Softmax - 11 vs 13¶
- Softmax11 → Softmax13 +10 -19
Softmax11 → Softmax13
RENAMED
@@ -1 +1 @@
|
|
1
|
-
The operator computes the
|
1
|
+
The operator computes the normalized exponential values for the given input:
|
2
|
+
Softmax(input, axis) = Exp(input) / ReduceSum(Exp(input), axis=axis, keepdims=1)
|
3
|
+
|
2
|
-
|
4
|
+
The "axis" attribute indicates the dimension along which Softmax
|
3
|
-
The input does not need to explicitly be a 2D vector; rather, it will be
|
4
|
-
coerced into one. For an arbitrary n-dimensional tensor
|
5
|
-
input in [a_0, a_1, ..., a_{k-1}, a_k, ..., a_{n-1}] and k is
|
6
|
-
the axis provided, then input will be coerced into a 2-dimensional tensor with
|
7
|
-
dimensions [a_0 * ... * a_{k-1}, a_k * ... * a_{n-1}]. For the default
|
8
|
-
case where axis=1, this means the input tensor will be coerced into a 2D tensor
|
9
|
-
of dimensions [a_0, a_1 * ... * a_{n-1}], where a_0 is often the batch size.
|
10
|
-
In this situation, we must have a_0 = N and a_1 * ... * a_{n-1} = D.
|
11
|
-
Each of these dimensions must be matched correctly, or else the operator
|
12
|
-
will
|
5
|
+
will be performed. The output tensor has the same shape
|
13
|
-
and contains the
|
6
|
+
and contains the Softmax values of the corresponding input.
|
14
7
|
**Attributes**
|
15
8
|
* **axis**:
|
9
|
+
Describes the dimension Softmax will be performed on. Negative
|
16
|
-
Describes the axis of the inputs when coerced to 2D; defaults to one
|
17
|
-
because the 0th axis most likely describes the batch_size. Negative
|
18
10
|
value means counting dimensions from the back. Accepted range is
|
19
11
|
[-r, r-1] where r = rank(input).
|
20
12
|
**Inputs**
|
21
13
|
* **input** (heterogeneous) - **T**:
|
14
|
+
The input tensor of rank >= axis.
|
22
|
-
The input tensor that's coerced into a 2D matrix of size (NxD) as
|
23
|
-
described above.
|
24
15
|
**Outputs**
|
25
16
|
* **output** (heterogeneous) - **T**:
|
26
|
-
The output values with the same shape as input tensor
|
17
|
+
The output values with the same shape as the input tensor.
|
27
|
-
size without coercion).
|
28
18
|
**Type Constraints**
|
29
19
|
* **T** in (
|
20
|
+
tensor(bfloat16),
|
30
21
|
tensor(double),
|
31
22
|
tensor(float),
|
32
23
|
tensor(float16)
|
33
24
|
):
|
34
25
|
Constrain input and output types to float tensors.
|