o W hB%@sddlZddlZddlmZddlZddlmZddlmZm Z ddl m Z ddl m Z ddlmZddlmZmZddlmZmZmZdd lmZmZGd d d e ZdS) N)copy)build_dataloaderbuild_yolo_dataset) BaseTrainer)yolo)DetectionModel)LOGGERRANK) plot_images plot_labels plot_results) de_paralleltorch_distributed_zero_firstcseZdZdZdddZd dd Zd d Zd d Zd!ddZddZ d"ddZ ddZ ddZ ddZ ddZfddZZS)#DetectionTrainera A class extending the BaseTrainer class for training based on a detection model. This trainer specializes in object detection tasks, handling the specific requirements for training YOLO models for object detection. Attributes: model (DetectionModel): The YOLO detection model being trained. data (dict): Dictionary containing dataset information including class names and number of classes. loss_names (Tuple[str]): Names of the loss components used in training (box_loss, cls_loss, dfl_loss). Methods: build_dataset: Build YOLO dataset for training or validation. get_dataloader: Construct and return dataloader for the specified mode. preprocess_batch: Preprocess a batch of images by scaling and converting to float. set_model_attributes: Set model attributes based on dataset information. get_model: Return a YOLO detection model. get_validator: Return a validator for model evaluation. label_loss_items: Return a loss dictionary with labeled training loss items. progress_string: Return a formatted string of training progress. plot_training_samples: Plot training samples with their annotations. plot_metrics: Plot metrics from a CSV file. plot_training_labels: Create a labeled training plot of the YOLO model. auto_batch: Calculate optimal batch size based on model memory requirements. Examples: >>> from ultralytics.models.yolo.detect import DetectionTrainer >>> args = dict(model="yolo11n.pt", data="coco8.yaml", epochs=3) >>> trainer = DetectionTrainer(overrides=args) >>> trainer.train() trainNc CsBtt|jr t|jjndd}t|j|||j||dk|dS)a Build YOLO Dataset for training or validation. 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 object configured for the specified mode. r val)moderectstride)maxintmodelr rrargsdata)selfimg_pathrbatchgsrX/var/www/vscode/kcb/lib/python3.10/site-packages/ultralytics/models/yolo/detect/train.py build_dataset4s$ zDetectionTrainer.build_datasetrcCs|dvs Jd|dt|||||}Wdn1s"wY|dk}t|ddr:|r:tdd}|dkrB|jjn|jjd }t|||||S) a Construct and return dataloader for the specified mode. Args: dataset_path (str): Path to the dataset. batch_size (int): Number of images per batch. rank (int): Process rank for distributed training. mode (str): 'train' for training dataloader, 'val' for validation dataloader. Returns: (DataLoader): PyTorch dataloader object. >rrz#Mode must be 'train' or 'val', not .NrrFzJ'rect=True' is incompatible with DataLoader shuffle, setting shuffle=False)rr!getattrrwarningrworkersr)r dataset_path batch_sizerankrdatasetshuffler'rrr get_dataloaderCs  zDetectionTrainer.get_dataloadercs|djjddd|d<jjr_|d}ttjjdtjjdj j j }|t |j ddd kr[fd d |j ddD}t j j||d d d}||d<|S)a Preprocess a batch of images by scaling and converting to float. Args: batch (dict): Dictionary containing batch data with 'img' tensor. Returns: (dict): Preprocessed batch with normalized images. imgT) non_blockingg?g?r$Ncs&g|]}t|jjqSr)mathceilr.0xrsfrr nsz5DetectionTrainer.preprocess_batch..bilinearF)sizer align_corners)todevicefloatr multi_scalerandom randrangerimgszrrshapenn functional interpolate)rrimgssznsrr7r preprocess_batchZs" (  z!DetectionTrainer.preprocess_batchcCs*|jd|j_|jd|j_|j|j_dS)z2Set model attributes based on dataset information.ncnamesN)rrrLrMrrrrr set_model_attributesusz%DetectionTrainer.set_model_attributesTcCs6t||jd|jd|otdkd}|r|||S)aF Return a YOLO detection model. Args: cfg (str, optional): Path to model configuration file. weights (str, optional): Path to model weights. verbose (bool): Whether to display model information. Returns: (DetectionModel): YOLO detection model. rLchannels)rLchverbose)rrr load)rcfgweightsrSrrrr get_models$  zDetectionTrainer.get_modelcCs&d|_tjj|j|jt|j|jdS)z6Return a DetectionValidator for YOLO model validation.)box_losscls_lossdfl_loss)save_dirr _callbacks) loss_namesrdetectDetectionValidator test_loaderr[rr callbacksrNrrr get_validatorszDetectionTrainer.get_validatorcs<fdd|jD}|durdd|D}tt||S|S)ah Return a loss dict with labeled training loss items tensor. Args: loss_items (List[float], optional): List of loss values. prefix (str): Prefix for keys in the returned dictionary. Returns: (Dict | List): Dictionary of labeled loss items if loss_items is provided, otherwise list of keys. csg|] }d|qS)/rr4prefixrr r9z5DetectionTrainer.label_loss_items..NcSsg|] }tt|dqS))roundr?r4rrr r9rf)r]dictzip)r loss_itemsrekeysrrdr label_loss_itemss  z!DetectionTrainer.label_loss_itemscCs.dddt|jddg|jddRS)z`Return a formatted string of training progress with epoch, GPU memory, loss, instances and size. z%11sEpochGPU_mem InstancesSize)lenr]rNrrr progress_stringsz DetectionTrainer.progress_stringc CsDt|d|d|dd|d|d|jd|d|jd d S) z Plot training samples with their annotations. Args: batch (dict): Dictionary containing batch data. ni (int): Number of iterations. r. batch_idxclsrQbboxesim_file train_batchz.jpg)imagesrvrwrxpathsfnameon_plotN)r squeezer[r~)rrnirrr plot_training_sampless  z&DetectionTrainer.plot_training_samplescCst|j|jddS)zPlot metrics from a CSV file.)filer~N)r csvr~rNrrr plot_metricsszDetectionTrainer.plot_metricscCs\tdd|jjjDd}tdd|jjjDd}t|||jd|j|j ddS)z1Create a labeled training plot of the YOLO model.cSg|]}|dqS)rxrr5lbrrr r9z9DetectionTrainer.plot_training_labels..rcSr)rwrrrrr r9rrM)rMr[r~N) np concatenate train_loaderr+labelsr rrr[r~)rboxesrwrrr plot_training_labelss$z%DetectionTrainer.plot_training_labelscs6|j|jddd}tdd|jDd}t|S)z Get optimal batch size by calculating memory occupation of model. Returns: (int): Optimal batch size. rr")rrcss|] }t|dVqdS)rwN)rt)r5labelrrr sz.DetectionTrainer.auto_batch..ro)r!trainsetrrsuper auto_batch)r train_dataset max_num_obj __class__rr rs zDetectionTrainer.auto_batch)rN)r"rr)NNT)Nr)__name__ __module__ __qualname____doc__r!r-rKrOrWrbrmrurrrr __classcell__rrrr rs    r)r2rArnumpyrtorch.nnrEultralytics.datarrultralytics.engine.trainerrultralytics.modelsrultralytics.nn.tasksrultralytics.utilsrr ultralytics.utils.plottingr r r ultralytics.utils.torch_utilsr rrrrrr s