o Zh>@sRdZddlZddlmZddlmZmZmZmZddl Z ddl m Z m Z ddlmZmZmZddlmZmZmZmZmZmZmZmZmZdd lmZmZmZm Z e !e"Z#er^dd l$m%Z%ed d  d de&de&de&de&de'dee&e&ff ddZ(de j)de&de j)fddZ*d!de j)de&de&dee j)e j)ffddZ+Gddde Z,dgZ-dS)"z"Image processor class for SigLIP2.N) lru_cache)ListOptionalTupleUnion)BaseImageProcessor BatchFeature)convert_to_rgbresizeto_channel_dimension_format) ChannelDimension ImageInputPILImageResamplinginfer_channel_dimension_formatis_scaled_imagemake_flat_list_of_imagesto_numpy_array valid_imagesvalidate_preprocess_arguments) TensorTypefilter_out_non_signature_kwargsis_vision_availablelogging)Image)maxsizeh㈵> image_height image_width patch_sizemax_num_patchesepsreturnc Csdtdtdtdtfdd}|dd}}|||krC||d }||||} ||||} | || |} | |kr;|}n|}|||ks|}||||} ||||} | | fS) a" Determine image size based on max number of patches, ensure dimensions are divisible by patch size and image is at least 1 patch. Args: image_height (`int`): Original image height. image_width (`int`): Original image width. patch_size (`int`): Patch size for processing. max_num_patches (`int`): Maximum number of patches. eps (`float`): Small threshold for binary search. Returns: Tuple: (target_height, target_width) scalesizer r#cSs,||}t|||}t||}t|S)N)mathceilmaxint)r$r%r Z scaled_sizer*c/var/www/auris/lib/python3.10/site-packages/transformers/models/siglip2/image_processing_siglip2.pyget_scaled_image_sizeIs zAget_image_size_for_max_num_patches..get_scaled_image_size gY@)floatr)) rrr r!r"r,Z scale_minZ scale_maxr$Z target_heightZ target_widthZ num_patchesr*r*r+"get_image_size_for_max_num_patches2s       r0imagecCsT|j\}}}||}||}||||||}|ddddd}|||d}|S)z Convert 3D array image of shape (image_height, image_width, num_channels) into 2D array of patches of shape (num_patches_height * num_patches_width, patch_size * patch_size * num_channels). rr.r)shapeZreshapeZ transpose)r1r rrZ num_channelsnum_patches_heightnum_patches_widthZ patched_imager*r*r+convert_image_to_patchesbs r8array target_length pad_valuecCsn|jd}||}tj|ftjd}|dkr3d|fgdg|jd}tj||d|d}d|| d<||fS)z2 Pad the array along the first dimension. r)Zdtype)rrr2Zconstant)modeZconstant_valuesN)r5npZonesZint32ndimpad)r9r:r;Zcurrent_lengthZpadding_lengthmaskZpaddingsr*r*r+pad_along_first_dimps rAcsHeZdZdZgdZdejddddddddf ded d d ed ed ede e ee efde e ee efde ede de ffdd Z e            ddede ed e d d e ed e ed e ede e ee efde e ee efde e eefde e eefde ede e de e ddfddZZS)Siglip2ImageProcessora3 Constructs a SigLIP2 image processor. Args: do_resize (`bool`, *optional*, defaults to `True`): Whether to resize the image's dimensions to fit `max_num_patches` according to given `patch_size`. Can be overridden by `do_resize` in the `preprocess` method. resample (`PILImageResampling`, *optional*, defaults to `Resampling.BILINEAR`): Resampling filter to use if resizing the image. Can be overridden by `resample` in the `preprocess` method. do_rescale (`bool`, *optional*, defaults to `True`): Whether to rescale the image by the specified scale `rescale_factor`. Can be overridden by `do_rescale` in the `preprocess` method. rescale_factor (`int` or `float`, *optional*, defaults to `1/255`): Scale factor to use if rescaling the image. Can be overridden by `rescale_factor` in the `preprocess` method. do_normalize (`bool`, *optional*, defaults to `True`): Whether to normalize the image by the specified mean and standard deviation. Can be overridden by `do_normalize` in the `preprocess` method. image_mean (`float` or `List[float]`, *optional*, defaults to `[0.5, 0.5, 0.5]`): Mean to use if normalizing the image. This is a float or list of floats the length of the number of channels in the image. Can be overridden by the `image_mean` parameter in the `preprocess` method. image_std (`float` or `List[float]`, *optional*, defaults to `[0.5, 0.5, 0.5]`): Standard deviation to use if normalizing the image. This is a float or list of floats the length of the number of channels in the image. Can be overridden by the `image_std` parameter in the `preprocess` method. Can be overridden by the `image_std` parameter in the `preprocess` method. do_convert_rgb (`bool`, *optional*, defaults to `True`): Whether to convert the image to RGB. patch_size (`int`, *optional*, defaults to 16): The size (resolution) of each patch the image will be split to. max_num_patches (`int`, *optional*, defaults to 256): The image will be resized to have at most this number of patches, and then padded in "patch" dimension to match this number exactly.  pixel_valuesZpixel_attention_maskspatial_shapesTgp?Nr do_resizeresampler do_rescalerescale_factor do_normalize image_mean image_stddo_convert_rgbr r!c  sztjdi| |dur|ngd}|dur|ngd}||_||_||_||_||_||_||_||_ | |_ | |_ dS)N)?rOrOr*) super__init__rGrHrIrJrKrLrMrNr r!) selfrGrHrIrJrKrLrMrNr r!kwargs __class__r*r+rQs zSiglip2ImageProcessor.__init__imagesreturn_tensorsinput_data_formatr#z Image.ImagecCs&|dur|n|j}|dur|n|j}|dur|n|j}|dur!|n|j}|dur*|n|j}|dur3|n|j}|dur<|n|j}| durE| n|j} | durN| n|j} | durW| n|j } t j }t |}t |sitdt|||||d| r{dd|D}dd|D}|rt|drtd| durt|d} g}g}g}|D]e}t||| d }|rt|jd|jd | | d \}}t|||f||d }|r|j|||d }|r|j||||d}t|| }t|| \}}|jd| }|jd | }|||f||||qt|||d| d}|S)a Preprocess an image or batch of images. Args: images (`ImageInput`): Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If passing in images with pixel values between 0 and 1, set `do_rescale=False`. do_resize (`bool`, *optional*, defaults to `self.do_resize`): Whether to resize the image. size (`Dict[str, int]`, *optional*, defaults to `self.size`): Size of the image after resizing. resample (`int`, *optional*, defaults to `self.resample`): Resampling filter to use if resizing the image. This can be one of the enum `PILImageResampling`. Only has an effect if `do_resize` is set to `True`. do_rescale (`bool`, *optional*, defaults to `self.do_rescale`): Whether to rescale the image. rescale_factor (`float`, *optional*, defaults to `self.rescale_factor`): Rescale factor to rescale the image by if `do_rescale` is set to `True`. do_normalize (`bool`, *optional*, defaults to `self.do_normalize`): Whether to normalize the image. image_mean (`float` or `List[float]`, *optional*, defaults to `self.image_mean`): Image mean to use for normalization. Only has an effect if `do_normalize` is set to `True`. image_std (`float` or `List[float]`, *optional*, defaults to `self.image_std`): Image standard deviation to use for normalization. Only has an effect if `do_normalize` is set to `True`. return_tensors (`str` or `TensorType`, *optional*): The type of tensors to return. Can be one of: - Unset: Return a list of `np.ndarray`. - `TensorType.TENSORFLOW` or `'tf'`: Return a batch of type `tf.Tensor`. - `TensorType.PYTORCH` or `'pt'`: Return a batch of type `torch.Tensor`. - `TensorType.NUMPY` or `'np'`: Return a batch of type `np.ndarray`. - `TensorType.JAX` or `'jax'`: Return a batch of type `jax.numpy.ndarray`. input_data_format (`ChannelDimension` or `str`, *optional*): The channel dimension format for the input image. If unset, the channel dimension format is inferred from the input image. Can be one of: - `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format. - `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format. - `"none"` or `ChannelDimension.NONE`: image in (height, width) format. do_convert_rgb (`bool`, *optional*, defaults to `self.do_convert_rgb`): Whether to convert the image to RGB. patch_size (`int`, *optional*, defaults to `self.patch_size`): Patch size for processing, same as the patch size used in the model. max_num_patches (`int`, *optional*, defaults to `self.max_num_patches`): Maximum number of patches per image, the image will be resized to have at most this number of patches. NzkInvalid image type. Must be of type PIL.Image.Image, numpy.ndarray, torch.Tensor, tf.Tensor or jax.ndarray.)rIrJrKrLrMcSg|]}t|qSr*)r .0r1r*r*r+ z4Siglip2ImageProcessor.preprocess..cSrYr*)rrZr*r*r+r\ r]rzIt looks like you are trying to rescale already rescaled images. If the input images have pixel values between 0 and 1, set `do_rescale=False` to avoid rescaling them again.)Zinput_channel_dimr2)rrr r!)r1r%rHrX)r1r$rX)r1meanZstdrXrC)dataZ tensor_type)rGrHrIrJrKrLrMrNr r!r ZLASTrr ValueErrorrrloggerZ warning_oncerr r0r5r Zrescale normalizer8rAappendr )rRrVrGrHrIrJrKrLrMrWrXrNr r!Z data_formatZ pixel_masksrDrEr1heightwidthZpatchesr@r6r7Z batch_featurer*r*r+ preprocesss>      z Siglip2ImageProcessor.preprocess) NNNNNNNNNNNN)__name__ __module__ __qualname____doc__Zmodel_input_namesrZBILINEARboolr/rrrr)rQrrstrrr rf __classcell__r*r*rTr+rB~s"        rB)r)r).rjr& functoolsrtypingrrrrnumpyr=Zimage_processing_utilsrr Zimage_transformsr r r Z image_utilsr rrrrrrrrutilsrrrrZ get_loggerrgraZPILrr)r/r0Zndarrayr8rArB__all__r*r*r*r+s@ ,    /* Z