|
NeuroTIC 0.0
|
Activation Function List
http://tituxdev.github.io/NeuroTIC/src/CPU/ntactivation.c.
More...

Go to the source code of this file.
Enumerations | |
| enum | ntact_function_id_t { NTACT_BOOLEAN , NTACT_SIGMOID , NTACT_TOTAL_FUNCTIONS } |
| Enumeration of supported activation functions. More... | |
Variables | |
| float(* | ntact_activation [NTACT_TOTAL_FUNCTIONS][2])(float) |
| Activation function and derivative dispatch tables. | |
| float | ntact_rand_range [NTACT_TOTAL_FUNCTIONS][2] |
| Random initialization range table for activation functions. | |
Activation Function List
http://tituxdev.github.io/NeuroTIC/src/CPU/ntactivation.c.
@copydetauls ntactivation.c
Definition in file ntactivation.h.
| enum ntact_function_id_t |
Enumeration of supported activation functions.
This enumeration defines identifiers for the activation functions implemented in the NeuroTIC framework.
Each identifier corresponds to a specific activation function and its derivative, which are stored in the ntact_activation dispatch table.
The NTACT_TOTAL_FUNCTIONS entry is used to indicate the total number of activation functions available, which is useful for iterating over the dispatch table and for validating activation function identifiers.
Additional activation functions can be added to this enumeration as needed, along with their corresponding implementations in the ntact_activation dispatch table and appropriate entries in the ntact_rand_range for weight initialization.
| Enumerator | |
|---|---|
| NTACT_BOOLEAN | Boolean step activation function. |
| NTACT_SIGMOID | Sigmoid activation function. |
| NTACT_TOTAL_FUNCTIONS | Total number of activation functions. |
|
extern |
Activation function and derivative dispatch tables.
Defines the mapping between activation identifiers and their associated execution functions.
Indexing:
The table is initialized with the supported activation functions, and can be extended to include additional functions as needed.
Referenced by activate().
|
extern |
Random initialization range table for activation functions.
randnet function to ensure that weights are initialized within appropriate bounds for each activation function, which can help improve training performance and convergence. Referenced by randnet().