View source

金铨达配资Compatibility utility required to allow for both V1 and V2 behavior in TF.


  • tf.compat.v1.compat.dimension_at_index
  • tf.compat.v1.dimension_at_index
  • tf.compat.v2.compat.dimension_at_index

Until the release of TF 2.0, we need the legacy behavior of TensorShape to coexist with the new behavior. This utility is a bridge between the two.

If you want to retrieve the Dimension instance corresponding to a certain index in a TensorShape instance, use this utility, like this:

# If you had this in your V1 code:
dim = tensor_shape[i]

# Use `dimension_at_index` as direct replacement compatible with both V1 & V2:
dim = dimension_at_index(tensor_shape, i)

# Another possibility would be this, but WARNING: it only works if the
# tensor_shape instance has a defined rank.
dim = tensor_shape.dims[i]  # `dims` may be None if the rank is undefined!

# In native V2 code, we recommend instead being more explicit:
if tensor_shape.rank is None:
  dim = Dimension(None)
  dim = tensor_shape.dims[i]

# Being more explicit will save you from the following trap (present in V1):
# you might do in-place modifications to `dim` and expect them to be reflected
# in `tensor_shape[i]`, but they would not be (as the Dimension object was
# instantiated on the fly.


  • shape: A TensorShape instance.
  • index: An integer index.


A dimension object.

results matching ""

    No results matching ""