tensorflow - Can't parse serialized Example. TFRecord

Question

I extracted training and testing data from GEE to TFRecord using the below code:(after using sampleRegions)

Export.table.toCloudStorage({
collection:trainingPartition,
description:'Training_Export',
fileNamePrefix:trainFilePrefix,
bucket:outputBucket,
fileFormat:'TFRecord'});

Later, I created a dataset from the TFRecord file on colab and tried to parse the data:

size = 128
feature_columns = [
tf.io.FixedLenFeature(shape=[size, size], dtype=tf.float32) for k in featureNames
]

features_dict = dict(zip(featureNames, feature_columns))


def parse_tfrecord(example_proto):

parsed_features = tf.io.parse_single_example(example_proto, features_dict)
labels = parsed_features.pop(label)
return parsed_features, tf.cast(labels, tf.int32)

# Map the function over the dataset.
parsedDataset = trainDataset.map(parse_tfrecord, num_parallel_calls=4)

from pprint import pprint

# Print the first parsed record to check.
pprint(iter(parsedDataset).next())

Error I'm getting:

InvalidArgumentError: Key: 3_B12_min.  Can't parse serialized Example.
 [[{{node ParseSingleExample/ParseExample/ParseExampleV2}}]]

This only happens when I set the size to any value larger than 1

question from:https://stackoverflow.com/questions/66060130/cant-parse-serialized-example-tfrecord

Answer 1

Waitting for answers