๐ฅ๐ฅSuper-Hard Poll folks๐ฅ๐ฅ
๐ This dilemma is driving me crazy. Vote: https://www.linkedin.com/posts/visionarynet_activity-7421974594917588992-YNAG
(and of course comment here)
๐ This dilemma is driving me crazy. Vote: https://www.linkedin.com/posts/visionarynet_activity-7421974594917588992-YNAG
(and of course comment here)
โค4๐1๐ฅ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐ปMLLMs Fine Segmentation๐ป
๐SimpleSeg: MLLMs with native pixel-level perception. Repo & Model available๐
๐Review https://t.ly/eVguh
๐Paper arxiv.org/pdf/2601.19228
๐Project simpleseg.github.io/
๐Repo github.com/songtianhui/SimpleSeg
๐SimpleSeg: MLLMs with native pixel-level perception. Repo & Model available๐
๐Review https://t.ly/eVguh
๐Paper arxiv.org/pdf/2601.19228
๐Project simpleseg.github.io/
๐Repo github.com/songtianhui/SimpleSeg
๐ฅ4๐3โค2๐1
๐ฅ DeepSeek-OCR 2 is out ๐ฅ
๐DeepSeek-AI announced the new version of its powerful SOTA OCR. A new architectural approach with the potential to achieve genuine 2D reasoning. Codes & weights๐
๐Review https://t.ly/gX4bX
๐Paper https://arxiv.org/pdf/2601.20552
๐Repo github.com/deepseek-ai/DeepSeek-OCR-2
๐DeepSeek-AI announced the new version of its powerful SOTA OCR. A new architectural approach with the potential to achieve genuine 2D reasoning. Codes & weights๐
๐Review https://t.ly/gX4bX
๐Paper https://arxiv.org/pdf/2601.20552
๐Repo github.com/deepseek-ai/DeepSeek-OCR-2
๐ฅ7โค4๐1
This media is not supported in your browser
VIEW IN TELEGRAM
๐ SOTA Style Transfer ๐
๐TeleAI unveils TeleStyle, a lightweight yet effective model for image/video stylization. Built upon Qwen-Image-Edit, TeleStyle leverages the base modelโs robust capabilities in content preservation & style customization. Code & Model released๐
๐Review https://t.ly/viVR0
๐Paper arxiv.org/pdf/2601.20175
๐Project tele-ai.github.io/TeleStyle/
๐Repo github.com/Tele-AI/TeleStyle
๐TeleAI unveils TeleStyle, a lightweight yet effective model for image/video stylization. Built upon Qwen-Image-Edit, TeleStyle leverages the base modelโs robust capabilities in content preservation & style customization. Code & Model released๐
๐Review https://t.ly/viVR0
๐Paper arxiv.org/pdf/2601.20175
๐Project tele-ai.github.io/TeleStyle/
๐Repo github.com/Tele-AI/TeleStyle
โค10๐2๐ฅ1๐คฏ1๐คฃ1
This media is not supported in your browser
VIEW IN TELEGRAM
๐ Metric Anything is out ๐
๐Metric Anything (Li Auto inc.) is a simple and scalable pretraining framework that learns metric depth from noisy, diverse 3D sources without manually engineered prompts, camera-specific modeling, or task-specific architectures. Impressive. Code announced ๐
๐Review https://t.ly/54Ccr
๐Paper arxiv.org/pdf/2601.22054
๐Project metric-anything.github.io/metric-anything-io/
๐Repo github.com/metric-anything/metric-anything
๐Metric Anything (Li Auto inc.) is a simple and scalable pretraining framework that learns metric depth from noisy, diverse 3D sources without manually engineered prompts, camera-specific modeling, or task-specific architectures. Impressive. Code announced ๐
๐Review https://t.ly/54Ccr
๐Paper arxiv.org/pdf/2601.22054
๐Project metric-anything.github.io/metric-anything-io/
๐Repo github.com/metric-anything/metric-anything
๐ฅ10โค5๐1
โค4
This media is not supported in your browser
VIEW IN TELEGRAM
๐Segment Any Events by Language๐
๐SEAL (by NUS) is the first Semantic-aware Segment Any Events framework that addresses Open-Vocabulary Event Instance Segmentation. Code announced๐
๐Review https://t.ly/1ZMF0
๐Paper https://arxiv.org/pdf/2601.23159
๐Project https://0nandon.github.io/SEAL/
๐Repo https://github.com/0nandon/SEAL
๐SEAL (by NUS) is the first Semantic-aware Segment Any Events framework that addresses Open-Vocabulary Event Instance Segmentation. Code announced๐
๐Review https://t.ly/1ZMF0
๐Paper https://arxiv.org/pdf/2601.23159
๐Project https://0nandon.github.io/SEAL/
๐Repo https://github.com/0nandon/SEAL
๐ฅ6โค3๐1๐คฏ1
๐RAM prices skyrocketing
๐Me acting like a rich kid.
Let's talk: https://www.linkedin.com/posts/visionarynet_ai-ram-ddr5-activity-7424127924020072448-NbaO
๐Me acting like a rich kid.
Let's talk: https://www.linkedin.com/posts/visionarynet_ai-ram-ddr5-activity-7424127924020072448-NbaO
๐คฃ18โค4๐ฅ1๐1
This media is not supported in your browser
VIEW IN TELEGRAM
๐ฎCoWTracker: Track-Warping๐ฎ
๐CoWTracker (VGG + META) is a novel dense point tracker that eschews cost volumes in favor of warping. Code/Models under FAIR NC๐
๐Review https://t.ly/6bAn9
๐Paper https://arxiv.org/pdf/2602.04877
๐Project https://cowtracker.github.io/
๐Repo https://github.com/facebookresearch/cowtracker
๐CoWTracker (VGG + META) is a novel dense point tracker that eschews cost volumes in favor of warping. Code/Models under FAIR NC๐
๐Review https://t.ly/6bAn9
๐Paper https://arxiv.org/pdf/2602.04877
๐Project https://cowtracker.github.io/
๐Repo https://github.com/facebookresearch/cowtracker
โค2๐ฅ2
This media is not supported in your browser
VIEW IN TELEGRAM
๐TrajVG Trajectory-Geometry๐
๐TrajVG is a novel reconstruction framework that makes cross-frame 3D correspondence an explicit prediction by estimating camera-coordinate 3D trajectories. Code announced๐
๐Review https://t.ly/yVi01
๐Paper arxiv.org/pdf/2602.04439
๐Project xingy038.github.io/TrajVG/
๐Repo github.com/xingy038/TrajVG
๐TrajVG is a novel reconstruction framework that makes cross-frame 3D correspondence an explicit prediction by estimating camera-coordinate 3D trajectories. Code announced๐
๐Review https://t.ly/yVi01
๐Paper arxiv.org/pdf/2602.04439
๐Project xingy038.github.io/TrajVG/
๐Repo github.com/xingy038/TrajVG
๐ฅ2โค1