正在加载视频...

视频加载失败

返回首页

Curious whether video generation models (like #SORA) qualify as world models? We conduct a systematic study to answer this question by investigating whether a video gen model is able to learn physical laws. Three are three key messages to take home: 1⃣The model generalises perfectly for in-distribution data, but...

bingyikang's profile picture

Bingyi Kang

3,518 subscribers

606,632 次观看 • 1 年前 •via X (Twitter)

教育科学技术#SORA
LIVE

Anya Rossi• Live Now

Private livecam show

10 条评论

Bingyi Kang 的头像
Bingyi Kang1 年前

The video was created by @raylu_THU, who consistently provides insightful discussions for analyzing experiments. Great Work!

Bingyi Kang 的头像
Bingyi Kang1 年前

The huggingface paper page:

Tianyuan Zhang 的头像
Tianyuan Zhang1 年前

Very cool paper. I guess one reason behind color > size > velocity > shape, is in your dataset, the color attributes affects lots of pixels and influence the L2 diffusion loss a lot.

Bingyi Kang 的头像
Bingyi Kang1 年前

Yeah, we do have a similar hypothesis, check the Open Discussion section ( on our project page for details.

Tommy Nic 的头像
Tommy Nic1 年前

The study shows we're not there yet. These models don’t grasp the ‘rules’ of physics as a true world model would. But early text models had similar limits – stuck in mimicry until they broke through to real generalization. There’s a good chance future video models will follow the same path.

iandanforth 🦋 @iandanforth.bsky.social 的头像
iandanforth 🦋 @iandanforth.bsky.social1 年前

Were the models also given task prompts as text to 'explain' the task to the model? Large image generation models have demonstrated language task ability so it's possible latent understanding / steer-ability exists in video generation models as well.

Bingyi Kang 的头像
Bingyi Kang1 年前

actually not, as each model is often trained for one task. However, we did try to use internal states (e.g., language description of size and velocity) of a physical event as prompt to the model. They often give worse ood generalization.

Yang Yue 的头像
Yang Yue1 年前

Thrilled to see this work come to life after 7-8 months of deep thinking about #SORA and its connection to physical laws. This paper has been my most demanding yet rewarding project. Proud to have been part of this journey! Check out our findings via the video, website and paer.

BensenHsu 的头像
BensenHsu1 年前

The paper explores whether video generation models can discover fundamental physical laws by merely observing visual data, without any human priors. This is an important question as video generation is seen as a promising path towards building scalable world models that can accurately simulate the physical world. The researchers' analysis revealed two key insights about the generalization mechanisms of the video generation models: 1. The models rely more on memorization and case-based imitation, rather than abstracting universal physical rules. 2. The models prioritize certain attributes (color > size > velocity > shape) when referencing training data during generalization, which may explain their difficulties in maintaining object consistency. full paper:

david glukhov 的头像
david glukhov1 年前

Extremely reminiscent of this earlier work which observed many of the same issues in a simpler setting and domain

相关视频

Meta announces Movie Gen A Cast of Media Foundation Models We present Movie Gen, a cast of foundation models that generates high-quality, 1080p HD videos with different aspect ratios and synchronized audio. We also show additional capabilities such as precise instruction-based video editing and generation of personalized videos based on a user’s image. Our models set a new state-of-the-art on multiple tasks: text-to-video synthesis, video personalization, video editing, video-to-audio generation, and text-to-audio generation. Our largest video generation model is a 30B parameter transformer trained with a maximum context length of 73K video tokens, corresponding to a generated video of 16 seconds at 16 frames-per-second. We show multiple technical innovations and simplifications on the architecture, latent spaces, training objectives and recipes, data curation, evaluation protocols, parallelization techniques, and inference optimizations that allow us to reap the benefits of scaling pre-training data, model size, and training compute for training large scale media generation models. We hope this paper helps the research community to accelerate progress and innovation in media generation models
0:30

Meta announces Movie Gen A Cast of Media Foundation Models We present Movie Gen, a cast of foundation models that generates high-quality, 1080p HD videos with different aspect ratios and synchronized audio. We also show additional capabilities such as precise instruction-based video editing and generation of personalized videos based on a user’s image. Our models set a new state-of-the-art on multiple tasks: text-to-video synthesis, video personalization, video editing, video-to-audio generation, and text-to-audio generation. Our largest video generation model is a 30B parameter transformer trained with a maximum context length of 73K video tokens, corresponding to a generated video of 16 seconds at 16 frames-per-second. We show multiple technical innovations and simplifications on the architecture, latent spaces, training objectives and recipes, data curation, evaluation protocols, parallelization techniques, and inference optimizations that allow us to reap the benefits of scaling pre-training data, model size, and training compute for training large scale media generation models. We hope this paper helps the research community to accelerate progress and innovation in media generation models

AK

62,719 次观看 • 1 年前

Small Language Models (SML) are the future of AI. "Small" (SML) instead of "Large" (LLM). These small models are highly specialized models with superhuman abilities on specific tasks. Here are two techniques to build these models: • Spectrum • Model Merging I give you a short introduction in the attached video, but here is a quick summary: Spectrum helps us identify the most relevant layers to solve one specific task. We can ignore everything else and focus on fine-tuning these layers. Using Spectrum, we can fine-tune models in a heartbeat. Model Merging combines multiple models into a unique, much better model than any of the individual input models. You can also combine models specialized in different tasks and get a model with multiple abilities. This is the state of the art of productizing models. It's what Arcee.ai's platform does behind the scenes. Arcee collaborated with me on this post and is sponsoring it. There are three main steps to produce a model for your particular use case: 1. You create a dataset by uploading your data. 2. You train a model. At this step, Arcee uses Spectrum and Model Merging to produce a highly specialized model for your task. 3. You can deploy that model to any environment you want. Three important notes: • Training process is 2x faster and 2x cheaper than regular fine-tuning. • Resultant models are smaller and have higher accuracy. • They create these specialized models from open-source models. Check this site so you can fully appreciate how this works: If you want to fine-tune an open-source model, consider Arcee's platform. This is the state of the art.
2:52

Small Language Models (SML) are the future of AI. "Small" (SML) instead of "Large" (LLM). These small models are highly specialized models with superhuman abilities on specific tasks. Here are two techniques to build these models: • Spectrum • Model Merging I give you a short introduction in the attached video, but here is a quick summary: Spectrum helps us identify the most relevant layers to solve one specific task. We can ignore everything else and focus on fine-tuning these layers. Using Spectrum, we can fine-tune models in a heartbeat. Model Merging combines multiple models into a unique, much better model than any of the individual input models. You can also combine models specialized in different tasks and get a model with multiple abilities. This is the state of the art of productizing models. It's what Arcee.ai's platform does behind the scenes. Arcee collaborated with me on this post and is sponsoring it. There are three main steps to produce a model for your particular use case: 1. You create a dataset by uploading your data. 2. You train a model. At this step, Arcee uses Spectrum and Model Merging to produce a highly specialized model for your task. 3. You can deploy that model to any environment you want. Three important notes: • Training process is 2x faster and 2x cheaper than regular fine-tuning. • Resultant models are smaller and have higher accuracy. • They create these specialized models from open-source models. Check this site so you can fully appreciate how this works: If you want to fine-tune an open-source model, consider Arcee's platform. This is the state of the art.

Santiago

164,162 次观看 • 2 年前

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
3:58

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

61,662 次观看 • 1 年前

Every home is different. That means that to build a useful home robot, we must be able to perform zero-shot generalization on a wide range of tasks. Humanoid company 1X has a solution: world models. 1X Director of Evaluations Daniel Ho joins us on RoboPapers to talk about: - why world models are the future for scaling robot learning - how to use world models for robot control - what world models unlock for evaluating robot model performance - how we can hill-climb from here to general purpose robots Watch Episode #61 of RoboPapers, with Michael Cho - Rbt/Acc and Chris Paxton, now!
54:52

Every home is different. That means that to build a useful home robot, we must be able to perform zero-shot generalization on a wide range of tasks. Humanoid company 1X has a solution: world models. 1X Director of Evaluations Daniel Ho joins us on RoboPapers to talk about: - why world models are the future for scaling robot learning - how to use world models for robot control - what world models unlock for evaluating robot model performance - how we can hill-climb from here to general purpose robots Watch Episode #61 of RoboPapers, with Michael Cho - Rbt/Acc and Chris Paxton, now!

RoboPapers

27,567 次观看 • 5 个月前

Google presents Still-Moving Customized Video Generation without Customized Video Data Customizing text-to-image (T2I) models has seen tremendous progress recently, particularly in areas such as personalization, stylization, and conditional generation. However, expanding this progress to video generation is still in its infancy, primarily due to the lack of customized video data. In this work, we introduce Still-Moving, a novel generic framework for customizing a text-to-video (T2V) model, without requiring any customized video data. The framework applies to the prominent T2V design where the video model is built over a text-to-image (T2I) model (e.g., via inflation). We assume access to a customized version of the T2I model, trained only on still image data (e.g., using DreamBooth or StyleDrop). Naively plugging in the weights of the customized T2I model into the T2V model often leads to significant artifacts or insufficient adherence to the customization data. To overcome this issue, we train lightweight Spatial Adapters that adjust the features produced by the injected T2I layers. Importantly, our adapters are trained on "frozen videos" (i.e., repeated images), constructed from image samples generated by the customized T2I model. This training is facilitated by a novel Motion Adapter module, which allows us to train on such static videos while preserving the motion prior of the video model. At test time, we remove the Motion Adapter modules and leave in only the trained Spatial Adapters. This restores the motion prior of the T2V model while adhering to the spatial prior of the customized T2I model. We demonstrate the effectiveness of our approach on diverse tasks including personalized, stylized, and conditional generation. In all evaluated scenarios, our method seamlessly integrates the spatial prior of the customized T2I model with a motion prior supplied by the T2V model.
1:43

Google presents Still-Moving Customized Video Generation without Customized Video Data Customizing text-to-image (T2I) models has seen tremendous progress recently, particularly in areas such as personalization, stylization, and conditional generation. However, expanding this progress to video generation is still in its infancy, primarily due to the lack of customized video data. In this work, we introduce Still-Moving, a novel generic framework for customizing a text-to-video (T2V) model, without requiring any customized video data. The framework applies to the prominent T2V design where the video model is built over a text-to-image (T2I) model (e.g., via inflation). We assume access to a customized version of the T2I model, trained only on still image data (e.g., using DreamBooth or StyleDrop). Naively plugging in the weights of the customized T2I model into the T2V model often leads to significant artifacts or insufficient adherence to the customization data. To overcome this issue, we train lightweight Spatial Adapters that adjust the features produced by the injected T2I layers. Importantly, our adapters are trained on "frozen videos" (i.e., repeated images), constructed from image samples generated by the customized T2I model. This training is facilitated by a novel Motion Adapter module, which allows us to train on such static videos while preserving the motion prior of the video model. At test time, we remove the Motion Adapter modules and leave in only the trained Spatial Adapters. This restores the motion prior of the T2V model while adhering to the spatial prior of the customized T2I model. We demonstrate the effectiveness of our approach on diverse tasks including personalized, stylized, and conditional generation. In all evaluated scenarios, our method seamlessly integrates the spatial prior of the customized T2I model with a motion prior supplied by the T2V model.

AK

40,474 次观看 • 2 年前

An interactive world model developed by NVIDIA in collaboration with academic partners. - DreamDojo turns egocentric human video data into physical intelligence. - Human data is more scalable than robotics data but lacks action labels. - To solve this, a dedicated action model extracts latent actions by identifying physics and motion deltas between frames. Training - A massive 44k hours of video data are used for pre-training. - Post-training on small-scale robot datasets maps human physics to specific robot embodiments. - An additional distillation stage converts the model into an autoregressive, few-step diffusion model, enabling real-time, action-controllable simulation. Primary Use Cases - Live Teleoperation: Controlling a robot inside a world simulation in real-time. - Model-based Planning: Previewing and curating the best actions for improved success. - Policy Evaluation: Testing robot policies in realistic, out-of-distribution scenarios. Everything that's open-sourced: weights, code, post-training dataset, eval set, and details to reproduce.
1:14

An interactive world model developed by NVIDIA in collaboration with academic partners. - DreamDojo turns egocentric human video data into physical intelligence. - Human data is more scalable than robotics data but lacks action labels. - To solve this, a dedicated action model extracts latent actions by identifying physics and motion deltas between frames. Training - A massive 44k hours of video data are used for pre-training. - Post-training on small-scale robot datasets maps human physics to specific robot embodiments. - An additional distillation stage converts the model into an autoregressive, few-step diffusion model, enabling real-time, action-controllable simulation. Primary Use Cases - Live Teleoperation: Controlling a robot inside a world simulation in real-time. - Model-based Planning: Previewing and curating the best actions for improved success. - Policy Evaluation: Testing robot policies in realistic, out-of-distribution scenarios. Everything that's open-sourced: weights, code, post-training dataset, eval set, and details to reproduce.

The Humanoid Hub

11,575 次观看 • 4 个月前

AI agents are about to redefine the internet. The mistake we made with Large Language Models? We let a handful of corporations capture all the value. Action Model is building a different path. By training through our extension, users gain fractional ownership in the Large Action Model, giving them a real stake in the future of AI. When LLMs emerged, the upside flowed to Big Tech. This time, it doesn’t have to. They’re building AI on our data, and keeping the upside for themselves. Community-owned Large Action Model is how we take it back.
0:49

AI agents are about to redefine the internet. The mistake we made with Large Language Models? We let a handful of corporations capture all the value. Action Model is building a different path. By training through our extension, users gain fractional ownership in the Large Action Model, giving them a real stake in the future of AI. When LLMs emerged, the upside flowed to Big Tech. This time, it doesn’t have to. They’re building AI on our data, and keeping the upside for themselves. Community-owned Large Action Model is how we take it back.

Action Model

76,752 次观看 • 4 个月前

Today we’re releasing V-JEPA, a method for teaching machines to understand and model the physical world by watching videos. This work is another important step towards Yann LeCun’s outlined vision of AI models that use a learned understanding of the world to plan, reason and accomplish complex tasks. Details ➡️ We're releasing a collection of V-JEPA vision models trained with a feature prediction objective using self-supervised learning. The models are able to understand and predict what is going on in a video, even with limited information. It learns by predicting missing or obscured parts of a video in its internal feature space. Unlike generative approaches that fill in missing pixels, this flexible approach enables up to 6x improvements in training and sample efficiency. The models were pre-trained on entirely unlabeled data, and a small amount of labeled data can be used to train a task-specific prediction head on top after pre-training. Our results show that, using a frozen backbone, our top V-JEPA models achieve 82.0% on Kinetics-400, 72.2% on Something-Something-v2 and 77.9% on ImageNet1K — competitive with or exceeding previous leading video models. We believe that this work is an important milestone on the path to advancing machine intelligence.
1:07

Today we’re releasing V-JEPA, a method for teaching machines to understand and model the physical world by watching videos. This work is another important step towards Yann LeCun’s outlined vision of AI models that use a learned understanding of the world to plan, reason and accomplish complex tasks. Details ➡️ We're releasing a collection of V-JEPA vision models trained with a feature prediction objective using self-supervised learning. The models are able to understand and predict what is going on in a video, even with limited information. It learns by predicting missing or obscured parts of a video in its internal feature space. Unlike generative approaches that fill in missing pixels, this flexible approach enables up to 6x improvements in training and sample efficiency. The models were pre-trained on entirely unlabeled data, and a small amount of labeled data can be used to train a task-specific prediction head on top after pre-training. Our results show that, using a frozen backbone, our top V-JEPA models achieve 82.0% on Kinetics-400, 72.2% on Something-Something-v2 and 77.9% on ImageNet1K — competitive with or exceeding previous leading video models. We believe that this work is an important milestone on the path to advancing machine intelligence.

AI at Meta

703,647 次观看 • 2 年前

A breakthrough in real-time video generation. As a research preview developed with NVIDIA and shared at NVIDIAGTC this week, we trained a new real-time video model running on Vera Rubin. HD videos generate instantly, with time-to-first-frame under 100ms. Unlocking an entirely new creative paradigm and bolstering the foundations of our General World Model, GWM-1. Real-time generation opens a fundamentally different design space for video models and world simulation. We're investing in co-designing our models alongside advances in hardware to keep pushing this frontier.
0:54

A breakthrough in real-time video generation. As a research preview developed with NVIDIA and shared at NVIDIAGTC this week, we trained a new real-time video model running on Vera Rubin. HD videos generate instantly, with time-to-first-frame under 100ms. Unlocking an entirely new creative paradigm and bolstering the foundations of our General World Model, GWM-1. Real-time generation opens a fundamentally different design space for video models and world simulation. We're investing in co-designing our models alongside advances in hardware to keep pushing this frontier.

Runway

1,162,192 次观看 • 3 个月前

Tencent presents GameGen-O Open-world Video Game Generation We introduce GameGen-O, the first diffusion transformer model tailored for the generation of open-world video games. This model facilitates high-quality, open-domain generation by simulating a wide array of game engine features, such as innovative characters, dynamic environments, complex actions, and diverse events. Additionally, it provides interactive controllability, thus allowing for the gameplay simulation. The development of GameGen-O involves a comprehensive data collection and processing effort from scratch. We collect and build the first Open-World Video Game Dataset (OGameData), amassed extensive data from over a hundred of next-generation open-world games, employing a proprietary data pipeline for efficient sorting, scoring, filtering, and decoupled captioning. This robust and extensive OGameData forms the foundation of our model's training process. GameGen-O undergoes a two-stage training process, consisting of foundation model pretraining and instruction tuning. In the first phase, the model is pre-trained on the OGameData via the text-to-video and video continuation, endowing GameGen-O with the capability for open-domain video game generation. In the second phase, the pre-trained model is frozen, and we fine-tuned using a trainable InstructNet, which enables the production of subsequent frames based on multimodal structural instructions. This whole training process imparts the model with the ability to generate and interactively control content. In summary, GameGen-O represents a notable initial step forward in the realm of open-world video game generation via generative models. It underscores the potential of generative models to serve as an alternative to rendering techniques, which can efficiently combine creative generation with interactive capabilities.
0:51

Tencent presents GameGen-O Open-world Video Game Generation We introduce GameGen-O, the first diffusion transformer model tailored for the generation of open-world video games. This model facilitates high-quality, open-domain generation by simulating a wide array of game engine features, such as innovative characters, dynamic environments, complex actions, and diverse events. Additionally, it provides interactive controllability, thus allowing for the gameplay simulation. The development of GameGen-O involves a comprehensive data collection and processing effort from scratch. We collect and build the first Open-World Video Game Dataset (OGameData), amassed extensive data from over a hundred of next-generation open-world games, employing a proprietary data pipeline for efficient sorting, scoring, filtering, and decoupled captioning. This robust and extensive OGameData forms the foundation of our model's training process. GameGen-O undergoes a two-stage training process, consisting of foundation model pretraining and instruction tuning. In the first phase, the model is pre-trained on the OGameData via the text-to-video and video continuation, endowing GameGen-O with the capability for open-domain video game generation. In the second phase, the pre-trained model is frozen, and we fine-tuned using a trainable InstructNet, which enables the production of subsequent frames based on multimodal structural instructions. This whole training process imparts the model with the ability to generate and interactively control content. In summary, GameGen-O represents a notable initial step forward in the realm of open-world video game generation via generative models. It underscores the potential of generative models to serve as an alternative to rendering techniques, which can efficiently combine creative generation with interactive capabilities.

AK

367,000 次观看 • 1 年前

"One of the very confusing things about the models right now: how to reconcile the fact that they are doing so well on evals. And you look at the evals and you go, 'Those are pretty hard evals.' But the economic impact seems to be dramatically behind. There is [a possible] explanation. Back when people were doing pre-training, the question of what data to train on was answered, because that answer was everything. So you don't have to think if it's going to be this data or that data. When people do RL training, they say, 'Okay, we want to have this kind of RL training for this thing and that kind of RL training for that thing.' You say, 'Hey, I would love our model to do really well when we release it. I want the evals to look great. What would be RL training that could help on this task?' If you combine this with generalization of the models actually being inadequate, that has the potential to explain a lot of what we are seeing, this disconnect between eval performance and actual real-world performance"
2:21

"One of the very confusing things about the models right now: how to reconcile the fact that they are doing so well on evals. And you look at the evals and you go, 'Those are pretty hard evals.' But the economic impact seems to be dramatically behind. There is [a possible] explanation. Back when people were doing pre-training, the question of what data to train on was answered, because that answer was everything. So you don't have to think if it's going to be this data or that data. When people do RL training, they say, 'Okay, we want to have this kind of RL training for this thing and that kind of RL training for that thing.' You say, 'Hey, I would love our model to do really well when we release it. I want the evals to look great. What would be RL training that could help on this task?' If you combine this with generalization of the models actually being inadequate, that has the potential to explain a lot of what we are seeing, this disconnect between eval performance and actual real-world performance"

Dwarkesh Patel

502,060 次观看 • 7 个月前

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!
3:54

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

85,657 次观看 • 2 年前

Tracking Anything with Decoupled Video Segmentation paper page: Training data for video segmentation are expensive to annotate. This impedes extensions of end-to-end algorithms to new video segmentation tasks, especially in large-vocabulary settings. To 'track anything' without training on video data for every individual task, we develop a decoupled video segmentation approach (DEVA), composed of task-specific image-level segmentation and class/task-agnostic bi-directional temporal propagation. Due to this design, we only need an image-level model for the target task (which is cheaper to train) and a universal temporal propagation model which is trained once and generalizes across tasks. To effectively combine these two modules, we use bi-directional propagation for (semi-)online fusion of segmentation hypotheses from different frames to generate a coherent segmentation. We show that this decoupled formulation compares favorably to end-to-end approaches in several data-scarce tasks including large-vocabulary video panoptic segmentation, open-world video segmentation, referring video segmentation, and unsupervised video object segmentation.
0:32

Tracking Anything with Decoupled Video Segmentation paper page: Training data for video segmentation are expensive to annotate. This impedes extensions of end-to-end algorithms to new video segmentation tasks, especially in large-vocabulary settings. To 'track anything' without training on video data for every individual task, we develop a decoupled video segmentation approach (DEVA), composed of task-specific image-level segmentation and class/task-agnostic bi-directional temporal propagation. Due to this design, we only need an image-level model for the target task (which is cheaper to train) and a universal temporal propagation model which is trained once and generalizes across tasks. To effectively combine these two modules, we use bi-directional propagation for (semi-)online fusion of segmentation hypotheses from different frames to generate a coherent segmentation. We show that this decoupled formulation compares favorably to end-to-end approaches in several data-scarce tasks including large-vocabulary video panoptic segmentation, open-world video segmentation, referring video segmentation, and unsupervised video object segmentation.

AK

305,633 次观看 • 2 年前

Emerging as the global frontier of model-reasoning has been a core vision of Affine since its inception. In pursuit of this vision, we've been working relentlessly to forge a repeatable training loop: a mechanism that uses benchmark driven post-training methodology to facilitate the improvement of reasoning itself. Through applying our training methodology onto Qwen3-32B as our base model, Affine’s miners have been able to push the Pareto frontier outward. We’ve now reached saturation; our models have become so good at reasoning that any further improvements are producing diminishing returns. As a result we now move our base model to Qwen3.6-35B-A3B. It provides a stronger basis, offering an extended context window, as well as an improved foundation for agentic work. The same training methodology will be applied going forward. Our objective is no longer simply to improve a single model, but to demonstrate that the process in itself is repeatable, scalable, and model agnostic. We view the commoditization of reasoning as vital to ensure our technology is not only transparent, but also co-owned by humanity.
0:29

Emerging as the global frontier of model-reasoning has been a core vision of Affine since its inception. In pursuit of this vision, we've been working relentlessly to forge a repeatable training loop: a mechanism that uses benchmark driven post-training methodology to facilitate the improvement of reasoning itself. Through applying our training methodology onto Qwen3-32B as our base model, Affine’s miners have been able to push the Pareto frontier outward. We’ve now reached saturation; our models have become so good at reasoning that any further improvements are producing diminishing returns. As a result we now move our base model to Qwen3.6-35B-A3B. It provides a stronger basis, offering an extended context window, as well as an improved foundation for agentic work. The same training methodology will be applied going forward. Our objective is no longer simply to improve a single model, but to demonstrate that the process in itself is repeatable, scalable, and model agnostic. We view the commoditization of reasoning as vital to ensure our technology is not only transparent, but also co-owned by humanity.

Affine

14,416 次观看 • 24 天前

Today we're announcing #GAIA1: a 9B parameter world model, trained on 4,700 hours of driving data, able to simulate complex and diverse driving scenes from video, text and action inputs. This model is 480x larger than the preview we shared earlier this year and the results are incredible. These videos are entirely synthetically generated by Wayve's generative AI, GAIA-1. But there is more here than just generating videos, GAIA is an entire world model. A world model allows us to simulate the future, conditioned on video, text and action inputs, which can be leveraged for making informed decisions when driving. Why is this game-changing for autonomous driving? 1. Safety. One limitation with AI systems like today's Large Language Models is that they are autoregressive, next-word prediction algorithms, but aren't necessarily aware of the implications of their decisions. A world model allows us to give our AI the capability to be aware of its decisions, by simulating the future, which is important for self-driving safety. 2. Synthetic training data. I believe synthetic training data is the future for AI, because it is safer, cheaper, and infinitely scalable. GAIA-1 unlocks unprecedented realism and diversity of synthetic data for self-driving. 3. Long-tail robustness. One of the biggest challenges for self-driving is long-tail robustness: dealing with the enormous magnitude of edge cases we see on the road. An advantage of generative AI is its incredible ability to recombine experiences in new ways. This is exciting for self-driving as it means we can learn from two edge case scenarios, and combine them to become a corner case. For example, we can experience driving in fog, and experience of jay-walking pedestrians, and GAIA can learn from these experiences to understand how to generate a fog+jay walking scenario. Check out many more videos in our blog or further technical details in our paper: Or come chat with our team who are at the International Conference on Computer Vision (#ICCV2023) this week in Paris in Booth 32 Jamie Shotton
0:41

Today we're announcing #GAIA1: a 9B parameter world model, trained on 4,700 hours of driving data, able to simulate complex and diverse driving scenes from video, text and action inputs. This model is 480x larger than the preview we shared earlier this year and the results are incredible. These videos are entirely synthetically generated by Wayve's generative AI, GAIA-1. But there is more here than just generating videos, GAIA is an entire world model. A world model allows us to simulate the future, conditioned on video, text and action inputs, which can be leveraged for making informed decisions when driving. Why is this game-changing for autonomous driving? 1. Safety. One limitation with AI systems like today's Large Language Models is that they are autoregressive, next-word prediction algorithms, but aren't necessarily aware of the implications of their decisions. A world model allows us to give our AI the capability to be aware of its decisions, by simulating the future, which is important for self-driving safety. 2. Synthetic training data. I believe synthetic training data is the future for AI, because it is safer, cheaper, and infinitely scalable. GAIA-1 unlocks unprecedented realism and diversity of synthetic data for self-driving. 3. Long-tail robustness. One of the biggest challenges for self-driving is long-tail robustness: dealing with the enormous magnitude of edge cases we see on the road. An advantage of generative AI is its incredible ability to recombine experiences in new ways. This is exciting for self-driving as it means we can learn from two edge case scenarios, and combine them to become a corner case. For example, we can experience driving in fog, and experience of jay-walking pedestrians, and GAIA can learn from these experiences to understand how to generate a fog+jay walking scenario. Check out many more videos in our blog or further technical details in our paper: Or come chat with our team who are at the International Conference on Computer Vision (#ICCV2023) this week in Paris in Booth 32 Jamie Shotton

Alex Kendall

631,856 次观看 • 2 年前

Not the flashiest demos, but what’s under the hood represents a foundational shift for general-purpose robotics. World models are the next-gen foundation of Physical AI, not the VLM backbones found in typical VLAs. DreamZero is a 14B-parameter World Action Model (WAM) by NVIDIA that treats robotics as a joint video-and-action prediction task. Unlike traditional Vision-Language-Action (VLA) models that map images directly to motor commands, DreamZero leverages a pretrained video diffusion backbone to predict future world states and actions simultaneously. - achieves 2× better zero-shot generalization to unseen tasks and environments compared to state-of-the-art VLAs. - learns effectively from heterogeneous, non-repetitive data (500 hours), breaking the need for thousands of repeated demonstrations. - adapts to new robot embodiments with just 30 minutes of play data. - enables 7Hz closed-loop control via system optimizations and "DreamZero-Flash," making high-capacity diffusion models viable for real-time use.
1:01

Not the flashiest demos, but what’s under the hood represents a foundational shift for general-purpose robotics. World models are the next-gen foundation of Physical AI, not the VLM backbones found in typical VLAs. DreamZero is a 14B-parameter World Action Model (WAM) by NVIDIA that treats robotics as a joint video-and-action prediction task. Unlike traditional Vision-Language-Action (VLA) models that map images directly to motor commands, DreamZero leverages a pretrained video diffusion backbone to predict future world states and actions simultaneously. - achieves 2× better zero-shot generalization to unseen tasks and environments compared to state-of-the-art VLAs. - learns effectively from heterogeneous, non-repetitive data (500 hours), breaking the need for thousands of repeated demonstrations. - adapts to new robot embodiments with just 30 minutes of play data. - enables 7Hz closed-loop control via system optimizations and "DreamZero-Flash," making high-capacity diffusion models viable for real-time use.

The Humanoid Hub

35,204 次观看 • 5 个月前

Building a truly private AI model isn’t as simple as the tech world wants you to believe. despite all the marketing around one click deployments, RAG systems, fine tuning services, and memory features, we still don’t have models that are genuinely trained only on your personal data with complete privacy from third parties. the reality is that most AI solutions today require you to trust someone else with your books, notes, journals, and training data. even when companies promise privacy, there’s usually a way for the provider or other parties to access your information during training or inference. Phala is taking a different approach by building the infrastructure needed for real data ownership. They’re developing confidential runtime environments where your AI training happens inside secure enclaves with attestation guarantees. The key innovation is their in-enclave keying system through dstack, which means that once your model is training, even Phala themselves cannot see your data or model weights.
0:34

Building a truly private AI model isn’t as simple as the tech world wants you to believe. despite all the marketing around one click deployments, RAG systems, fine tuning services, and memory features, we still don’t have models that are genuinely trained only on your personal data with complete privacy from third parties. the reality is that most AI solutions today require you to trust someone else with your books, notes, journals, and training data. even when companies promise privacy, there’s usually a way for the provider or other parties to access your information during training or inference. Phala is taking a different approach by building the infrastructure needed for real data ownership. They’re developing confidential runtime environments where your AI training happens inside secure enclaves with attestation guarantees. The key innovation is their in-enclave keying system through dstack, which means that once your model is training, even Phala themselves cannot see your data or model weights.

soulman 🎮

11,702 次观看 • 9 个月前

NVIDIA just released a new open source transcription model, Nemotron Speech ASR, designed from the ground up for low-latency use cases like voice agents. Here's a voice agent built with this new model. 24ms transcription finalization and total voice-to-voice inference time under 500ms. This agent actually uses *three* NVIDIA open source models: - Nemotron Speech ASR - Nemotron 3 Nano 30GB in a 4-bit quant (released in December) - A preview checkpoint of the upcoming Magpie text-to-speech model These models are all truly open source: weights, training data, training code, and inference code. This is a big deal! Jensen said in the CES keynote yesterday that he expects open source models to catch up to proprietary models this year in a number of categories. NVIDIA is putting their weight behind making this happen. (As Alan Kay said, the best way to predict the future is to invent it.) The code for this agent is open source too, of course. You can deploy it to production with Modal and Pipecat AI cloud, or run locally on an NVIDIA DGX Spark or RTX 5090.
1:24

NVIDIA just released a new open source transcription model, Nemotron Speech ASR, designed from the ground up for low-latency use cases like voice agents. Here's a voice agent built with this new model. 24ms transcription finalization and total voice-to-voice inference time under 500ms. This agent actually uses *three* NVIDIA open source models: - Nemotron Speech ASR - Nemotron 3 Nano 30GB in a 4-bit quant (released in December) - A preview checkpoint of the upcoming Magpie text-to-speech model These models are all truly open source: weights, training data, training code, and inference code. This is a big deal! Jensen said in the CES keynote yesterday that he expects open source models to catch up to proprietary models this year in a number of categories. NVIDIA is putting their weight behind making this happen. (As Alan Kay said, the best way to predict the future is to invent it.) The code for this agent is open source too, of course. You can deploy it to production with Modal and Pipecat AI cloud, or run locally on an NVIDIA DGX Spark or RTX 5090.

kwindla

274,403 次观看 • 6 个月前

𝐃𝐞𝐬𝐭𝐫𝐚 𝐎𝐂𝐀𝐈 (𝐎𝐧𝐞-𝐂𝐥𝐢𝐜𝐤 𝐀𝐈) 𝐌𝐨𝐝𝐞𝐥𝐬 - $DSYNC 𝐄𝐚𝐫𝐥𝐲 𝐀𝐜𝐜𝐞𝐬𝐬 𝐏𝐫𝐞𝐯𝐢𝐞𝐰 We are training and building a wide range of AI models for the Destra OCAI platform. Today, we will be showcasing an early preview of one of the initial models we are working on—a bullish/bearish prediction model designed to 𝐟𝐨𝐫𝐞𝐜𝐚𝐬𝐭 𝐭𝐡𝐞 𝐦𝐚𝐫𝐤𝐞𝐭 𝐝𝐢𝐫𝐞𝐜𝐭𝐢𝐨𝐧 𝐨𝐟 𝐬𝐩𝐞𝐜𝐢𝐟𝐢𝐜 𝐭𝐨𝐤𝐞𝐧𝐬. This is just one of the initial models we are working on; as we progress, the models will become more complex and intricate. In this behind-the-scenes video, we demonstrate how the model predicts the future market trend of the BTC-USD pair with an 𝐚𝐜𝐜𝐮𝐫𝐚𝐜𝐲 𝐨𝐟 𝟔𝟓%. This is the barebones structure of the models we are training, which will soon be available to users with a single click through the Destra OCAI platform. The OCAI platform includes a wide array of Destra Exclusive AI models, specifically designed to simplify life for crypto audiences, that are envisioned and built from scratch at Destra Labs.
0:54

𝐃𝐞𝐬𝐭𝐫𝐚 𝐎𝐂𝐀𝐈 (𝐎𝐧𝐞-𝐂𝐥𝐢𝐜𝐤 𝐀𝐈) 𝐌𝐨𝐝𝐞𝐥𝐬 - $DSYNC 𝐄𝐚𝐫𝐥𝐲 𝐀𝐜𝐜𝐞𝐬𝐬 𝐏𝐫𝐞𝐯𝐢𝐞𝐰 We are training and building a wide range of AI models for the Destra OCAI platform. Today, we will be showcasing an early preview of one of the initial models we are working on—a bullish/bearish prediction model designed to 𝐟𝐨𝐫𝐞𝐜𝐚𝐬𝐭 𝐭𝐡𝐞 𝐦𝐚𝐫𝐤𝐞𝐭 𝐝𝐢𝐫𝐞𝐜𝐭𝐢𝐨𝐧 𝐨𝐟 𝐬𝐩𝐞𝐜𝐢𝐟𝐢𝐜 𝐭𝐨𝐤𝐞𝐧𝐬. This is just one of the initial models we are working on; as we progress, the models will become more complex and intricate. In this behind-the-scenes video, we demonstrate how the model predicts the future market trend of the BTC-USD pair with an 𝐚𝐜𝐜𝐮𝐫𝐚𝐜𝐲 𝐨𝐟 𝟔𝟓%. This is the barebones structure of the models we are training, which will soon be available to users with a single click through the Destra OCAI platform. The OCAI platform includes a wide array of Destra Exclusive AI models, specifically designed to simplify life for crypto audiences, that are envisioned and built from scratch at Destra Labs.

Destra Network

37,135 次观看 • 1 年前