how to get string value out of tf.tensor which dtype is string

Question

I want to use tf.data.Dataset.list_files function to feed my datasets.
But because the file is not image, I need to load it manually.
The problem is tf.data.Dataset.list_files pass variable as tf.tensor and my python code can not handle tensor.

How can I get string value from tf.tensor. The dtype is string.

train_dataset = tf.data.Dataset.list_files(PATH+'clean_4s_val/*.wav')
train_dataset = train_dataset.map(lambda x: load_audio_file(x))

def load_audio_file(file_path):
  print("file_path: ", file_path)
  # i want do something like string_path = convert_tensor_to_string(file_path)

file_path is Tensor("arg0:0", shape=(), dtype=string)

I use tensorflow 1.13.1 and eager mode.

thanks in advance

Answer 1

You can use tf.py_func to wrap load_audio_file().

import tensorflow as tf

tf.enable_eager_execution()

def load_audio_file(file_path):
    # you should decode bytes type to string type
    print("file_path: ",bytes.decode(file_path),type(bytes.decode(file_path)))
    return file_path

train_dataset = tf.data.Dataset.list_files('clean_4s_val/*.wav')
train_dataset = train_dataset.map(lambda x: tf.py_func(load_audio_file, [x], [tf.string]))

for one_element in train_dataset:
    print(one_element)

file_path:  clean_4s_val/1.wav <class 'str'>
(<tf.Tensor: id=32, shape=(), dtype=string, numpy=b'clean_4s_val/1.wav'>,)
file_path:  clean_4s_val/3.wav <class 'str'>
(<tf.Tensor: id=34, shape=(), dtype=string, numpy=b'clean_4s_val/3.wav'>,)
file_path:  clean_4s_val/2.wav <class 'str'>
(<tf.Tensor: id=36, shape=(), dtype=string, numpy=b'clean_4s_val/2.wav'>,)

UPDATE for TF 2

The above solution will not work with TF 2 (tested with 2.2.0), even when replacing tf.py_func with tf.py_function, giving

InvalidArgumentError: TypeError: descriptor 'decode' requires a 'bytes' object but received a 'tensorflow.python.framework.ops.EagerTensor'

To make it work in TF 2, make the following changes:

• Remove tf.enable_eager_execution() (eager is enabled by default in TF 2, which you can verify with tf.executing_eagerly() returning True)
• Replace tf.py_func with tf.py_function
• Replace all in-function references of file_path with file_path.numpy()

Answer 2

If you want to do something completely custom, then wrapping your code in tf.py_function is what you should do. Keep in mind that this will result in poor performance. See documentation and examples here:

https://www.tensorflow.org/api_docs/python/tf/data/Dataset#map

On the other hand if you are doing something generic, then you don't need to wrap your code in py_function instead use any of the methods provided in tf.strings module. These methods are made to work on string tensors and provide many common methods like split, join, len etc. These will not negatively effect performance, they will work on the tensor directly and return a modified tensor.

See documentation of tf.strings here: https://www.tensorflow.org/api_docs/python/tf/strings

For example lets say you wanted to extract the name of the label from the file name you could then write code like this:

ds.map(lambda x: tf.strings.split(x, sep='$')[1])

The above assumes that the label is separated by a $.

Answer 3

if you really want to unwrap Tensor to its string content only - you need to serialize TFRecord in order to use tf_example.SerializeToString() - to get (printable) string value - see here