FijddlmZddlmZddlmZddlmZddlmZm Z ddl m Z m Z Gdd eZ y ) ) annotations)copy)DetectionTrainer)RTDETRDetectionModel)RANKcolorstr) RTDETRDatasetRTDETRValidatorc*eZdZdZdddZdd dZdZy) RTDETRTraineraTrainer class for the RT-DETR model developed by Baidu for real-time object detection. This class extends the DetectionTrainer class for YOLO to adapt to the specific features and architecture of RT-DETR. The model leverages Vision Transformers and has capabilities like IoU-aware query selection and adaptable inference speed. Attributes: loss_names (tuple): Names of the loss components used for training. data (dict): Dataset configuration containing class count and other parameters. args (dict): Training arguments and hyperparameters. save_dir (Path): Directory to save training results. test_loader (DataLoader): DataLoader for validation/testing data. Methods: get_model: Initialize and return an RT-DETR model for object detection tasks. build_dataset: Build and return an RT-DETR dataset for training or validation. get_validator: Return a DetectionValidator suitable for RT-DETR model validation. Examples: >>> from ultralytics.models.rtdetr.train import RTDETRTrainer >>> args = dict(model="rtdetr-l.yaml", data="coco8.yaml", imgsz=640, epochs=3) >>> trainer = RTDETRTrainer(overrides=args) >>> trainer.train() Notes: - F.grid_sample used in RT-DETR does not support the `deterministic=True` argument. - AMP training can lead to NaN outputs and may produce errors during bipartite graph matching. Nct||jd|jd|xr tdk(}|r|j||S)aWInitialize and return an RT-DETR model for object detection tasks. Args: cfg (dict, optional): Model configuration. weights (str, optional): Path to pre-trained model weights. verbose (bool): Verbose logging if True. Returns: (RTDETRDetectionModel): Initialized model. ncchannels)rchverbose)rdatarload)selfcfgweightsrmodels a/var/www/pod-logistic/pod-ai/venv/lib/python3.12/site-packages/ultralytics/models/rtdetr/train.py get_modelzRTDETRTrainer.get_model,sF%STYYt_:AV`g`vlptvlvw  JJw  cft||jj||dk(|jd|jjxsd|jjxsdt |d|jj |j|dk(r|jj Sd S)asBuild and return an RT-DETR dataset for training or validation. Args: img_path (str): Path to the folder containing images. mode (str): Dataset mode, either 'train' or 'val'. batch (int, optional): Batch size for rectangle training. Returns: (RTDETRDataset): Dataset object for the specific mode. trainFNz: g?) img_pathimgsz batch_sizeaugmenthyprectcache single_clsprefixclassesrfraction) r argsr r%r&rr(rr))rrmodebatchs r build_datasetzRTDETRTrainer.build_dataset<s))//GO ))//)Tyy++4utfBK(II%%+/7?TYY''  AD  rczd|_t|j|jt |j S)z@Return an RTDETRValidator suitable for RT-DETR model validation.) giou_losscls_lossl1_loss)save_dirr*) loss_namesr test_loaderr2rr*)rs r get_validatorzRTDETRTrainer.get_validatorVs-<t//$--dSWS\S\o^^r)NNT)rz dict | Nonerz str | Nonerbool)valN)rstrr+r8r,z int | None)__name__ __module__ __qualname____doc__rr-r5rrr r s:  4_rr N) __future__rrultralytics.models.yolo.detectrultralytics.nn.tasksrultralytics.utilsrrr7r r r r=rrrBs*#;5,/K_$K_r