CODEWITHSUNDEEP
pythonpython-3.xtensorflow
Yuval Atzmon
Yuval Atzmon

Reputation: 5945

How to apply dynamic shape with tensorflow `scatter_nd`

In tensorflow, how can I apply dynamic shape to scatter_nd?

When I use an input tensor with a dynamic shape, I get the following error:

ValueError: Cannot convert a partially known TensorShape to a Tensor: (20, ?)

Here is the function I use. It works when tensor has a static shape. But with a dynamic shape (e.g. (?, 7)) it fails. 

def tf_zero_pad_columns(tensor, columns_list, num_output_columns):
    assert(tensor.shape.as_list()[1] == len(columns_list))
    assert(num_output_columns >= len(columns_list))

    tensor = tf.transpose(tensor)
    columns = tf.constant(np.array([columns_list]).T.astype('int32'))
    shape=tf.TensorShape((num_output_columns, tensor.get_shape()[1]))
    scattered = tf.scatter_nd(columns, tensor, shape=shape)
    return tf.transpose(scattered)

I also tried replacing tensor.get_shape()[1] by -1 but this produces a different error during training:

InvalidArgumentError: Dimension -1 must be >= 0 [[Node: lambda_40/ScatterNd ....

EDIT:

Example input with a dynamic shape (this reproduces the error):

tensor = tf.placeholder(tf.float32, shape=(None, 7))
tf_zero_pad_columns(tensor, [11,12,13,4,5,6,7], 20)

Example input with a static shape:

import numpy as np
tensor_np = np.tile(range(7), (4, 1)) + np.array(range(4))[:, None]
tensor = tf.constant(tensor_np)


tf_zero_pad_columns(tensor, [11,12,13,4,5,6,7], 20)

Output is:

array([[0, 0, 0, 0, 3, 4, 5, 6, 0, 0, 0, 0, 1, 2, 0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 4, 5, 6, 7, 0, 0, 0, 1, 2, 3, 0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 5, 6, 7, 8, 0, 0, 0, 2, 3, 4, 0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 6, 7, 8, 9, 0, 0, 0, 3, 4, 5, 0, 0, 0, 0, 0, 0]])

Upvotes: 1

Views: 491

Answers (1)

gorjan
gorjan

Reputation: 5555

This works for me:

def tf_zero_pad_columns(tensor, columns_list, num_output_columns):
    assert(tensor.shape.as_list()[1] == len(columns_list))
    assert(num_output_columns >= len(columns_list))

    tensor = tf.transpose(tensor)
    columns = tf.constant(np.array([columns_list]).T.astype('int32'))
    tensor_shape = tf.shape(tensor)[1]
    scattered = tf.scatter_nd(columns, tensor, shape=(num_output_columns, tensor_shape))
    return tf.transpose(scattered)

Upvotes: 1

Related Questions