正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

In-context learning as the mysterious ability in LMs. We propose ✨Deep-thinking✨ to boost ICL by iterative forward tuning. It is possible to tune LMs without backpropagation! 🤯 Paper: Gradio Demo:

Binyuan Hui

43,478 subscribers

53,403 次观看 • 3 年前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

1 条评论

RayTan 的头像

RayTan1 年前

The Huggingface Demo page return 404 now. can you help to update it?

相关视频

✨DESTINYS CHILD✨ RALSEI LMS RELEASE THANK YOU ALL FOR THE PATIENCE ON THIS LMS RELEASE ! We hope to bring you some more musical releases soon ! 👀 For now we give you the toothpaste boy HIMSELF !! 🐐

✨DESTINYS CHILD✨ RALSEI LMS RELEASE THANK YOU ALL FOR THE PATIENCE ON THIS LMS RELEASE ! We hope to bring you some more musical releases soon ! 👀 For now we give you the toothpaste boy HIMSELF !! 🐐

TENNA: TV Torment

11,068 次观看 • 1 天前

News: Possible Aberrant Custom LMS For characters like c00lkidd and 1x1x1x1 who have LMS with specific characters they will still keep those original LMS. But Alex does have an idea of a remix of Eternal Hope Eternal Fight using other LMS themes as a global LMS #Forsaken

News: Possible Aberrant Custom LMS For characters like c00lkidd and 1x1x1x1 who have LMS with specific characters they will still keep those original LMS. But Alex does have an idea of a remix of Eternal Hope Eternal Fight using other LMS themes as a global LMS #Forsaken

Forsaken Wiki (Fandom)

50,114 次观看 • 10 个月前

PROCEED❤️ PERMAFROST 🧊 2/2 OF NOELLE'S LMS THEMES We hope you enjoy this snowstorm of a LMS and tune into more tracks on the T:TT channel ‼️

PROCEED❤️ PERMAFROST 🧊 2/2 OF NOELLE'S LMS THEMES We hope you enjoy this snowstorm of a LMS and tune into more tracks on the T:TT channel ‼️

TENNA: TV Torment

18,071 次观看 • 26 天前

// idea If noli makes it to a lms with a another character that’s supposed to have a unique LMS with another killer. It will play that unique LMS for a couple seconds before cutting back to “ grave soul now run”

// idea If noli makes it to a lms with a another character that’s supposed to have a unique LMS with another killer. It will play that unique LMS for a couple seconds before cutting back to “ grave soul now run”

lumberjack slasher

167,484 次观看 • 26 天前

Robot AI brains, aka Vision-Language-Action models, cannot adapt to new tasks as easily as LLMs like Gemini, ChatGPT, or Grok. LLMs can adapt quickly with their in-context learning (ICL) capabilities. But can we inject ICL abilities into a pre-trained VLA like pi0? Yes! Introducing RICL (Retraining for In-Context Learning), our Conference on Robot Learning (CoRL) 2025 paper. Our RICL-pi0 model can adapt to unseen objects, novel motions, and new scenes with just ICL and RAG (retrieval-augmented generation). RICL-pi0 also boosts performance on the long-tail of tasks. A quick 1 minute video summary:

Robot AI brains, aka Vision-Language-Action models, cannot adapt to new tasks as easily as LLMs like Gemini, ChatGPT, or Grok. LLMs can adapt quickly with their in-context learning (ICL) capabilities. But can we inject ICL abilities into a pre-trained VLA like pi0? Yes! Introducing RICL (Retraining for In-Context Learning), our Conference on Robot Learning (CoRL) 2025 paper. Our RICL-pi0 model can adapt to unseen objects, novel motions, and new scenes with just ICL and RAG (retrieval-augmented generation). RICL-pi0 also boosts performance on the long-tail of tasks. A quick 1 minute video summary:

Kaustubh Sridhar

52,897 次观看 • 1 年前

Eggman LMS in IRL

Eggman LMS in IRL

RivCaX

86,223 次观看 • 9 天前

I decided to make a little video to where if 2011x was in Forsaken but as a c00lkid skin, I had fun voicing these! lms is by Vasalto #forsaken #outcomememories #sonicexe

I decided to make a little video to where if 2011x was in Forsaken but as a c00lkid skin, I had fun voicing these! lms is by Vasalto #forsaken #outcomememories #sonicexe

gooftroop

52,127 次观看 • 10 个月前

Still thinking about this lms i had the luck to witness Both 1x and shedletsky were very good so seeing them fight had me on the edge of my seat Plus the matching skins adds to how cool this lms was

Still thinking about this lms i had the luck to witness Both 1x and shedletsky were very good so seeing them fight had me on the edge of my seat Plus the matching skins adds to how cool this lms was

Voidrabbit 🌌🐇

75,275 次观看 • 1 年前

NeurIPS 2025 Paper: LLMs are Reinforcement Learners 🤯! Surprisingly, we show that LLMs can solve RL tasks without any external component! We introduce Prompted Policy Search (ProPS), an RL method based only LLMs and in-context learning. [Paper]

NeurIPS 2025 Paper: LLMs are Reinforcement Learners 🤯! Surprisingly, we show that LLMs can solve RL tasks without any external component! We introduce Prompted Policy Search (ProPS), an RL method based only LLMs and in-context learning. [Paper]

Heni Ben Amor

51,248 次观看 • 11 个月前

imagine this right no announcement of a new sonic lms theme pure radio silence about it not even mentioned in the patch notes so you hop in the game and test a sonic lms to check right and for like 2 seconds is pure silence and then you hear this

imagine this right no announcement of a new sonic lms theme pure radio silence about it not even mentioned in the patch notes so you hop in the game and test a sonic lms to check right and for like 2 seconds is pure silence and then you hear this

MacheteLu 🇵🇾 | #Dess4BiteByNight

29,676 次观看 • 2 个月前

We got Josh Lujan addicted to LMS BOYS OSRS PVP HYPEEEE

We got Josh Lujan addicted to LMS BOYS OSRS PVP HYPEEEE

Savix

65,245 次观看 • 1 年前

Introducing PC-ALM, a local-learning alternative to backpropagation. Our method trains 1000-layer neural nets using only local dynamics, and without backprop. Blog: Standard deep learning relies on backpropagation. The brain, however, cannot implement backpropagation, at least not exactly. How can a physical system, such as the brain, solve multilayer credit assignment without explicit use of backprop? We look for inspiration in two related fields: distributed optimization and NeuroAI. In NeuroAI, predictive coding asks each neuron activation to solve an energy-based inference problem instead of using a standard forward pass. That inference step can be implemented as energy-minimization dynamics on local prediction errors. This perspective -- each layer as a dynamical system -- has proven promising, but performance of predictive coding hasn't scaled well with depth. Credit signals at far ends of the network struggle to diffuse into internal layers. We turn to distributed optimization, generalizing predictive coding to use an augmented Lagrangian instead of energy. This motivation stems back to a classic 1988 paper by LeCun, showing that the Lagrange multipliers of a deep network can be identified with gradients of a supervised loss. The augmented Lagrangian then bridges LeCun's perspective to the standard predictive coding that is used in NeuroAI. We find that this new perspective yields a natural PC-like alternative to backpropagation, resulting in a method we call PC-ALM. PC-ALM differs from PC in that it introduces dual neurons (Lagrange multipliers) as part of the layer-local dynamics, resulting in each layer acting as a PI feedback control system to minimize local prediction errors. We find that PC-ALM is capable of propagating signals to seemingly arbitrary depth, especially in deep narrow networks where standard PC struggles to learn. Ultimately, our motivation here is to understand how distributed physical systems, such as the brain, can compute credit signals using only local coupling and local dynamics. PC-ALM may also inform deep learning in neuromorphic hardware, where dynamics are cheaper than on GPUs. Paper: Code:

Introducing PC-ALM, a local-learning alternative to backpropagation. Our method trains 1000-layer neural nets using only local dynamics, and without backprop. Blog: Standard deep learning relies on backpropagation. The brain, however, cannot implement backpropagation, at least not exactly. How can a physical system, such as the brain, solve multilayer credit assignment without explicit use of backprop? We look for inspiration in two related fields: distributed optimization and NeuroAI. In NeuroAI, predictive coding asks each neuron activation to solve an energy-based inference problem instead of using a standard forward pass. That inference step can be implemented as energy-minimization dynamics on local prediction errors. This perspective -- each layer as a dynamical system -- has proven promising, but performance of predictive coding hasn't scaled well with depth. Credit signals at far ends of the network struggle to diffuse into internal layers. We turn to distributed optimization, generalizing predictive coding to use an augmented Lagrangian instead of energy. This motivation stems back to a classic 1988 paper by LeCun, showing that the Lagrange multipliers of a deep network can be identified with gradients of a supervised loss. The augmented Lagrangian then bridges LeCun's perspective to the standard predictive coding that is used in NeuroAI. We find that this new perspective yields a natural PC-like alternative to backpropagation, resulting in a method we call PC-ALM. PC-ALM differs from PC in that it introduces dual neurons (Lagrange multipliers) as part of the layer-local dynamics, resulting in each layer acting as a PI feedback control system to minimize local prediction errors. We find that PC-ALM is capable of propagating signals to seemingly arbitrary depth, especially in deep narrow networks where standard PC struggles to learn. Ultimately, our motivation here is to understand how distributed physical systems, such as the brain, can compute credit signals using only local coupling and local dynamics. PC-ALM may also inform deep learning in neuromorphic hardware, where dynamics are cheaper than on GPUs. Paper: Code:

Sakana AI

455,486 次观看 • 2 个月前

This skin is AMAZING but I have three big wishes for it in the future. 1) His voice lines are replaced with animalistic growls similar to how he sounds in ITP 2) He gets an original chase theme and LMS 3) The LMS references Voices From The Pit (credits song from ITP)

This skin is AMAZING but I have three big wishes for it in the future. 1) His voice lines are replaced with animalistic growls similar to how he sounds in ITP 2) He gets an original chase theme and LMS 3) The LMS references Voices From The Pit (credits song from ITP)

💜 Alex 💜

20,374 次观看 • 6 个月前

Almost all animals sleep. Why don’t LMs? Introducing our new work on language model sleep. tl;dr : A periodic, recurrent “sleep” phase allows LMs to digest their context and transfer it into their weights, improving recall and reasoning on challenging tasks.

Almost all animals sleep. Why don’t LMs? Introducing our new work on language model sleep. tl;dr : A periodic, recurrent “sleep” phase allows LMs to digest their context and transfer it into their weights, improving recall and reasoning on challenging tasks.

Sangyun Lee

126,452 次观看 • 5 个月前

I'm begging the devs to keep this bug and make it actually be the LMS for the Milestone 3/4 skins. That or at LEAST make this the LMS for Aberrant.

Sensitive content

I'm begging the devs to keep this bug and make it actually be the LMS for the Milestone 3/4 skins. That or at LEAST make this the LMS for Aberrant.

Slasher

53,098 次观看 • 10 个月前

The Way It Ends - SCP Survivor VS SCP Killer LMS Made by: Emmy! |

The Way It Ends - SCP Survivor VS SCP Killer LMS Made by: Emmy! |

Forsaken News

25,957 次观看 • 5 个月前

“We noticed that along with extended thinking, with GPT-5.5, we were able to do something that we didn't think was possible,” Shopify’s goal is to have more entrepreneurs in the world and the way they do that is by removing as much friction as possible from the builder journey. There was life before Shopify with GPT-5.5, and after. We prefer after. ✨

“We noticed that along with extended thinking, with GPT-5.5, we were able to do something that we didn't think was possible,” Shopify’s goal is to have more entrepreneurs in the world and the way they do that is by removing as much friction as possible from the builder journey. There was life before Shopify with GPT-5.5, and after. We prefer after. ✨

Farhan Thawar

13,857 次观看 • 4 个月前

Storing too many tools in your context window increases latency and can lead to wrong tool selection. In this demo, we used LFM2.5-ColBERT-350M as a filter to only select the five most relevant tools among 151 options. It's fast and reliable, even without any specific fine-tuning. Try the demo on Hugging Face! And learn more on our blog:

Storing too many tools in your context window increases latency and can lead to wrong tool selection. In this demo, we used LFM2.5-ColBERT-350M as a filter to only select the five most relevant tools among 151 options. It's fast and reliable, even without any specific fine-tuning. Try the demo on Hugging Face! And learn more on our blog:

Liquid AI

33,824 次观看 • 4 个月前