uv-scripts/sam3
SAM3 Vision Scripts Detect and segment objects in images using Meta's SAM3 (Segment Anything Model 3) with text prompts. Process HuggingFace datasets with zero-shot detection and segmentation using natural language descriptions. Script What it does Output detect-objects.py Object detection with bounding boxes objects column with bbox, category, score segment-objects.py Pixel-level segmentation masks Segmentation maps or per-instance masks Browse results… See the full description on the dataset page: https://huggingface.co/datasets/uv-scripts/sam3.
Sync from GitHub via hub-sync
Add segment-objects.py for pixel-level image segmentation
Add hf-jobs tag to README frontmatter
Update README.md
Improve formatting in README for better readability and consistency
Fix formatting in detection output message for better readability
Update example detection image in README for better responsiveness and visibility
Increase height of example detection image in README for improved visibility
Reduce height of example detection image in README for better display
Update example detection image in README to use HTML for better display
Update example detection image in README to use HTML for better display
Add example output section and image to README for object detection results
Add README.md for SAM3 object detection with usage instructions and examples
Add dataset card creation and visualization script for object detection results
Convert boxes and scores to float32 before numpy conversion in process_batch for compatibility
Refactor imports for better organization and enable high-performance mode
Refactor process_batch to return detections in list format for improved consistency
Refactor output format to use dict-of-lists for object detections and update output schema accordingly
Uncomment features parameter in object detection function call
Comment out features parameter in object detection function call
Refactor output schema to use list for objects and simplify feature definitions
Refactor output schema creation to define features before processing dataset
Enhance batch processing to set tensor dtype for model compatibility
Refactor image processing in batch to handle grayscale images and unify prompts for model input
Refactor object detection script to use single class detection and update argument naming
Enhance input processing by moving tensors to the model's device and converting to the appropriate dtype
Refactor
Comment out torchvision and accelerate dependencies in detect-objects.py
draft]
initial commit