MCPcopy Index your code
hub / github.com/huggingface/diffusers / numpy_to_pil

Method numpy_to_pil

src/diffusers/image_processor.py:995–1016  ·  view source on GitHub ↗

r""" Convert a NumPy image or a batch of images to a list of PIL images. Args: images (`np.ndarray`): The input NumPy array of images, which can be a single image or a batch. Returns: `list[PIL.Image.Image]`: A list of

(images: np.ndarray)

Source from the content-addressed store, hash-verified

993
994 @staticmethod
995 def numpy_to_pil(images: np.ndarray) -> list[PIL.Image.Image]:
996 r"""
997 Convert a NumPy image or a batch of images to a list of PIL images.
998
999 Args:
1000 images (`np.ndarray`):
1001 The input NumPy array of images, which can be a single image or a batch.
1002
1003 Returns:
1004 `list[PIL.Image.Image]`:
1005 A list of PIL images converted from the input NumPy array.
1006 """
1007 if images.ndim == 3:
1008 images = images[None, ...]
1009 images = (images * 255).round().astype("uint8")
1010 if images.shape[-1] == 1:
1011 # special case for grayscale (single channel) images
1012 pil_images = [Image.fromarray(image.squeeze(), mode="L") for image in images]
1013 else:
1014 pil_images = [Image.fromarray(image[:, :, :3]) for image in images]
1015
1016 return pil_images
1017
1018 @staticmethod
1019 def depth_pil_to_numpy(images: list[PIL.Image.Image] | PIL.Image.Image) -> np.ndarray:

Callers 1

postprocessMethod · 0.95

Calls

no outgoing calls

Tested by

no test coverage detected