𝗮𝗿𝗱 inference, and 𝗻𝗼 𝗽𝗼𝘀𝘁-𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻. Yet it matches or surpasses strong optimization-based pipelines. (1/5) Google DeepMind Berkeley AI Research

Junyi Zhang

562,850 просмотров • 5 месяцев назад

The World Computer should be measured by the innovations it enables, specifically through the applications that bring its potential to life Our core principle at MegaETH has always been to empower builders to create truly “zero-to-one” decentralized applications on crypto rails, free from any performance constraints is a culmination of that. The most cracked of the cracked, shipping new experiences that will inspire the next generation of dapps We’ll make Ethereum so great again that it will hurt

The World Computer should be measured by the innovations it enables, specifically through the applications that bring its potential to life Our core principle at MegaETH has always been to empower builders to create truly “zero-to-one” decentralized applications on crypto rails, free from any performance constraints is a culmination of that. The most cracked of the cracked, shipping new experiences that will inspire the next generation of dapps We’ll make Ethereum so great again that it will hurt

MegaMafia

344,702 просмотров • 1 год назад