o W h@s\ddlmZddlmZddlmZddlmZmZddlm Z m Z Gdddej j Z dS) )copy)yolo)SegmentationModel) DEFAULT_CFGRANK) plot_images plot_resultscsJeZdZdZeddffdd ZdddZdd Zd d Zd d Z Z S)SegmentationTrainera A class extending the DetectionTrainer class for training based on a segmentation model. This trainer specializes in handling segmentation tasks, extending the detection trainer with segmentation-specific functionality including model initialization, validation, and visualization. Attributes: loss_names (Tuple[str]): Names of the loss components used during training. Examples: >>> from ultralytics.models.yolo.segment import SegmentationTrainer >>> args = dict(model="yolo11n-seg.pt", data="coco8-seg.yaml", epochs=3) >>> trainer = SegmentationTrainer(overrides=args) >>> trainer.train() Ncs(|duri}d|d<t|||dS)a} Initialize a SegmentationTrainer object. This initializes a trainer for segmentation tasks, extending the detection trainer with segmentation-specific functionality. It sets the task to 'segment' and prepares the trainer for training segmentation models. Args: cfg (dict): Configuration dictionary with default training settings. Defaults to DEFAULT_CFG. overrides (dict, optional): Dictionary of parameter overrides for the default configuration. _callbacks (list, optional): List of callback functions to be executed during training. Examples: >>> from ultralytics.models.yolo.segment import SegmentationTrainer >>> args = dict(model="yolov8n-seg.pt", data="coco8-seg.yaml", epochs=3) >>> trainer = SegmentationTrainer(overrides=args) >>> trainer.train() Nsegmenttask)super__init__)selfcfg overrides _callbacks __class__Y/var/www/vscode/kcb/lib/python3.10/site-packages/ultralytics/models/yolo/segment/train.pyr szSegmentationTrainer.__init__TcCs6t||jd|jd|otdkd}|r|||S)a Initialize and return a SegmentationModel with specified configuration and weights. Args: cfg (dict | str | None): Model configuration. Can be a dictionary, a path to a YAML file, or None. weights (str | Path | None): Path to pretrained weights file. verbose (bool): Whether to display model information during initialization. Returns: (SegmentationModel): Initialized segmentation model with loaded weights if specified. Examples: >>> trainer = SegmentationTrainer() >>> model = trainer.get_model(cfg="yolov8n-seg.yaml") >>> model = trainer.get_model(weights="yolov8n-seg.pt", verbose=False) ncchannels)rchverbose)rdatarload)rrweightsrmodelrrr get_model3s$ zSegmentationTrainer.get_modelcCs&d|_tjj|j|jt|j|jdS)zIReturn an instance of SegmentationValidator for validation of YOLO model.)box_lossseg_losscls_lossdfl_loss)save_dirargsr) loss_namesrr SegmentationValidator test_loaderr$rr% callbacksrrrr get_validatorJsz!SegmentationTrainer.get_validatorc CsJt|d|d|dd|d|d|d|jd|d |jd d S) a# Plot training sample images with labels, bounding boxes, and masks. This method creates a visualization of training batch images with their corresponding labels, bounding boxes, and segmentation masks, saving the result to a file for inspection and debugging. Args: batch (dict): Dictionary containing batch data with the following keys: 'img': Images tensor 'batch_idx': Batch indices for each box 'cls': Class labels tensor (squeezed to remove last dimension) 'bboxes': Bounding box coordinates tensor 'masks': Segmentation masks tensor 'im_file': List of image file paths ni (int): Current training iteration number, used for naming the output file. Examples: >>> trainer = SegmentationTrainer() >>> batch = { ... "img": torch.rand(16, 3, 640, 640), ... "batch_idx": torch.zeros(16), ... "cls": torch.randint(0, 80, (16, 1)), ... "bboxes": torch.rand(16, 4), ... "masks": torch.rand(16, 640, 640), ... "im_file": ["image1.jpg", "image2.jpg"], ... } >>> trainer.plot_training_samples(batch, ni=5) img batch_idxclsrbboxesmasksim_file train_batchz.jpg)r0pathsfnameon_plotN)rsqueezer$r5)rbatchnirrrplot_training_samplesQs  z)SegmentationTrainer.plot_training_samplescCst|jd|jddS)zPlots training/val metrics.T)filer r5N)rcsvr5r*rrr plot_metricsysz SegmentationTrainer.plot_metrics)NNT) __name__ __module__ __qualname____doc__rr rr+r9r< __classcell__rrrrr s (r N)rultralytics.modelsrultralytics.nn.tasksrultralytics.utilsrrultralytics.utils.plottingrrdetectDetectionTrainerr rrrrs