aboutsummaryrefslogtreecommitdiff
path: root/chapters/activation_funcs.adoc
diff options
context:
space:
mode:
Diffstat (limited to 'chapters/activation_funcs.adoc')
-rw-r--r--chapters/activation_funcs.adoc49
1 files changed, 9 insertions, 40 deletions
diff --git a/chapters/activation_funcs.adoc b/chapters/activation_funcs.adoc
index 46fa19d..3bbeb30 100644
--- a/chapters/activation_funcs.adoc
+++ b/chapters/activation_funcs.adoc
@@ -30,24 +30,17 @@ for_each(index in shape) {
==== SIGMOID
-Applies the sigmoid logistic function to each element of the input tensor.
+Sigmoid function: output = 1 / (1 + exp(-input))
-// sigmoid(x) = \frac{1}{1 + e^{-x}}
+For quantized integer data types, the TABLE operator should be used instead with
+the following definition.
-.Calculation for the sigmoid function
-image::sigmoid.svg["Sigmoid definition"]
+The sigmoid table has 513 entries each of 16-bit precision and covering the input range -16.0 to +16.0 in steps of 1/16.
-For quantized integer data types, the TABLE operator should be used instead.
-Each implementation may choose an appropriate TABLE given the scale and zero point of the input data.
-Eight or sixteen bit precision tables may be used based on the input tensor to the sigmoid function.
-Below we give an example table generation for 16-bit sigmoid.
-This sigmoid table has 513 entries each of 16-bit precision and covering the input range -16.0 to +16.0 in steps of 1/16.
-
-.Code for generating 16-bit sigmoid table
[source,c++]
----
int16_t sigmoid_reference(int16_t x) { // input x range is -256 to + 256 inclusive
- fp64_t v = (fp64_t)x / (fp64_t)16;
+ F64 v = (double)x / (double)16;
v = 1.0/(1.0 + exp(-v));
return round_to_nearest_int(32768.0 * v);
}
@@ -57,34 +50,19 @@ generate_lookup_table(&sigmoid_table, &sigmoid_reference);
include::{generated}/operators/SIGMOID.adoc[]
-[source,c++]
-----
-for_each(index in shape) {
- in_out_t value1 = tensor_read<in_out_t>(input, shape, index);
- value = sigmoid<in_out_t>(value1);
- tensor_write<in_out_t>(output, shape, index, value);
-}
-----
-
==== TANH
Parameterized hyperbolic tangent.
-// tanh(x) = \frac{1 - e^{-2x}}{1 + e^{-2x}}
-.Calculation for the sigmoid function
-image::tanh.svg["Hyperbolic tangent definition"]
+For quantized integer data types, the TABLE operator should be used instead with
+the following definition.
-For quantized integer data types, the TABLE operator should be used instead.
-Each implementation may choose an appropriate TABLE given the scale and zero point of the input data.
-Eight or sixteen bit precision tables may be used based on the input tensor to the sigmoid function.
-Below we give an example table generation for 16-bit hyperbolic tangent.
-This tanh_table has 513 entries each of 16-bit precision and covering the input range -8.0 to +8.0 in steps of 1/32.
+The tanh_table has 513 entries each of 16-bit precision and covering the input range -8.0 to +8.0 in steps of 1/32. The table is specified by:
-.Calculation of an example 16-bit tanh table
[source,c++]
----
int16_t tanh_reference(int16_t x) { // input x range is -256 to +256 inclusive
- fp64_t v = (fp64_t)x/(fp64_t)32;
+ F64 v = (double)x/(double)32;
v = exp(-2.0*v);
v = (1.0-v)/(1.0+v);
return round_to_nearest_int(32768.0 * v);
@@ -94,12 +72,3 @@ generate_lookup_table(&tanh_table, &tanh_reference);
----
include::{generated}/operators/TANH.adoc[]
-
-[source,c++]
-----
-for_each(index in shape) {
- in_out_t value1 = tensor_read<in_out_t>(input, shape, index);
- value = tanh<in_out_t>(value1);
- tensor_write<in_out_t>(output, shape, index, value);
-}
-----