Python mmdet.datasets.CocoDataset() Examples

The following are 17 code examples of mmdet.datasets.CocoDataset(). You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may also want to check out all available functions/classes of the module mmdet.datasets , or try the search function .
Example #1
Source File: train.py    From mmdetection_with_SENet154 with Apache License 2.0 5 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    data_loaders = [
        build_dataloader(
            dataset,
            cfg.data.imgs_per_gpu,
            cfg.data.workers_per_gpu,
            dist=True)
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())
    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)
    # register hooks
    optimizer_config = DistOptimizerHook(**cfg.optimizer_config)
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(CocoDistEvalRecallHook(val_dataset_cfg))
        else:
            dataset_type = getattr(datasets, val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(CocoDistEvalmAPHook(val_dataset_cfg))
            else:
                runner.register_hook(DistEvalmAPHook(val_dataset_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #2
Source File: train.py    From kaggle-imaterialist with MIT License 5 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    data_loaders = [
        build_dataloader(
            dataset,
            cfg.data.imgs_per_gpu,
            cfg.data.workers_per_gpu,
            dist=True)
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())
    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)
    # register hooks
    optimizer_config = DistOptimizerHook(**cfg.optimizer_config)
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(CocoDistEvalRecallHook(val_dataset_cfg))
        else:
            dataset_type = getattr(datasets, val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(CocoDistEvalmAPHook(val_dataset_cfg))
            else:
                runner.register_hook(DistEvalmAPHook(val_dataset_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #3
Source File: train.py    From Grid-R-CNN with Apache License 2.0 5 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    data_loaders = [
        build_dataloader(
            dataset,
            cfg.data.imgs_per_gpu,
            cfg.data.workers_per_gpu,
            dist=True)
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())
    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)
    # register hooks
    optimizer_config = DistOptimizerHook(**cfg.optimizer_config)
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(CocoDistEvalRecallHook(val_dataset_cfg))
        else:
            dataset_type = getattr(datasets, val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(CocoDistEvalmAPHook(val_dataset_cfg))
            else:
                runner.register_hook(DistEvalmAPHook(val_dataset_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #4
Source File: train.py    From AerialDetection with Apache License 2.0 5 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    data_loaders = [
        build_dataloader(
            dataset,
            cfg.data.imgs_per_gpu,
            cfg.data.workers_per_gpu,
            dist=True)
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())
    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)
    # register hooks
    optimizer_config = DistOptimizerHook(**cfg.optimizer_config)
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(CocoDistEvalRecallHook(val_dataset_cfg))
        else:
            dataset_type = getattr(datasets, val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(CocoDistEvalmAPHook(val_dataset_cfg))
            else:
                runner.register_hook(DistEvalmAPHook(val_dataset_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #5
Source File: divide_dataset.py    From FNA with Apache License 2.0 5 votes vote down vote up 
def build_divide_dataset(cfg, part_1_ratio=0.5, seed=520):
    """Need to change `coco.getImgIds()` output, i.e., 
    `self.imgs` in a COCO instance. 
    `self.imgs` is created by `self.dataset['images']`.
    Thus, hacking `self.dataset['images']` when initializing a 
    COCO class.
    """
    logging.info('Prepare datasets.')
    data_type = cfg.train.type
    if data_type == 'RepeatDataset':
        train_cfg = cfg.train.dataset
    else:
        train_cfg = cfg.train
    assert train_cfg.pop('type') == 'CocoDataset', 'Only support COCO.'
    annotations = mmcv.load(train_cfg.pop('ann_file'))
    images = annotations.pop('images')
    part_1_annotations = copy.copy(annotations)
    part_2_annotations = copy.copy(annotations)

    part_1_length = int(part_1_ratio * len(images))
    if seed is not None:
        random.seed(seed)
        random.shuffle(images)
    part_1_images = images[:part_1_length]
    part_2_images = images[part_1_length:]

    part_1_annotations['images'] = part_1_images
    part_2_annotations['images'] = part_2_images

    part_1_coco = COCOFromDict(part_1_annotations)
    part_2_coco = COCOFromDict(part_2_annotations)
    part_1_dataset = InitDatasetFromCOCOClass(**train_cfg, ann_file=part_1_coco)
    part_2_dataset = InitDatasetFromCOCOClass(**cfg.val, ann_file=part_2_coco)
    if data_type == 'RepeatDataset':
        part_1_dataset = RepeatDataset(part_1_dataset, cfg.train.times)
    logging.info(f'Finished preparing datasets.')

    return part_1_dataset, part_2_dataset
Example #6
Source File: train.py    From IoU-Uniform-R-CNN with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    dataset = dataset if isinstance(dataset, (list, tuple)) else [dataset]
    data_loaders = [
        build_dataloader(
            ds, cfg.data.imgs_per_gpu, cfg.data.workers_per_gpu, dist=True)
        for ds in dataset
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = DATASETS.get(val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #7
Source File: train.py    From ttfnet with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    dataset = dataset if isinstance(dataset, (list, tuple)) else [dataset]
    data_loaders = [
        build_dataloader(
            ds, cfg.data.imgs_per_gpu, cfg.data.workers_per_gpu, dist=True)
        for ds in dataset
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = DATASETS.get(val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #8
Source File: train.py    From CenterNet with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    data_loaders = [
        build_dataloader(
            dataset,
            cfg.data.imgs_per_gpu,
            cfg.data.workers_per_gpu,
            dist=True)
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = getattr(datasets, val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #9
Source File: train.py    From Feature-Selective-Anchor-Free-Module-for-Single-Shot-Object-Detection with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    dataset = dataset if isinstance(dataset, (list, tuple)) else [dataset]
    data_loaders = [
        build_dataloader(
            ds, cfg.data.imgs_per_gpu, cfg.data.workers_per_gpu, dist=True)
        for ds in dataset
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = DATASETS.get(val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #10
Source File: train.py    From Cascade-RPN with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    dataset = dataset if isinstance(dataset, (list, tuple)) else [dataset]
    data_loaders = [
        build_dataloader(
            ds, cfg.data.imgs_per_gpu, cfg.data.workers_per_gpu, dist=True)
        for ds in dataset
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, (RPN, CascadeRPN)):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = DATASETS.get(val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #11
Source File: train.py    From FoveaBox with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    dataset = dataset if isinstance(dataset, (list, tuple)) else [dataset]
    data_loaders = [
        build_dataloader(
            ds, cfg.data.imgs_per_gpu, cfg.data.workers_per_gpu, dist=True)
        for ds in dataset
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = DATASETS.get(val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #12
Source File: train.py    From Libra_R-CNN with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    data_loaders = [
        build_dataloader(
            dataset,
            cfg.data.imgs_per_gpu,
            cfg.data.workers_per_gpu,
            dist=True)
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = getattr(datasets, val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #13
Source File: train.py    From RDSNet with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    dataset = dataset if isinstance(dataset, (list, tuple)) else [dataset]
    data_loaders = [
        build_dataloader(
            ds, cfg.data.imgs_per_gpu, cfg.data.workers_per_gpu, dist=True)
        for ds in dataset
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = DATASETS.get(val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #14
Source File: train.py    From kaggle-kuzushiji-recognition with MIT License 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    dataset = dataset if isinstance(dataset, (list, tuple)) else [dataset]
    data_loaders = [
        build_dataloader(
            ds, cfg.data.imgs_per_gpu, cfg.data.workers_per_gpu, dist=True)
        for ds in dataset
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = DATASETS.get(val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalF1Hook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #15
Source File: train.py    From PolarMask with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    dataset = dataset if isinstance(dataset, (list, tuple)) else [dataset]
    data_loaders = [
        build_dataloader(
            ds, cfg.data.imgs_per_gpu, cfg.data.workers_per_gpu, dist=True)
        for ds in dataset
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = DATASETS.get(val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #16
Source File: train.py    From mmdetection-annotated with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    data_loaders = [
        build_dataloader(
            dataset,
            cfg.data.imgs_per_gpu,
            cfg.data.workers_per_gpu,
            dist=True)
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = getattr(datasets, val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)
Example #17
Source File: train.py    From GCNet with Apache License 2.0 4 votes vote down vote up 
def _dist_train(model, dataset, cfg, validate=False):
    # prepare data loaders
    data_loaders = [
        build_dataloader(
            dataset,
            cfg.data.imgs_per_gpu,
            cfg.data.workers_per_gpu,
            dist=True)
    ]
    # put model on gpus
    model = MMDistributedDataParallel(model.cuda())

    # build runner
    optimizer = build_optimizer(model, cfg.optimizer)
    runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
                    cfg.log_level)

    # fp16 setting
    fp16_cfg = cfg.get('fp16', None)
    if fp16_cfg is not None:
        optimizer_config = Fp16OptimizerHook(**cfg.optimizer_config,
                                             **fp16_cfg)
    else:
        optimizer_config = DistOptimizerHook(**cfg.optimizer_config)

    # register hooks
    runner.register_training_hooks(cfg.lr_config, optimizer_config,
                                   cfg.checkpoint_config, cfg.log_config)
    runner.register_hook(DistSamplerSeedHook())
    # register eval hooks
    if validate:
        val_dataset_cfg = cfg.data.val
        eval_cfg = cfg.get('evaluation', {})
        if isinstance(model.module, RPN):
            # TODO: implement recall hooks for other datasets
            runner.register_hook(
                CocoDistEvalRecallHook(val_dataset_cfg, **eval_cfg))
        else:
            dataset_type = getattr(datasets, val_dataset_cfg.type)
            if issubclass(dataset_type, datasets.CocoDataset):
                runner.register_hook(
                    CocoDistEvalmAPHook(val_dataset_cfg, **eval_cfg))
            else:
                runner.register_hook(
                    DistEvalmAPHook(val_dataset_cfg, **eval_cfg))

    if cfg.resume_from:
        runner.resume(cfg.resume_from)
    elif cfg.load_from:
        runner.load_checkpoint(cfg.load_from)
    runner.run(data_loaders, cfg.workflow, cfg.total_epochs)