ScatterElements - 11 vs 16¶
Next section compares an older to a newer version of the same operator after both definition are converted into markdown text. Green means an addition to the newer version, red means a deletion. Anything else is unchanged.
ScatterElements11 → ScatterElements16
RENAMED
@@ -1 +1 @@
|
|
1
1
|
ScatterElements takes three inputs data, updates, and indices of the same
|
2
2
|
rank r >= 1 and an optional attribute axis that identifies an axis of data
|
3
3
|
(by default, the outer-most axis, that is axis 0). The output of the operation
|
4
4
|
is produced by creating a copy of the input data, and then updating its value
|
5
5
|
to values specified by updates at specific index positions specified by
|
6
6
|
indices. Its output shape is the same as the shape of data.
|
7
|
-
|
8
7
|
For each entry in updates, the target index in data is obtained by combining
|
9
8
|
the corresponding entry in indices with the index of the entry itself: the
|
10
9
|
index-value for dimension = axis is obtained from the value of the corresponding
|
11
10
|
entry in indices and the index-value for dimension != axis is obtained from the
|
12
11
|
index of the entry itself.
|
13
|
-
|
12
|
+
reduction allows specification of an optional reduction operation, which is applied to all values in updates
|
13
|
+
tensor into output at the specified indices.
|
14
|
+
In cases where reduction is set to "none", indices should not have duplicate entries: that is, if idx1 != idx2,
|
14
|
-
For instance, in a 2-D tensor case, the update
|
15
|
+
then indices[idx1] != indices[idx2]. For instance, in a 2-D tensor case, the update
|
15
|
-
is performed as below:
|
16
|
+
corresponding to the [i][j] entry is performed as below:
|
16
17
|
output[indices[i][j]][j] = updates[i][j] if axis = 0,
|
17
18
|
output[i][indices[i][j]] = updates[i][j] if axis = 1,
|
19
|
+
When reduction is set to "add", the update corresponding to the [i][j] entry is performed as below:
|
20
|
+
|
21
|
+
output[indices[i][j]][j] += updates[i][j] if axis = 0,
|
22
|
+
output[i][indices[i][j]] += updates[i][j] if axis = 1,
|
23
|
+
|
24
|
+
When reduction is set to "mul", the update corresponding to the [i][j] entry is performed as below:
|
25
|
+
|
26
|
+
output[indices[i][j]][j] *= updates[i][j] if axis = 0,
|
27
|
+
output[i][indices[i][j]] *= updates[i][j] if axis = 1,
|
18
28
|
This operator is the inverse of GatherElements. It is similar to Torch's Scatter operation.
|
19
|
-
|
20
29
|
Example 1:
|
21
30
|
data = [
|
22
31
|
[0.0, 0.0, 0.0],
|
23
32
|
[0.0, 0.0, 0.0],
|
24
33
|
[0.0, 0.0, 0.0],
|
25
34
|
]
|
26
35
|
indices = [
|
27
36
|
[1, 0, 2],
|
28
37
|
[0, 2, 1],
|
29
38
|
]
|
30
39
|
updates = [
|
31
40
|
[1.0, 1.1, 1.2],
|
32
41
|
[2.0, 2.1, 2.2],
|
33
42
|
]
|
34
43
|
output = [
|
35
44
|
[2.0, 1.1, 0.0]
|
36
45
|
[1.0, 0.0, 2.2]
|
37
46
|
[0.0, 2.1, 1.2]
|
38
47
|
]
|
39
48
|
Example 2:
|
40
49
|
data = [[1.0, 2.0, 3.0, 4.0, 5.0]]
|
41
50
|
indices = [[1, 3]]
|
42
51
|
updates = [[1.1, 2.1]]
|
43
52
|
axis = 1
|
44
53
|
output = [[1.0, 1.1, 3.0, 2.1, 5.0]]
|
45
54
|
### Attributes
|
46
55
|
* **axis - INT** (default is '0'):
|
47
56
|
Which axis to scatter on. Negative value means counting dimensions from the back. Accepted range is [-r, r-1] where r = rank(data).
|
57
|
+
* **reduction - STRING** (default is 'none'):
|
58
|
+
|
59
|
+
Type of reduction to apply: none (default), add, mul. 'none': no reduction applied. 'add': reduction using the addition operation. 'mul': reduction using the multiplication operation.
|
60
|
+
|
48
61
|
### Inputs
|
49
62
|
- **data** (heterogeneous) - **T**:
|
50
63
|
Tensor of rank r >= 1.
|
51
64
|
- **indices** (heterogeneous) - **Tind**:
|
52
65
|
Tensor of int32/int64 indices, of r >= 1 (same rank as input). All index values are expected to be within bounds [-s, s-1] along axis of size s. It is an error if any of the index values are out of bounds.
|
53
66
|
- **updates** (heterogeneous) - **T**:
|
54
67
|
Tensor of rank r >=1 (same rank and shape as indices)
|
55
68
|
### Outputs
|
56
69
|
- **output** (heterogeneous) - **T**:
|
57
70
|
Tensor of rank r >= 1 (same rank as input).
|
58
71
|
### Type Constraints
|
59
|
-
* **T** in ( tensor(bool), tensor(complex128), tensor(complex64), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ):
|
72
|
+
* **T** in ( tensor(bfloat16), tensor(bool), tensor(complex128), tensor(complex64), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ):
|
60
73
|
Input and output types can be of any tensor type.
|
61
74
|
* **Tind** in ( tensor(int32), tensor(int64) ):
|
62
75
|
Constrain indices to integer types
|