如何解决val_dataloader步骤中的‘Module is not callable’错误

错误原因 #

你代码中调用的pytorchvideo.data.ucf101是一个Python模块，而非可实例化的类或可调用函数，直接对模块加括号调用会触发该错误。

解决方案 #

方法1：直接调用模块内的Ucf101类 #

将所有pytorchvideo.data.ucf101(替换为pytorchvideo.data.ucf101.Ucf101(，修正后的UCFDataModule代码如下：

class UCFDataModule(pytorch_lightning.LightningDataModule):

  _CLIP_DURATION = 2  # Duration of sampled clip for each video
  _BATCH_SIZE = 8
  _NUM_WORKERS = 8  # Number of parallel processes fetching data
    
  def train_dataloader(self):

    train_dataset = pytorchvideo.data.ucf101.Ucf101(
        data_path = os.path.join(dataset_root_path, "train"),
        clip_sampler = pytorchvideo.data.make_clip_sampler("random", self._CLIP_DURATION),
        transform = train_transform
    )

    return torch.utils.data.DataLoader(
        train_dataset,
        batch_size = self._BATCH_SIZE,
        num_workers = self._NUM_WORKERS,
    )


  def val_dataloader(self):

    val_dataset = pytorchvideo.data.ucf101.Ucf101(
        data_path = os.path.join(dataset_root_path, "val"),
        clip_sampler = pytorchvideo.data.make_clip_sampler("uniform", self._CLIP_DURATION),
        decode_audio = False,
        transform = val_transform
    )

    return torch.utils.data.DataLoader(
        val_dataset,
        batch_size = self._BATCH_SIZE,
        num_workers = self._NUM_WORKERS
    )

  def test_dataloader(self):

    test_dataset = pytorchvideo.data.ucf101.Ucf101(
        data_path = os.path.join(dataset_root_path, "test"),
        clip_sampler = pytorchvideo.data.make_clip_sampler("uniform", self._CLIP_DURATION),
        decode_audio = False,
        transform = val_transform 
    )

    return torch.utils.data.DataLoader(
        test_dataset,
        batch_size = self._BATCH_SIZE,
        num_workers = self._NUM_WORKERS
    )

方法2：提前导入类简化代码 #

在文件开头添加导入语句：

from pytorchvideo.data.ucf101 import Ucf101

之后三个dataloader函数中直接使用Ucf101(即可，示例：

train_dataset = Ucf101(
    data_path = os.path.join(dataset_root_path, "train"),
    clip_sampler = pytorchvideo.data.make_clip_sampler("random", self._CLIP_DURATION),
    transform = train_transform
)

额外注意事项 #

确保代码中dataset_root_path、train_transform、val_transform这些变量已正确定义并初始化，避免后续出现未定义相关的错误。

内容的提问来源于stack exchange，提问作者Ninalol