TopK - 10 vs 11

Files changed (1) hide show
  1. TopK10 → TopK11 +23 -5
TopK10 → TopK11 RENAMED
@@ -1 +1 @@
1
- Retrieve the top-K elements along a specified axis. Given an input tensor of
1
+ Retrieve the top-K largest or smallest elements along a specified axis. Given an input tensor of
2
2
  shape [a_1, a_2, ..., a_n, r] and integer argument k, return two outputs:
3
3
  -Value tensor of shape [a_1, a_2, ..., a_{axis-1}, k, a_{axis+1}, ... a_n]
4
4
  which contains the values of the top k elements along the specified axis
5
5
  -Index tensor of shape [a_1, a_2, ..., a_{axis-1}, k, a_{axis+1}, ... a_n] which
6
6
  contains the indices of the top k elements (original indices from the input
7
7
  tensor).
8
+ If "largest" is 1 (the default value) then the k largest elements are returned.
9
+ If "sorted" is 1 (the default value) then the resulting k elements will be sorted.
10
+ If "sorted" is 0, order of returned 'Values' and 'Indices' are undefined.
11
+
8
- Given two equivalent values, this operator uses the indices along the axis as
12
+ Given two equivalent values, this operator uses the indices along the axis as
9
13
  a tiebreaker. That is, the element with the lower index will appear first.
10
14
  **Attributes**
11
15
  * **axis**:
12
- Dimension on which to do the sort.
16
+ Dimension on which to do the sort. Negative value means counting
17
+ dimensions from the back. Accepted range is [-r, r-1] where r =
18
+ rank(input).
19
+ * **largest**:
20
+ Whether to return the top-K largest or smallest elements.
21
+ * **sorted**:
22
+ Whether to return the elements in sorted order.
13
23
  **Inputs**
14
24
  * **X** (heterogeneous) - **T**:
15
25
  Tensor of shape [a_1, a_2, ..., a_n, r]
16
26
  * **K** (heterogeneous) - **tensor(int64)**:
17
27
  A 1-D tensor containing a single positive value corresponding to the
18
28
  number of top elements to retrieve
19
29
  **Outputs**
20
30
  * **Values** (heterogeneous) - **T**:
21
31
  Tensor of shape [a_1, a_2, ..., a_{axis-1}, k, a_{axis+1}, ... a_n]
22
32
  containing top K values from the input tensor
23
33
  * **Indices** (heterogeneous) - **I**:
24
34
  Tensor of shape [a_1, a_2, ..., a_{axis-1}, k, a_{axis+1}, ... a_n]
25
35
  containing the corresponding input tensor indices for the top K
26
36
  values.
27
37
  **Type Constraints**
28
38
  * **T** in (
29
39
  tensor(double),
30
40
  tensor(float),
31
- tensor(float16)
41
+ tensor(float16),
42
+ tensor(int16),
43
+ tensor(int32),
44
+ tensor(int64),
45
+ tensor(int8),
46
+ tensor(uint16),
47
+ tensor(uint32),
48
+ tensor(uint64),
49
+ tensor(uint8)
32
50
  ):
33
- Constrain input and output types to float tensors.
51
+ Constrain input and output types to numeric tensors.
34
52
  * **I** in (
35
53
  tensor(int64)
36
54
  ):
37
55
  Constrain index tensor to int64