/j3ddlmZddlmZddlmZddlZddlZddl m cm Z ddl mZddlmZmZddlmZddlmZmZGd d eZdS) ) annotations)Path)AnyN)DetectionValidator)LOGGERops)check_requirements)SegmentMetricsmask_iouceZdZdZd)d*fd Zd+fd Zd,fd Zd-dZd.fd Zd/fd Z d0fd Z d1fd Z d2d#Z d3fd% Z d4fd& Zd5fd( ZxZS)6SegmentationValidatoraA class extending the DetectionValidator class for validation based on a segmentation model. This validator handles the evaluation of segmentation models, processing both bounding box and mask predictions to compute metrics such as mAP for both detection and segmentation tasks. Attributes: plot_masks (list): List to store masks for plotting. process (callable): Function to process masks based on save_json and save_txt flags. args (SimpleNamespace): Arguments for the validator. metrics (SegmentMetrics): Metrics calculator for segmentation tasks. stats (dict): Dictionary to store statistics during validation. Examples: >>> from ultralytics.models.yolo.segment import SegmentationValidator >>> args = dict(model="yolo26n-seg.pt", data="coco8-seg.yaml") >>> validator = SegmentationValidator(args=args) >>> validator() N _callbacks dict | NonereturnNonect||||d|_d|j_t |_dS)aInitialize SegmentationValidator and set task to 'segment', metrics to SegmentMetrics. Args: dataloader (torch.utils.data.DataLoader, optional): DataLoader to use for validation. save_dir (Path, optional): Directory to save results. args (dict, optional): Arguments for the validator. _callbacks (dict, optional): Dictionary of callback functions. Nsegment)super__init__processargstaskr metrics)self dataloadersave_dirrr __class__s h/home/longshao/multi-rider-rag/.venv/lib/python3.11/site-packages/ultralytics/models/yolo/segment/val.pyrzSegmentationValidator.__init__&sD XtZ@@@ " %'' batchdict[str, Any]ct|}|d|d<|S)zPreprocess batch of images for YOLO segmentation validation. Args: batch (dict[str, Any]): Batch containing images and annotations. Returns: (dict[str, Any]): Preprocessed batch. masks)r preprocessfloat)rr rs rr$z SegmentationValidator.preprocess4s:""5))w--//g rmodeltorch.nn.Modulect||jjrt d|jjs |jjr t jn t j|_ dS)zInitialize metrics and select mask processing function based on save_json flag. Args: model (torch.nn.Module): Model to validate. zfaster-coco-eval>=1.6.7N) r init_metricsr save_jsonr save_txtrprocess_mask_native process_maskr)rr&rs rr)z"SegmentationValidator.init_metricsAsd U### 9  : 8 9 9 926)2EqI[qs..adaq rstrc ddzS)z5Return a formatted description of evaluation metrics.z,%22s%11s%11s%11s%11s%11s%11s%11s%11s%11s%11s) ClassImages InstanceszBox(PRmAP50 mAP50-95)zMask(Pr3r4r5)rs rget_desczSegmentationValidator.get_descMs$ )  rpredslist[torch.Tensor]list[dict[str, torch.Tensor]]cLt|dtr|ddn|d}t|d}d|jddD}t |D]\}}|d}|jdr%|||||d|nQtj dg|jtj ur|n|jddRtj |dj |d <|S) aPost-process YOLO predictions and return output detections with proto. Args: preds (list[torch.Tensor]): Raw predictions from the model. Returns: (list[dict[str, torch.Tensor]]): Processed detection predictions with masks. rcg|]}d|zSr6).0xs r z5SegmentationValidator.postprocess..hs0001Q000rNextrabboxes)shape)dtypedevicer#) isinstancetupler postprocessrF enumeratepoprtorchzerosrr,uint8rH)rr8protoimgszipred coefficientrs rrKz!SegmentationValidator.postprocess]s* *%(E::Ha a##E!H--00 ABB000 ''  GAt((7++K$Q' U1X{DN% PPP[a4<33J#J#J%%PUP[\]\^\^P_aa+>0 MM rsiintc@t||}|djd}jjr[|d|}t jd|dz|j|dd}||k }n|d|d|k}|r`fd|dD}|jdd |kr9tj |d |d d d}| d }||d<|S)a:Prepare a batch for validation by processing images and targets. 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. clsrr#r<)rH batch_idxcDg|]}jtjur|n|dzSr>)rrr,)r@srs rrBz8SegmentationValidator._prepare_batch..s2sssVWdlc.EEE1PQ6sssrrRNbilinearF)mode align_cornersg?) r_prepare_batchrFr overlap_maskrNarangerHviewr%F interpolategt_) rrVr prepared_batchnlr#index mask_sizers ` rr`z$SegmentationValidator._prepare_batchvs(//E:: E " ( + 9 ! ='N2&ELBF5<@@@EEb!QOOEe^**,,EE'N5#5#;>> preds = {"cls": torch.tensor([1, 0]), "masks": torch.rand(2, 640, 640), "bboxes": torch.rand(2, 4)} >>> batch = {"cls": torch.tensor([1, 0]), "masks": torch.rand(2, 640, 640), "bboxes": torch.rand(2, 4)} >>> correct_preds = validator._process_batch(preds, batch) Notes: - This method computes IoU between predicted and ground truth masks. - Overlapping masks are handled based on the overlap_mask argument setting. rYrrGr#r<tp_m)r_process_batchrFnprOniouboolr flattenr%match_predictionscpunumpyupdate)rr8r tpgt_clsroiours rrpz$SegmentationValidator._process_batchs&WW # #E5 1 1u <?a  5<#5a#8A#=#=8U5\/2DI>dKKKDD5>11!44eGn6L6LQ6O6O6U6U6W6WXXC))%,DDHHJJPPRRD 64.!!! rnic|D]}|d}|jd|jjkr"tjd|jjdt j|d|jjt j|d<t ||||jjdS)aPlot batch predictions with masks and bounding boxes. 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. r#rz&Limiting validation plots to 'max_det=z' items.Nrn)max_det) rFrr~rwarningrN as_tensorrPrvrplot_predictions)rr r8r|pr#rs rrz&SegmentationValidator.plot_predictionss ^ ^AgJE{1~ 111c HYcccddd/B1B/B)C5;WWW[[]]AgJJ   r49;L MMMMMrpredn save_confrsrFtuple[int, int]filerc ddlm}|tj|d|dftjd|jt j|d|dd|d dgd t j |d t j  || dS)aSave YOLO detections to a txt file in normalized coordinates in a specific format. Args: predn (dict[str, torch.Tensor]): Prediction dictionary containing 'bboxes', 'conf', 'cls', and 'masks' keys. save_conf (bool): Whether to save confidence scores. shape (tuple[int, int]): Shape of the original image. file (Path): File path to save the detections. r)Resultsr<rnNrEconfrY)dimr#)pathnamesboxesr#)r) ultralytics.engine.resultsrrqrOrPrrNcat unsqueezerr+)rrrrFrrs r save_one_txtz"SegmentationValidator.save_one_txts 766666 HeAha) : : :*)U8_eFm.E.Eb.I.I5QV.to_stringsF3v;;''  q NNq55VAE]+++A DA!GA*+T@AGGaD"T GAMM#a&&))) 776?? "rpixels torch.Tensorc|ddddf|ddddfk}tj|\}}|dz}g}t|jdD] }|||k}t |rtj|}|d|d| t |||dz nt ||g}||ddkrdg|}| | |S)aIConvert multiple binary masks using Run-Length Encoding (RLE). Args: pixels (torch.Tensor): A 2D tensor where each row represents a flattened binary mask with shape [N, H*W]. Returns: (list[list[int]]): A list of RLE counts for each mask. Nr<rr) rNwhererrFrdifftolistinsertitemr)r transitionsrow_idxcol_idxrrS positionscounts r multi_encodez8SegmentationValidator.pred_to_json..multi_encodesV!ABB-6!!!SbS&>9K${;77 GWkGF6<?++ % %#GqL1 y>>-!Jy1188::ELLIaL$5$5$7$7888LLVAY)B-2D2D2F2F!FGGGG ^^,E!9Q<$$&&!++KKE e$$$$Mrr#rCr<r)sizer segmentationN)rrrr.)rrrr) transpose contiguousrcrrFrr pred_to_jsonrLjdict)rrrrr pred_maskshwrrlesrrSrrs rrz"SegmentationValidator.pred_to_jsons6 # # # #B    @7^--a33>>@@EEc%PW.FYFY[]^^ W~#AaC(1j)) B BA KK!Q99Q<<@@ A A A A UF+++dOO ; ;DAq9:DJD zA~ &~ 6 6 ; ;rcit||dtj|dd|d|ddiS)z.Scales predictions to the original image size.r#N ori_shape ratio_pad)rr)r scale_predsr scale_masksbyte)rrrrs rrz!SegmentationValidator.scale_preds"so gg!!%00 S_U7^D%96+;NZ`alZmnnndff    rstatsc|jdz }|jddz |jrdnd|jjdz }t |||ddgd d g S) z;Return COCO-style instance segmentation evaluation metrics.zpredictions.jsonrrzinstances_val2017.jsonlvis_v1_z.jsonbboxsegmBoxMask)suffix)rdatais_cocorsplitr coco_evaluate)rr pred_json anno_jsonrs r eval_jsonzSegmentationValidator.eval_json+s~M$66 If  +/<^''=^ =^=^=^ `  ww$$UIy66BR\aci[j$kkkr)NNNN)rrrr)r r!rr!)r&r'rr)rr.)r8r9rr:)rVrWr r!rr!)r8rkr r!rrl)r r!r8r:r|rWrr) rrkrrsrFrrrrr)rrkrr!rr)rrkrr!rrk)rr!rr!)__name__ __module__ __qualname____doc__rr$r)r7rKr`rprrrrr __classcell__)rs@rr r s& ( ( ( ( ( ( (       r r r r r r    24: N N N N N N....&Q;Q;Q;Q;Q;Q;f      llllllllllrr ) __future__rpathlibrtypingrrwrqrNtorch.nn.functionalnn functionalrdultralytics.models.yolo.detectrultralytics.utilsrrultralytics.utils.checksr ultralytics.utils.metricsr r r r6rrrs #""""" ======))))))))777777>>>>>>>>alalalalal.alalalalalr