Tool Text To Seg

Performs text-to-instance segmentation to detect and segment objects based on text descriptions.

Args: data (FormData): The form data containing the following fields: - prompt (str): Text description of objects to detect and segment. - model (TextToSegModel): The segmentation model to use. Currently supports florence2sam2. - image: Image file to analyze for segmentation. - video: Video file to analyze for temporal segmentation. - chunk_length_frames: Number of frames per chunk for video processing. - iou_threshold: IoU threshold for filtering detections. Range: 0.1-1.0. - nms_threshold: Non-maximum suppression threshold. Range: 0.1-1.0. - job_id (UUID, optional): Unique identifier for tracking the job. baseten_inference_sender: Dependency Injection for sending inference requests to the Baseten model server for text-to-segmentation processing.

Returns: TextToInstanceSegmentationResponse | JSONResponse Success: TextToInstanceSegmentationResponse containing segmented instances with masks. Error: JSONResponse with error details and appropriate status code.