o ZhZ @sddlmZmZmZmZddlmZddlmZddl m Z m Z m Z e r)ddl Z e eZ ddeeefd ed eed ed eeeff d dZGdddeZdgZdS))ListOptionalTupleUnion)DetrImageProcessorFast)center_to_corners_format) TensorTypeis_torch_availableloggingN image_sizesizemax_sizemod_sizereturnc CsJ|\}}d}|dur.tt||f}tt||f}||||kr.|||}tt|}||krN|} |durE|durEt|||} n:t|||} n1||krV||ks^||krd||krd||} } n|} |durw|durwt|||} nt|||} |durtt| |} tt| |} | | } | | } | | fS)a Computes the output image size given the input image size and the desired output size with multiple of divisible_size. Args: image_size (`Tuple[int, int]`): The input image size. size (`int`): The desired output size. max_size (`int`, *optional*): The maximum allowed output size. mod_size (`int`, *optional*): The size to make multiple of mod_size. N) floatminmaxintroundtorch remainderZtensoritem) r rrrheightwidthZraw_sizeZmin_original_sizeZmax_original_sizeowZohZow_modZoh_modrV/var/www/auris/lib/python3.10/site-packages/transformers/models/yolos/modular_yolos.pyget_size_with_aspect_ratios2    rc@sjeZdZddZ ddedeeeefde fd d Z d d Z d dZ ddZ ddZddZddZdS)YolosImageProcessorFastc Cstd|j|j}}t|t|krtd|jddkr#td|}tj | |jddddd \}}|}tj ||jdd d } ||jd} t |} t | d| dddd } |d\} } tj| | | | gdd }| |d d d d d f} ddt|| | D}|S)a[ Converts the raw output of [`YolosForObjectDetection`] into final bounding boxes in (top_left_x, top_left_y, bottom_right_x, bottom_right_y) format. Only supports PyTorch. Args: outputs ([`YolosObjectDetectionOutput`]): Raw outputs of the model. target_sizes (`torch.Tensor` of shape `(batch_size, 2)`): Tensor containing the size (height, width) of each image of the batch. For evaluation, this must be the original image size (before any data augmentation). For visualization, this should be the image size after data augment, but before padding. Returns: `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image in the batch as predicted by the model. z`post_process` is deprecated and will be removed in v5 of Transformers, please use `post_process_object_detection` instead, with `threshold=0.` for equivalent results.TMake sure that you pass in as many target sizes as the batch dimension of the logitszTEach element of target_sizes must contain the size (h, w) of each image of the batchrddimfloorZ rounding_modeNcSsg|] \}}}|||dqS)scoreslabelsboxesr).0slbrrr osz8YolosImageProcessorFast.post_process..)loggerZ warning_oncelogits pred_boxeslen ValueErrorshapesigmoidrtopkviewdivrgather unsqueezerepeatunbindstackzip)selfoutputs target_sizes out_logitsout_bboxprob topk_values topk_indexesr, topk_boxesr-r.img_himg_w scale_fctresultsrrr post_processFs("z$YolosImageProcessorFast.post_process?Nr% thresholdrFtop_kc Cs|j|j}}|durt|t|krtd|}||jdd}t||d}t j ||dd\} } | } t j | |jddd } | |jd} t |}t |d| dddd }|durt|tr{t d d |D}t d d |D}n|d\}}t j||||gdd|j}||dddddf}g}t| | |D] \}}}|||k}|||k}|||k}||||dq|S)a  Converts the raw output of [`YolosForObjectDetection`] into final bounding boxes in (top_left_x, top_left_y, bottom_right_x, bottom_right_y) format. Only supports PyTorch. Args: outputs ([`YolosObjectDetectionOutput`]): Raw outputs of the model. threshold (`float`, *optional*): Score threshold to keep object detection predictions. target_sizes (`torch.Tensor` or `List[Tuple[int, int]]`, *optional*): Tensor of shape `(batch_size, 2)` or list of tuples (`Tuple[int, int]`) containing the target size (height, width) of each image in the batch. If left to None, predictions will not be resized. top_k (`int`, *optional*, defaults to 100): Keep only top k bounding boxes before filtering by thresholding. Returns: `List[Dict]`: A list of dictionaries, each dictionary containing the scores, labels and boxes for an image in the batch as predicted by the model. Nr!rr$r"r&r#r(r)r*cSg|]}|dqS)rrr/irrrr3zIYolosImageProcessorFast.post_process_object_detection..cSrU)r"rrVrrrr3rXr+)r5r6r7r8r:r<r9rrrr;r=rr>r?r@ isinstancerZTensorrArBtoZdevicerCappend)rDrErSrFrTrGrHrIZk_valuerJrKr,rLr-r.rMrNrOrPr0r1r2Zscorelabelboxrrrpost_process_object_detectionss:    z5YolosImageProcessorFast.post_process_object_detectioncCtdNzHSegmentation post-processing is not implemented for Deformable DETR yet.NotImplementedErrorrrrrpost_process_segmentationz1YolosImageProcessorFast.post_process_segmentationcCr_)NzDInstance post-processing is not implemented for Deformable DETR yet.rarrrrpost_process_instancerdz-YolosImageProcessorFast.post_process_instancecCr_)NzDPanoptic post-processing is not implemented for Deformable DETR yet.rarrrrpost_process_panopticrdz-YolosImageProcessorFast.post_process_panopticcCr_r`rarrrr"post_process_instance_segmentationrdz:YolosImageProcessorFast.post_process_instance_segmentationcCr_)NzQSemantic segmentation post-processing is not implemented for Deformable DETR yet.rarrrr"post_process_semantic_segmentationrdz:YolosImageProcessorFast.post_process_semantic_segmentationcCr_)NzQPanoptic segmentation post-processing is not implemented for Deformable DETR yet.rarrrr"post_process_panoptic_segmentationrdz:YolosImageProcessorFast.post_process_panoptic_segmentation)rRNr%)__name__ __module__ __qualname__rQrrr rrrr^rcrerfrgrhrirrrrr Es . ; r )Nr )typingrrrrZ3transformers.models.detr.image_processing_detr_fastrZimage_transformsrutilsr r r rZ get_loggerrjr4rrr __all__rrrrs*      1 |