Autogluon: Error on demo code execution -- in get_dataset if '.rec' in path: TypeError: argument of type 'NoneType' is not iterable

[UNIcodeX]

The following code yields an error. I'm using a Conda environment with Python 3.6 because it is required for the ConfigSpace Python package.

import autogluon as ag
from autogluon import ImageClassification as task

# Load a toy image dataset:
filename = ag.download('http://autogluon-hackathon.s3-website-us-west-2.amazonaws.com/data.zip')
ag.unzip(filename)
dataset = task.Dataset(train_path='data/train')

# Fit classification models:
classifier = task.fit(dataset, epochs=2)

# Evaluate predictions on test data:

test_dataset = task.Dataset('data/test', train=False)
test_acc = classifier.evaluate(test_dataset)

This also gives the error:

from autogluon import ImageClassification as task
dataset = task.Dataset(train_path='data/train',
                       test_path='data/test')
classifier = task.fit(dataset,
                      net=ag.space.Categorical('ResNet50_v1b', 'ResNet18_v1b'),
                      time_limits=None,
                      # ngpus_per_trial=1,
                      num_trials = 4)
image = 'data/test/BabyShirt/BabyShirt_323.jpg'
ind, prob = classifier.predict(image)

The error:

scheduler: FIFOScheduler(
DistributedResourceManager{
(Remote: Remote REMOTE_ID: 0, 
        <Remote: 'inproc://172.18.30.18/28461/1' processes=1 threads=4, memory=7.28 GB>, Resource: NodeResourceManager(4 CPUs, 0 GPUs))
})

Starting Experiments
Num of Finished Tasks is 0
Num of Pending Tasks is 4
 25%|█████████████████████▌                                                                | 1/4 [00:00<00:00,  8.97it/s]Process Process-2:
Traceback (most recent call last):
  File "/home/z/anaconda3/envs/IC2/lib/python3.6/multiprocessing/process.py", line 258, in _bootstrap
    self.run()
  File "/home/z/anaconda3/envs/IC2/lib/python3.6/multiprocessing/process.py", line 93, in run
    self._target(*self._args, **self._kwargs)
  File "/home/z/anaconda3/envs/IC2/lib/python3.6/site-packages/autogluon/scheduler/scheduler.py", line 125, in _worker
    ret = fn(**kwargs)
  File "/home/z/anaconda3/envs/IC2/lib/python3.6/site-packages/autogluon/core/decorator.py", line 52, in __call__
    args = sample_config(args, new_config)
  File "/home/z/anaconda3/envs/IC2/lib/python3.6/site-packages/autogluon/core/decorator.py", line 38, in sample_config
    args_dict[k] = v.init()
  File "/home/z/anaconda3/envs/IC2/lib/python3.6/site-packages/autogluon/core/space.py", line 110, in init
    return self.sample(**config)
  File "/home/z/anaconda3/envs/IC2/lib/python3.6/site-packages/autogluon/core/decorator.py", line 185, in sample
    return self.func(*self.args, **kwargs)
  File "/home/z/anaconda3/envs/IC2/lib/python3.6/site-packages/autogluon/task/image_classification/dataset.py", line 108, in get_dataset
    if '.rec' in path:
TypeError: argument of type 'NoneType' is not iterable

[Innixma]

@zhanghang1989 Could you respond to this issue?

[zhanghang1989]

I believe this issue has been resolved.
Could you try the newest tutorial?

https://autogluon.mxnet.io/tutorials/image_classification/beginner.html

[usufu]

the error still there with the latest tutorial.

[Innixma]

@zhreshold Could you take a look at this?

[zhreshold]

@UNIcodeX The problem is that you used the wrong argument name train_path instead of the correct path arg in task.Dataset.

I found it's actually due to the wrong documentation. To avoid similar problems in the future, I filed a PR to fix the docstrings: https://github.com/awslabs/autogluon/pull/436

[UNIcodeX]

@zhreshold LMAO! Alright thanks!