:jCddlmZddlmZmZmZmZddlZddl m Z ddl m Z ddl mZddlmZdd lmZGd d eeZy) ) annotations)DictOptionalSequenceUnionN)Protocol)BaseWaveformTransform)Metric)Task))SupervisedRepresentationLearningTaskMixincjeZdZdZ d dfd ZdZxZS)+SupervisedRepresentationLearningWithArcFaceaSupervised representation learning with ArcFace loss Representation learning is the task of ... Parameters ---------- protocol : Protocol pyannote.database protocol duration : float, optional Chunks duration in seconds. Defaults to two seconds (2.). min_duration : float, optional Sample training chunks duration uniformly between `min_duration` and `duration`. Defaults to `duration` (i.e. fixed length chunks). num_classes_per_batch : int, optional Number of classes per batch. Defaults to 32. num_chunks_per_class : int, optional Number of chunks per class. Defaults to 1. margin : float, optional Margin. Defaults to 28.6. scale : float, optional Scale. Defaults to 64. num_workers : int, optional Number of workers used for generating training samples. Defaults to multiprocessing.cpu_count() // 2. pin_memory : bool, optional If True, data loaders will copy tensors into CUDA pinned memory before returning them. See pytorch documentation for more details. Defaults to False. augmentation : BaseWaveformTransform, optional torch_audiomentations waveform transform, used by dataloader during training. metric : optional Validation metric(s). Can be anything supported by torchmetrics.MetricCollection. Defaults to AUROC (area under the ROC curve). c ||_||_||_||_t |||||j || | | y)N)duration min_duration batch_size num_workers pin_memory augmentationmetric)num_chunks_per_classnum_classes_per_batchmarginscalesuper__init__r) selfprotocolrrrrrrrrrr __class__s {/home/mcse/projects/srt_converter/srt-converter-venv/lib/python3.12/site-packages/pyannote/audio/tasks/embedding/arcface.pyrz4SupervisedRepresentationLearningWithArcFace.__init__RsQ%9!%:"    %#!%  c.|j|jjj\}}tjj t |jj||j|j|j_ y)N)rr) modelexample_input_arrayshapepytorch_metric_learninglosses ArcFaceLosslenspecificationsclassesrr loss_func)r_embedding_sizes r!setup_loss_funcz;SupervisedRepresentationLearningWithArcFace.setup_loss_funcrsm JJtzz'E'EFLL>6==II ##++ , ;;** J  r") Ng@ r g<@gP@NFNN)rrrzOptional[float]rfloatrintrr3rr2rr2rz Optional[int]rboolrzOptional[BaseWaveformTransform]rz2Union[Metric, Sequence[Metric], Dict[str, Metric]])__name__ __module__ __qualname____doc__rr0 __classcell__)r s@r!rr&s"V)-%'$%%) 8rCs40#22%&Q)=U -U r"