experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[tf.data] Minor fix to remove unnecessary difference between the implementations of the batch and padded batch reducers.

Browse Source 
PiperOrigin-RevId: 211706766
This commit is contained in:
Jiri Simsa 2018-09-05 15:07:07 -07:00 committed by TensorFlower Gardener
parent 7288b3da07
commit 7e2577b098
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
tensorflow/contrib/data/python/ops/batching.py
View File

@ -272,9 +272,9 @@ def _padded_batch_dense_window(dataset, padded_shape, padding_value=None):
padding_value = 0 padding_value = 0
def batch_init_fn(_): def batch_init_fn(_):
return array_ops.fill( batch_shape = array_ops.concat(
array_ops.concat([np.array([0], dtype=np.int32), padded_shape], 0), [np.array([0], dtype=np.int32), padded_shape], 0)
constant_op.constant(padding_value, dtype=dataset.output_types)) return gen_array_ops.empty(batch_shape, dtype=dataset.output_types)
def batch_reduce_fn(state, value): def batch_reduce_fn(state, value):
return array_ops.concat([state, [value]], 0) return array_ops.concat([state, [value]], 0)