Hardmax - 1 vs 11

Files changed (1) hide show
  1. Hardmax1 → Hardmax11 +7 -6
Hardmax1 → Hardmax11 RENAMED
@@ -1 +1 @@
1
1
  The operator computes the hardmax (1 for the first maximum value, and 0 for all others) values for each layer in the batch
2
+ of the given input.
2
- of the given input. The input is a 2-D tensor (Tensor<float>) of size
3
- (batch_size x input_feature_dimensions). The output tensor has the same shape
4
- and contains the hardmax values of the corresponding input.
5
- Input does not need to explicitly be a 2D vector; rather, it will be
3
+ The input does not need to explicitly be a 2D vector; rather, it will be
6
4
  coerced into one. For an arbitrary n-dimensional tensor
7
5
  input in [a_0, a_1, ..., a_{k-1}, a_k, ..., a_{n-1}] and k is
8
6
  the axis provided, then input will be coerced into a 2-dimensional tensor with
9
7
  dimensions [a_0 * ... * a_{k-1}, a_k * ... * a_{n-1}]. For the default
10
8
  case where axis=1, this means the input tensor will be coerced into a 2D tensor
11
9
  of dimensions [a_0, a_1 * ... * a_{n-1}], where a_0 is often the batch size.
12
10
  In this situation, we must have a_0 = N and a_1 * ... * a_{n-1} = D.
13
11
  Each of these dimensions must be matched correctly, or else the operator
14
- will throw errors.
12
+ will throw errors. The output tensor has the same shape
13
+ and contains the hardmax values of the corresponding input.
15
14
  **Attributes**
16
15
  * **axis**:
17
16
  Describes the axis of the inputs when coerced to 2D; defaults to one
18
- because the 0th axis most likely describes the batch_size
17
+ because the 0th axis most likely describes the batch_size. Negative
18
+ value means counting dimensions from the back. Accepted range is
19
+ [-r, r-1] where r = rank(input).
19
20
  **Inputs**
20
21
  * **input** (heterogeneous) - **T**:
21
22
  The input tensor that's coerced into a 2D matrix of size (NxD) as
22
23
  described above.
23
24
  **Outputs**
24
25
  * **output** (heterogeneous) - **T**:
25
26
  The output values with the same shape as input tensor (the original
26
27
  size without coercion).
27
28
  **Type Constraints**
28
29
  * **T** in (
29
30
  tensor(double),
30
31
  tensor(float),
31
32
  tensor(float16)
32
33
  ):
33
34
  Constrain input and output types to float tensors.