Failed to find data adapter that can handle input

[philzetter]

Hi @pythonlessons, I'm trying to use the image to word Tutorial.

I changed the train.py a bit in order to read my images and labels better. The only change was in def read_annotation_file:

Old Code:

def read_annotation_file(annotation_path):
   dataset, vocab, max_len = [], set(), 0
    with open(annotation_path, "r") as f:
        for line in tqdm(f.readlines()):
            line = line.split()
            image_path = data_path + line[0][1:]
            label = line[0].split("_")[1]
            dataset.append([image_path, label])
            vocab.update(list(label))
            max_len = max(max_len, len(label))
    return dataset, sorted(vocab), max_len

New Code:

def read_annotation_file(annotation_path):
    dataset, vocab, max_len = [], set(), 0
    with open(annotation_path, "r") as f:
        for line in tqdm(f.readlines()):
            line = line.split(' ')
            image_path = data_path + line[0]
            label = line[1]
            dataset.append([image_path, label])
            vocab.update(list(label))
            max_len = max(max_len, len(label))
    return dataset, sorted(vocab), max_len

I also changed something in line 91 due to an error:

Old Code: metrics=[CWERMetric()],

New Code: metrics=[CWERMetric('accuracy')],

This is my error:

Traceback (most recent call last): File "/path/to/mltu/Tutorials/01_image_to_word/train.py", line 111, in model.fit( File "/path/to/.local/lib/python3.10/site-packages/keras/utils/traceback_utils.py", line 70, in error_handler raise e.with_traceback(filtered_tb) from None File "/path/to/.local/lib/python3.10/site-packages/keras/engine/data_adapter.py", line 1083, in select_data_adapter raise ValueError( ValueError: Failed to find data adapter that can handle input: <class 'mltu.dataProvider.DataProvider'>, <class 'NoneType'>

How do I need to change train_data_provider or train_dataset or is it a version problem ?

[pythonlessons]

Hi, if you followed my tutorial from pylessons.com then you may face a problem with code.

Use newest mltu library version. And use the right ImageReader, example code:

# Create training data provider
train_data_provider = DataProvider(
    dataset=train_dataset,
    skip_validation=True,
    batch_size=configs.batch_size,
    data_preprocessors=[ImageReader(CVImage)],
    transformers=[
        ImageResizer(configs.width, configs.height),
        LabelIndexer(configs.vocab),
        LabelPadding(max_word_length=configs.max_text_length, padding_value=len(configs.vocab))
        ],
)

[philzetter]

Thank you for the fast response ! I'm using exactly the same code. My train_dataset looks like this:

[['CINCMC/images/one.jpg', 'one']]

[pythonlessons]

try to iterate train_data_provider in for loop and check if you receive correct data there

[philzetter]

Could you sent me an example of your annotation files. I just need a few lines of the txt files. I think that's the problem.

[philzetter]

I still get this error:

Traceback (most recent call last):
  File "/path/to/mltu/Tutorials/01_image_to_word/train.py", line 102, in <module>
    model.fit(
  File "/path/to/python3.10/site-packages/keras/utils/traceback_utils.py", line 70, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "/path/to/python3.10/site-packages/keras/engine/data_adapter.py", line 1083, in select_data_adapter
    raise ValueError(
ValueError: Failed to find data adapter that can handle input: <class 'mltu.dataProvider.DataProvider'>, <class 'NoneType'>

My dataset in read_annotation_file looks like this:

[['CINCMC/train_images/word_1001.png', 'KSZU0005183'], ['CINCMC/train_images/word_1002.png', '25G2'], ['CINCMC/train_images/word_1003.png', 'SLRU0001112'], ['CINCMC/train_images/word_1004.png', '25U2'], ['CINCMC/train_images/word_1005.png', 'SLRU0001540'], ['CINCMC/train_images/word_1006.png', 'SLRU0001072'], ['CINCMC/train_images/word_1007.png', '25U2'], ['CINCMC/train_images/word_1008.png', '25U2'], ['CINCMC/train_images/word_1009.png', 'SLRU0000939'], ['CINCMC/train_images/word_1010.png', '25U2'], ['CINCMC/train_images/word_1011.png', '25U2'], ['CINCMC/train_images/word_1012.png', 'SLRU0001046'], ['CINCMC/train_images/word_1013.png', '25U2'], ['CINCMC/train_images/word_1014.png', 'SLRU0000923'], ['CINCMC/train_images/word_1015.png', '25U2'], ['CINCMC/train_images/word_1016.png', 'SLRU0001133'], ['CINCMC/train_images/word_1017.png', '25U2'], ['CINCMC/train_images/word_1018.png', 'SLRU0001386'], ['CINCMC/train_images/word_1019.png', 'SLRU0001473'], ['CINCMC/train_images/word_1020.png', '25U2'], ['CINCMC/train_images/word_1021.png', '25U2'], ['CINCMC/train_images/word_1022.png', 'SLRU0001046'], ['CINCMC/train_images/word_1023.png', 'SLRU0001410'], ['CINCMC/train_images/word_1024.png', '25U2'], ['CINCMC/train_images/word_1025.png', '25U2'], ['CINCMC/train_images/word_1026.png', '25U2'], ['CINCMC/train_images/word_1027.png', 'SLRU0001616'], ['CINCMC/train_images/word_1028.png', 'SLRU0001128'], ['CINCMC/train_images/word_1029.png', 'SLRU0001149'], ['CINCMC/train_images/word_1030.png', '25U2'], ['CINCMC/train_images/word_1031.png', '25U2'], ['CINCMC/train_images/word_1032.png', 'SLRU0001365'], ['CINCMC/train_images/word_1033.png', '25U2'], ['CINCMC/train_images/word_1034.png', 'SLRU0001318'], ['CINCMC/train_images/word_1035.png', '25U2'], ['CINCMC/train_images/word_1036.png', '25U2'], ['CINCMC/train_images/word_1037.png', 'SLRU0001262'], ['CINCMC/train_images/word_1038.png', '25U2']]

Versions: keras 2.13.1 tensorflow 2.13.0

[pythonlessons]

you tried to run this:

for data in train_data_provider:
    print(data[0], data[1])

Check if it doesn't return None, because it seems like that

[philzetter]

The result of:

for data in train_data_provider:
    print(data[0], data[1])

looks like this:

[[[[102 102 102]
   [103 103 103]
   [104 104 104]
   ...
   [134 134 134]
   [138 138 138]
   [140 140 140]]

  [[102 102 102]
   [104 104 104]
   [105 105 105]
   ...
   [134 134 134]
   [138 138 138]
   [140 140 140]]

  [[100 100 100]
   [103 103 103]
   [104 104 104]
   ...
   [135 135 135]
   [138 138 138]
   [140 140 140]]

  ...

  [[ 96  96  96]
   [ 99  99  99]
   [100 100 100]
   ...
   [135 135 135]
   [139 139 139]
   [139 139 139]]

  [[ 98  98  98]
   [ 99  99  99]
   [ 99  99  99]
   ...
   [135 135 135]
   [139 139 139]
   [139 139 139]]

  [[101 101 101]
   [ 99  99  99]
   [100 100 100]
   ...
   [136 136 136]
   [139 139 139]
   [139 139 139]]]

 [[[ 98  98  98]
   [100 100 100]
   [ 99  99  99]
   ...
   [175 175 175]
   [175 175 175]
   [174 174 174]]

  [[ 98  98  98]
   [ 99  99  99]
   [ 99  99  99]
   ...
   [197 197 197]
   [195 195 195]
   [189 189 189]]

  [[ 99  99  99]
   [100 100 100]
   [100 100 100]
   ...
   [197 197 197]
   [197 197 197]
   [203 203 203]]

  ...

  [[ 99  99  99]
   [100 100 100]
   [103 103 103]
   ...
   [202 202 202]
   [206 206 206]
   [203 203 203]]

  [[ 99  99  99]
   [102 102 102]
   [102 102 102]
   ...
   [200 200 200]
   [200 200 200]
   [197 197 197]]

  [[100 100 100]
   [100 100 100]
   [101 101 101]
   ...
   [182 182 182]
   [182 182 182]
   [181 181 181]]]

 [[[127 127 127]
   [127 127 127]
   [127 127 127]
   ...
   [137 137 137]
   [137 137 137]
   [134 134 134]]

  [[127 127 127]
   [126 126 126]
   [126 126 126]
   ...
   [136 136 136]
   [136 136 136]
   [134 134 134]]

  [[127 127 127]
   [126 126 126]
   [125 125 125]
   ...
   [136 136 136]
   [136 136 136]
   [136 136 136]]

  ...

  [[126 126 126]
   [125 125 125]
   [125 125 125]
   ...
   [134 134 134]
   [136 136 136]
   [139 139 139]]

  [[126 126 126]
   [126 126 126]
   [126 126 126]
   ...
   [136 136 136]
   [138 138 138]
   [138 138 138]]

  [[127 127 127]
   [127 127 127]
   [127 127 127]
   ...
   [138 138 138]
   [140 140 140]
   [138 138 138]]]

 ...

 [[[113 113 113]
   [112 112 112]
   [114 114 114]
   ...
   [128 128 128]
   [124 124 124]
   [128 128 128]]

  [[115 115 115]
   [115 115 115]
   [116 116 116]
   ...
   [146 146 146]
   [135 135 135]
   [129 129 129]]

  [[118 118 118]
   [117 117 117]
   [118 118 118]
   ...
   [173 173 173]
   [148 148 148]
   [129 129 129]]

  ...

  [[115 115 115]
   [116 116 116]
   [121 121 121]
   ...
   [127 127 127]
   [128 128 128]
   [127 127 127]]

  [[113 113 113]
   [116 116 116]
   [121 121 121]
   ...
   [127 127 127]
   [127 127 127]
   [127 127 127]]

  [[117 117 117]
   [117 117 117]
   [119 119 119]
   ...
   [127 127 127]
   [127 127 127]
   [127 127 127]]]

 [[[119 119 119]
   [118 118 118]
   [117 117 117]
   ...
   [177 177 177]
   [182 182 182]
   [177 177 177]]

  [[119 119 119]
   [120 120 120]
   [121 121 121]
   ...
   [173 173 173]
   [155 155 155]
   [137 137 137]]

  ...

  [[118 118 118]
   [116 116 116]
   [118 118 118]
   ...
   [125 125 125]
   [127 127 127]
   [127 127 127]]

  [[120 120 120]
   [117 117 117]
   [116 116 116]
   ...
   [125 125 125]
   [126 126 126]
   [126 126 126]]

  [[121 121 121]
   [118 118 118]
   [116 116 116]
   ...
   [124 124 124]
   [125 125 125]
   [125 125 125]]]] [[ 4  5 16 ... 39 39 39]
 [28 12 23 ...  7  3 39]
 [ 4  5 16 ... 39 39 39]
 ...
 [ 2  5 16 ... 39 39 39]
 [20 27 33 ...  2  8 39]
 [ 2  5 16 ... 39 39 39]]

[pythonlessons]

Looks ok, try to print following:

for data in train_data_provider:
    if data[0]==None or data[1]==None:
        print("sometwing wrong in data")
    else:
        print(data[0].shape, data[1].shape)

[philzetter]

Tried the following code:

for data in train_data_provider:
    if data[0] is None or data[1] is None:
        print("something wrong in data")
    else:
        print(data[0].shape, data[1].shape)

Result is:

(1000, 32, 128, 3) (1000, 12)

But this was the only line that was printed out

[pythonlessons]

your batch_size is 1000?

[philzetter]

1000 is the size of my dataset and I'm loading it as described on to the dataset. I already sent the dataset above. Could you maybe show me a few lines of your example annotation file maybe I'm loading the wrong way ....