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

Change references to TensorArray.pack, TensorArray.unpack to TensorArray.stack

Browse Source 
and TensorArray.unstack since pack and unpack are getting deprecated. Also, I switched a few references to tf.pack/tf.unpack to tf.stack and tf.unstack.
Change: 141619037
This commit is contained in:
A. Unique TensorFlower 2016-12-09 17:23:37 -08:00 committed by TensorFlower Gardener
parent 6da1cfa157
commit 35a58c8141
5 changed files with 26 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
tensorflow/python/kernel_tests/rnn_test.py
View File

@ -781,7 +781,7 @@ class RawRNNTest(tf.test.TestCase):
dtype=tf.float32)
sequence_length = tf.placeholder(shape=(batch_size,), dtype=tf.int32)
inputs_ta = tf.TensorArray(dtype=tf.float32, size=tf.shape(inputs)[0])
inputs_ta = inputs_ta.unpack(inputs)
inputs_ta = inputs_ta.unstack(inputs)
cell = tf.contrib.rnn.LSTMCell(num_units, state_is_tuple=True)
@ -804,7 +804,7 @@ class RawRNNTest(tf.test.TestCase):
outputs_ta, final_state, _ = tf.nn.raw_rnn(
cell, loop_fn, scope=reuse_scope)
outputs = outputs_ta.pack()
outputs = outputs_ta.stack()
reuse_scope.reuse_variables()
outputs_dynamic_rnn, final_state_dynamic_rnn = tf.nn.dynamic_rnn(
@ -874,7 +874,7 @@ class RawRNNTest(tf.test.TestCase):
inputs = np.random.randn(max_time, batch_size, input_depth)
inputs_ta = tf.TensorArray(dtype=tf.float32, size=tf.shape(inputs)[0])
inputs_ta = inputs_ta.unpack(inputs)
inputs_ta = inputs_ta.unstack(inputs)
cell = tf.contrib.rnn.LSTMCell(num_units, state_is_tuple=True)
@ -909,7 +909,7 @@ class RawRNNTest(tf.test.TestCase):
inputs = np.random.randn(max_time, batch_size, input_depth)
inputs_ta = tf.TensorArray(dtype=tf.float32, size=tf.shape(inputs)[0])
inputs_ta = inputs_ta.unpack(inputs)
inputs_ta = inputs_ta.unstack(inputs)
cell = tf.contrib.rnn.LSTMCell(num_units, state_is_tuple=True)
def loop_fn(time_, cell_output, cell_state, loop_state):
@ -935,7 +935,7 @@ class RawRNNTest(tf.test.TestCase):
r = tf.nn.raw_rnn(cell, loop_fn)
loop_state = r[-1]
loop_state = loop_state.pack()
loop_state = loop_state.stack()
self.assertAllEqual([1, 2, 2 + 2, 4 + 3, 7 + 4], loop_state.eval())
def testEmitDifferentStructureThanCellOutput(self):
@ -947,7 +947,7 @@ class RawRNNTest(tf.test.TestCase):
inputs = np.random.randn(max_time, batch_size, input_depth)
inputs_ta = tf.TensorArray(dtype=tf.float32, size=tf.shape(inputs)[0])
inputs_ta = inputs_ta.unpack(inputs)
inputs_ta = inputs_ta.unstack(inputs)
cell = tf.contrib.rnn.LSTMCell(num_units, state_is_tuple=True)
def loop_fn(time_, cell_output, cell_state, _):
@ -972,7 +972,7 @@ class RawRNNTest(tf.test.TestCase):
output_ta = r[0]
self.assertEqual(2, len(output_ta))
self.assertEqual([tf.int32, tf.int64], [ta.dtype for ta in output_ta])
output = [ta.pack() for ta in output_ta]
output = [ta.stack() for ta in output_ta]
output_vals = sess.run(output)
self.assertAllEqual(
np.ones((max_time, batch_size, 2, 3), np.int32), output_vals[0])
@ -1010,7 +1010,7 @@ class RawRNNTest(tf.test.TestCase):
dtype=tf.float32)
sequence_length = tf.placeholder(shape=(batch_size,), dtype=tf.int32)
inputs_ta = tf.TensorArray(dtype=tf.float32, size=tf.shape(inputs)[0])
inputs_ta = inputs_ta.unpack(inputs)
inputs_ta = inputs_ta.unstack(inputs)
cell = tf.contrib.rnn.LSTMCell(num_units, state_is_tuple=True)
def loop_fn(time_, cell_output, cell_state, unused_loop_state):

2
tensorflow/python/ops/control_flow_ops.py
View File

@ -2977,7 +2977,7 @@ def case(pred_fn_pairs, default, exclusive=False, name="case"):
return prev_case
if exclusive:
preds_c = array_ops.pack(preds, name="preds_c")
preds_c = array_ops.stack(preds, name="preds_c")
num_true_conditions = math_ops.reduce_sum(
math_ops.cast(preds_c, dtypes.int32), name="num_true_conds")
at_most_one_true_condition = math_ops.less(

4
tensorflow/python/ops/control_flow_ops_test.py
View File

@ -192,7 +192,7 @@ class SwitchTestCase(TensorFlowTestCase):
_, outputs = tf.while_loop(Cond, Body, [initial_i, initial_outputs])
outputs = tf.reduce_sum(outputs.pack())
outputs = tf.reduce_sum(outputs.stack())
r = tf.gradients([outputs], [inputs])[0]
grad_wr_inputs = ops.convert_to_tensor(r)
o, grad = sess.run([outputs, grad_wr_inputs],
@ -218,7 +218,7 @@ class SwitchTestCase(TensorFlowTestCase):
_, outputs = tf.while_loop(Cond, Body, [initial_i, initial_outputs])
outputs = tf.reduce_sum(outputs.pack())
outputs = tf.reduce_sum(outputs.stack())
r = tf.gradients([outputs], [inputs])[0]
grad_wr_inputs = ops.convert_to_tensor(r)
o, grad = sess.run([outputs, grad_wr_inputs],

12
tensorflow/python/ops/functional_ops.py
View File

@ -106,7 +106,7 @@ def foldl(fn, elems, initializer=None, parallel_iterations=10, back_prop=True,
elems_ta = tensor_array_ops.TensorArray(dtype=elems.dtype, size=n,
dynamic_size=False,
infer_shape=True)
elems_ta = elems_ta.unpack(elems)
elems_ta = elems_ta.unstack(elems)
if initializer is None:
a = elems_ta.read(0)
@ -186,7 +186,7 @@ def foldr(fn, elems, initializer=None, parallel_iterations=10, back_prop=True,
elems_ta = tensor_array_ops.TensorArray(dtype=elems.dtype, size=n,
dynamic_size=False,
infer_shape=True)
elems_ta = elems_ta.unpack(elems)
elems_ta = elems_ta.unstack(elems)
if initializer is None:
i = n - 1
@ -354,7 +354,7 @@ def map_fn(fn, elems, dtype=None, parallel_iterations=10, back_prop=True,
for elem in elems_flat]
# Unpack elements
elems_ta = [
elem_ta.unpack(elem) for elem_ta, elem in zip(elems_ta, elems_flat)]
elem_ta.unstack(elem) for elem_ta, elem in zip(elems_ta, elems_flat)]
i = constant_op.constant(0)
@ -390,7 +390,7 @@ def map_fn(fn, elems, dtype=None, parallel_iterations=10, back_prop=True,
parallel_iterations=parallel_iterations,
back_prop=back_prop,
swap_memory=swap_memory)
results_flat = [r.pack() for r in r_a]
results_flat = [r.stack() for r in r_a]
n_static = elems_flat[0].get_shape().with_rank_at_least(1)[0]
for elem in elems_flat[1:]:
@ -536,7 +536,7 @@ def scan(fn, elems, initializer=None, parallel_iterations=10, back_prop=True,
for elem in elems_flat]
# Unpack elements
elems_ta = [
elem_ta.unpack(elem) for elem_ta, elem in zip(elems_ta, elems_flat)]
elem_ta.unstack(elem) for elem_ta, elem in zip(elems_ta, elems_flat)]
if initializer is None:
a_flat = [elem.read(0) for elem in elems_ta]
@ -586,7 +586,7 @@ def scan(fn, elems, initializer=None, parallel_iterations=10, back_prop=True,
parallel_iterations=parallel_iterations,
back_prop=back_prop, swap_memory=swap_memory)
results_flat = [r.pack() for r in r_a]
results_flat = [r.stack() for r in r_a]
n_static = elems_flat[0].get_shape().with_rank_at_least(1)[0]
for elem in elems_flat[1:]:

18
tensorflow/python/ops/rnn.py
View File

@ -191,7 +191,7 @@ def rnn(cell, inputs, initial_state=None, dtype=None,
# convert int to TensorShape if necessary
size = _state_size_with_prefix(output_size, prefix=[batch_size])
output = array_ops.zeros(
array_ops.pack(size), _infer_state_dtype(dtype, state))
array_ops.stack(size), _infer_state_dtype(dtype, state))
shape = _state_size_with_prefix(
output_size, prefix=[fixed_batch_size.value])
output.set_shape(tensor_shape.TensorShape(shape))
@ -471,7 +471,7 @@ def _reverse_seq(input_seq, lengths):
input_.set_shape(input_shape)
# Join into (time, batch_size, depth)
s_joined = array_ops.pack(sequence)
s_joined = array_ops.stack(sequence)
# TODO(schuster, ebrevdo): Remove cast when reverse_sequence takes int32
if lengths is not None:
@ -480,7 +480,7 @@ def _reverse_seq(input_seq, lengths):
# Reverse along dimension 0
s_reversed = array_ops.reverse_sequence(s_joined, lengths, 0, 1)
# Split again into list
result = array_ops.unpack(s_reversed)
result = array_ops.unstack(s_reversed)
for r, flat_result in zip(result, flat_results):
r.set_shape(input_shape)
flat_result.append(r)
@ -835,7 +835,7 @@ def dynamic_rnn(cell, inputs, sequence_length=None, initial_state=None,
def _assert_has_shape(x, shape):
x_shape = array_ops.shape(x)
packed_shape = array_ops.pack(shape)
packed_shape = array_ops.stack(shape)
return control_flow_ops.Assert(
math_ops.reduce_all(math_ops.equal(x_shape, packed_shape)),
["Expected shape for Tensor %s is " % x.name,
@ -947,7 +947,7 @@ def _dynamic_rnn_loop(cell,
def _create_zero_arrays(size):
size = _state_size_with_prefix(size, prefix=[batch_size])
return array_ops.zeros(
array_ops.pack(size), _infer_state_dtype(dtype, state))
array_ops.stack(size), _infer_state_dtype(dtype, state))
flat_zero_output = tuple(_create_zero_arrays(output)
for output in flat_output_size)
@ -974,7 +974,7 @@ def _dynamic_rnn_loop(cell,
input_ta = tuple(_create_ta("input_%d" % i, flat_input[0].dtype)
for i in range(len(flat_input)))
input_ta = tuple(ta.unpack(input_)
input_ta = tuple(ta.unstack(input_)
for ta, input_ in zip(input_ta, flat_input))
def _time_step(time, output_ta_t, state):
@ -1027,7 +1027,7 @@ def _dynamic_rnn_loop(cell,
swap_memory=swap_memory)
# Unpack final output if not using output tuples.
final_outputs = tuple(ta.pack() for ta in output_final_ta)
final_outputs = tuple(ta.stack() for ta in output_final_ta)
# Restore some shape information
for output, output_size in zip(final_outputs, flat_output_size):
@ -1092,7 +1092,7 @@ def raw_rnn(cell, loop_fn,
dtype=tf.float32)
sequence_length = tf.placeholder(shape=(batch_size,), dtype=tf.int32)
inputs_ta = tf.TensorArray(dtype=tf.float32, size=max_time)
inputs_ta = inputs_ta.unpack(inputs)
inputs_ta = inputs_ta.unstack(inputs)
cell = tf.contrib.rnn.LSTMCell(num_units)
@ -1113,7 +1113,7 @@ def raw_rnn(cell, loop_fn,
emit_output, next_loop_state)
outputs_ta, final_state, _ = raw_rnn(cell, loop_fn)
outputs = outputs_ta.pack()
outputs = outputs_ta.stack()
```
Args: