Fi\ddlmZddlZddlmZddlmZddlZddl Z ddl m Z ddl mZmZmZddlmZddlmZmZmZmZddlmZdd lmZmZmZdd lmZGd d eZ y) ) annotationsN)Path)Any)build_dataloaderbuild_yolo_dataset converter) BaseValidator)LOGGERRANKnmsops)check_requirements)ConfusionMatrix DetMetricsbox_iou) plot_imagesc eZdZdZddfd ZddZddZddZddZddZ d dZ d!d Z d"d Z d"d Z d"d Zd#d Zd"dZd$dZd%d&dZd'dZd(dZ d) d*dZd+dZd,dZd-dZd.dZ d/ d0dZxZS)1DetectionValidatorayA 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. lb (list[Any]): List for storing ground truth labels for hybrid saving. 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() c t|||||d|_d|_d|_d|j _tjddd|_ |jj|_ t|_ y)aInitialize 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. FNdetectg?gffffff? )super__init__is_cocois_lvis class_mapargstasktorchlinspaceiouvnumelniourmetrics)self dataloadersave_dirr _callbacks __class__s d/var/www/pod-logistic/pod-ai/venv/lib/python3.12/site-packages/ultralytics/models/yolo/detect/val.pyrzDetectionValidator.__init__-sg XtZ@  ! NN3b1 IIOO% !| c^|jD]W\}}t|tjs!|j |j |j j dk(||<Y|jjr|djn|djdz |d<|S)zPreprocess batch of images for YOLO validation. Args: batch (dict[str, Any]): Batch containing images and annotations. Returns: (dict[str, Any]): Preprocessed batch. cuda) non_blockingimg) items isinstancerTensortodevicetyperhalffloat)r%batchkvs r* preprocesszDetectionValidator.preprocess?sKKM VDAq!U\\*44 $++:J:Jf:T4Ua V04yy~~e ))+5 |_ y ) zInitialize evaluation metrics for YOLO detection validation. Args: model (torch.nn.Module): Model to validate. cocoz val2017.txtztest-dev2017.txtlvisend2endFr)names save_matchesN)!datagetrsplitr2strendswithosseprrrcoco80_to_coco91_classlistrangelenrCr save_jsonvaltrainingncgetattrrBseenjdictr$ clear_statsclear_image_metricsrplots visualizeconfusion_matrix)r%modelrQs r* init_metricszDetectionValidator.init_metricsNs iimmDIIOOR0 sC  d#  d 45bP`Fa9b "#s+R# RdllBR ?C||99;QUV[\]_bchcncn_ors_sVtQu tyy}}e$,,2N$,,eX\XeXeTee[[ ekk"ui7   "[[    " ((* /ekkPTPYPYP_P_Pwdhdmdmdwdw xr+c ddzS)zBReturn a formatted string summarizing class metrics of YOLO model.z%22s%11s%11s%11s%11s%11s%11s)ClassImages InstanceszBox(PRmAP50z mAP50-95)r%s r*get_desczDetectionValidator.get_descgs#'kkkr+c tj||jj|jj|jj dk(rdn |j d|jjxs|jj|jj|j|jj dk( }|Dcgc])}|ddddf|dddf|dddf|ddd dfd +c}Scc}w) aEApply 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. rrTobb)rS multi_labelagnosticmax_detrBrotatedN)bboxesconfclsextra) r non_max_suppressionrrqiourrS single_cls agnostic_nmsrkrB)r%predsoutputsxs r* postprocesszDetectionValidator.postprocessks))  IINN IIMMIINNh.qDGGYY))CTYY-C-CII%%LLIINNe+  cjj]^1QU8Qq!tWQq!tWqQRTUTVQVxXjjjs.DcH|d|k(}|d|jd}|d|}|d|}|djdd}|d |}|jd r yy""!!,7yy!!!! II'';'MMH,$vi7H2I2N2N1Ot/TT A% r+cD|jjr9dD]4}|jj|j||j 6|j |j_|j|j_|j|j_y)z8Set final values for metrics speed and confusion matrix.)TF)r' normalizeon_plotN)rrYr[plotr'rspeedr$)r%rs r*finalize_metricsz#DetectionValidator.finalize_metricssv 99??( n %%**DMMY`d`l`l*m n!ZZ (,(=(= % $  r+cntdk(rpdgtjz}tj|j|d|j |D] }|s|jj |"ytdkDr3tj|jdd|j yy)zBGather per-image metrics from all GPUs for a single metric object.rNdst)r distget_world_size gather_object image_metricsrXupdate)r%metricgathered_image_metricsrs r*_gather_image_metricsz(DetectionValidator._gather_image_metricss 19&*Vd.A.A.C%C "   v335KQR S  & & (!7 ?  ((// > ?AX   v33Tq A  & & (r+ctdk(rYdgtjz}tj|jj |d|jj j Dcic]}|g}}|D] }|D]}||j||"dgtjz}tj|j|dg|_|D]}|jj|||j_|j|jjt|jj|_ytdkDrtj|jj ddtj|jdd|j|jjg|_|jjyycc}w)zGather stats from all GPUs.rNr)r rrrr$statskeysextendrVrboxrOr&datasetrUrW)r%gathered_statskey merged_stats stats_dictgathered_jdictrVs r* gather_statszDetectionValidator.gather_statss 19"Vd&9&9&;;N   t||11>q I/3||/A/A/F/F/HICGILI, > '>C %,,Z_=> >#Vd&9&9&;;N   tzz>q ADJ' ) !!%( )!-DLL   & &t||'7'7 8DOO334DI AX   t||114Q ?   tzz4Q 7  & &t||'7'7 8DJ LL $ $ & Js5 Hc|jj|j|jj|j |jj |jjS)zCalculate and return metrics statistics. Returns: (dict[str, Any]): Dictionary containing metrics results. )r'rr)r$processr'rrYrrW results_dictres r* get_statszDetectionValidator.get_statssP dmm$))//SWS_S_`   "||(((r+cddt|jjzz}tj|d|j |jj jg|jjz|jj jdk(r-tjd|jjd|jjr|js|jdkDrt|jjrt!|jj"D]w\}}tj||j$||jj&||jj |g|jj)|zyyyyyy) z0Print training/validation set metrics per class.z %22s%11i%11iz%11.3gallrzno labels found in z+ set, cannot compute metrics without labelsrAN)rOr$rr inforU nt_per_classsum mean_resultswarningrrverboserRrSrrap_class_indexrC nt_per_image class_result)r%pfics r* print_resultsz DetectionValidator.print_resultss] 8c$,,2C2C.D#D D B%DLL,E,E,I,I,KjdllNgNgNijjk << $ $ ( ( *a / NN00@@kl m 99  T]]tww{s4< 1  1  11!4 11!42215  H_{] r+cZ|djddk(s|djddk(r9dtj|djd|jftiSt |d|d}d|j |d|d|jjiS)aReturn 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. rrrtpdtyperp) rrrr#boolrmatch_predictionsrr)r%rxr9rus r*rz!DetectionValidator._process_batch$s <  a A %u););A)>!)C"((E%L$6$6q$9499#ETRS SeHouX7d,,U5\5<MQQSYY[\\r+c`t|j|||j||jS)acBuild YOLO Dataset. Args: img_path (str): Path to the folder containing images. mode (str): `train` mode or `val` mode, users are able to customize different augmentations for each mode. batch (int, optional): Size of batches, this is for `rect`. Returns: (Dataset): YOLO dataset. )modestride)rrrEr)r%img_pathrr9s r* build_datasetz DetectionValidator.build_dataset4s("$))Xudiid[_[f[fggr+c |j||d}t|||jjdd|jj|j S)zConstruct 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. rQ)r9rFr~)shufflerank drop_last pin_memory)rrrworkerscompilerR)r% dataset_path batch_sizers r*get_dataloaderz!DetectionValidator.get_dataloaderAsU$$\%$P   II  ii''}}  r+ctt||d|jd|dz |j|jy)zPlot validation image samples. Args: batch (dict[str, Any]): Batch containing images and annotations. ni (int): Batch index. r val_batchz _labels.jpg)rpathsfnamerCrN)rr'rCr)r%r9nis r*plot_val_samplesz#DetectionValidator.plot_val_samplesVs:  "--IbT"==**LL  r+c |syt|D]#\}}tj|d|z|d<%|dj}|xs|jj }|D cic].}|tj |D cgc] } | |d| c} d0} }} tj| d| d<t|d| |d|jd |d z |j|j ycc} wcc} }w) aiPlot 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. Nrqr}rdimrpr/rrz _pred.jpg)imagesrrrrCr) rr ones_likerrrkcatr xyxy2xywhrr'rCr) r%r9rxrrkrrrr:rz batched_predss r*plot_predictionsz#DetectionValidator.plot_predictionses  ' BGAt %V = AD  BQx}}.TYY..W[\RSEIIu&E!qtHW~&E1MM\ \"%-- h0G"H h<  "--IbT";;**LL  'F\s'C:C5 C:5C:c 2ddlm}|tj|d|dftjd|j t j|d|djd|d jdgd  j|| y) aSave 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)ResultsrArNrprqr~rrr)pathrCboxes)r) ultralytics.engine.resultsrrruint8rCrr unsqueezer)r%rrrfilers r*rzDetectionValidator.save_one_txts 7 HHeAha) :**))U8_eFm.E.Eb.I5QV>> result = { ... "image_id": 42, ... "file_name": "42.jpg", ... "category_id": 18, ... "bbox": [258.15, 41.29, 348.26, 243.78], ... "score": 0.236, ... } rrpNrrqrrrn)image_id file_name category_idrscore) rr isnumericintr rziptolistrVappendrrround) r%rrrrr rbsrrzs r*rzDetectionValidator.pred_to_jsons"F9%&yy $ 03t9dmmE(O, ArrE c!QR%j1n$ 3::<v)=)=)?uATATAVW GAq! JJ   (!%#'>>#aΛ4QU1a[4"1a[    5s/Dc ti|dtj|d|dj|d|diS)z.Scales predictions to the original image size.rprrr)r)r scale_boxesclone)r%rrs r*rzDetectionValidator.scale_predssN  coowh%%'{# -   r+c|jdz }|jddz |jrdnd|jjdz }|j |||S)aEvaluate 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. zpredictions.jsonrrzinstances_val2017.jsonlvis_v1_z.json)r'rErrrG coco_evaluate)r%r pred_json anno_jsons r* eval_jsonzDetectionValidator.eval_jsonshMM$66 IIf  +/<<'x GXX]=^ `  !!%I>>r+c |jjr|js |jr}t |j rgt jd|d|d ||fD]}|jrJ|dt|tr|gn|}t|tr|gn|}tdddl m }m}||} | j|} t!|D]\} } || | | |jt j} |j"j$j&Dcgc] }t)t+|j,"c}| j._| j3| j5| j7| j8d |d || dd <| j8d |d || dd <| j8d|d<| j8d|d<| j8d|d<d| j8d zd| j8d zz|d<|jsY| j8d|d|| dd <| j8d|d|| dd <| j8d|d|| dd <|jr|d|d<|S|Scc}w#t:$r#}t j<d|Yd}~|Sd}~wwxYw) amEvaluate COCO/LVIS metrics using faster-coco-eval library. Performs evaluation using the faster-coco-eval library to compute mAP metrics for object detection. Updates the provided stats dictionary with computed metrics including mAP50, mAP50-95, and LVIS-specific metrics if applicable. Args: stats (dict[str, Any]): Dictionary to store computed metrics and statistics. pred_json (str | Path): Path to JSON file containing predictions in COCO format. anno_json (str | Path): Path to JSON file containing ground truth annotations in COCO format. iou_types (str | list[str]): IoU type(s) for evaluation. Can be single string or list of strings. Common values include "bbox", "segm", "keypoints". Defaults to "bbox". suffix (str | list[str]): Suffix to append to metric names in stats dictionary. Should correspond to iou_types if multiple types provided. Defaults to "Box". Returns: (dict[str, Any]): Updated stats dictionary containing the computed COCO/LVIS evaluation metrics. z' Evaluating faster-coco-eval mAP using z and z...z file not foundzfaster-coco-eval>=1.6.7r)COCOCOCOeval_faster)iouType lvis_styleprint_functionAP_50zmetrics/mAP50()AP_allzmetrics/mAP50-95(AP_smallzmetrics/mAP_small(B) AP_mediumzmetrics/mAP_medium(B)AP_largezmetrics/mAP_large(B)g?g?fitnessAPrz metrics/APr(APcz metrics/APc(APfz metrics/APf(zmetrics/mAP50-95(B)z faster-coco-eval unable to run: N)rrPrrrOrVr ris_filer2rHrfaster_coco_evalrr loadResrr&rim_filesrrrparamsimgIdsevaluate accumulate summarize stats_as_dict Exceptionr)r%rrr iou_typessuffixrzrr annorriou_typerQes r*rz DetectionValidator.coco_evaluates4 99  DLLDLLc$**o KKB9+US\R]]`a b% G"I->A99;=1#_(==;>+5i+EYK9 %/%<&&"#<=BI||I.#,Y#7YKAx)dH^d^i^iCEIOOD[D[DdDd(eqT!W\\):(eCJJ%LLNNN$MMO?B>O>OPW>XEN6!9Q<.:;ADARARS[A\E-fQil^1=>474E4Ej4QE01585F5F{5SE12474E4Ej4QE01'*S->->x-H'H3QTQbQbcjQkKk'kE)$||@C@Q@QRW@X VAYq\N!<=@C@Q@QRW@X VAYq\N!<=@C@Q@QRW@X VAYq\N!<=-Y0<<',-B'CE)$ u 1)f, G!A!EFF  Gs9$J5KK!)NNNN)r(z dict | NonereturnNone)r9dict[str, Any]r>r@)r\ztorch.nn.Moduler>r?)r>rH)rxz torch.Tensorr>list[dict[str, torch.Tensor]])rrr9r@r>r@)rdict[str, torch.Tensor]r>rB)rxrAr9r@r>r?)r>r?)r>r@)rxrBr9r@r>zdict[str, np.ndarray])rQN)rrHrrHr9 int | Noner>ztorch.utils.data.Dataset)rrHrrr>ztorch.utils.data.DataLoader)r9r@rrr>r?)N) r9r@rxrArrrkrCr>r?) rrBrrrztuple[int, int]rrr>r?)rrBrr@r>r?)rrBrr@r>rB)rr@r>r@)rBox) rr@rrHrrHr9str | list[str]r:rEr>r@)__name__ __module__ __qualname____doc__rr<r]rfr{rrrrrrrrrrrrrrrrrr __classcell__)r)s@r*rrs.$$ y2lk. 6 ,\. )'0)(] h *  ko # ,I OR ]g  8.$B  ?,&,"' BBB B # B  B Br+r)! __future__rrJpathlibrtypingrrrrtorch.distributed distributedrultralytics.datarrrultralytics.engine.validatorr ultralytics.utilsr r r r ultralytics.utils.checksrultralytics.utils.metricsrrrultralytics.utils.plottingrrrdr+r*rVsE#  LL6447JJ2~~r+