def reduce_sum(input_tensor, reduction_indices=None, keep_dims=False, name=None):
"""Computes the sum of elements across dimensions of a tensor.
Reduces `input_tensor` along the dimensions given in `reduction_indices`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `reduction_indices`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.
If `reduction_indices` has no entries, all dimensions are reduced, and a
tensor with a single element is returned.
For example:
```python
# 'x' is [[1, 1, 1]
# [1, 1, 1]]
tf.reduce_sum(x) ==> 6
tf.reduce_sum(x, 0) ==> [2, 2, 2]
tf.reduce_sum(x, 1) ==> [3, 3]
tf.reduce_sum(x, 1, keep_dims=True) ==> [[3], [3]]
tf.reduce_sum(x, [0, 1]) ==> 6
```
Args:
input_tensor: The tensor to reduce. Should have numeric type.
reduction_indices: The dimensions to reduce. If `None` (the default),
reduces all dimensions.
keep_dims: If true, retains reduced dimensions with length 1.
name: A name for the operation (optional).
Returns:
The reduced tensor.
"""
return gen_math_ops._sum(input_tensor, _ReductionDims(input_tensor, reduction_indices), keep_dims, name=name)
def reduce_sum(input_tensor,
reduction_indices=None,
keep_dims=False,
name=None):
"""Computes the sum of elements across dimensions of a tensor.
Reduces `input_tensor` along the dimensions given in `reduction_indices`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `reduction_indices`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.
If `reduction_indices` has no entries, all dimensions are reduced, and a
tensor with a single element is returned.
For example:
```python
# 'x' is [[1, 1, 1]]
# [1, 1, 1]]
tf.reduce_sum(x) ==> 6
tf.reduce_sum(x, 0) ==> [2, 2, 2]
tf.reduce_sum(x, 1) ==> [3, 3]
tf.reduce_sum(x, 1, keep_dims=True) ==> [[3], [3]]
tf.reduce_sum(x, [0, 1]) ==> 6
```
Args:
input_tensor: The tensor to reduce. Should have numeric type.
reduction_indices: The dimensions to reduce. If `None` (the defaut),
reduces all dimensions.
keep_dims: If true, retains reduced dimensions with length 1.
name: A name for the operation (optional).
Returns:
The reduced tensor.
"""
return gen_math_ops._sum(input_tensor,
_ReductionDims(input_tensor, reduction_indices),
keep_dims,
name=name)
def sum1(self, input_itensor):
"""Create a specialized op that sums over 1 dimensional vector.
This avoids having to create Rank/Range ops that initialize indices
in the default tf.reduce_sum."""
op_type_name = "sum1"
tf_dtype = input_itensor.dtype
current_device = get_current_device_string(self.g)
current_device_sanitized = current_device.replace(":", "")
key = (op_type_name, tf_dtype.name, current_device_sanitized)
if key in self.op_cache:
if self.PRINT_CACHE_HITS:
print("Imperative cache hit for %s"%(str(key)))
op = self.op_cache[key]
else:
if self.PRINT_CACHE_MISSES:
print("Imperative cache miss for %s"%(str(key)))
with self.g.as_default():
op_prefix = op_type_name + "." + tf_dtype.name
holder, tensor = session_ops.get_session_tensor(
input_itensor.tf_handle, input_itensor.dtype, name=op_prefix+".0")
input_holders = {"input": holder}
reduction_indices = constant_op.constant([0], dtype=dtypes.int32,
name=op_prefix+".1")
output = gen_math_ops._sum(input=tensor,
reduction_indices=reduction_indices,
keep_dims=False, name=op_prefix+".op")
op_prefix = op_prefix+".out"
output_handle = session_ops.get_session_handle(output,
op_prefix+".handle")
op = Op(self, input_holders, output_handle)
self.cache_add(key, op)
return op(input=input_itensor)
def fast_sum(tensor, name=None):
return gen_math_ops._sum(input=tensor,
reduction_indices=[],
keep_dims=False,
name=name)
def fast_sum(tensor, name=None):
return gen_math_ops._sum(input=tensor,
reduction_indices=[],
keep_dims=False,
name=name)
