Fix: Align model and input data types to float16 for inference

README.md

@@ -62,7 +62,9 @@ input_points = [[[450, 600]]] # 2D localization of a window
 
 
 ```python
+model = model.to("cuda").half()
 inputs = processor(raw_image, input_points=input_points, return_tensors="pt").to("cuda")
+inputs = {k: v.half() if isinstance(v, torch.Tensor) and v.dtype == torch.float32 else v for k, v in inputs.items()}
 outputs = model(**inputs)
 masks = processor.image_processor.post_process_masks(outputs.pred_masks.cpu(), inputs["original_sizes"].cpu(), inputs["reshaped_input_sizes"].cpu())
 scores = outputs.iou_scores