描述bug

尝试重现MNIST演示。

重现步骤

重现行为的方法：
请按照以下步骤操作：https://uber.github.io/ludwig/examples/#image-classification-mnist
我的目录如下所示：

$ pwd
/Users/john/git/ludwig-sandbox/MNIST/mnist_png/mnist_png

$ ls
training
testing
mnist_dataset_training.csv
mnist_dataset_testing.csv
model_definition.yaml
results

time ludwig train   --data_train_csv mnist_dataset_training.csv   --data_test_csv  mnist_dataset_testing.csv   --model_definition_file model_definition.yaml

Traceback (most recent call last):
  File "/Users/john/.venv/ludwig/bin/ludwig", line 11, in <module>
    load_entry_point('ludwig==0.2.2.4', 'console_scripts', 'ludwig')()
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/cli.py", line 118, in main
    CLI()
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/cli.py", line 64, in __init__
    getattr(self, args.command)()
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/cli.py", line 74, in train
    train.cli(sys.argv[2:])
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/train.py", line 806, in cli
    full_train(**vars(args))
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/train.py", line 301, in full_train
    random_seed=random_seed
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/data/preprocessing.py", line 339, in preprocess_for_training
    random_seed=random_seed
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/data/preprocessing.py", line 528, in preprocess_for_training_by_type
    random_seed=random_seed
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/data/preprocessing.py", line 655, in _preprocess_csv_for_training
    random_seed=random_seed
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/data/preprocessing.py", line 96, in build_dataset_df
    global_preprocessing_parameters
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/data/preprocessing.py", line 171, in build_data
    preprocessing_parameters
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/features/image_feature.py", line 269, in add_feature_data
    preprocessing_parameters, first_image_path
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/ludwig/features/image_feature.py", line 179, in _finalize_preprocessing_parameters
    first_image = imread(first_image_path)
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/skimage/io/_io.py", line 62, in imread
    img = call_plugin('imread', fname, plugin=plugin, **plugin_args)
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/skimage/io/manage_plugins.py", line 214, in call_plugin
    return func(*args, **kwargs)
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/skimage/io/_plugins/pil_plugin.py", line 37, in imread
    return pil_to_ndarray(im, dtype=dtype, img_num=img_num)
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/skimage/io/_plugins/pil_plugin.py", line 67, in pil_to_ndarray
    image.seek(i)
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/PIL/PngImagePlugin.py", line 748, in seek
    self._seek(f)
  File "/Users/john/.venv/ludwig/lib/python3.7/site-packages/PIL/PngImagePlugin.py", line 791, in _seek
    cid, pos, length = self.png.read()
AttributeError: 'NoneType' object has no attribute 'read'

请提供代码、yaml定义文件和一个示例数据，以便完全重现问题。
无法重现的问题将被忽略。

预期行为

我希望它能运行实验。

环境信息(请填写以下信息):

• 操作系统：OSX
• 版本：10.15.4
• Python版本：Python 3.7.3
• Ludwig版本：ludwig==0.2.2.4
  附加上下文

我在pil插件中设置了一个断点，并检查了它正在使用的路径，这似乎完全有效。我认为所有目录的权限都是有效的...

$ ll /Users/john/git/ludwig-sandbox/MNIST/mnist_png/mnist_png/training/0/16585.png 
-rw-r----- 1 john staff 315 Dec 10  2015 /Users/john/git/ludwig-sandbox/MNIST/mnist_png/mnist_png/training/0/16585.png