Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

We found a general recipe for Solves: scale pretraining, then hill-climb with minimal in-house data. For the first time, one fine-tuning example can teach a new behavior that generalizes. Below: 4 folding strategies, learned from one example each and tested on held-out setups.

Tony Zhao

148,842 subscribers

170,577 Aufrufe • vor 1 Monat •via X (Twitter)

Wissenschaft & Technologie

Anya Rossi• Live Now

Private livecam show

10 Kommentare

Profilbild von Tony Zhao

Tony Zhaovor 1 Monat

We call this a Solve. Progress in robotics is difficult to measure because demos vary by setup. Demo ≠ Solved. A Solve declares two boundaries: scope and adaptation cost. Without both, 99% has no context.

Profilbild von Tony Zhao

Tony Zhaovor 1 Monat

Tons of emergent behavior along the way: - Picking up clothes off the ground - Handling baby wear to 8XL shirts - Robustness against adversarial disturbances and lighting

Profilbild von Tony Zhao

Tony Zhaovor 1 Monat

We completed the most rigorous generalization test to date. Across 785 trials in 31 unseen environments, ACT-2 achieved 99.1% success in laundry folding. ACT-2 also achieved human-level fold quality: receiving an average rating of 4.72/5, with 98.3% earning four or five stars.

Profilbild von Tony Zhao

Tony Zhaovor 1 Monat

Introducing ACT-2 Preview The first robotics model to unify broad generalization with high reliability. A single fine-tuning example can teach Memo a new behavior that generalizes. Zero shot, real unseen homes, 99% success rate.

Profilbild von Tony Zhao

Tony Zhaovor 1 Monat

Quantitatively, we measure the generalization gap as the difference between in-domain and out-of-domain performance. As we scale up pretraining, the gap falls sharply. This makes in-house performance a reliable predictor of performance in the wild.

Profilbild von Tony Zhao

Tony Zhaovor 1 Monat

This property allows us to hill-climb performance in our office, and trust those gains to hold in unseen homes Our fleet of Memos runs in parallel to rapidly advance reliability, quality, and speed. Left: fleet-scale improvement in-house Right: Memo working across unseen homes

Profilbild von Tony Zhao

Tony Zhaovor 1 Monat

Laundry is our first Solve of many. Our recipe is so general that scaling data and compute gives us predictable improvements. Unlocking one Solve accelerates the next Solve. The same ACT-2 model is learning to vacuum, organize toys, zip clothing, and turn pants inside out.

Profilbild von Tony Zhao

Tony Zhaovor 1 Monat

This fall, ACT-2’s first Solve enters homes through our Beta Program, the final step towards fully autonomous home robot deployment. Full technical report:

Profilbild von Pierre Delecto Too 🙈🙉🙊

Pierre Delecto Too 🙈🙉🙊vor 1 Monat

Can it take my laundry downstairs, load it into the washer & dryer, and haul everything back upstairs?

Profilbild von Tensor Templar

Tensor Templarvor 1 Monat

How big is it

Ähnliche Videos

Introducing ACT-2 Preview The first robotics model to unify broad generalization with high reliability. A single fine-tuning example can teach Memo a new behavior that generalizes. Zero shot, real unseen homes, 99% success rate.

Introducing ACT-2 Preview The first robotics model to unify broad generalization with high reliability. A single fine-tuning example can teach Memo a new behavior that generalizes. Zero shot, real unseen homes, 99% success rate.

Tony Zhao

1,043,045 Aufrufe • vor 1 Monat

Introducing GEN-1.5, a one-shot learner. It can learn new tasks in a few seconds. Show it what to do, and it generalizes. This capability emerged from pretraining on physical data at scale, as a step towards our mission of building general intelligence for the physical world.

Introducing GEN-1.5, a one-shot learner. It can learn new tasks in a few seconds. Show it what to do, and it generalizes. This capability emerged from pretraining on physical data at scale, as a step towards our mission of building general intelligence for the physical world.

Generalist

3,394,735 Aufrufe • vor 24 Tagen

New short course on Pretraining LLMs! Developed with Upstage and taught by their CEO Sung Kim and CSO Lucy Park. While prompting or fine-tuning existing models works well for many general language tasks, pretraining is valuable for specialized domains or languages with limited representation in current models. This course walks you through the LLM pretraining pipeline: 1. Data preparation: Learn to source, clean, and prepare training data using HuggingFace. 2. Model architecture: Configure transformer networks, including modifying existing models. 3. Training: Set up and run training using open-source libraries. 4. Evaluation: Benchmark performance using popular evaluation strategies. As an example use case, you'll also compare the output of a base model with its fine-tuned and further pretrained variants, to see the impact of pretraining on a model's ability to write Python. The course also explores an innovative technique called depth up-scaling, which Upstage used to train their Solar model family, reducing pretraining compute costs by up to 70%. This technique works by first duplicating layers of a smaller pretrained model to form a larger model, and then further pretraining the result. Sign up here!

New short course on Pretraining LLMs! Developed with Upstage and taught by their CEO Sung Kim and CSO Lucy Park. While prompting or fine-tuning existing models works well for many general language tasks, pretraining is valuable for specialized domains or languages with limited representation in current models. This course walks you through the LLM pretraining pipeline: 1. Data preparation: Learn to source, clean, and prepare training data using HuggingFace. 2. Model architecture: Configure transformer networks, including modifying existing models. 3. Training: Set up and run training using open-source libraries. 4. Evaluation: Benchmark performance using popular evaluation strategies. As an example use case, you'll also compare the output of a base model with its fine-tuned and further pretrained variants, to see the impact of pretraining on a model's ability to write Python. The course also explores an innovative technique called depth up-scaling, which Upstage used to train their Solar model family, reducing pretraining compute costs by up to 70%. This technique works by first duplicating layers of a smaller pretrained model to form a larger model, and then further pretraining the result. Sign up here!

Andrew Ng

85,714 Aufrufe • vor 2 Jahren

Can we scale 4D pretraining to learn general space-time representations that reconstruct an object from a few views at any time to any view at any other time? Introducing 4D-LRM: a Large Space-Time Reconstruction Model that ... 🔹 Predicts 4D Gaussian primitives directly from multi-view tokens (no motion vectors, no HexPlane); 🔹 Uses a clean, minimal Transformer backbone; 🔹 Generalizes with fast, high-quality feedforward rendering at any view and infinite frame rate. Check out more interactive demos and scaling behaviors on our homepage/paper. 👉Website: 👉Paper:

Can we scale 4D pretraining to learn general space-time representations that reconstruct an object from a few views at any time to any view at any other time? Introducing 4D-LRM: a Large Space-Time Reconstruction Model that ... 🔹 Predicts 4D Gaussian primitives directly from multi-view tokens (no motion vectors, no HexPlane); 🔹 Uses a clean, minimal Transformer backbone; 🔹 Generalizes with fast, high-quality feedforward rendering at any view and infinite frame rate. Check out more interactive demos and scaling behaviors on our homepage/paper. 👉Website: 👉Paper:

Martin Ziqiao Ma

21,833 Aufrufe • vor 1 Jahr

Revolutionizing Move Programming with OpenLedger In this demo, we showcase how Move datasets contributed by data providers to OpenLedger’s datanets are used to fine-tune specialized models with LoRA fine-tuning. As seen in the video, we showcase an example on how builders can deploy a Move-specialized model that powers Co-pilot agents using our no-code model fine-tuning platform. This is the future of AI and Web3 innovation. Watch this space to see more specialised models and data feeds being built for next generation agents on top of OpenLedger #Move

Revolutionizing Move Programming with OpenLedger In this demo, we showcase how Move datasets contributed by data providers to OpenLedger’s datanets are used to fine-tune specialized models with LoRA fine-tuning. As seen in the video, we showcase an example on how builders can deploy a Move-specialized model that powers Co-pilot agents using our no-code model fine-tuning platform. This is the future of AI and Web3 innovation. Watch this space to see more specialised models and data feeds being built for next generation agents on top of OpenLedger #Move

OpenLedger

61,693 Aufrufe • vor 1 Jahr

Ever feel like God can’t forgive your failures? He can and He does. I’m thankful for that! Peter’s life is a beautiful example. It’s just one of the stories in my new book #NothingIsImpossibleWithGod - out one week from today and available for pre-order now!

Ever feel like God can’t forgive your failures? He can and He does. I’m thankful for that! Peter’s life is a beautiful example. It’s just one of the stories in my new book #NothingIsImpossibleWithGod - out one week from today and available for pre-order now!

Shannon Bream

16,024 Aufrufe • vor 6 Monaten

A British surgeon tells the BBC what he saw in Gaza. "I can remember one child for example who was 7, Amir... He heard a drone above him, turned round, and then it fired on him. And we had many many reports like that from children and adults."

A British surgeon tells the BBC what he saw in Gaza. "I can remember one child for example who was 7, Amir... He heard a drone above him, turned round, and then it fired on him. And we had many many reports like that from children and adults."

Paul Graham

377,953 Aufrufe • vor 2 Monaten

How can you solve complex tasks using a Large Language Model? Here is a 2-minute introduction to everything you need to know to 10x the quality of your results. Let's talk about three techniques, in order of complexity, starting with the easiest one: • In-Context Learning • Indexing + In-Context Learning • Fine-tuning In-Context Learning The team that trained GPT-3 found something they couldn't explain: You can condition a model using examples of how you want it to behave. I included an example prompt in the attached video. You can "teach" the model how you want it to interpret questions, select the correct answers, and format the results by giving a few examples. You can also give specific knowledge to the model that will be helpful when formulating answers. We call this approach "grounding the model." There's another example in the video. Indexing + In-Context Learning Unfortunately, there is a limit to how much data you can include in a prompt. We call this the "context size." One version of GPT-4 supports a context of approximately 6,000 words, while the other supports 25,000 words. Although this sounds like a lot, many applications need more than that. Imagine you wrote a book and want to build an application to answer any questions about your story. What happens if your book is longer than the context? That's where Indexing comes in. Using a model, you can turn every book passage into an embedding. These are vectors, numbers that "encode" the passage's text. You can then store these embeddings in a particular database that supports fast retrieval of these vectors. You can then turn any question into an embedding and search the database for the list of passages that are similar to that query. Instead of using the entire book to ask the model, you can now use the relevant passages as in-context information, effectively working around the context size limitation. Fine-tuning Fine-tuning can give you an extra boost to get reliable outputs from your LLM. It is, however, the most complex approach on the list. There are different approaches to fine-tuning a model with your data. A popular technique is to process your data with your LLM and use the outputs to train a new classifier that solves your specific task. Notice that here you aren't modifying the LLM. Instead, you are chaining it with your trained classifier. Another approach is to modify the parameters of the LLM using your data. Think of this as "rewiring" the model in a way that solves your particular task. The results and costs will vary depending on how many layers you want to fine-tune from the original model. Many companies think that fine-tuning is the solution to their problems. In my experience, many will benefit from exploring the other two approaches. I love explaining Machine Learning and Artificial Intelligence ideas. If you enjoy in-depth content like this, follow me Santiago so you don't miss what comes next.

How can you solve complex tasks using a Large Language Model? Here is a 2-minute introduction to everything you need to know to 10x the quality of your results. Let's talk about three techniques, in order of complexity, starting with the easiest one: • In-Context Learning • Indexing + In-Context Learning • Fine-tuning In-Context Learning The team that trained GPT-3 found something they couldn't explain: You can condition a model using examples of how you want it to behave. I included an example prompt in the attached video. You can "teach" the model how you want it to interpret questions, select the correct answers, and format the results by giving a few examples. You can also give specific knowledge to the model that will be helpful when formulating answers. We call this approach "grounding the model." There's another example in the video. Indexing + In-Context Learning Unfortunately, there is a limit to how much data you can include in a prompt. We call this the "context size." One version of GPT-4 supports a context of approximately 6,000 words, while the other supports 25,000 words. Although this sounds like a lot, many applications need more than that. Imagine you wrote a book and want to build an application to answer any questions about your story. What happens if your book is longer than the context? That's where Indexing comes in. Using a model, you can turn every book passage into an embedding. These are vectors, numbers that "encode" the passage's text. You can then store these embeddings in a particular database that supports fast retrieval of these vectors. You can then turn any question into an embedding and search the database for the list of passages that are similar to that query. Instead of using the entire book to ask the model, you can now use the relevant passages as in-context information, effectively working around the context size limitation. Fine-tuning Fine-tuning can give you an extra boost to get reliable outputs from your LLM. It is, however, the most complex approach on the list. There are different approaches to fine-tuning a model with your data. A popular technique is to process your data with your LLM and use the outputs to train a new classifier that solves your specific task. Notice that here you aren't modifying the LLM. Instead, you are chaining it with your trained classifier. Another approach is to modify the parameters of the LLM using your data. Think of this as "rewiring" the model in a way that solves your particular task. The results and costs will vary depending on how many layers you want to fine-tune from the original model. Many companies think that fine-tuning is the solution to their problems. In my experience, many will benefit from exploring the other two approaches. I love explaining Machine Learning and Artificial Intelligence ideas. If you enjoy in-depth content like this, follow me Santiago so you don't miss what comes next.

Santiago

384,510 Aufrufe • vor 3 Jahren

For most of history, expertise was scarce, constrained by time and reach: one person, one career, one lifetime. Now, for the first time, we can encode, evaluate, and scale it. We believe the wisdom that once took a lifetime to build shouldn’t take a lifetime to find. Today, we’re excited to announce that AfterQuery has raised a $30M Series A at a $300M valuation and that we’ve since surpassed $100M in annual revenue run rate, to build the data layer of professional AI.

For most of history, expertise was scarce, constrained by time and reach: one person, one career, one lifetime. Now, for the first time, we can encode, evaluate, and scale it. We believe the wisdom that once took a lifetime to build shouldn’t take a lifetime to find. Today, we’re excited to announce that AfterQuery has raised a $30M Series A at a $300M valuation and that we’ve since surpassed $100M in annual revenue run rate, to build the data layer of professional AI.

Spencer Mateega

275,411 Aufrufe • vor 5 Monaten

Watch King reveals that a random first-time customer bought a $1,700,000 one-of-one AP watch “Random client came in for the first time, nobody, not a referral, at least as far as we know. Never interacted with him, and they asked us for a one-of-one AP, a truly unique piece, and it’s $1.7 million for the watch” “Essentially, I had my team on it, and within three days we found where in the world the watch was. We flew somebody there, we inspected it, we brought it back, we had international armed security, which is a whole thing, and we delivered the watch to the guy at his house, in under a week”

Watch King reveals that a random first-time customer bought a $1,700,000 one-of-one AP watch “Random client came in for the first time, nobody, not a referral, at least as far as we know. Never interacted with him, and they asked us for a one-of-one AP, a truly unique piece, and it’s $1.7 million for the watch” “Essentially, I had my team on it, and within three days we found where in the world the watch was. We flew somebody there, we inspected it, we brought it back, we had international armed security, which is a whole thing, and we delivered the watch to the guy at his house, in under a week”

Mikli

220,923 Aufrufe • vor 3 Monaten

Jay Abernathy's game-breaking speed is standing out this summer in the Cape League. One great example of what his speed can do came in a game on Sunday: Abernathy singled with one out, stole second, stole third, and then came home on a routine ground out to the right side.

Jay Abernathy's game-breaking speed is standing out this summer in the Cape League. One great example of what his speed can do came in a game on Sunday: Abernathy singled with one out, stole second, stole third, and then came home on a routine ground out to the right side.

D1Baseball

18,835 Aufrufe • vor 2 Monaten

One of the biggest arbitrage opportunities in AI is to build a "software factory." Here is the example: - I run an AI First Index with my Fortune 500 clients. - We interview their executives and score them on 16 dimensions. - Then we made a public version of it. Option one: open Claude Code, or Codex Say "build out the product." Spend hours refining and iterating until you finally release it Option two: go one level up and "build the factory for the thing." Ours has 5 parts: 1) Primitives. - Login. - Payments. - Social sharing. - You build each one only one time. 2) Promotion. - The owned media that sells the product is part of the machine 3) Memory - We map the roadblocks we had in a shared memory bank so never make that mistake again. - For example, we had a web hook problem. - "I want to learn through the mess." 4) Humans. - These loops are not autonomous. They are "very heavy-handed with humans." 5) Speed. - The first product is already profitable. - The next product is faster, better and stronger. The catch: This process is a "measure twice cut once", so make sure you take the time to build it right the first time Then use the same factory for your content engine, or for your new leads.

One of the biggest arbitrage opportunities in AI is to build a "software factory." Here is the example: - I run an AI First Index with my Fortune 500 clients. - We interview their executives and score them on 16 dimensions. - Then we made a public version of it. Option one: open Claude Code, or Codex Say "build out the product." Spend hours refining and iterating until you finally release it Option two: go one level up and "build the factory for the thing." Ours has 5 parts: 1) Primitives. - Login. - Payments. - Social sharing. - You build each one only one time. 2) Promotion. - The owned media that sells the product is part of the machine 3) Memory - We map the roadblocks we had in a shared memory bank so never make that mistake again. - For example, we had a web hook problem. - "I want to learn through the mess." 4) Humans. - These loops are not autonomous. They are "very heavy-handed with humans." 5) Speed. - The first product is already profitable. - The next product is faster, better and stronger. The catch: This process is a "measure twice cut once", so make sure you take the time to build it right the first time Then use the same factory for your content engine, or for your new leads.

The Startup Ideas Podcast (SIP) 🧃

40,176 Aufrufe • vor 29 Tagen

The one thing we can always agree on is pizza. For a limited time, get two large one-topping pizzas from Little Caesars Pizza for just $4.99 each with code: PIZZAPIZZA online only.

The one thing we can always agree on is pizza. For a limited time, get two large one-topping pizzas from Little Caesars Pizza for just $4.99 each with code: PIZZAPIZZA online only.

Trent McDuffie

36,896 Aufrufe • vor 1 Jahr

📢📢📢 Excited to release ManipTrans: Efficient Dexterous Bimanual Manipulation Transfer via Residual Learning (CVPR25). 🤏🤙✌️With ManipTrans, we can transfer dexterous manipulation skills into robotic hands in simulation and deploy them on a real robot, using a residual policy learned for dex manipulation. 🤖🤖🤖The video below illustrates how the MoCap data can be transferred to Inspire, Shadow, Xhand, Allegro, and Mano. With ManipTrans, we can scale up dex manip data greatly with minimal effort. For more details, please check our -webpage: -code: -huggingface:

📢📢📢 Excited to release ManipTrans: Efficient Dexterous Bimanual Manipulation Transfer via Residual Learning (CVPR25). 🤏🤙✌️With ManipTrans, we can transfer dexterous manipulation skills into robotic hands in simulation and deploy them on a real robot, using a residual policy learned for dex manipulation. 🤖🤖🤖The video below illustrates how the MoCap data can be transferred to Inspire, Shadow, Xhand, Allegro, and Mano. With ManipTrans, we can scale up dex manip data greatly with minimal effort. For more details, please check our -webpage: -code: -huggingface:

Siyuan Huang

20,983 Aufrufe • vor 1 Jahr

I put 150 film techniques on one page. Camera, lighting, composition, editing, genre, storytelling, VFX. Each one comes with an example shot, a short explanation, and the prompt written out for the moment you know the shot you want and the words won't come.

I put 150 film techniques on one page. Camera, lighting, composition, editing, genre, storytelling, VFX. Each one comes with an example shot, a short explanation, and the prompt written out for the moment you know the shot you want and the words won't come.

Ivan Flugelman — VVSVS™

138,889 Aufrufe • vor 3 Tagen

1/ Presenting #Shoshin, a Minimal Viable Reality with Asynchronous Onchain Strategies, this Saturday, July 22nd in Paris for the first time. . Apply to join the playtest. Click the link below to apply. Acceptance will be sent out on a rolling basis. . Please enable video sound.

1/ Presenting #Shoshin, a Minimal Viable Reality with Asynchronous Onchain Strategies, this Saturday, July 22nd in Paris for the first time. . Apply to join the playtest. Click the link below to apply. Acceptance will be sent out on a rolling basis. . Please enable video sound.

Topology

46,724 Aufrufe • vor 3 Jahren

#TowerofFantasy [The Little Rascal and the Foodie] New Simulacrum Debut #Carrot Ver.4.8 #InterstellarVisitor will be available from 4/8! Follow Tower of Fantasy and Repost for a chance to win Amazon Gift Card and exclusive in-game rewards! Don't miss out! One egg, one carrot! One spoonful of flour, one carrot! Half a carton of milk, one carrot!... Ta-da! Behold—my secret recipe: Super Yummy Carrot Cake! 🎂🥕 One slice for you, one slice for me! One for the Authorizer, one for me! One for Sister Hilda, one for me! And Black Dog… fine, you can have one too—but behave, alright? 😆🐾 HD Version >> #ToF

#TowerofFantasy [The Little Rascal and the Foodie] New Simulacrum Debut #Carrot Ver.4.8 #InterstellarVisitor will be available from 4/8! Follow Tower of Fantasy and Repost for a chance to win Amazon Gift Card and exclusive in-game rewards! Don't miss out! One egg, one carrot! One spoonful of flour, one carrot! Half a carton of milk, one carrot!... Ta-da! Behold—my secret recipe: Super Yummy Carrot Cake! 🎂🥕 One slice for you, one slice for me! One for the Authorizer, one for me! One for Sister Hilda, one for me! And Black Dog… fine, you can have one too—but behave, alright? 😆🐾 HD Version >> #ToF

Tower of Fantasy

67,643 Aufrufe • vor 1 Jahr

We’ve added a new, faster way to fine-tune belt tension on your Prusa CORE One. ✨ The new stroboscope-based Manual Belt Tuning helps you find the optimal setting for both belts quickly and precisely - resulting in smoother motion and cleaner prints. Available from firmware 6.4.0. 🦾 Try it out now in Settings → Manual Belt Tuning.

We’ve added a new, faster way to fine-tune belt tension on your Prusa CORE One. ✨ The new stroboscope-based Manual Belt Tuning helps you find the optimal setting for both belts quickly and precisely - resulting in smoother motion and cleaner prints. Available from firmware 6.4.0. 🦾 Try it out now in Settings → Manual Belt Tuning.

Prusa3D

29,815 Aufrufe • vor 10 Monaten