Discover all of the brand-new features and improvements to existing capabilities in the Dataiku 11.3 updateLET'S GO

Deep Learning Training Error

Forermx
Level 1
Deep Learning Training Error

 I'm getting the following error when I try to begin training the deep learning model

Traceback (most recent call last):
  File "/app/dataiku_design/dataiku-dss-9.0.3/python/dataiku/doctor/server.py", line 46, in serve
    ret = api_command(arg)
  File "/app/dataiku_design/dataiku-dss-9.0.3/python/dataiku/doctor/dkuapi.py", line 45, in aux
    return api(**kwargs)
  File "/app/dataiku_design/dataiku-dss-9.0.3/python/dataiku/doctor/commands.py", line 424, in train_prediction_keras
    pipeline.generated_features_mapping)
  File "/app/dataiku_design/dataiku-dss-9.0.3/python/dataiku/doctor/prediction_entrypoints.py", line 462, in prediction_train_model_keras
    generated_features_mapping)
  File "/app/dataiku_design/dataiku-dss-9.0.3/python/dataiku/doctor/deep_learning/keras_support.py", line 117, in get_keras_model
    base_callbacks)
  File "<string>", line 48, in fit_model
  File "/app/dataiku_design/dssdata-9.0.3/code-envs/python/tissue_segmentation/lib/python3.6/site-packages/keras/legacy/interfaces.py", line 91, in wrapper
    return func(*args, **kwargs)
  File "/app/dataiku_design/dssdata-9.0.3/code-envs/python/tissue_segmentation/lib/python3.6/site-packages/keras/engine/training.py", line 1732, in fit_generator
    initial_epoch=initial_epoch)
  File "/app/dataiku_design/dssdata-9.0.3/code-envs/python/tissue_segmentation/lib/python3.6/site-packages/keras/engine/training_generator.py", line 185, in fit_generator
    generator_output = next(output_generator)
  File "/app/dataiku_design/dssdata-9.0.3/code-envs/python/tissue_segmentation/lib/python3.6/site-packages/keras/utils/data_utils.py", line 625, in get
    six.reraise(*sys.exc_info())
  File "/app/dataiku_design/dssdata-9.0.3/code-envs/python/tissue_segmentation/lib/python3.6/site-packages/six.py", line 719, in reraise
    raise value
  File "/app/dataiku_design/dssdata-9.0.3/code-envs/python/tissue_segmentation/lib/python3.6/site-packages/keras/utils/data_utils.py", line 610, in get
    inputs = future.get(timeout=30)
  File "/utilities/python/3.6.13/lib/python3.6/multiprocessing/pool.py", line 644, in get
    raise self._value
  File "/utilities/python/3.6.13/lib/python3.6/multiprocessing/pool.py", line 119, in worker
    result = (True, func(*args, **kwds))
  File "/app/dataiku_design/dssdata-9.0.3/code-envs/python/tissue_segmentation/lib/python3.6/site-packages/keras/utils/data_utils.py", line 406, in get_index
    return _SHARED_SEQUENCES[uid][i]
  File "/app/dataiku_design/dataiku-dss-9.0.3/python/dataiku/doctor/deep_learning/sequences.py", line 186, in __getitem__
    return self.process_batch_func(self.original_sequence[index])
  File "/app/dataiku_design/dataiku-dss-9.0.3/python/dataiku/doctor/deep_learning/sequences.py", line 219, in _process_batch_func
    new_X_batch[inp] = DataAugmentationSequence.duplicate_rows(X_batch, n_augmentation)
  File "/app/dataiku_design/dataiku-dss-9.0.3/python/dataiku/doctor/deep_learning/sequences.py", line 192, in duplicate_rows
    return np.tile(array, tuple([n_augmentation] + [1] * (len(array.shape) - 1)))
AttributeError: 'dict' object has no attribute 'shape'

 

 I can't find an instance where I may have set up something wrong so I'm not sure what is wrong


Operating system used: Windows 10 Pro

0 Kudos
1 Reply
SarinaS
Dataiker

Hi @Forermx,

Thank you for reporting this! This is unfortunately a known issue when using a DataAugmentationSequence for deep learning when you also have a non-augmented feature in your training. The main option for now would be to remove the DataAugmentationSequence or the non-augmented feature from the training. This has been reported to our engineering team. Please let me know if you have any questions in the meantime. 

Thank you,
Sarina 

0 Kudos