§
    l¬¸j¸p  ã                  óÞ   — d dl mZ d dlZd dlmZ d dlmZ d dlZd dl	Z	d dl
mZ d dlmZmZmZ d dlmZ d dlmZ d dlmZmZmZmZmZ d d	lmZ d d
lmZmZm Z  d dl!m"Z"  G d„ de¦  «        Z#dS )é    )ÚannotationsN)ÚPath)ÚAny)Úbuild_dataloaderÚbuild_yolo_datasetÚ	converter)Úget_split_fraction)ÚBaseValidator)ÚDEFAULT_CFGÚLOGGERÚRANKÚnmsÚops)Úcheck_requirements)ÚConfusionMatrixÚ
DetMetricsÚbox_iou)Úplot_imagesc                  óú   ‡ — e Zd ZdZdJdKˆ fd„ZedLd
„¦   «         ZdMd„ZdNd„ZdOd„Z	dPd„Z
dQd„ZdRd„ZdSd„ZdTd„ZdTd„ZdTd „ZdUd!„ZdTd"„ZdVd$„ZdWdXd*„ZdYd.„ZdZd0„Z	 d[d\d2„Zd]d:„Zd^d<„Zd_d=„Zd`d?„Z	 	 dadbdI„Zˆ xZS )cÚDetectionValidatora)  A class extending the BaseValidator class for validation based on a detection model.

    This class implements validation functionality specific to object detection tasks, including metrics calculation,
    prediction processing, and visualization of results.

    Attributes:
        is_coco (bool): Whether the dataset is COCO.
        is_lvis (bool): Whether the dataset is LVIS.
        class_map (list[int]): Mapping from model class indices to dataset class indices.
        metrics (DetMetrics): Object detection metrics calculator.
        iouv (torch.Tensor): IoU thresholds for mAP calculation.
        niou (int): Number of IoU thresholds.
        jdict (list[dict[str, Any]]): List for storing JSON detection results.
        stats (dict[str, list[torch.Tensor]]): Dictionary for storing statistics during validation.

    Examples:
        >>> from ultralytics.models.yolo.detect import DetectionValidator
        >>> args = dict(model="yolo26n.pt", data="coco8.yaml")
        >>> validator = DetectionValidator(args=args)
        >>> validator()
    NÚ
_callbacksúdict | NoneÚreturnÚNonec                ó¶  •— t          |t          ¦  «        r|                     d¦  «        nt          |dd¦  «        }|€dn|| _        t          ¦   «                              ||||¦  «         d| _        d| _        d| _	        d| j
        _        t          j        ddd¦  «        | _        | j                             ¦   «         | _        t#          ¦   «         | _        dS )	a–  Initialize detection validator with necessary variables and settings.

        Args:
            dataloader (torch.utils.data.DataLoader, optional): DataLoader to use for validation.
            save_dir (Path, optional): Directory to save results.
            args (dict[str, Any], optional): Arguments for the validator.
            _callbacks (dict, optional): Dictionary of callback functions.
        ÚconfNg      Ð?FÚdetectg      à?gffffffî?é
   )Ú
isinstanceÚdictÚgetÚgetattrÚconfusion_matrix_confÚsuperÚ__init__Úis_cocoÚis_lvisÚ	class_mapÚargsÚtaskÚtorchÚlinspaceÚiouvÚnumelÚniour   Úmetrics)ÚselfÚ
dataloaderÚsave_dirr)   r   r   Ú	__class__s         €úr/root/private/project/pallet_detection_v2/.venv/lib/python3.11/site-packages/ultralytics/models/yolo/detect/val.pyr%   zDetectionValidator.__init__-   sº   ø€ õ $.¨dµDÑ#9Ô#9ÐZˆt�xŠx˜ÑÔÐ½wÀtÈVÐUYÑ?ZÔ?ZˆØ-1¨\ T T¸tˆÔ"Ý‰Œ×Ò˜ X¨t°ZÑ@Ô@Ð@ØˆŒØˆŒØˆŒØ!ˆŒ	ŒÝ”N 3¨¨bÑ1Ô1ˆŒ	Ø”I—O’OÑ%Ô%ˆŒ	Ý!‘|”|ˆŒˆˆó    Údatasetsú#dict[str, torch.utils.data.Dataset]c                óÈ  — d„ |                      ¦   «         D ¦   «         }t          |                     ¦   «         ¦  «        }|| j        k    rdS d                     d„ |                      ¦   «         D ¦   «         ¦  «        }d|› d|› d| j        › d�}| j        t
          j        k    r|| _        |d	|› d
�z  }n|d| j        › d�z  }t          dv rt          j        |¦  «         dS dS )zcWarn when dataset object counts exceed max_det and raise the default limit to the observed maximum.c                ól   — i | ]1\  }}|t          d „ t          |d|g¦  «        D ¦   «         d¬¦  «        “Œ2S )c              3  óî   K  — | ]p}t          |d ¦  «        r^|j        D ]V}t          |t          ¦  «        r?t	          |                     d¦  «        dd¦  «        dk    ¯?t          |d         ¦  «        V — ŒWŒqdS )ÚlabelsÚclsÚndimr   N)Úhasattrr<   r   r    r"   r!   Úlen)Ú.0ÚsubsetÚlabels      r5   ú	<genexpr>z?DetectionValidator._check_max_det.<locals>.<dictcomp>.<genexpr>F   s¡   è è € ð ð àÝ˜v xÑ0Ô0ðð "(¤ð	ð ð Ý! %­Ñ.Ô.ðõ
 4;¸5¿9º9ÀUÑ;KÔ;KÈVÐUVÑ3WÔ3WÐZ[Ò3[Ð3[õ	 ˜˜eœÑ%Ô%ð 4\Ð3[Ð3[Ð3[Ð3[ðð r6   r7   r   )Údefault)Úmaxr"   )rA   ÚsplitÚdatasets      r5   ú
<dictcomp>z5DetectionValidator._check_max_det.<locals>.<dictcomp>D   sn   € ð 
ð 
ð 
ñ ��wð •3ðð å")¨'°:À¸yÑ"IÔ"Iðñ ô ð ð	ñ 	ô 	ð
ð 
ð 
r6   Nz, c              3  ó*   K  — | ]\  }}|› d |› �V — ŒdS )ú=N© )rA   rG   Úcounts      r5   rD   z4DetectionValidator._check_max_det.<locals>.<genexpr>U   s4   è è € Ð WÐ W¹¸¸u EÐ!3Ð!3¨EÐ!3Ð!3Ð WÐ WÐ WÐ WÐ WÐ Wr6   zDataset images contain up to z
 objects (z), but max_det=z². This mismatch can cap recall and produce invalid validation metrics. Raising it may increase validation cost but cannot increase model or export capacity, which may cap recall.z Setting max_det=z to match the observed maximum.z$ Keeping the user-specified max_det=ú.>   r   éÿÿÿÿ)	ÚitemsrF   ÚvaluesÚmax_detÚjoinr   r   r   Úwarning)r)   r7   ÚmaximaÚobservedÚsplit_countsÚmessages         r5   Ú_check_max_detz!DetectionValidator._check_max_detA   s+  € ð
ð 
ð #+§.¢.Ñ"2Ô"2ð
ñ 
ô 
ˆõ �v—}’}‘”Ñ'Ô'ˆØ�t”|Ò#Ð#ØˆFà—y’yÐ WÐ WÈÏÊÉÌÐ WÑ WÔ WÑWÔWˆð{¨Hð {ð {Àð {ð {Ð]aÔ]ið {ð {ð {ð 	ð
 Œ<�;Ô.Ò.Ð.Ø#ˆDŒLØÐT¨8ÐTÐTÐTÑTˆGˆGàÐM¸d¼lÐMÐMÐMÑMˆGÝ�7ˆ?ˆ?ÝŒN˜7Ñ#Ô#Ð#Ð#Ð#ð ˆ?r6   Úbatchúdict[str, Any]c                óX  — |                      ¦   «         D ]J\  }}t          |t          j        ¦  «        r+|                     | j        | j        j        dv¬¦  «        ||<   ŒK| j        j        dk    r|d          	                    ¦   «         n|d          
                    ¦   «         dz  |d<   |S )zØPreprocess batch of images for YOLO validation.

        Args:
            batch (dict[str, Any]): Batch containing images and annotations.

        Returns:
            (dict[str, Any]): Preprocessed batch.
        >   ÚcpuÚmps)Únon_blockingé   Úimgéÿ   )rP   r   r+   ÚTensorÚtoÚdeviceÚtyper)   ÚquantizeÚhalfÚfloat)r1   rZ   ÚkÚvs       r5   Ú
preprocesszDetectionValidator.preprocessc   s¦   € ð —K’K‘M”Mð 	bð 	b‰DˆAˆqÝ˜!�Uœ\Ñ*Ô*ð bØŸ4š4 ¤¸$¼+Ô:JÐR`Ð:`˜4ÑaÔa��a‘øØ/3¬yÔ/AÀRÒ/GÐ/G˜˜eœ×)Ò)Ñ+Ô+Ð+ÈUÐSXÌ\×M_ÒM_ÑMaÔMaÐehÑhˆˆe‰Øˆr6   Úmodelútorch.nn.Modulec                óZ  — | j         s3|                      | j        | j        j        pd| j        j        i¦  «         | j                             | j        j        d¦  «        }t          |t          ¦  «        o4d|v o0| 
                    t          j        › d�t          j        › d�f¦  «        | _        t          |t          ¦  «        od|v o| j         | _        | j        rt          j        ¦   «         n1t#          t%          dt'          |j        ¦  «        dz   ¦  «        ¦  «        | _        | j        xj        | j        j        o| j        s| j        o| j          z  c_        |j        | _        t'          |j        ¦  «        | _        t3          |dd	¦  «        | _        t3          |d
d¦  «        dk    r|j        n|}| j        r;t9          |d¦  «        r+|                     | j        j        | j        j        ¬¦  «         d| _         g | _!        | j        j        o| j        j"        dk    o| j        p| j         | _#        | j#        rt3          | dd¦  «        nd| _$        | j#        o| j$        du | _%        | j#        rt#          | j        j&        ¦  «        nd| _'        g | _(        | j%        rg g d„ | j        D ¦   «         dœ| _$        |j        | j)        _        | j)         *                    ¦   «          | j)         +                    ¦   «          tY          |j        | j        j-        o| j        j.        ¬¦  «        | _/        dS )zŒInitialize evaluation metrics for YOLO detection validation.

        Args:
            model (torch.nn.Module): Model to validate.
        ÚvalÚ Úcocozval2017.txtztest-dev2017.txtÚlvisé   Úend2endFÚformatNÚptÚset_head_attr)rR   Úagnostic_nmsr   r   Úgdictc                ó   — g | ]}d |i‘ŒS )ÚidrL   ©rA   Úxs     r5   ú
<listcomp>z3DetectionValidator.init_metrics.<locals>.<listcomp>‘   s   € ÐIlÐIlÐIlÐXYÈ4ÐQRÈ)ÐIlÐIlÐIlr6   )Úimagesr   Ú
categories)ÚnamesÚsave_matches)0ÚtrainingrY   r)   rG   r2   rH   Údatar!   r   ÚstrÚendswithÚosÚsepr&   r'   r   Úcoco80_to_coco91_classÚlistÚranger@   r‚   r(   Ú	save_jsonrp   Úncr"   ru   rm   r?   rx   rR   ry   ÚseenÚjdictr*   Úis_custom_jsonrz   Úbuild_gdictÚsamplerÚeval_idsÚpred_countsr0   Úclear_statsÚclear_image_metricsr   ÚplotsÚ	visualizeÚconfusion_matrix)r1   rm   rp   Únative_models       r5   Úinit_metricszDetectionValidator.init_metricsr   s  € ð Œ}ð 	`Ø×Ò ¤	¨D¬I¬OÐ,D¸uÀdÄoÔF]Ð+^Ñ_Ô_Ð_ØŒi�mŠm˜DœIœO¨RÑ0Ô0ˆå�s�CÑ Ô ð VØ˜#�ðVà—’¥"¤&Ð5Ð5Ð5½"¼&Ð7RÐ7RÐ7RÐSÑTÔTð 	Œõ
 " #¥sÑ+Ô+ÐR°¸#°ÐRÀdÄlÐBRˆŒØ?C¼|Ðu�Ô9Ñ;Ô;Ð;ÕQUÕV[Ð\]Õ_bÐchÔcnÑ_oÔ_oÐrsÑ_sÑVtÔVtÑQuÔQuˆŒØŒ	ÐÔ˜tœyœ}Ðe°$´,Ð2NÀ$Ä,ÐeÐX\ÔXeÐTeÑeÐÔØ”[ˆŒ
Ý�e”kÑ"Ô"ˆŒÝ˜u i°Ñ7Ô7ˆŒÝ&-¨e°X¸tÑ&DÔ&DÈÒ&LÐ&L�u”{�{ÐRWˆØŒ<ð 	g�G L°/ÑBÔBð 	gØ×&Ò&¨t¬yÔ/@ÈtÌyÔOeÐ&ÑfÔfÐfØˆŒ	ØˆŒ
Ø"œiÔ1Ðw°d´i´nÈÒ6PÐwÐZ^ÔZfÐZvÐjnÔjvÐUwˆÔØ59Ô5HÐR•W˜T 7¨DÑ1Ô1Ð1ÈdˆŒ
ØÔ.ÐE°4´:ÀÐ3EˆÔØ9=Ô9LÐV�˜Tœ_Ô4Ñ5Ô5Ð5ÐRVˆŒØˆÔØÔð 	nØ$&°rÐIlÐIlÐ]aÔ]kÐIlÑIlÔIlÐmÐmˆDŒJØ"œ[ˆŒÔØŒ× Ò Ñ"Ô"Ð"ØŒ×(Ò(Ñ*Ô*Ð*Ý /°e´kÐPTÔPYÔP_ÐPwÐdhÔdmÔdwÐ xÑ xÔ xˆÔÐÐr6   r†   c                ó   — ddz  S )zBReturn a formatted string summarizing class metrics of YOLO model.z%22s%11s%11s%11s%11s%11s%11s)ÚClassÚImagesÚ	InstanceszBox(PÚRÚmAP50z	mAP50-95)rL   )r1   s    r5   Úget_desczDetectionValidator.get_desc—   s   € à#Ð'kÑkÐkr6   Úpredsútorch.Tensorúlist[dict[str, torch.Tensor]]c                ó  — t          j        || j        j        | j        j        | j        j        dk    rdn| j        d| j        j        p| j        j        | j        j	        | j
        | j        j        dk    ¬¦	  «	        }d„ |D ¦   «         S )aE  Apply Non-maximum suppression to prediction outputs.

        Args:
            preds (torch.Tensor): Raw predictions from the model.

        Returns:
            (list[dict[str, torch.Tensor]]): Processed predictions after NMS, where each dict contains 'bboxes', 'conf',
                'cls', and 'extra' tensors.
        r   r   TÚobb)rŽ   Úmulti_labelÚagnosticrR   ru   Úrotatedc           	     ót   — g | ]5}|d d …d d…f         |d d …df         |d d …df         |d d …dd …f         dœ‘Œ6S )Né   é   é   )Úbboxesr   r=   ÚextrarL   r}   s     r5   r   z2DetectionValidator.postprocess.<locals>.<listcomp>°   sd   € ÐjÐjÐjÐ]^˜1˜Q˜Q˜Q   ˜Uœ8¨Q¨q¨q¨q°!¨t¬W¸Q¸q¸q¸qÀ!¸t¼WÈqÐQRÐQRÐQRÐTUÐTVÐTVÐQVÌxÐXÐXÐjÐjÐjr6   )r   Únon_max_suppressionr)   r   Úiour*   rŽ   Ú
single_clsry   rR   ru   )r1   r¤   Úoutputss      r5   ÚpostprocesszDetectionValidator.postprocess›   s‰   € õ Ô)ØØŒIŒNØŒIŒMØ”I”N hÒ.Ð.ˆqˆq°D´GØØ”YÔ)ÐC¨T¬YÔ-CØ”IÔ%Ø”LØ”I”N eÒ+ð

ñ 

ô 

ˆð kÐjÐbiÐjÑjÔjÐjr6   ÚsiÚintc                óŽ  — |d         |k    }|d         |                               d¦  «        }|d         |         }|d         |         }|d         j        dd…         }|d	         |         }|j        d
         r8t          j        |¦  «        t	          j        || j        ¬¦  «        g d¢         z  }||||||d         |         dœS )a3  Prepare a batch of images and annotations for validation.

        Args:
            si (int): Sample index within the batch.
            batch (dict[str, Any]): Batch data containing images and annotations.

        Returns:
            (dict[str, Any]): Prepared batch with processed annotations.
        Ú	batch_idxr=   rO   r°   Ú	ori_shapera   é   NÚ	ratio_padr   ©re   )rt   r   rt   r   Úim_file)r=   r°   r»   Úimgszr½   r¿   )ÚsqueezeÚshaper   Ú	xywh2xyxyr+   Útensorre   )	r1   r·   rZ   Úidxr=   Úbboxr»   rÀ   r½   s	            r5   Ú_prepare_batchz!DetectionValidator._prepare_batch²   sÛ   € ð �KÔ  BÒ&ˆØ�EŒl˜3Ô×'Ò'¨Ñ+Ô+ˆØ�XŒ˜sÔ#ˆØ˜+Ô& rÔ*ˆ	Ø�e”Ô" 1 2 2Ô&ˆØ˜+Ô& rÔ*ˆ	ØŒ9�QŒ<ð 	_Ý”= Ñ&Ô&­¬°eÀDÄKÐ)PÑ)PÔ)PÐQ]ÐQ]ÐQ]Ô)^Ñ^ˆDàØØ"ØØ"Ø˜YÔ'¨Ô+ð
ð 
ð 	
r6   Úpredúdict[str, torch.Tensor]c                ó>   — | j         j        r|dxx         dz  cc<   |S )a  Prepare predictions for evaluation against ground truth.

        Args:
            pred (dict[str, torch.Tensor]): Post-processed predictions from the model.

        Returns:
            (dict[str, torch.Tensor]): Prepared predictions in native space.
        r=   r   )r)   r´   )r1   rÈ   s     r5   Ú_prepare_predz DetectionValidator._prepare_predÍ   s,   € ð Œ9Ôð 	Ø�ˆKˆKŒK˜1ÑˆKˆK‰KØˆr6   c                óª  ‡ ‡— t          |¦  «        D �]¿\  }}‰ xj        dz  c_        ‰                      ||¦  «        }|d                              ¦   «                              ¦   «         }‰ j        r‰ j        ‰ j        dz
           ndŠ‰ j        rÏt          j	        t          j
        |d         |d                              ¦   «         |d         |d         ¦  «        ¦  «                             ¦   «         }‰ j        d                              d	‰i¦  «         ‰ j        d
                              ˆˆ fd„t          t!          ||¦  «        ¦  «        D ¦   «         ¦  «         ‰                      |¦  «        }‰ j        r-‰ j                             t'          |d         ¦  «        ¦  «         |d         j        d         dk    }	‰ j                             i ‰                      ||¦  «        ¥|t1          j        |¦  «        |	rt1          j        d¦  «        n+|d                              ¦   «                              ¦   «         |	rt1          j        d¦  «        n+|d                              ¦   «                              ¦   «         t7          |d         ¦  «        j        dœ¥¦  «         ‰ j        j        rw‰ j                              ||‰ j!        ¬¦  «         ‰ j        j"        rI‰ j         #                    |d         |         |d         ‰ j$        ‰ j        j%        ‰ j        j&        ¦  «         |	r�Œ‰ j        j'        s‰ j        j(        r‰  )                    ||¦  «        }
‰ j        j'        r‰  *                    |
|¦  «         ‰ j        j(        rN‰  +                    |
‰ j        j,        |d         ‰ j$        dz  t7          |d         ¦  «        j-        › d�z  ¦  «         �ŒÁdS )zìUpdate metrics with new predictions and ground truth.

        Args:
            preds (list[dict[str, torch.Tensor]]): List of predictions from the model.
            batch (dict[str, Any]): Batch data containing ground truth.
        rt   r=   NrÀ   r°   r»   r½   r€   r|   r   c              3  ó–   •K  — | ]C\  }\  }}‰d z  |z  dz   ‰‰j         t          |¦  «                 ||d         |d         z  ddœV — ŒDdS )é    rt   r¼   é   r   )r|   Úimage_idÚcategory_idrÆ   ÚareaÚiscrowdN)r(   r¸   )rA   ÚiÚbÚcÚim_idxr1   s       €€r5   rD   z4DetectionValidator.update_metrics.<locals>.<genexpr>ë   s�   øè è € ð 
1ð 
1ñ "˜™6˜A˜qð  &¨™|¨aÑ/°1Ñ4Ø$*Ø'+¤~µc¸!±f´fÔ'=Ø !Ø ! !¤ q¨¤t¡Ø#$ðð ð
1ð 
1ð 
1ð 
1ð 
1ð 
1r6   r   r   r¿   )Ú
target_clsÚ
target_imgr   Úpred_clsÚim_name)r   ra   r<   z.txt).Ú	enumerater�   rÇ   r]   Únumpyr‘   r”   r’   r   Ú	xyxy2ltwhÚscale_boxesÚcloneÚtolistrz   ÚappendÚextendÚziprË   r•   r@   rÂ   r0   Úupdate_statsÚ_process_batchÚnpÚuniqueÚzerosr   Únamer)   r˜   rš   Úprocess_batchr#   r™   Úplot_matchesr3   Úshow_labelsÚ	show_confr�   Úsave_txtÚscale_predsÚpred_to_jsonÚsave_one_txtÚ	save_confÚstem)r1   r¤   rZ   r·   rÈ   Úpbatchr=   ÚboxesÚprednÚno_predÚpredn_scaledr×   s   `          @r5   Úupdate_metricsz!DetectionValidator.update_metricsÚ   sÉ  øø€ õ " %Ñ(Ô(ð <	ñ <	‰HˆB�ØˆIŒI˜‰NˆIŒIØ×(Ò(¨¨UÑ3Ô3ˆFØ˜”-×#Ò#Ñ%Ô%×+Ò+Ñ-Ô-ˆCØ59Ô5HÐR�T”] 4¤9¨q¡=Ô1Ð1ÈdˆFØÔð ÝœÝ”O F¨7¤O°V¸HÔ5E×5KÒ5KÑ5MÔ5MÈvÐVaÔObÐdjÐkvÔdwÑxÔxñô ç’&‘(”(ð ð ”
˜8Ô$×+Ò+¨T°6¨NÑ;Ô;Ð;Ø”
˜=Ô)×0Ò0ð 
1ð 
1ð 
1ð 
1ð 
1õ &/­s°5¸#©¬Ñ%?Ô%?ð
1ñ 
1ô 
1ñ 
ô 
ð 
ð ×&Ò& tÑ,Ô,ˆEØÔ"ð ;ØÔ ×'Ò'­¨E°%¬LÑ(9Ô(9Ñ:Ô:Ð:à˜E”lÔ(¨Ô+¨qÒ0ˆGØŒL×%Ò%ðØ×)Ò)¨%°Ñ8Ô8ðà"%Ý"$¤)¨C¡.¤.Ø+2ÐS�BœH Q™KœK˜K¸¸f¼×8IÒ8IÑ8KÔ8K×8QÒ8QÑ8SÔ8SØ/6Ð V¥¤¨¡¤ ¸EÀ%¼L×<LÒ<LÑ<NÔ<N×<TÒ<TÑ<VÔ<VÝ# F¨9Ô$5Ñ6Ô6Ô;ðð ð ñ	ô 	ð 	ð ŒyŒð 	ØÔ%×3Ò3°E¸6ÈÔHbÐ3ÑcÔcÐcØ”9Ô&ð ØÔ)×6Ò6Ø˜eœ RÔ(Ø˜yÔ)ØœØœ	Ô-Øœ	Ô+ñô ð ð ð ÙàŒyÔ"ð ? d¤iÔ&8ð ?Ø#×/Ò/°°vÑ>Ô>�ØŒyÔ"ð 8Ø×!Ò! ,°Ñ7Ô7Ð7ØŒyÔ!ð Ø×!Ò!Ø Ø”IÔ'Ø˜;Ô'Ø”M HÑ,µ$°v¸iÔ7HÑ2IÔ2IÔ2NÐ/TÐ/TÐ/TÑTñ	ô ð ùðo<	ð <	r6   c                óÜ   — | j         j        r,dD ])}| j                             | j        || j        ¬¦  «         Œ*| j        | j        _        | j        | j        _        | j        | j        _        dS )z8Set final values for metrics speed and confusion matrix.)TF)r3   Ú	normalizeÚon_plotN)r)   r˜   rš   Úplotr3   rý   Úspeedr0   )r1   rü   s     r5   Úfinalize_metricsz#DetectionValidator.finalize_metrics  sr   € àŒ9Œ?ð 	nØ(ð nð n�	ØÔ%×*Ò*°D´MÈYÐ`dÔ`lÐ*ÑmÔmÐmÐmØ!œZˆŒÔØ(,Ô(=ˆŒÔ%Ø $¤ˆŒÔÐÐr6   c                ój  — t           dk    rjdgt          j        ¦   «         z  }t          j        |j        |d¬¦  «         |                     ¦   «          |D ]}|r|j                             |¦  «         ŒdS t           dk    r2t          j        |j        dd¬¦  «         |                     ¦   «          dS dS )zBGather per-image metrics from all GPUs for a single metric object.r   N©Údst)r   ÚdistÚget_world_sizeÚgather_objectÚimage_metricsr—   Úupdate)r1   ÚmetricÚgathered_image_metricsr  s       r5   Ú_gather_image_metricsz(DetectionValidator._gather_image_metrics(  sÎ   € å�1Š9ˆ9Ø&* V­dÔ.AÑ.CÔ.CÑ%CÐ"ÝÔ˜vÔ3Ð5KÐQRÐSÑSÔSÐSØ×&Ò&Ñ(Ô(Ð(Ø!7ð ?ð ?�Ø ð ?ØÔ(×/Ò/°Ñ>Ô>Ð>øð?ð ?õ �AŠXˆXÝÔ˜vÔ3°T¸qÐAÑAÔAÐAØ×&Ò&Ñ(Ô(Ð(Ð(Ð(ð ˆXr6   c                ó^  ‡— t           dk    �r_dgt          j        ¦   «         z  }t          j        | j        j        |d¬¦  «         d„ | j        j        D ¦   «         }|D ]7}|                     ¦   «         D ] \  Š}|‰                              |¦  «         Œ!Œ8dgt          j        ¦   «         z  }t          j        | j        | j	        r| j
        nd| j        f|d¬¦  «         d„ |D ¦   «         | _        d„ |D ¦   «         | _        | j	        rdD ]Šˆfd„|D ¦   «         | j
        ‰<   Œ|| j        _        |                      | j        j        ¦  «         t          | j        j        ¦  «        | _        n�t           dk    r’t          j        | j        j        dd¬¦  «         t          j        | j        | j	        r| j
        nd| j        fdd¬¦  «         |                      | j        j        ¦  «         g | _        | j                             ¦   «          | j        j        r“t           d	k    rŠt+          j        | j        j        | j        ¬
¦  «        }t          j        |dt          j        j        ¬¦  «         t           dk    r6|                     ¦   «                              ¦   «         | j        _        dS dS dS dS )zGather stats from all GPUs.r   Nr  c                ó   — i | ]}|g “ŒS rL   rL   )rA   Úkeys     r5   rI   z3DetectionValidator.gather_stats.<locals>.<dictcomp>:  s   € ÐBÐBÐB¨˜C ÐBÐBÐBr6   c                ó$   — g | ]\  }}}|D ]}|‘ŒŒS rL   rL   )rA   r�   Ú_r~   s       r5   r   z3DetectionValidator.gather_stats.<locals>.<listcomp>B  s+   € ÐLÐLÐL¡  q¨!ÀeÐLÐLÀ˜!ÐLÐLÐLÐLr6   c                ó$   — g | ]\  }}}|D ]}|‘ŒŒS rL   rL   )rA   r  Úcountsr~   s       r5   r   z3DetectionValidator.gather_stats.<locals>.<listcomp>C  s+   € ÐTÐTÐT¡l a¨¨FÈVÐTÐTÈ ÐTÐTÐTÐTr6   )r€   r   c                ó2   •— g | ]\  }}}|‰         D ]}|‘ŒŒS rL   rL   )rA   r  rz   r~   r  s       €r5   r   z3DetectionValidator.gather_stats.<locals>.<listcomp>F  s3   ø€ Ð&^Ð&^Ð&^©[¨Q°°qÐSXÐY\ÔS]Ð&^Ð&^Èa qÐ&^Ð&^Ð&^Ð&^r6   rO   r¾   )r  Úop)r   r  r  r  r0   ÚstatsrP   rã   r�   r’   rz   r•   r  Úboxr@   r2   rH   r�   r–   r)   r˜   r+   Ú	as_tensorrš   Úmatrixre   ÚreduceÚReduceOpÚSUMr]   rÝ   )r1   Úgathered_statsÚmerged_statsÚ
stats_dictÚvalueÚgathered_jsonr  r  s          @r5   Úgather_statszDetectionValidator.gather_stats5  s¾  ø€ å�1Š9‰9Ø"˜V¥dÔ&9Ñ&;Ô&;Ñ;ˆNÝÔ˜tœ|Ô1°>ÀqÐIÑIÔIÐIØBÐB¨t¬|Ô/AÐBÑBÔBˆLØ,ð 4ð 4�
Ø",×"2Ò"2Ñ"4Ô"4ð 4ð 4‘J�C˜Ø  Ô%×,Ò,¨UÑ3Ô3Ð3Ð3ð4à!˜F¥TÔ%8Ñ%:Ô%:Ñ:ˆMÝÔØ”¨4Ô+;ÐE˜TœZ˜ZÀÀtÔGWÐXÐZgÐmnðñ ô ð ð MÐL¨}ÐLÑLÔLˆDŒJØTÐT°mÐTÑTÔTˆDÔØÔð _Ø2ð _ð _�CØ&^Ð&^Ð&^Ð&^¸MÐ&^Ñ&^Ô&^�D”J˜s‘O�OØ!-ˆDŒLÔØ×&Ò& t¤|Ô'7Ñ8Ô8Ð8Ý˜DœOÔ3Ñ4Ô4ˆDŒIˆIÝ�AŠXˆXÝÔ˜tœ|Ô1°4¸QÐ?Ñ?Ô?Ð?ÝÔ ¤
¸$Ô:JÐ,T¨D¬J¨JÐPTÐVZÔVfÐgÐimÐstÐuÑuÔuÐuØ×&Ò& t¤|Ô'7Ñ8Ô8Ð8ØˆDŒJØŒL×$Ò$Ñ&Ô&Ð&ØŒ9Œ?ð 	D�t bšy˜yÝ”_ TÔ%:Ô%AÈ$Ì+ÐVÑVÔVˆFÝŒK˜ A­$¬-Ô*;Ð<Ñ<Ô<Ð<Ý�qŠyˆyØ/5¯zªz©|¬|×/AÒ/AÑ/CÔ/C�Ô%Ô,Ð,Ð,ð		Dð 	D˜y˜yð ˆyr6   c                ó`  — | j                              | j        | j        j        | j        ¬¦  «         | j         j        }| j        j        rK| j        j        dk    r;| 	                    d„ dD ¦   «         ¦  «         | j
        r|                      |¦  «        }| j                              ¦   «          |S )zˆCalculate and return metrics statistics.

        Returns:
            (dict[str, Any]): Dictionary containing metrics results.
        )r3   rþ   rý   r   c                ó   — i | ]	}d |› d�d“Œ
S )úmetrics/mAP_z(B)g        rL   r}   s     r5   rI   z0DetectionValidator.get_stats.<locals>.<dictcomp>_  s&   € Ð[Ð[Ð[¸Ð/¨Ð/Ð/Ð/°Ð[Ð[Ð[r6   ©ÚsmallÚmediumÚlarge)r0   Úprocessr3   r)   r˜   rý   Úresults_dictr�   r*   r  r„   Ú	eval_jsonr–   )r1   r  s     r5   Ú	get_statszDetectionValidator.get_statsV  s¦   € ð 	Œ×Ò d¤m¸$¼)¼/ÐSWÔS_ÐÑ`Ô`Ð`Ø”Ô)ˆØŒ9Ôð 	. 4¤9¤>°XÒ#=Ð#=Ø�LŠLÐ[Ð[Ð>ZÐ[Ñ[Ô[Ñ\Ô\Ð\ØŒ}ð .ØŸš uÑ-Ô-�ØŒ× Ò Ñ"Ô"Ð"Øˆr6   c                ó¼  — ddt          | j        j        ¦  «        z  z   }t          j        |d| j        | j        j                             ¦   «         g| j                             ¦   «         ¢R z  ¦  «         | j        j                             ¦   «         dk    r"t          j	        d| j
        j        › d�¦  «         | j
        j        r‘| j        sŒ| j        dk    rƒt          | j        j        ¦  «        D ]k\  }}t          j        || j        |         | j        j        |         | j        j        |         g| j                             |¦  «        ¢R z  ¦  «         ŒfdS dS dS dS )	z0Print training/validation set metrics per class.z%22s%11i%11iz%11.3gÚallr   zno labels found in z+ set, cannot compute metrics without labelsrt   N)r@   r0   Úkeysr   Úinfor�   Únt_per_classÚsumÚmean_resultsrT   r)   r*   Úverboser„   rŽ   rÜ   Úap_class_indexr‚   Únt_per_imageÚclass_result)r1   ÚpfrÔ   rÖ   s       r5   Úprint_resultsz DetectionValidator.print_resultse  sk  € à  8­c°$´,Ô2CÑ.DÔ.DÑ#DÑDˆÝŒ�B˜% ¤¨D¬LÔ,E×,IÒ,IÑ,KÔ,KÐjÈdÌl×NgÒNgÑNiÔNiÐjÐjÑjÑkÔkÐkØŒ<Ô$×(Ò(Ñ*Ô*¨aÒ/Ð/ÝŒNÐl°´´ÐlÐlÐlÑmÔmÐmð Œ9Ôð 
	 T¤]ð 
	°t´wÀ²{°{Ý! $¤,Ô"=Ñ>Ô>ð 	ð 	‘��1Ý”Øàœ
 1œØœÔ1°!Ô4ØœÔ1°!Ô4ðð œ×2Ò2°1Ñ5Ô5ð	ð ññô ð ð ð
	ð 
	ð 
	ð 
	°{°{ð	ð 	r6   údict[str, np.ndarray]c                ó’  — |d         j         d         dk    s|d         j         d         dk    r5dt          j        |d         j         d         | j        ft          ¬¦  «        iS t          |d         |d         ¦  «        }d|                      |d         |d         |¦  «                             ¦   «                              ¦   «         iS )aÃ  Return correct prediction matrix.

        Args:
            preds (dict[str, torch.Tensor]): Dictionary containing prediction data with 'bboxes' and 'cls' keys.
            batch (dict[str, Any]): Batch dictionary containing ground truth data with 'bboxes' and 'cls' keys.

        Returns:
            (dict[str, np.ndarray]): Dictionary containing 'tp' key with correct prediction matrix of shape (N, 10) for
                10 IoU levels.
        r=   r   Útp©Údtyper°   )	rÂ   rç   ré   r/   Úboolr   Úmatch_predictionsr]   rÝ   )r1   r¤   rZ   r³   s       r5   ræ   z!DetectionValidator._process_batchy  s²   € ð �Œ<Ô˜aÔ  AÒ%Ð%¨¨u¬Ô);¸AÔ)>À!Ò)CÐ)CØ�"œ( E¨%¤LÔ$6°qÔ$9¸4¼9Ð#EÍTÐRÑRÔRÐSÐSÝ�e˜H”o u¨X¤Ñ7Ô7ˆØ�d×,Ò,¨U°5¬\¸5À¼<ÈÑMÔM×QÒQÑSÔS×YÒYÑ[Ô[Ð\Ð\r6   rp   Úimg_pathÚmodeú
int | Noneútorch.utils.data.Datasetc           	     ó˜   — t          | j        j        | j        j        pd¦  «        }t	          | j        ||| j        || j        |¬¦  «        S )ac  Build YOLO Dataset.

        Args:
            img_path (str): Path to the folder containing images.
            mode (str): `train` mode or `val` mode, users are able to customize different augmentations for each mode.
            batch (int, optional): Size of batches, this is for `rect`.

        Returns:
            (Dataset): YOLO dataset.
        rp   )rB  ÚstrideÚfraction)r	   r)   rG  rG   r   r…   rF  )r1   rA  rB  rZ   rG  s        r5   Úbuild_datasetz DetectionValidator.build_dataset‰  sN   € õ & d¤iÔ&8¸$¼)¼/Ð:RÈUÑSÔSˆÝ!ØŒI�x ¨¬	¸ÀTÄ[Ð[cð
ñ 
ô 
ð 	
r6   Údataset_pathÚ
batch_sizeútorch.utils.data.DataLoaderc           
     óÆ  — |                       ||d¬¦  «        }| j        rŽt          | j        ¦  «        | j        d         k     rkt	          j        dt          | j        ¦  «        › d�¦  «         |                     t          t          t          | j        ¦  «        ¦  «        ¦  «        ¦  «         t          ||| j
        j        dd| j
        j        | j        | j        ¬¦  «        S )	z÷Construct and return dataloader.

        Args:
            dataset_path (str): Path to the dataset.
            batch_size (int): Size of each batch.

        Returns:
            (torch.utils.data.DataLoader): DataLoader for validation.
        rp   )rZ   rB  rŽ   z)Skipping labels of classes the model (nc=z) cannot predict.FrO   )ÚshuffleÚrankÚ	drop_lastÚ
pin_memoryre   )rH  r‚   r@   r…   r   rT   Úupdate_labelsr‹   rŒ   r   r)   ÚworkersÚcompiler„   re   )r1   rI  rJ  rH   s       r5   Úget_dataloaderz!DetectionValidator.get_dataloader™  sÎ   € ð ×$Ò$ \¸È%Ð$ÑPÔPˆØŒ:ð 	@�#˜dœj™/œ/¨D¬I°d¬OÒ;Ð;ÝŒNÐiÅsÈ4Ì:ÁÄÐiÐiÐiÑjÔjÐjØ×!Ò!¥$¥u­S°´©_¬_Ñ'=Ô'=Ñ">Ô">Ñ?Ô?Ð?ÝØØØŒIÔØØØ”iÔ'Ø”}Ø”;ð	
ñ 	
ô 	
ð 		
r6   Únic                óf   — t          ||d         | j        d|› d�z  | j        | j        ¬¦  «         dS )z¦Plot validation image samples.

        Args:
            batch (dict[str, Any]): Batch containing images and annotations.
            ni (int): Batch index.
        r¿   Ú	val_batchz_labels.jpg)r<   ÚpathsÚfnamer‚   rý   N)r   r3   r‚   rý   )r1   rZ   rU  s      r5   Úplot_val_samplesz#DetectionValidator.plot_val_samples²  sP   € õ 	ØØ˜	Ô"Ø”-Ð"=¨bÐ"=Ð"=Ð"=Ñ=Ø”*Ø”Lð	
ñ 	
ô 	
ð 	
ð 	
ð 	
r6   rR   c           	     ó’  ‡‡— ‰sdS t          ‰¦  «        D ]%\  }}t          j        |d         ¦  «        |z  |d<   Œ&‰d                              ¦   «         }‰p| j        j        Šˆˆfd„|D ¦   «         }t          j        |d         ¦  «        |d<   t          |d         ||d         | j	        d	|› d
�z  | j
        | j        ¬¦  «         dS )ai  Plot predicted bounding boxes on input images and save the result.

        Args:
            batch (dict[str, Any]): Batch containing images and annotations.
            preds (list[dict[str, torch.Tensor]]): List of predictions from the model.
            ni (int): Batch index.
            max_det (int | None): Maximum number of detections to plot.
        Nr   rº   r   c                óZ   •‡— i | ]&Š‰t          j        ˆˆfd „‰D ¦   «         d¬¦  «        “Œ'S )c                ó0   •— g | ]}|‰         d ‰…         ‘ŒS ©NrL   )rA   r~   rj   rR   s     €€r5   r   zBDetectionValidator.plot_predictions.<locals>.<dictcomp>.<listcomp>Ò  s%   ø€ Ð&EÐ&EÐ&E¸! q¨¤t¨H¨W¨H¤~Ð&EÐ&EÐ&Er6   r   ©Údim)r+   Úcat)rA   rj   rR   r¤   s    @€€r5   rI   z7DetectionValidator.plot_predictions.<locals>.<dictcomp>Ò  sD   øø€ Ð\Ð\Ð\ÐRS˜�EœIÐ&EÐ&EÐ&EÐ&EÐ&E¸uÐ&EÑ&EÔ&EÈ1ÐMÑMÔMÐ\Ð\Ð\r6   r°   ra   r¿   rW  z	_pred.jpg)r€   r<   rX  rY  r‚   rý   )rÜ   r+   Ú	ones_liker/  r)   rR   r   Ú	xyxy2xywhr   r3   r‚   rý   )	r1   rZ   r¤   rU  rR   rÔ   rÈ   r/  Úbatched_predss	     ` `    r5   Úplot_predictionsz#DetectionValidator.plot_predictionsÁ  sù   øø€ ð ð 	ØˆFÝ  Ñ'Ô'ð 	Bð 	B‰GˆAˆtÝ %¤°°V´Ñ =Ô =ÀÑ AˆD�ÑÐØ�QŒx�}Š}‰ŒˆØÐ.˜TœYÔ.ˆØ\Ð\Ð\Ð\Ð\ÐW[Ð\Ñ\Ô\ˆÝ"%¤-°¸hÔ0GÑ"HÔ"Hˆ�hÑÝØ˜”<Ø Ø˜	Ô"Ø”-Ð";¨bÐ";Ð";Ð";Ñ;Ø”*Ø”Lð	
ñ 	
ô 	
ð 	
ð 	
ð 	
r6   r÷   ró   r?  rÂ   útuple[int, int]Úfiler   c                óZ  — ddl m}  |t          j        |d         |d         ft          j        ¬¦  «        d| j        t          j        |d         |d                              d¦  «        |d	                              d¦  «        gd¬
¦  «        ¬¦  «         	                    ||¬¦  «         dS )a®  Save YOLO detections to a txt file in normalized coordinates in a specific format.

        Args:
            predn (dict[str, torch.Tensor]): Dictionary containing predictions with keys 'bboxes', 'conf', and 'cls'.
            save_conf (bool): Whether to save confidence scores.
            shape (tuple[int, int]): Shape of the original image (height, width).
            file (Path): File path to save the detections.
        r   )ÚResultsrt   r=  Nr°   r   rO   r=   r_  )Úpathr‚   rö   )ró   )
Úultralytics.engine.resultsri  rç   ré   Úuint8r‚   r+   ra  Ú	unsqueezerï   )r1   r÷   ró   rÂ   rg  ri  s         r5   rò   zDetectionValidator.save_one_txtÝ  s³   € ð 	7Ð6Ð6Ð6Ð6Ð6àˆÝŒH�e˜A”h  a¤Ð)µ´Ð:Ñ:Ô:ØØ”*Ý”)˜U 8œ_¨e°F¬m×.EÒ.EÀbÑ.IÔ.IÈ5ÐQVÌ<×KaÒKaÐbdÑKeÔKeÐfÐlmÐnÑnÔnð		
ñ 	
ô 	
÷
 Š(�4 9ˆ(Ñ
-Ô
-Ð
-Ð
-Ð
-r6   rõ   c                ód  — t          |d         ¦  «        }|j        }|                     ¦   «         rt          |¦  «        n|}t	          j        |d         ¦  «        }|dd…dd…fxx         |dd…dd…f         dz  z  cc<   t          |                     ¦   «         |d                              ¦   «         |d                              ¦   «         ¦  «        D ][\  }}}	| j         	                    ||j
        | j        t          |	¦  «                 d„ |D ¦   «         t          |d¦  «        d	œ¦  «         Œ\dS )
a¡  Serialize YOLO predictions to COCO json format.

        Args:
            predn (dict[str, torch.Tensor]): Predictions dictionary containing 'bboxes', 'conf', and 'cls' keys with
                bounding box coordinates, confidence scores, and class predictions.
            pbatch (dict[str, Any]): Batch dictionary containing 'imgsz', 'ori_shape', 'ratio_pad', and 'im_file'.

        Examples:
             >>> result = {
             ...     "image_id": 42,
             ...     "file_name": "42.jpg",
             ...     "category_id": 18,
             ...     "bbox": [258.15, 41.29, 348.26, 243.78],
             ...     "score": 0.236,
             ... }
        r¿   r°   Nr¼   r   r=   c                ó.   — g | ]}t          |d ¦  «        ‘ŒS )rÏ   )Úroundr}   s     r5   r   z3DetectionValidator.pred_to_json.<locals>.<listcomp>  s    € Ð4Ð4Ð4¨Q�U 1 a™[œ[Ð4Ð4Ð4r6   r®   )rÐ   Ú	file_namerÑ   rÆ   Úscore)r   rô   Ú	isnumericr¸   r   rc  rä   rá   r�   râ   rê   r(   rp  )
r1   r÷   rõ   rj  rô   rÐ   r  rÕ   ÚsrÖ   s
             r5   rñ   zDetectionValidator.pred_to_jsonï  s4  € õ" �F˜9Ô%Ñ&Ô&ˆØŒyˆØ $§¢Ñ 0Ô 0Ð:•3�t‘9”9�9°dˆÝŒm˜E (œOÑ,Ô,ˆØˆAˆAˆAˆr�ˆrˆEˆ
ˆ
Œ
�c˜!˜!˜!˜Q˜R˜R˜%”j 1‘nÑ$ˆ
ˆ
‰
Ý˜3Ÿ:š:™<œ<¨¨v¬×)=Ò)=Ñ)?Ô)?ÀÀuÄ×ATÒATÑAVÔAVÑWÔWð 		ð 		‰GˆAˆq�!ØŒJ×Òà (Ø!%¤Ø#'¤>µ#°a±&´&Ô#9Ø4Ð4°!Ð4Ñ4Ô4Ý" 1 a™[œ[ðð ñô ð ð ð		ð 		r6   c           	     ó’   — i |¥dt          j        |d         |d                              ¦   «         |d         |d         ¬¦  «        i¥S )z.Scales predictions to the original image size.r°   rÀ   r»   r½   )r½   )r   rß   rà   )r1   r÷   rõ   s      r5   rð   zDetectionValidator.scale_preds  s^   € ð
Øð
à•c”oØ�w”Ø�h”×%Ò%Ñ'Ô'Ø�{Ô#Ø  Ô-ð	ñ ô ð
ð 
ð 	
r6   r  c                óT  ‡— | j         rAt          | j        ¦  «        Šˆfd„t          | j         d         | j        ¦  «        D ¦   «         }n| j        r| j        n	| j        dz  }| j         p)| j        d         dz  | j        rdnd| j	        j
        › d�z  }|                      |||¦  «        S )	a  Evaluate YOLO output in JSON format and return performance statistics.

        Args:
            stats (dict[str, Any]): Current statistics dictionary.

        Returns:
            (dict[str, Any]): Updated statistics dictionary with COCO/LVIS evaluation results.
        c                óp   •— g | ]2\  }}t          |¦  «        D ]}i t          ‰¦  «        ¥d |d         i¥‘ŒŒ3S )rÐ   r|   )rŒ   Únext)rA   ÚimagerM   r  Úpredictionss       €r5   r   z0DetectionValidator.eval_json.<locals>.<listcomp>'  se   ø€ ð ð ð á �E˜5Ý˜u™œðð ð ð ?•4˜Ñ$Ô$Ð> j°%¸´+Ð>Ð>ðð ð ð r6   r€   zpredictions.jsonrj  r   zinstances_val2017.jsonÚlvis_v1_z.json)rz   Úiterr�   rä   r•   r„   r3   r…   r&   r)   rG   Úcoco_evaluate)r1   r  Ú	pred_jsonÚ	anno_jsonrz  s       @r5   r+  zDetectionValidator.eval_json  sÚ   ø€ ð Œ:ð 	\Ý˜tœzÑ*Ô*ˆKðð ð ð å$'¨¬
°8Ô(<¸dÔ>NÑ$OÔ$Oðñ ô ˆIˆIð '+¤mÐ[˜œ
˜
¸¼ÐI[Ñ9[ˆIØ”Jð 
ØŒI�fÔØñà+/¬<Ð^Ð'Ð'Ð=^ÈÌ	ÌÐ=^Ð=^Ð=^ñ`ð 	ð
 ×!Ò! %¨°IÑ>Ô>Ð>r6   rÆ   ÚBoxr~  ústr | Path | listr  ústr | Path | dictÚ	iou_typesústr | list[str]Úsuffixc           	     ój  — | j         j        �r%t          | j        ¦  «        �r| j        s| j        s| j        �rút          j        d¦  «         	 ||fD ]L}t          |t          t          f¦  «        r.t          |¦  «                             ¦   «         sJ |› d�¦   «         ‚ŒMt          |t          ¦  «        r|gn|}t          |t          ¦  «        r|gn|}t          d¦  «         ddlm}m} t#          | dd¦  «        p
 ||¦  «        }	|	| _        |	                     |¦  «        }
t)          |¦  «        D �]¿\  }} ||	|
|| j        t          j        ¬¦  «        }| j        r|	                     ¦   «         nd	„ | j        j        j        D ¦   «         |j        _        |                     ¦   «          |                     ¦   «          |                     ¦   «          | j        sr| j        s| j        rd|j        d
         |d||         d         › d�<   |j        d         |d||         d         › d�<   d|j        d         z  d|j        d
         z  z   |d<   dD ]4}d|› �|j        v r&|j        d|› �         |d|› d||         d         › d�<   Œ5| j        sg| j        r`|j        d         |d||         d         › d�<   |j        d         |d||         d         › d�<   |j        d         |d||         d         › d�<   �ŒÁ| j        r|d         |d<   n.# t@          $ r!}t          j!        d|› �¦  «         Y d}~nd}~ww xY w|S )aµ  Evaluate COCO/LVIS or custom COCO-format detection metrics using faster-coco-eval.

        Args:
            stats (dict[str, Any]): Dictionary to store computed metrics and statistics.
            pred_json (str | Path | list): Path or in-memory predictions in COCO format.
            anno_json (str | Path | dict): Path or in-memory ground truth in COCO format.
            iou_types (str | list[str]): IoU types to evaluate, such as "bbox", "segm", or "keypoints".
            suffix (str | list[str]): Metric suffixes corresponding to the IoU types.

        Returns:
            (dict[str, Any]): Updated stats dictionary containing the computed COCO-format evaluation metrics.
        z#
Evaluating faster-coco-eval mAP...z file not foundzfaster-coco-eval>=1.6.7r   )ÚCOCOÚCOCOeval_fasterÚ	_coco_apiN)ÚiouTypeÚ
lvis_styleÚprint_functionc                óP   — g | ]#}t          t          |¦  «        j        ¦  «        ‘Œ$S rL   )r¸   r   rô   r}   s     r5   r   z4DetectionValidator.coco_evaluate.<locals>.<listcomp>^  s(   € ÐZÐZÐZ°A�c¥$ q¡'¤'¤,Ñ/Ô/ÐZÐZÐZr6   ÚAP_50zmetrics/mAP50(ú)ÚAP_allzmetrics/mAP50-95(gÍÌÌÌÌÌì?gš™™™™™¹?Úfitnessr%  ÚAP_r$  ú(ÚAPrzmetrics/APr(ÚAPczmetrics/APc(ÚAPfzmetrics/APf(zmetrics/mAP50-95(B)z faster-coco-eval unable to run: )"r)   r�   r@   r�   r&   r'   rz   r   r0  r   r†   r   Úis_filer   Úfaster_coco_evalr‡  rˆ  r"   r‰  ÚloadResrÜ   Ú	getImgIdsr2   rH   Úim_filesÚparamsÚimgIdsÚevaluateÚ
accumulateÚ	summarizer„   Ústats_as_dictÚ	ExceptionrT   )r1   r  r~  r  rƒ  r…  r~   r‡  rˆ  ÚannorÈ   rÔ   Úiou_typerp   Úes                  r5   r}  z DetectionValidator.coco_evaluate5  sá  € ð( Œ9Ôñ *	G¥3 t¤z¡?¤?ñ *	G¸¼ð *	GÈÌð *	GÐX\ÔXbñ *	GÝŒKÐ>Ñ?Ô?Ð?ð(GØ" IÐ-ð Hð H�AÝ! !¥c­4 [Ñ1Ô1ð HÝ# A™wœwŸšÑ0Ô0ÐGÐG°QÐ2GÐ2GÐ2GÑGÔGÐ0øÝ+5°iÅÑ+EÔ+EÐT˜Y˜K˜KÈ9�	Ý%/°½Ñ%<Ô%<ÐH˜&˜˜À&�Ý"Ð#<Ñ=Ô=Ð=ØBÐBÐBÐBÐBÐBÐBÐBå˜t [°$Ñ7Ô7ÐJ¸4¸4À	¹?¼?�Ø!%�”Ø—|’| IÑ.Ô.�Ý#,¨YÑ#7Ô#7ð Yñ Y‘K�A�xØ)˜/Ø˜d¨HÀÄÕ^dÔ^iðñ ô �Cð
  œ:ð[˜ŸšÑ(Ô(Ð(àZÐZ¸¼Ô9PÔ9YÐZÑZÔZð ”JÔ%ð
 —L’L‘N”N�NØ—N’NÑ$Ô$Ð$Ø—M’M‘O”O�Oàœ=ð p¨d¬lð p¸d¼lð pØBEÔBSÐT[ÔB\˜Ð>¨v°a¬y¸¬|Ð>Ð>Ð>Ñ?ØEHÔEVÐW_ÔE`˜ÐA°&¸´)¸A´,ÐAÐAÐAÑBØ+.°Ô1BÀ8Ô1LÑ+LÈsÐUXÔUfÐgnÔUoÑOoÑ+o˜˜iÑ(Ø7ð eð e˜Ø$ ˜9˜9¨Ô(9Ð9Ð9ØHKÔHYÐZcÐ`aÐZcÐZcÔHd˜EÐ"D°Ð"DÐ"D°V¸A´Y¸q´\Ð"DÐ"DÐ"DÑEøØœ=ð Y¨T¬\ð YØ@CÔ@QÐRWÔ@X˜Ð<¨V°A¬Y°q¬\Ð<Ð<Ð<Ñ=Ø@CÔ@QÐRWÔ@X˜Ð<¨V°A¬Y°q¬\Ð<Ð<Ð<Ñ=Ø@CÔ@QÐRWÔ@X˜Ð<¨V°A¬Y°q¬\Ð<Ð<Ð<Ñ=ùà”<ð DØ',Ð-BÔ'C�E˜)Ñ$øøÝð Gð Gð GÝ”ÐEÀ!ÐEÐEÑFÔFÐFÐFÐFÐFÐFÐFøøøøðGøøøàˆs   ÁJ6L Ì
L0ÌL+Ì+L0)NNNN)r   r   r   r   )r7   r8   r   r   )rZ   r[   r   r[   )rm   rn   r   r   )r   r†   )r¤   r¥   r   r¦   )r·   r¸   rZ   r[   r   r[   )rÈ   rÉ   r   rÉ   )r¤   r¦   rZ   r[   r   r   )r   r   )r   r[   )r¤   rÉ   rZ   r[   r   r:  )rp   N)rA  r†   rB  r†   rZ   rC  r   rD  )rI  r†   rJ  r¸   r   rK  )rZ   r[   rU  r¸   r   r   r^  )
rZ   r[   r¤   r¦   rU  r¸   rR   rC  r   r   )
r÷   rÉ   ró   r?  rÂ   rf  rg  r   r   r   )r÷   rÉ   rõ   r[   r   r   )r÷   rÉ   rõ   r[   r   rÉ   )r  r[   r   r[   )rÆ   r€  )r  r[   r~  r�  r  r‚  rƒ  r„  r…  r„  r   r[   )Ú__name__Ú
__module__Ú__qualname__Ú__doc__r%   ÚstaticmethodrY   rl   rœ   r£   r¶   rÇ   rË   rú   r   r  r!  r,  r9  ræ   rH  rT  rZ  re  rò   rñ   rð   r+  r}  Ú__classcell__)r4   s   @r5   r   r      s_  ø€ € € € € ðð ð,$ð $ð $ð $ð $ð $ð $ð( ð$ð $ð $ñ „\ð$ðBð ð ð ð#yð #yð #yð #yðJlð lð lð lðkð kð kð kð.
ð 
ð 
ð 
ð6ð ð ð ðCð Cð Cð CðJ.ð .ð .ð .ð)ð )ð )ð )ðDð Dð Dð DðBð ð ð ðð ð ð ð(]ð ]ð ]ð ]ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð2
ð 
ð 
ð 
ð  koð
ð 
ð 
ð 
ð 
ð8.ð .ð .ð .ð$ð ð ð ðB

ð 

ð 

ð 

ð?ð ?ð ?ð ?ð< &,Ø"'ð?ð ?ð ?ð ?ð ?ð ?ð ?ð ?ð ?r6   r   )$Ú
__future__r   rˆ   Úpathlibr   Útypingr   rÝ   rç   r+   Útorch.distributedÚdistributedr  Úultralytics.datar   r   r   Úultralytics.data.utilsr	   Úultralytics.engine.validatorr
   Úultralytics.utilsr   r   r   r   r   Úultralytics.utils.checksr   Úultralytics.utils.metricsr   r   r   Úultralytics.utils.plottingr   r   rL   r6   r5   ú<module>r¸     sY  ðð #Ð "Ð "Ð "Ð "Ð "à 	€	€	€	Ø Ð Ð Ð Ð Ð Ø Ð Ð Ð Ð Ð à Ð Ð Ð Ø €€€Ø  Ð  Ð  Ð  Ð  Ð  à LÐ LÐ LÐ LÐ LÐ LÐ LÐ LÐ LÐ LØ 5Ð 5Ð 5Ð 5Ð 5Ð 5Ø 6Ð 6Ð 6Ð 6Ð 6Ð 6Ø AÐ AÐ AÐ AÐ AÐ AÐ AÐ AÐ AÐ AÐ AÐ AÐ AÐ AØ 7Ð 7Ð 7Ð 7Ð 7Ð 7Ø JÐ JÐ JÐ JÐ JÐ JÐ JÐ JÐ JÐ JØ 2Ð 2Ð 2Ð 2Ð 2Ð 2ð^	ð ^	ð ^	ð ^	ð ^	˜ñ ^	ô ^	ð ^	ð ^	ð ^	r6   