@VincentLogic: NVIDIA's newly open-sourced LocateAnything model is really impressive. The previous visual grounding models generated coordinates digit by digit (like squeezing toothpaste), slow and unstable. This new model uses "parallel bounding box decoding" to predict complete coordinates in one step, much faster and more accurate...

NVIDIA just open-sourced the LocateAnything model, which is really impressive. In the past, visual grounding models generated coordinates one digit at a time (like squeezing toothpaste), which was slow and unstable. This new model uses "parallel bounding box decoding" to directly predict complete coordinates in one step, making it much faster and more accurate. Whether it's locating objects in videos, recognizing UI interfaces, or OCR text, it can handle it all. Most importantly, the model is very small, with only 3B parameters (about 7.8GB), and can run locally on consumer-grade GPUs! If you work in computer vision or multimodal AI, you must try this. The project is open-source, first come first served!