Loading video...

Video Failed to Load

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

👀Humans compare images by looking back and forth. Many open-weight VLMs encode each image independently, and defer comparison to the LM. We introduce SVE: Stateful Visual Encoders for Vision-Language Models, where the visual encoder itself becomes change-aware. 🌐Project: 📰Paper: 💻Code: 1/n

Zirui "Colin" Wang

1,474 subscribers

63,610 views • 3 months ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

9 Comments

Zirui "Colin" Wang3 months ago

VLMs are increasingly used for interactive tasks that require comparative visual reasoning, yet their upstream visual encoders have largely remained stateless. Each image is encoded independently for dozens of layers before the LLM ever sees the visual features. We hypothesize that this creates a bottleneck for change-aware visual representations, so we make the visual encoder stateful. 2/n

Zirui "Colin" Wang3 months ago

We tested several ways to make visual encoders stateful: * extend past-image features into the current image’s K/V * use AdaLN-Zero-style feature modulation * add cross-attention layers over past features The strongest design adds a cross-attention layer with its own FFN. But optimization matters. For stability, we stop gradients to past features, initialize added attention/FFN layers from nearby pretrained layers, and zero-initialize the final matrix in each added block. This lets SVE start close to the original stateless model, then learn to use prior visual context. 3/n

Zirui "Colin" Wang3 months ago

Across input resolutions, model sizes, and VLM backbones, making the visual encoder stateful consistently improves over the stateless baseline. The gains are especially clear when the baseline is weaker, and they persist across Qwen3-VL, Qwen3.5, GLM, InternVL, and Gemma backbones. 4/n

Zirui "Colin" Wang3 months ago

SVE also improves real-world visual comparison. In longitudinal radiology, image editing, and remote sensing change detection, the stateless baseline often misses, flips, or hallucinates visual differences. With stateful visual encoding, the model better tracks what actually changed across the image pair. 5/n

Zirui "Colin" Wang3 months ago

This project was done during my time at @VoioInc. Huge thanks to @yujunwei04, @YalaTweets, @_dmchan, @profjoeyg, @trevordarrell, @berkeley_ai, and @BerkeleySky for making this project possible, and to everyone who gave feedback on the idea, experiments, writing, and figures. We hope SVE is a useful step toward VLMs that reason over visual change more directly. More details here: 📷 n/n

Nikhil Parthasarathy3 months ago

Very nice work! Our recent work might be of interest to you as well as your formulation of cross-attention state integration is similar to what we discovered as the best way to integrate temporal state across video frames:

Zirui "Colin" Wang3 months ago

Yes, I think this is very elegant and relevant! Thanks for sharing your work!! As we scale up to dozens of images (what we haven't done), keeping a fixed state than retrieving from all past images should def be the more efficient/scalable approach

EB1A Experts3 months ago

Great work!

ζ Pedram ζ3 months ago

Very cool

Related Videos

🚀 New Paper: Pixel Reasoner 🧠🖼️ How can Vision-Language Models (VLMs) perform chain-of-thought reasoning within the image itself? We introduce Pixel Reasoner, the first open-source framework that enables VLMs to “think in pixel space” through curiosity-driven reinforcement learning. Current VLMs reason only in text — even when grounded in rich images or videos, their logical steps are verbalized in natural language. This restricts their ability to interrogate visual evidence and demonstrate how conclusions are drawn. 🔍 So we ask: What if we could make VLMs "show their work" by reasoning directly in the pixel space? Inspired by GPT-o3’s "think-in-image" ability, we propose a framework where VLMs use interactive visual operations — zoom, select-frame, highlight — to reason through complex visual inputs. To do this, we design a two-stage training process: Instruction tuning with synthesized visual reasoning traces. Reinforcement learning with curiosity-driven reward to balance exploration between pixel and text reasoning ✨ With this, Pixel Reasoner achieves near-SoTA performance on many information-rich multimodal benchmarks: 📊 84% on InfographicsVQA 🧠 84% on V* benchmark 🧩 74% on TallyQA-Complex It also achieves strong accuracy of 68% on MVBench (a video benchmark). Website: Paper: Code: Demo: (coming soon)

🚀 New Paper: Pixel Reasoner 🧠🖼️ How can Vision-Language Models (VLMs) perform chain-of-thought reasoning within the image itself? We introduce Pixel Reasoner, the first open-source framework that enables VLMs to “think in pixel space” through curiosity-driven reinforcement learning. Current VLMs reason only in text — even when grounded in rich images or videos, their logical steps are verbalized in natural language. This restricts their ability to interrogate visual evidence and demonstrate how conclusions are drawn. 🔍 So we ask: What if we could make VLMs "show their work" by reasoning directly in the pixel space? Inspired by GPT-o3’s "think-in-image" ability, we propose a framework where VLMs use interactive visual operations — zoom, select-frame, highlight — to reason through complex visual inputs. To do this, we design a two-stage training process: Instruction tuning with synthesized visual reasoning traces. Reinforcement learning with curiosity-driven reward to balance exploration between pixel and text reasoning ✨ With this, Pixel Reasoner achieves near-SoTA performance on many information-rich multimodal benchmarks: 📊 84% on InfographicsVQA 🧠 84% on V* benchmark 🧩 74% on TallyQA-Complex It also achieves strong accuracy of 68% on MVBench (a video benchmark). Website: Paper: Code: Demo: (coming soon)

Wenhu Chen

82,829 views • 1 year ago

1/N Most Vision-Language-Action models need tons of data for finetuning, and still fail for new objects and instructions. Introducing OTTER, a lightweight, easy-to-train model that uses text-aware visual features to nail unseen tasks out of the box! Here's how it works 👇

1/N Most Vision-Language-Action models need tons of data for finetuning, and still fail for new objects and instructions. Introducing OTTER, a lightweight, easy-to-train model that uses text-aware visual features to nail unseen tasks out of the box! Here's how it works 👇

Fangchen Liu

68,852 views • 1 year ago

MagicScroll: Nontypical Aspect-Ratio Image Generation for Visual Storytelling via Multi-Layered Semantic-Aware Denoising paper page: Visual storytelling often uses nontypical aspect-ratio images like scroll paintings, comic strips, and panoramas to create an expressive and compelling narrative. While generative AI has achieved great success and shown the potential to reshape the creative industry, it remains a challenge to generate coherent and engaging content with arbitrary size and controllable style, concept, and layout, all of which are essential for visual storytelling. To overcome the shortcomings of previous methods including repetitive content, style inconsistency, and lack of controllability, we propose MagicScroll, a multi-layered, progressive diffusion-based image generation framework with a novel semantic-aware denoising process. The model enables fine-grained control over the generated image on object, scene, and background levels with text, image, and layout conditions. We also establish the first benchmark for nontypical aspect-ratio image generation for visual storytelling including mediums like paintings, comics, and cinematic panoramas, with customized metrics for systematic evaluation. Through comparative and ablation studies, MagicScroll showcases promising results in aligning with the narrative text, improving visual coherence, and engaging the audience. We plan to release the code and benchmark in the hope of a better collaboration between AI researchers and creative practitioners involving visual storytelling.

MagicScroll: Nontypical Aspect-Ratio Image Generation for Visual Storytelling via Multi-Layered Semantic-Aware Denoising paper page: Visual storytelling often uses nontypical aspect-ratio images like scroll paintings, comic strips, and panoramas to create an expressive and compelling narrative. While generative AI has achieved great success and shown the potential to reshape the creative industry, it remains a challenge to generate coherent and engaging content with arbitrary size and controllable style, concept, and layout, all of which are essential for visual storytelling. To overcome the shortcomings of previous methods including repetitive content, style inconsistency, and lack of controllability, we propose MagicScroll, a multi-layered, progressive diffusion-based image generation framework with a novel semantic-aware denoising process. The model enables fine-grained control over the generated image on object, scene, and background levels with text, image, and layout conditions. We also establish the first benchmark for nontypical aspect-ratio image generation for visual storytelling including mediums like paintings, comics, and cinematic panoramas, with customized metrics for systematic evaluation. Through comparative and ablation studies, MagicScroll showcases promising results in aligning with the narrative text, improving visual coherence, and engaging the audience. We plan to release the code and benchmark in the hope of a better collaboration between AI researchers and creative practitioners involving visual storytelling.

AK

22,379 views • 2 years ago

Today we’re releasing OpenEQA — the Open-Vocabulary Embodied Question Answering Benchmark. It measures an AI agent’s understanding of physical environments by probing it with open vocabulary questions like “Where did I leave my badge?” More details ➡️ All of today’s state-of-art vision+language models (VLMs) fall well short of human performance. In fact, for questions that require spatial understanding, today’s VLMs are nearly “blind” – access to visual content provides only minor improvements over language-only models. We hope that OpenEQA motivates additional research into helping AI understand and communicate about the world it sees.

Today we’re releasing OpenEQA — the Open-Vocabulary Embodied Question Answering Benchmark. It measures an AI agent’s understanding of physical environments by probing it with open vocabulary questions like “Where did I leave my badge?” More details ➡️ All of today’s state-of-art vision+language models (VLMs) fall well short of human performance. In fact, for questions that require spatial understanding, today’s VLMs are nearly “blind” – access to visual content provides only minor improvements over language-only models. We hope that OpenEQA motivates additional research into helping AI understand and communicate about the world it sees.

AI at Meta

407,198 views • 2 years ago

How far can a very simple eye go in solving vision tasks? Like a 1-pixel camera? Humans have one of the greatest eyes in nature, while many animals have significantly simpler eyes and visual systems yet show complex perceptual behavior. In an interesting project, we find that many computer vision tasks can be solved without a typical camera and with such simple 1-pixel sensors (photoreceptors). We also find that proper design (e.g., where to place the photoreceptors strategically) makes a big difference, so we developed a computational design method to find them. 🌐 👁️[Solving Vision Tasks with Simple Photoreceptors Instead of Cameras] 🧵1/n

How far can a very simple eye go in solving vision tasks? Like a 1-pixel camera? Humans have one of the greatest eyes in nature, while many animals have significantly simpler eyes and visual systems yet show complex perceptual behavior. In an interesting project, we find that many computer vision tasks can be solved without a typical camera and with such simple 1-pixel sensors (photoreceptors). We also find that proper design (e.g., where to place the photoreceptors strategically) makes a big difference, so we developed a computational design method to find them. 🌐 👁️[Solving Vision Tasks with Simple Photoreceptors Instead of Cameras] 🧵1/n

Amir Zamir

75,908 views • 2 years ago

[CLIP] by Hand ✍️ The CLIP (Contrastive Language–Image Pre-training) model, a groundbreaking work by OpenAI, redefines the intersection of computer vision and natural language processing. It is the basis of all the multi-modal foundation models we see today. How does CLIP work? Goal: 🟨 Learn a shared embedding space for text and image [1] Given ↳ A mini batch of 3 text-image pairs ↳ OpenAI used 400 million text-image pairs to train its original CLIP model. Process 1st pair: "big table" [2] 🟪 Text → 2 Vectors (3D) ↳ Look up word embedding vectors using word2vec. [3] 🟩 Image → 2 Vectors (4D) ↳ Divide the image into two patches. ↳ Flatten each patch [4] Process other pairs ↳ Repeat [2]-[3] [5] 🟪 Text Encoder & 🟩 Image Encoder ↳ Encode input vectors into feature vectors ↳ Here, both encoders are simple one layer perceptron (linear + ReLU) ↳ In practice, the encoders are usually transformer models. [6] 🟪 🟩 Mean Pooling: 2 → 1 vector ↳ Average 2 feature vectors into a single vector by averaging across the columns ↳ The goal is to have one vector to represent each image or text [7] 🟪 🟩 -> 🟨 Projection ↳ Note that the text and image feature vectors from the encoders have different dimensions (3D vs. 4D). ↳ Use a linear layer to project image and text vectors to a 2D shared embedding space. 🏋️ Contrastive Pre-training 🏋️ [8] Prepare for MatMul ↳ Copy text vectors (T1,T2,T3) ↳ Copy the transpose of image vectors (I1,I2,I3) ↳ They are all in the 2D shared embedding space. [9] 🟦 MatMul ↳ Multiply T and I matrices. ↳ This is equivalent to taking dot product between every pair of image and text vectors. ↳ The purpose is to use dot product to estimate the similarity between a pair of image-text. [10] 🟦 Softmax: e^x ↳ Raise e to the power of the number in each cell ↳ To simplify hand calculation, we approximate e^□ with 3^□. [11] 🟦 Softmax: ∑ ↳ Sum each row for 🟩 image→🟪 text ↳ Sum each column for 🟪 text→ 🟩 image [12] 🟦 Softmax: 1 / sum ↳ Divide each element by the column sum to obtain a similarity matrix for 🟪 text→🟩 image ↳ Divide each element by the row sum to obtain a similarity matrix for 🟩 image→🟪 text [13] 🟥 Loss Gradients ↳ The "Targets" for the similarity matrices are Identity Matrices. ↳ Why? If I and T come from the same pair (i=j), we want the highest value, which is 1, and 0 otherwise. ↳ Apply the simple equation of [Similarity - Target] to compute gradients of for both directions. ↳ Why so simple? Because when Softmax and Cross-Entropy Loss are used together, the math magically works out that way. ↳ These gradients kick off the backpropagation process to update weights and biases of the encoders and projection layers (red borders).

[CLIP] by Hand ✍️ The CLIP (Contrastive Language–Image Pre-training) model, a groundbreaking work by OpenAI, redefines the intersection of computer vision and natural language processing. It is the basis of all the multi-modal foundation models we see today. How does CLIP work? Goal: 🟨 Learn a shared embedding space for text and image [1] Given ↳ A mini batch of 3 text-image pairs ↳ OpenAI used 400 million text-image pairs to train its original CLIP model. Process 1st pair: "big table" [2] 🟪 Text → 2 Vectors (3D) ↳ Look up word embedding vectors using word2vec. [3] 🟩 Image → 2 Vectors (4D) ↳ Divide the image into two patches. ↳ Flatten each patch [4] Process other pairs ↳ Repeat [2]-[3] [5] 🟪 Text Encoder & 🟩 Image Encoder ↳ Encode input vectors into feature vectors ↳ Here, both encoders are simple one layer perceptron (linear + ReLU) ↳ In practice, the encoders are usually transformer models. [6] 🟪 🟩 Mean Pooling: 2 → 1 vector ↳ Average 2 feature vectors into a single vector by averaging across the columns ↳ The goal is to have one vector to represent each image or text [7] 🟪 🟩 -> 🟨 Projection ↳ Note that the text and image feature vectors from the encoders have different dimensions (3D vs. 4D). ↳ Use a linear layer to project image and text vectors to a 2D shared embedding space. 🏋️ Contrastive Pre-training 🏋️ [8] Prepare for MatMul ↳ Copy text vectors (T1,T2,T3) ↳ Copy the transpose of image vectors (I1,I2,I3) ↳ They are all in the 2D shared embedding space. [9] 🟦 MatMul ↳ Multiply T and I matrices. ↳ This is equivalent to taking dot product between every pair of image and text vectors. ↳ The purpose is to use dot product to estimate the similarity between a pair of image-text. [10] 🟦 Softmax: e^x ↳ Raise e to the power of the number in each cell ↳ To simplify hand calculation, we approximate e^□ with 3^□. [11] 🟦 Softmax: ∑ ↳ Sum each row for 🟩 image→🟪 text ↳ Sum each column for 🟪 text→ 🟩 image [12] 🟦 Softmax: 1 / sum ↳ Divide each element by the column sum to obtain a similarity matrix for 🟪 text→🟩 image ↳ Divide each element by the row sum to obtain a similarity matrix for 🟩 image→🟪 text [13] 🟥 Loss Gradients ↳ The "Targets" for the similarity matrices are Identity Matrices. ↳ Why? If I and T come from the same pair (i=j), we want the highest value, which is 1, and 0 otherwise. ↳ Apply the simple equation of [Similarity - Target] to compute gradients of for both directions. ↳ Why so simple? Because when Softmax and Cross-Entropy Loss are used together, the math magically works out that way. ↳ These gradients kick off the backpropagation process to update weights and biases of the encoders and projection layers (red borders).

Tom Yeh

67,896 views • 2 years ago

LLaVA-3D A Simple yet Effective Pathway to Empowering LMMs with 3D-awareness Recent advancements in Large Multimodal Models (LMMs) have greatly enhanced their proficiency in 2D visual understanding tasks, enabling them to effectively process and understand images and videos. However, the development of LMMs with 3D-awareness for 3D scene understanding has been hindered by the lack of large-scale 3D vision-language datasets and powerful 3D encoders. In this paper, we introduce a simple yet effective framework called LLaVA-3D. Leveraging the strong 2D understanding priors from LLaVA, our LLaVA-3D efficiently adapts LLaVA for 3D scene understanding without compromising 2D understanding capabilities. To achieve this, we employ a simple yet effective representation, 3D Patch, which connects 2D CLIP patch features with their corresponding positions in 3D space. By integrating the 3D Patches into 2D LMMs and employing joint 2D and 3D vision-language instruction tuning, we establish a unified architecture for both 2D image understanding and 3D scene understanding. Experimental results show that LLaVA-3D converges 3.5x faster than existing 3D LMMs when trained on 3D vision-language datasets. Moreover, LLaVA-3D not only achieves state-of-the-art performance across various 3D tasks but also maintains comparable 2D image understanding and vision-language conversation capabilities with LLaVA.

LLaVA-3D A Simple yet Effective Pathway to Empowering LMMs with 3D-awareness Recent advancements in Large Multimodal Models (LMMs) have greatly enhanced their proficiency in 2D visual understanding tasks, enabling them to effectively process and understand images and videos. However, the development of LMMs with 3D-awareness for 3D scene understanding has been hindered by the lack of large-scale 3D vision-language datasets and powerful 3D encoders. In this paper, we introduce a simple yet effective framework called LLaVA-3D. Leveraging the strong 2D understanding priors from LLaVA, our LLaVA-3D efficiently adapts LLaVA for 3D scene understanding without compromising 2D understanding capabilities. To achieve this, we employ a simple yet effective representation, 3D Patch, which connects 2D CLIP patch features with their corresponding positions in 3D space. By integrating the 3D Patches into 2D LMMs and employing joint 2D and 3D vision-language instruction tuning, we establish a unified architecture for both 2D image understanding and 3D scene understanding. Experimental results show that LLaVA-3D converges 3.5x faster than existing 3D LMMs when trained on 3D vision-language datasets. Moreover, LLaVA-3D not only achieves state-of-the-art performance across various 3D tasks but also maintains comparable 2D image understanding and vision-language conversation capabilities with LLaVA.

AK

41,747 views • 2 years ago

Got the vibe but not the words? Starting today, you can simply show or tell AI Mode about what you're looking for and get rich, visual results to explore. ✨ This experience works by bringing together Search’s world-class visual understanding from Lens and Image search with Gemini 2.5’s advanced multimodal capabilities. When you show AI Mode what you’re searching for it uses our new “visual search fan-out” technique to get a deeper understanding of what’s in an image (including subtle details and secondary objects), and runs multiple queries in the background. This helps it understand the visual context and nuance of your request to deliver relevant visual responses. You can also tell AI Mode what vibe you’re going for (no image needed). For example, ask AI Mode for bedroom design inspo and go back and forth to refine your vision. We’re rolling out this experience in English in the U.S. starting this week. Try it out and let us know what you think!

Got the vibe but not the words? Starting today, you can simply show or tell AI Mode about what you're looking for and get rich, visual results to explore. ✨ This experience works by bringing together Search’s world-class visual understanding from Lens and Image search with Gemini 2.5’s advanced multimodal capabilities. When you show AI Mode what you’re searching for it uses our new “visual search fan-out” technique to get a deeper understanding of what’s in an image (including subtle details and secondary objects), and runs multiple queries in the background. This helps it understand the visual context and nuance of your request to deliver relevant visual responses. You can also tell AI Mode what vibe you’re going for (no image needed). For example, ask AI Mode for bedroom design inspo and go back and forth to refine your vision. We’re rolling out this experience in English in the U.S. starting this week. Try it out and let us know what you think!

Google AI

68,285 views • 11 months ago

🚀 Meta FAIR is releasing several new research artifacts on our road to advanced machine intelligence (AMI). These latest advancements are transforming our understanding of perception. 1️⃣ Meta Perception Encoder: A large-scale vision encoder that excels across several image & video tasks. 2️⃣ Meta Perception Language Model: A fully open & reproducible vision-language model designed to tackle visual recognition tasks. 3️⃣ Meta Locate 3D: An end-to-end model for accurate object localization in 3D environments. 4️⃣ Releasing model weights for our 8B-parameter Dynamic Byte Latent Transformer, an alternative to traditional tokenization methods with the potential to redefine the standards for language model efficiency and reliability. 5️⃣Collaborative Reasoner: A framework for evaluating & improving collaborative reasoning skills in language models. Download the code, datasets, and research papers and learn more about how these artifacts are paving the way for more efficient and accurate AI systems.➡️

🚀 Meta FAIR is releasing several new research artifacts on our road to advanced machine intelligence (AMI). These latest advancements are transforming our understanding of perception. 1️⃣ Meta Perception Encoder: A large-scale vision encoder that excels across several image & video tasks. 2️⃣ Meta Perception Language Model: A fully open & reproducible vision-language model designed to tackle visual recognition tasks. 3️⃣ Meta Locate 3D: An end-to-end model for accurate object localization in 3D environments. 4️⃣ Releasing model weights for our 8B-parameter Dynamic Byte Latent Transformer, an alternative to traditional tokenization methods with the potential to redefine the standards for language model efficiency and reliability. 5️⃣Collaborative Reasoner: A framework for evaluating & improving collaborative reasoning skills in language models. Download the code, datasets, and research papers and learn more about how these artifacts are paving the way for more efficient and accurate AI systems.➡️

AI at Meta

163,490 views • 1 year ago

Vision-Language Models (VLMs) can describe the environment, but can they refer within it? Our findings reveal a critical gap: VLMs fall short of pragmatic optimality. We identify 3 key failures of pragmatic competence in referring expression generation with VLMs: (1) cannot uniquely refer to the referent, (2) include excessive or irrelevant information, and (3) misalign with human pragmatic preferences. We introduce RefOI, a new dataset of 1.5k objects, each with 3 written and 2 spoken human-produced referring expressions. We also release RefOI-TLHF, a large dataset of token-level human feedback for 10.6k referring expressions. 👀 📄 Excited to colead this project with the amazing JaneDing, and huge thanks to the dream team Xuejun Zhang Dezhi Luo Diiikee @6SihanXu Yuchen Huang Roihn Run Peng 彭润 Michigan SLED Lab.

Vision-Language Models (VLMs) can describe the environment, but can they refer within it? Our findings reveal a critical gap: VLMs fall short of pragmatic optimality. We identify 3 key failures of pragmatic competence in referring expression generation with VLMs: (1) cannot uniquely refer to the referent, (2) include excessive or irrelevant information, and (3) misalign with human pragmatic preferences. We introduce RefOI, a new dataset of 1.5k objects, each with 3 written and 2 spoken human-produced referring expressions. We also release RefOI-TLHF, a large dataset of token-level human feedback for 10.6k referring expressions. 👀 📄 Excited to colead this project with the amazing JaneDing, and huge thanks to the dream team Xuejun Zhang Dezhi Luo Diiikee @6SihanXu Yuchen Huang Roihn Run Peng 彭润 Michigan SLED Lab.

Martin Ziqiao Ma

20,701 views • 1 year ago

We've officially released and open-sourced HunyuanImage 2.1, our latest text-to-image model. The new model delivers on our commitment to balancing performance and quality. With native 2K image generation, HunyuanImage 2.1 is an advanced open-source text-to-image model.🎨 ✨ New in 2.1: 🔹Advanced Semantics: Supports ultra-long and complex prompts of up to 1000 tokens, and precisely controls the generation of multiple subjects in a single image. 🔹Precise Chinese and English Text Rendering with seamless image–text integration: The model naturally integrates text into images, making it suitable for a wide range of applications such as product covers, illustrations, and poster design to meet the needs of various fields. 🔹Rich Styles and High Aesthetic: Capable of generating images in various styles—including photorealistic portraits, comics, and vinyl figures—it delivers outstanding visual appeal and artistic quality. 🔹High-Quality Generation: Efficiently produces ultra-high-definition (2K) images in the same time other models take to generate a 1K image. HunyuanImage 2.1 uses two text encoders: a multimodal large language model (MLLM) to improve the model's image and text alignment capabilities, and a multi-language character-aware encoder to improve text rendering capabilities. The model is a single- and double-stream diffusion transformer with 17B parameters. We've also open-sourced the weights of the the accelerated version with meanflow which reduces inference steps from 100 to just 8, and PromptEnhancer, the first industrial-grade rewriting model that enhances your prompts for more nuanced and expressive image generation. Now, creators turn complex ideas—like posters with slogans or multi-panel comics—into visuals faster than ever. We’re just getting started. Stay tuned for our native multimodal image generation model coming soon. 🌐Website: 🔗Github: 🤗Hugging Face: ✨Hugging Face Demo:

We've officially released and open-sourced HunyuanImage 2.1, our latest text-to-image model. The new model delivers on our commitment to balancing performance and quality. With native 2K image generation, HunyuanImage 2.1 is an advanced open-source text-to-image model.🎨 ✨ New in 2.1: 🔹Advanced Semantics: Supports ultra-long and complex prompts of up to 1000 tokens, and precisely controls the generation of multiple subjects in a single image. 🔹Precise Chinese and English Text Rendering with seamless image–text integration: The model naturally integrates text into images, making it suitable for a wide range of applications such as product covers, illustrations, and poster design to meet the needs of various fields. 🔹Rich Styles and High Aesthetic: Capable of generating images in various styles—including photorealistic portraits, comics, and vinyl figures—it delivers outstanding visual appeal and artistic quality. 🔹High-Quality Generation: Efficiently produces ultra-high-definition (2K) images in the same time other models take to generate a 1K image. HunyuanImage 2.1 uses two text encoders: a multimodal large language model (MLLM) to improve the model's image and text alignment capabilities, and a multi-language character-aware encoder to improve text rendering capabilities. The model is a single- and double-stream diffusion transformer with 17B parameters. We've also open-sourced the weights of the the accelerated version with meanflow which reduces inference steps from 100 to just 8, and PromptEnhancer, the first industrial-grade rewriting model that enhances your prompts for more nuanced and expressive image generation. Now, creators turn complex ideas—like posters with slogans or multi-panel comics—into visuals faster than ever. We’re just getting started. Stay tuned for our native multimodal image generation model coming soon. 🌐Website: 🔗Github: 🤗Hugging Face: ✨Hugging Face Demo:

Tencent Hy

89,392 views • 1 year ago

Do Vision-Language Models represent space, and how? Spatial terms like "left" or "right" may not be enough to match images with spatial descriptions, as we often overlook the different frames of reference (FoR) used by speakers and listeners. See Figure 1 for examples! Introducing the COnsistent Multilingual Frame Of Reference Test (COMFORT), an evaluation protocol to assess the spatial reasoning capabilities of VLMs. COMFORT includes systematically designed datasets and metrics that evaluate model performance, and their deeper linguistic competence, specifically the spatial knowledge encoded in their internal representations. Find out more in the video teaser! Almost all VLMs prefer the egocentric relative FoR with reflected transform, similar to English. Yet, we reveal significant shortcomings of VLMs: notably, the models (1) exhibit poor robustness and consistency, (2) lack the flexibility to accommodate multiple FoRs, and (3) fail to adhere to language-specific or culture-specific conventions in cross-lingual tests, as English tends to dominate other languages. A shortened version will appear in Pluralistic Alignment Workshop Pluralistic Alignment Workshop #NeurIPS2024. It seems that the ArXiv moderators put it on hold and are eager to give it a thorough read first🤣! So here is the Paper/Code/Data: This collaboration turns out to be amazing, jointly led by Brian Zheyuan Zhang, @Hu_FY_ Jayjun Lee, with so many contributions and insights from Freda Shi, Parisa Kordjamshidi Michigan SLED Lab. With a growing effort to align vision-language models with human cognitive intuitions, we call for more attention to the ambiguous nature and cross-cultural diversity of spatial reasoning!

Do Vision-Language Models represent space, and how? Spatial terms like "left" or "right" may not be enough to match images with spatial descriptions, as we often overlook the different frames of reference (FoR) used by speakers and listeners. See Figure 1 for examples! Introducing the COnsistent Multilingual Frame Of Reference Test (COMFORT), an evaluation protocol to assess the spatial reasoning capabilities of VLMs. COMFORT includes systematically designed datasets and metrics that evaluate model performance, and their deeper linguistic competence, specifically the spatial knowledge encoded in their internal representations. Find out more in the video teaser! Almost all VLMs prefer the egocentric relative FoR with reflected transform, similar to English. Yet, we reveal significant shortcomings of VLMs: notably, the models (1) exhibit poor robustness and consistency, (2) lack the flexibility to accommodate multiple FoRs, and (3) fail to adhere to language-specific or culture-specific conventions in cross-lingual tests, as English tends to dominate other languages. A shortened version will appear in Pluralistic Alignment Workshop Pluralistic Alignment Workshop #NeurIPS2024. It seems that the ArXiv moderators put it on hold and are eager to give it a thorough read first🤣! So here is the Paper/Code/Data: This collaboration turns out to be amazing, jointly led by Brian Zheyuan Zhang, @Hu_FY_ Jayjun Lee, with so many contributions and insights from Freda Shi, Parisa Kordjamshidi Michigan SLED Lab. With a growing effort to align vision-language models with human cognitive intuitions, we call for more attention to the ambiguous nature and cross-cultural diversity of spatial reasoning!

Martin Ziqiao Ma

35,595 views • 1 year ago

If you want a vision encoder for dexterous manipulation, what should be the most important part to model? 🤔 Current standard models like CLIP, SigLIP, and DINOv2 have an incredible grasp of semantics and spatial details. But they lack the action-centric structure needed for downstream visuomotor control. But collecting annotated robotic trajectories at scale is SUPER expensive and largely unrealistic. So, how do we bridge this gap? We introduce CAIP (Contrastive Action-Image Pre-training) ⬇️ 🔸 Action-centric upstream: we align visual observations with action chunks through a contrastive objective. 🔸 Human video as a proxy: we represent 3D human hand poses analogously to robotic end-effector actions, tapping into a massive source of human demonstrations. 🔸 Massive scale: pre-trained on over 32,000 hours of manipulation video, driving both sample efficiency and robust generalization. 🔸 Hardware proven: achieves a 76% average success rate on a real-world Dexmate Vega bimanual Dexmate manipulator with dual 22-DoF Sharpa Wave hands Sharpa . 🔸 State-of-the-art: significantly outperforms strong baselines like DINOv2, SigLIP, MVP, and Qwen3.5 ViT across complex tasks, even under unexpected lighting changes and visual distractors. 🌐 Project: 📝 Blog: 📄 Paper: 💻 Model:

If you want a vision encoder for dexterous manipulation, what should be the most important part to model? 🤔 Current standard models like CLIP, SigLIP, and DINOv2 have an incredible grasp of semantics and spatial details. But they lack the action-centric structure needed for downstream visuomotor control. But collecting annotated robotic trajectories at scale is SUPER expensive and largely unrealistic. So, how do we bridge this gap? We introduce CAIP (Contrastive Action-Image Pre-training) ⬇️ 🔸 Action-centric upstream: we align visual observations with action chunks through a contrastive objective. 🔸 Human video as a proxy: we represent 3D human hand poses analogously to robotic end-effector actions, tapping into a massive source of human demonstrations. 🔸 Massive scale: pre-trained on over 32,000 hours of manipulation video, driving both sample efficiency and robust generalization. 🔸 Hardware proven: achieves a 76% average success rate on a real-world Dexmate Vega bimanual Dexmate manipulator with dual 22-DoF Sharpa Wave hands Sharpa . 🔸 State-of-the-art: significantly outperforms strong baselines like DINOv2, SigLIP, MVP, and Qwen3.5 ViT across complex tasks, even under unexpected lighting changes and visual distractors. 🌐 Project: 📝 Blog: 📄 Paper: 💻 Model:

Roei Herzig

26,068 views • 3 months ago

🚀 Exciting News! We're thrilled to announce the launch of Qwen Chat ( ) – your new go-to Web UI for interacting with Qwen models! 🌟 💬 Chat effortlessly with our flagship model Qwen2.5-Plus , explore vision-language capabilities with Qwen2-VL-Max , and dive into reasoning models like QwQ and QVQ, code with coding expert Qwen2.5-Coder-32B-Instruct, etc. ✨ Key Features: * Choose & compare multiple models in one interface * Upload documents & get answers based on them * Preview mode with HTML support * Upload images for visual understanding 🌟 Coming Soon: Web search, image generation, voice mode, and more! 🔗 Ready to experience the future? Start to chat with Qwen today 💬 Feel free to send us your feedback and suggestions!

🚀 Exciting News! We're thrilled to announce the launch of Qwen Chat ( ) – your new go-to Web UI for interacting with Qwen models! 🌟 💬 Chat effortlessly with our flagship model Qwen2.5-Plus , explore vision-language capabilities with Qwen2-VL-Max , and dive into reasoning models like QwQ and QVQ, code with coding expert Qwen2.5-Coder-32B-Instruct, etc. ✨ Key Features: * Choose & compare multiple models in one interface * Upload documents & get answers based on them * Preview mode with HTML support * Upload images for visual understanding 🌟 Coming Soon: Web search, image generation, voice mode, and more! 🔗 Ready to experience the future? Start to chat with Qwen today 💬 Feel free to send us your feedback and suggestions!

Qwen

215,434 views • 1 year ago

🔥Excited to introduce Tencent Hunyuan3D-Omni, the industry's first 3D asset creation system with multi-condition control. Built on the open-source Hunyuan3D 2.1 as the "ControlNet of 3D," it addresses single-view occlusions, and poor geometry by integrating up to four control conditions. Two core innovations power the system: a Lightweight Unified Control Encoder for efficient multimodal fusion, and a Progressive Difficulty-Aware Training Strategy that boosts robustness. ✅Under single-image conditions, skeletal input provides precise pose control for animation and avatars. ✅Point cloud input—full or depth-projected—removes visual ambiguity and yields accurate photorealistic geometry. ✅Bounding box control adjusts L/W/H ratios to match design scale. ✅Voxel control sculpts topology for engineering and creative workflows. We are open-sourcing inference code and weights to accelerate controllable 3D research and industry deployment. Code: Weights: Tech Report: #3DGenAI #TencentHunyuan #OpenSourceAI

🔥Excited to introduce Tencent Hunyuan3D-Omni, the industry's first 3D asset creation system with multi-condition control. Built on the open-source Hunyuan3D 2.1 as the "ControlNet of 3D," it addresses single-view occlusions, and poor geometry by integrating up to four control conditions. Two core innovations power the system: a Lightweight Unified Control Encoder for efficient multimodal fusion, and a Progressive Difficulty-Aware Training Strategy that boosts robustness. ✅Under single-image conditions, skeletal input provides precise pose control for animation and avatars. ✅Point cloud input—full or depth-projected—removes visual ambiguity and yields accurate photorealistic geometry. ✅Bounding box control adjusts L/W/H ratios to match design scale. ✅Voxel control sculpts topology for engineering and creative workflows. We are open-sourcing inference code and weights to accelerate controllable 3D research and industry deployment. Code: Weights: Tech Report: #3DGenAI #TencentHunyuan #OpenSourceAI

Tencent Hy

34,898 views • 11 months ago

1/🧠Humans are the best robot data source — but video alone misses one thing: force. 2/🙁Tactile gloves capture force — but they're costly and block the real touch manipulation depends on. 3/💪Maybe the future of touch lives on your wrist: surface EMG reads the muscles that cause force — tactile sensing without ever touching a tactile sensor. 4/🔥Want a fully open-source framework — hardware + software — to train your own force-aware learn-from-human-data robot policy? 🚀We introduce ForceBand: Learning Forceful Manipulation with sEMG -- bring force into human videos with sEMG, for force-aware manipulation ⬇️ ✦ Zero-Shot Human-to-Robot Transfer ✦ Force Beyond Vision ✦ Free-Hand Force Sensing ✦ Collect by Anyone, Anytime, Anywhere ✦ Deploy on Any Robot, Any Camera, Any Environment ✦ Open-Source & Low-Cost & Easy-to-Implement Let's squeeze every bit of signal out of human data, and let robots feel the force! 🌐 Website: 📄 Paper: 💻 Code: 🎥 Video: 🧵 1/n

1/🧠Humans are the best robot data source — but video alone misses one thing: force. 2/🙁Tactile gloves capture force — but they're costly and block the real touch manipulation depends on. 3/💪Maybe the future of touch lives on your wrist: surface EMG reads the muscles that cause force — tactile sensing without ever touching a tactile sensor. 4/🔥Want a fully open-source framework — hardware + software — to train your own force-aware learn-from-human-data robot policy? 🚀We introduce ForceBand: Learning Forceful Manipulation with sEMG -- bring force into human videos with sEMG, for force-aware manipulation ⬇️ ✦ Zero-Shot Human-to-Robot Transfer ✦ Force Beyond Vision ✦ Free-Hand Force Sensing ✦ Collect by Anyone, Anytime, Anywhere ✦ Deploy on Any Robot, Any Camera, Any Environment ✦ Open-Source & Low-Cost & Easy-to-Implement Let's squeeze every bit of signal out of human data, and let robots feel the force! 🌐 Website: 📄 Paper: 💻 Code: 🎥 Video: 🧵 1/n

Zhi (Leo) Wang

52,379 views • 2 months ago

3D-LLM: Injecting the 3D World into Large Language Models paper page: Large language models (LLMs) and Vision-Language Models (VLMs) have been proven to excel at multiple tasks, such as commonsense reasoning. Powerful as these models can be, they are not grounded in the 3D physical world, which involves richer concepts such as spatial relationships, affordances, physics, layout, and so on. In this work, we propose to inject the 3D world into large language models and introduce a whole new family of 3D-LLMs. Specifically, 3D-LLMs can take 3D point clouds and their features as input and perform a diverse set of 3D-related tasks, including captioning, dense captioning, 3D question answering, task decomposition, 3D grounding, 3D-assisted dialog, navigation, and so on. Using three types of prompting mechanisms that we design, we are able to collect over 300k 3D-language data covering these tasks. To efficiently train 3D-LLMs, we first utilize a 3D feature extractor that obtains 3D features from rendered multi- view images. Then, we use 2D VLMs as our backbones to train our 3D-LLMs. By introducing a 3D localization mechanism, 3D-LLMs can better capture 3D spatial information. Experiments on ScanQA show that our model outperforms state-of-the-art baselines by a large margin (e.g., the BLEU-1 score surpasses state-of-the-art score by 9%). Furthermore, experiments on our held-in datasets for 3D captioning, task composition, and 3D-assisted dialogue show that our model outperforms 2D VLMs. Qualitative examples also show that our model could perform more tasks beyond the scope of existing LLMs and VLMs.

3D-LLM: Injecting the 3D World into Large Language Models paper page: Large language models (LLMs) and Vision-Language Models (VLMs) have been proven to excel at multiple tasks, such as commonsense reasoning. Powerful as these models can be, they are not grounded in the 3D physical world, which involves richer concepts such as spatial relationships, affordances, physics, layout, and so on. In this work, we propose to inject the 3D world into large language models and introduce a whole new family of 3D-LLMs. Specifically, 3D-LLMs can take 3D point clouds and their features as input and perform a diverse set of 3D-related tasks, including captioning, dense captioning, 3D question answering, task decomposition, 3D grounding, 3D-assisted dialog, navigation, and so on. Using three types of prompting mechanisms that we design, we are able to collect over 300k 3D-language data covering these tasks. To efficiently train 3D-LLMs, we first utilize a 3D feature extractor that obtains 3D features from rendered multi- view images. Then, we use 2D VLMs as our backbones to train our 3D-LLMs. By introducing a 3D localization mechanism, 3D-LLMs can better capture 3D spatial information. Experiments on ScanQA show that our model outperforms state-of-the-art baselines by a large margin (e.g., the BLEU-1 score surpasses state-of-the-art score by 9%). Furthermore, experiments on our held-in datasets for 3D captioning, task composition, and 3D-assisted dialogue show that our model outperforms 2D VLMs. Qualitative examples also show that our model could perform more tasks beyond the scope of existing LLMs and VLMs.

AK

249,798 views • 3 years ago

This is THE moment of Physical AI! We are officially announcing Cosmos 3: Omnimodal World Models for Physical AI 🚀 - Cosmos 3 is an omnimodal world model: within a unified architecture, it can understand and generate language, images, video, audio, and actions. - It is not just a VLM, not just a video generator, not just an audio-visual generative model, and not just a physics simulator / world-action model. It can understand images and videos, generate images, videos, and audio, simulate future worlds, predict actions, and generate robot policies—enabling models to truly begin to “touch the world.” - Cosmos 3 is the #1 open-weight reasoner / T2I / I2V / robot policy across many benchmarks. Huge thanks to every teammate who fought side by side on this journey—from architecture, data, training, infra, serving, and evaluation to post-training. Every part of this project carries an incredible amount of hard work. This was my first time leading a project as Tech Lead, and I feel truly fortunate. The future of Physical AI needs models that can not only “see” and “describe” the world, but also “imagine,” “simulate,” and “act”—and eventually close the loop with the real world. I hope Cosmos 3 can become an important starting point for this direction, and I’m excited to push Physical AI into its next stage together with the open-source community. Welcome to the era of Physical AI. HuggingFace: Project Website: Code:

This is THE moment of Physical AI! We are officially announcing Cosmos 3: Omnimodal World Models for Physical AI 🚀 - Cosmos 3 is an omnimodal world model: within a unified architecture, it can understand and generate language, images, video, audio, and actions. - It is not just a VLM, not just a video generator, not just an audio-visual generative model, and not just a physics simulator / world-action model. It can understand images and videos, generate images, videos, and audio, simulate future worlds, predict actions, and generate robot policies—enabling models to truly begin to “touch the world.” - Cosmos 3 is the #1 open-weight reasoner / T2I / I2V / robot policy across many benchmarks. Huge thanks to every teammate who fought side by side on this journey—from architecture, data, training, infra, serving, and evaluation to post-training. Every part of this project carries an incredible amount of hard work. This was my first time leading a project as Tech Lead, and I feel truly fortunate. The future of Physical AI needs models that can not only “see” and “describe” the world, but also “imagine,” “simulate,” and “act”—and eventually close the loop with the real world. I hope Cosmos 3 can become an important starting point for this direction, and I’m excited to push Physical AI into its next stage together with the open-source community. Welcome to the era of Physical AI. HuggingFace: Project Website: Code:

Max Zhaoshuo Li 李赵硕

1,079,880 views • 3 months ago