char * layer_args_json_export(struct ArgsLayer *args)
Returns a json formatted string of the neural layer parameters.
size_t layer_args_load(struct ArgsLayer **largs, FILE *fp)
Loads neural network layer parameters.
void layer_args_init(struct ArgsLayer *args)
Sets layer parameters to default values.
struct ArgsLayer * layer_args_tail(struct ArgsLayer *head)
Returns the current output layer arguments.
uint32_t layer_args_opt(const struct ArgsLayer *args)
Returns a bitstring representing the permissions granted by a layer.
char * layer_args_json_import(struct ArgsLayer *largs, cJSON *json)
Sets the layer parameters from a cJSON object.
void layer_args_free(struct ArgsLayer **largs)
Frees memory used by a list of layer parameters and points to NULL.
size_t layer_args_save(const struct ArgsLayer *args, FILE *fp)
Saves neural network layer parameters.
void layer_args_validate(struct ArgsLayer *args)
Checks network layer arguments are valid.
struct ArgsLayer * layer_args_copy(const struct ArgsLayer *src)
Creates and returns a copy of specified layer parameters.
Parameters for initialising a neural network layer.
_Bool evolve_weights
Ability to evolve weights.
int n_init
Initial number of units / neurons / filters.
_Bool evolve_neurons
Ability to evolve number of units.
_Bool sgd_weights
Ability to update weights with gradient descent.
_Bool evolve_functions
Ability to evolve activation function.
double decay
Weight decay for gradient descent.
double momentum
Momentum for gradient descent.
_Bool evolve_eta
Ability to evolve gradient descent rate.
int max_neuron_grow
Maximum number neurons to add per mutation event.
double probability
Usage depends on layer implementation.
double eta
Gradient descent rate.
double eta_min
Current gradient descent rate.
int channels
Pool, Conv, and Upsample.
int n_max
Maximum number of units / neurons.
double scale
Usage depends on layer implementation.
int width
Pool, Conv, and Upsample.
int height
Pool, Conv, and Upsample.
int stride
Pool, Conv, and Upsample.
int n_inputs
Number of inputs.
int type
Layer type: CONNECTED, DROPOUT, etc.
struct ArgsLayer * next
Next layer parameters.
_Bool evolve_connect
Ability to evolve weight connectivity.
int recurrent_function
Recurrent activation function.
Utility functions for random number handling, etc.