| /external/tensorflow/tensorflow/python/data/util/ |
| D | sparse_test.py | 53 "classes": (ops.Tensor, sparse_tensor.SparseTensor),
57 "classes": (sparse_tensor.SparseTensor, sparse_tensor.SparseTensor),
62 "classes": (sparse_tensor.SparseTensor, ops.Tensor),
66 "classes": (((sparse_tensor.SparseTensor))),
96 "classes": sparse_tensor.SparseTensor,
106 "classes": (sparse_tensor.SparseTensor),
121 "classes": (sparse_tensor.SparseTensor, ()),
126 "classes": ((), sparse_tensor.SparseTensor),
136 "classes": (sparse_tensor.SparseTensor, (),
137 sparse_tensor.SparseTensor),
[all …]
|
| D | structure_test.py | 47 (lambda: sparse_tensor.SparseTensor(
58 "b": (sparse_tensor.SparseTensor(
60 sparse_tensor.SparseTensor(
82 (lambda: sparse_tensor.SparseTensor(
85 sparse_tensor.SparseTensor(
93 sparse_tensor.SparseTensor(
97 sparse_tensor.SparseTensor(
116 sparse_tensor.SparseTensor(
138 (lambda: sparse_tensor.SparseTensor(
143 "b": (sparse_tensor.SparseTensor(
[all …]
|
| D | sparse.py | 37 return any(c is sparse_tensor.SparseTensor for c in nest.flatten(classes))
53 tensor_shape.unknown_shape() if c is sparse_tensor.SparseTensor else shape
72 dtypes.variant if c is sparse_tensor.SparseTensor else ty
93 if c is sparse_tensor.SparseTensor else tensor
113 sparse_tensor.SparseTensor
114 if isinstance(tensor, sparse_tensor.SparseTensor) else ops.Tensor
149 if isinstance(tensor, sparse_tensor.SparseTensor) else tensor
|
| /external/tensorflow/tensorflow/core/api_def/base_api/ |
| D | api_def_DeserializeSparse.pbtxt | 6 The serialized `SparseTensor` objects. The last dimension
13 The `dtype` of the serialized `SparseTensor` objects.
16 summary: "Deserialize `SparseTensor` objects."
19 the last dimension stores serialized `SparseTensor` objects and the other N
21 `SparseTensor` objects must all match. When the final `SparseTensor` is
22 created, its rank is the rank of the incoming `SparseTensor` objects plus N;
26 The output `SparseTensor` object's shape values for the original dimensions
27 are the max across the input `SparseTensor` objects' shape values for the
30 The input `SparseTensor` objects' indices are assumed ordered in
35 original `SparseTensor` objects:
[all …]
|
| D | api_def_TakeManySparseFromTensorsMap.pbtxt | 6 1-D, The `N` serialized `SparseTensor` objects.
13 2-D. The `indices` of the minibatch `SparseTensor`.
19 1-D. The `values` of the minibatch `SparseTensor`.
25 1-D. The `shape` of the minibatch `SparseTensor`.
31 The `dtype` of the `SparseTensor` objects stored in the
54 original `SparseTensor` objects that went into the given input ops must all
55 match. When the final `SparseTensor` is created, it has rank one
56 higher than the ranks of the incoming `SparseTensor` objects
59 The output `SparseTensor` object's shape values for all dimensions but the
60 first are the max across the input `SparseTensor` objects' shape values
[all …]
|
| D | api_def_AddManySparseToTensorsMap.pbtxt | 6 2-D. The `indices` of the minibatch `SparseTensor`.
13 1-D. The `values` of the minibatch `SparseTensor`.
19 1-D. The `shape` of the minibatch `SparseTensor`.
26 1-D. The handles of the `SparseTensor` now stored in the
43 summary: "Add an `N`-minibatch `SparseTensor` to a `SparseTensorsMap`, return `N` handles."
45 A `SparseTensor` of rank `R` is represented by three tensors: `sparse_indices`,
50 An `N`-minibatch of `SparseTensor` objects is represented as a `SparseTensor`
54 The input `SparseTensor` must have rank `R` greater than 1, and the first
55 dimension is treated as the minibatch dimension. Elements of the `SparseTensor`
57 `SparseTensor` objects pointed to by each row of the output `sparse_handles`
[all …]
|
| D | api_def_SparseToSparseSetOperation.pbtxt | 6 2D `Tensor`, indices of a `SparseTensor`. Must be in row-major
13 1D `Tensor`, values of a `SparseTensor`. Must be in row-major
20 1D `Tensor`, shape of a `SparseTensor`. `set1_shape[0...n-1]` must
28 2D `Tensor`, indices of a `SparseTensor`. Must be in row-major
35 1D `Tensor`, values of a `SparseTensor`. Must be in row-major
42 1D `Tensor`, shape of a `SparseTensor`. `set2_shape[0...n-1]` must
50 2D indices of a `SparseTensor`.
56 1D values of a `SparseTensor`.
62 1D `Tensor` shape of a `SparseTensor`. `result_shape[0...n-1]` is
67 summary: "Applies set operation along last dimension of 2 `SparseTensor` inputs."
[all …]
|
| D | api_def_DeserializeManySparse.pbtxt | 6 2-D, The `N` serialized `SparseTensor` objects.
13 The `dtype` of the serialized `SparseTensor` objects.
20 `SerializeSparse`. The ranks of the original `SparseTensor` objects
21 must all match. When the final `SparseTensor` is created, it has rank one
22 higher than the ranks of the incoming `SparseTensor` objects
25 The output `SparseTensor` object's shape values for all dimensions but the
26 first are the max across the input `SparseTensor` objects' shape values
30 The input `SparseTensor` objects' indices are assumed ordered in
35 original `SparseTensor` objects:
50 then the final deserialized `SparseTensor` will be:
|
| D | api_def_SparseAdd.pbtxt | 6 2-D. The `indices` of the first `SparseTensor`, size `[nnz, ndims]` Matrix.
12 1-D. The `values` of the first `SparseTensor`, size `[nnz]` Vector.
18 1-D. The `shape` of the first `SparseTensor`, size `[ndims]` Vector.
24 2-D. The `indices` of the second `SparseTensor`, size `[nnz, ndims]` Matrix.
30 1-D. The `values` of the second `SparseTensor`, size `[nnz]` Vector.
36 1-D. The `shape` of the second `SparseTensor`, size `[ndims]` Vector.
46 summary: "Adds two `SparseTensor` objects to produce another `SparseTensor`."
48 The input `SparseTensor` objects' indices are assumed ordered in standard
52 By default, if two values sum to zero at some index, the output `SparseTensor`
|
| D | api_def_SparseCross.pbtxt | 6 2-D. Indices of each input `SparseTensor`.
12 1-D. values of each `SparseTensor`.
18 1-D. Shapes of each `SparseTensor`.
30 2-D. Indices of the concatenated `SparseTensor`.
37 `SparseTensor`.
43 1-D. Shape of the concatenated `SparseTensor`.
69 The op takes two lists, one of 2D `SparseTensor` and one of 2D `Tensor`, each
70 representing features of one feature column. It outputs a 2D `SparseTensor` with
75 inputs[0]: SparseTensor with shape = [2, 2]
80 inputs[1]: SparseTensor with shape = [2, 1]
|
| D | api_def_SerializeManySparse.pbtxt | 6 2-D. The `indices` of the minibatch `SparseTensor`.
12 1-D. The `values` of the minibatch `SparseTensor`.
18 1-D. The `shape` of the minibatch `SparseTensor`.
28 summary: "Serialize an `N`-minibatch `SparseTensor` into an `[N, 3]` `Tensor` object."
30 The `SparseTensor` must have rank `R` greater than 1, and the first dimension
31 is treated as the minibatch dimension. Elements of the `SparseTensor`
33 `SparseTensor` objects going into each row of `serialized_sparse` will have
|
| D | api_def_AddSparseToTensorsMap.pbtxt | 6 2-D. The `indices` of the `SparseTensor`.
12 1-D. The `values` of the `SparseTensor`.
18 1-D. The `shape` of the `SparseTensor`.
24 0-D. The handle of the `SparseTensor` now stored in the
41 summary: "Add a `SparseTensor` to a `SparseTensorsMap` return its handle."
43 A `SparseTensor` is represented by three tensors: `sparse_indices`,
46 This operator takes the given `SparseTensor` and adds it to a container
50 The `SparseTensor` can then be read out as part of a minibatch by passing
|
| D | api_def_SparseConcat.pbtxt | 6 2-D. Indices of each input `SparseTensor`.
12 1-D. Non-empty values of each `SparseTensor`.
18 1-D. Shapes of each `SparseTensor`.
24 2-D. Indices of the concatenated `SparseTensor`.
30 1-D. Non-empty values of the concatenated `SparseTensor`.
36 1-D. Shape of the concatenated `SparseTensor`.
43 where rank is the number of dimensions in each input `SparseTensor`.
46 summary: "Concatenates a list of `SparseTensor` along the specified dimension."
49 It is assumed that each input is a `SparseTensor` whose elements are ordered
|
| D | api_def_DenseToSparseSetOperation.pbtxt | 13 2D `Tensor`, indices of a `SparseTensor`. Must be in row-major
20 1D `Tensor`, values of a `SparseTensor`. Must be in row-major
27 1D `Tensor`, shape of a `SparseTensor`. `set2_shape[0...n-1]` must
35 2D indices of a `SparseTensor`.
41 1D values of a `SparseTensor`.
47 1D `Tensor` shape of a `SparseTensor`. `result_shape[0...n-1]` is
52 summary: "Applies set operation along last dimension of `Tensor` and `SparseTensor`."
56 Input `set2` is a `SparseTensor` represented by `set2_indices`, `set2_values`,
64 Output `result` is a `SparseTensor` represented by `result_indices`,
|
| D | api_def_SparseSliceGrad.pbtxt | 7 the non-empty values of the sliced `SparseTensor`.
13 2-D. The `indices` of the input `SparseTensor`.
25 2-D. The `indices` of the sliced `SparseTensor`.
31 1-D. The gradient with respect to the non-empty values of input `SparseTensor`.
37 the sliced `SparseTensor`, and outputs the gradients w.r.t.
38 the non-empty values of input `SparseTensor`.
|
| D | api_def_SparseSoftmax.pbtxt | 7 SparseTensor, in canonical ordering.
19 1-D. Shape of the input SparseTensor.
25 1-D. The `NNZ` values for the result `SparseTensor`.
28 summary: "Applies softmax to a batched N-D `SparseTensor`."
30 The inputs represent an N-D SparseTensor with logical shape `[..., B, C]`
43 Hence, the `SparseTensor` result has exactly the same non-zero indices and
|
| /external/tensorflow/tensorflow/core/util/sparse/ |
| D | sparse_tensor.h | 42 class SparseTensor {
48 const VarDimArray order, SparseTensor* result) { in Create()
77 *result = SparseTensor(ix, vals, shape, order); in Create()
82 SparseTensor* result) { in Create()
88 SparseTensor* result) { in Create()
93 const VarDimArray order, SparseTensor* result) { in Create()
97 SparseTensor() : dims_(0) {} in SparseTensor() function
100 SparseTensor(Tensor ix, Tensor vals, const TensorShape& shape) in SparseTensor() function
101 : SparseTensor(ix, vals, TensorShapeToVector(shape), in SparseTensor()
105 SparseTensor(Tensor ix, Tensor vals, const VarDimArray shape) in SparseTensor() function
[all …]
|
| D | sparse_tensor_test.cc | 97 SparseTensor result; in TEST()
99 EXPECT_EQ(SparseTensor::Create(ix, vals, TensorShape({10, 10, 10}), {0, 1, 2}, in TEST()
110 SparseTensor result; in TEST()
112 EXPECT_EQ(SparseTensor::Create(ix, vals, TensorShape({10, 10, 10}), {0, 1, 2}, in TEST()
123 SparseTensor result; in TEST()
125 EXPECT_EQ(SparseTensor::Create(ix, vals, TensorShape({10, 10, 10}), {0, 1, 2}, in TEST()
136 SparseTensor result; in TEST()
138 EXPECT_EQ(SparseTensor::Create(ix, vals, TensorShape({10, 10, 10}), {0, 1, 2}, in TEST()
149 SparseTensor result; in TEST()
152 SparseTensor::Create(ix, vals, TensorShape({10, 10, 10}), {0, 1}, &result) in TEST()
[all …]
|
| D | README.md | 1 SparseTensor chapter
47 If the SparseTensor is constructed without a provided order, then a
51 Resorting the SparseTensor in-place (which resorts the underlying index and
87 order of the SparseTensor does not match the dimensions you wish to group by.
89 SparseTensor with
97 to sort the SparseTensor before grouping.
104 SparseTensor sp(indices, vals, shape);
131 SparseTensor sp(indices, vals, shape);
142 Concatenates multiple SparseTensors and returns a new SparseTensor.
204 SparseTensor st1(ix1, vals1, TensorShape({10, 20, 5}), {1, 0, 2});
[all …]
|
| /external/tensorflow/tensorflow/python/ops/ragged/ |
| D | ragged_from_sparse_op_test.py | 36 st = sparse_tensor.SparseTensor(
45 st = sparse_tensor.SparseTensor(
54 st1 = sparse_tensor.SparseTensor(indices=[[0]], values=[0], dense_shape=[3])
58 st2 = sparse_tensor.SparseTensor(
64 st3 = sparse_tensor.SparseTensor(
73 st1 = sparse_tensor.SparseTensor(
77 st2 = sparse_tensor.SparseTensor(
93 st1 = sparse_tensor.SparseTensor(
99 st2 = sparse_tensor.SparseTensor(
105 st3 = sparse_tensor.SparseTensor(
|
| /external/tensorflow/tensorflow/contrib/layers/python/layers/ |
| D | feature_column_ops_test.py | 58 rating_tensor = sparse_tensor.SparseTensor(
76 rating_tensor = sparse_tensor.SparseTensor(
131 wire_tensor = sparse_tensor.SparseTensor(
154 wire_tensor = sparse_tensor.SparseTensor(
182 self.assertIsInstance(output, sparse_tensor.SparseTensor)
190 wire_tensor = sparse_tensor.SparseTensor(
216 wire_tensor = sparse_tensor.SparseTensor(
247 self.assertIsInstance(output, sparse_tensor.SparseTensor)
257 wire_tensor = sparse_tensor.SparseTensor(
287 self.assertIsInstance(output, sparse_tensor.SparseTensor)
[all …]
|
| /external/tensorflow/tensorflow/python/ops/ |
| D | sparse_ops.py | 66 return sparse_tensor.SparseTensor.from_value(sp_input)
67 if not isinstance(sp_input, sparse_tensor.SparseTensor):
153 return sparse_tensor.SparseTensor(
185 return sparse_tensor.SparseTensor(
334 return sparse_tensor.SparseTensor(output_ind, output_val, output_shape)
458 sparse_classes = (sparse_tensor.SparseTensor, sparse_tensor.SparseTensorValue)
480 return sparse_tensor.SparseTensor(output_ind, output_val, output_shape)
581 isinstance(i, sparse_tensor.SparseTensor) or isinstance(i, ops.Tensor)
586 i for i in inputs if isinstance(i, sparse_tensor.SparseTensor)
589 i for i in inputs if not isinstance(i, sparse_tensor.SparseTensor)
[all …]
|
| D | sets_impl.py | 52 if not isinstance(a, sparse_tensor.SparseTensor):
87 if (isinstance(a, sparse_tensor.SparseTensor) and
88 not isinstance(b, sparse_tensor.SparseTensor)):
118 if isinstance(a, sparse_tensor.SparseTensor):
119 if isinstance(b, sparse_tensor.SparseTensor):
127 elif isinstance(b, sparse_tensor.SparseTensor):
133 return sparse_tensor.SparseTensor(indices, values, shape)
|
| /external/tensorflow/tensorflow/core/kernels/ |
| D | sparse_tensors_map_ops.cc | 40 using sparse::SparseTensor;
54 Status AddSparseTensor(OpKernelContext* ctx, const SparseTensor& sp, in AddSparseTensor()
81 std::vector<SparseTensor>* sparse_tensors) { in RetrieveAndClearSparseTensors()
96 SparseTensor tensor; in RetrieveAndClearSparseTensors()
97 TF_RETURN_IF_ERROR(SparseTensor::Create(*ix, *values, shape, &tensor)); in RetrieveAndClearSparseTensors()
199 SparseTensor st; in Compute()
200 OP_REQUIRES_OK(context, SparseTensor::Create(*input_indices, *input_values, in Compute()
259 SparseTensor input_st; in Compute()
260 OP_REQUIRES_OK(context, SparseTensor::Create(*input_indices, *input_values, in Compute()
308 SparseTensor st_i; in Compute()
[all …]
|
| /external/tensorflow/tensorflow/python/framework/ |
| D | sparse_tensor_test.py | 40 sparse_tensor.SparseTensor(indices, values, shape),
41 sparse_tensor.SparseTensor.from_value(sp_value),
42 sparse_tensor.SparseTensor.from_value(
43 sparse_tensor.SparseTensor(indices, values, shape))]:
64 sparse_tensor.is_sparse(sparse_tensor.SparseTensor([[0]], [0], [1])))
71 sp = sparse_tensor.SparseTensor([[0, 0], [1, 2]], [1.0, 3.0], [3, 4])
100 st = sparse_tensor.SparseTensor.from_value(sparse_tensor_value)
|