The FANN library is designed to be very easy to use. A feedforward ann can be created by a simple fann_create_standard function, while other ANNs can be created just as easily. The ANNs can be trained by fann_train_on_file and executed by fann_run.
All of this can be done without much knowledge of the internals of ANNs, although the ANNs created will still be powerful and effective. If you have more knowledge about ANNs, and desire more control, almost every part of the ANNs can be parametrized to create specialized and highly optimal ANNs.
| FANN Creation/ | The FANN library is designed to be very easy to use. |
| Creation, Destruction & Execution | |
| fann_create_standard | Creates a standard fully connected backpropagation neural network. |
| fann_create_standard_array | Just like fann_create_standard, but with an array of layer sizes instead of individual parameters. |
| fann_create_sparse | Creates a standard backpropagation neural network, which is not fully connected. |
| fann_create_sparse_array | Just like fann_create_sparse, but with an array of layer sizes instead of individual parameters. |
| fann_create_shortcut | Creates a standard backpropagation neural network, which is fully connected and which also has shortcut connections. |
| fann_create_shortcut_array | Just like fann_create_shortcut, but with an array of layer sizes instead of individual parameters. |
| fann_destroy | Destroys the entire network, properly freeing all the associated memory. |
| fann_copy | Creates a copy of a fann structure. |
| fann_run | Will run input through the neural network, returning an array of outputs, the number of which being equal to the number of neurons in the output layer. |
| fann_randomize_weights | Give each connection a random weight between min_weight and max_weight |
| fann_init_weights | Initialize the weights using Widrow + Nguyen’s algorithm. |
| fann_print_connections | Will print the connections of the ann in a compact matrix, for easy viewing of the internals of the ann. |
| Parameters | |
| fann_print_parameters | Prints all of the parameters and options of the ANN |
| fann_get_num_input | Get the number of input neurons. |
| fann_get_num_output | Get the number of output neurons. |
| fann_get_total_neurons | Get the total number of neurons in the entire network. |
| fann_get_total_connections | Get the total number of connections in the entire network. |
| fann_get_network_type | Get the type of neural network it was created as. |
| fann_get_connection_rate | Get the connection rate used when the network was created |
| fann_get_num_layers | Get the number of layers in the network |
| fann_get_layer_array | Get the number of neurons in each layer in the network. |
| fann_get_bias_array | Get the number of bias in each layer in the network. |
| fann_get_connection_array | Get the connections in the network. |
| fann_set_weight_array | Set connections in the network. |
| fann_set_weight | Set a connection in the network. |
| fann_get_weights | Get all the network weights. |
| fann_set_weights | Set network weights. |
| fann_set_user_data | Store a pointer to user defined data. |
| fann_get_user_data | Get a pointer to user defined data that was previously set with fann_set_user_data. |
| fann_disable_seed_rand | Disables the automatic random generator seeding that happens in FANN. |
| fann_enable_seed_rand | Enables the automatic random generator seeding that happens in FANN. |
| fann_get_decimal_point | Returns the position of the decimal point in the ann. |
| fann_get_multiplier | returns the multiplier that fix point data is multiplied with. |
FANN_EXTERNAL struct fann *FANN_API fann_create_standard( unsigned int num_layers, ... )
Creates a standard fully connected backpropagation neural network.
There will be a bias neuron in each layer (except the output layer), and this bias neuron will be connected to all neurons in the next layer. When running the network, the bias nodes always emits 1.
To destroy a struct fann use the fann_destroy function.
| num_layers | The total number of layers including the input and the output layer. |
| ... | Integer values determining the number of neurons in each layer starting with the input layer and ending with the output layer. |
A pointer to the newly created struct fann.
// Creating an ANN with 2 input neurons, 1 output neuron, // and two hidden layers with 8 and 9 neurons struct fann *ann = fann_create_standard(4, 2, 8, 9, 1);
fann_create_standard_array, fann_create_sparse, fann_create_shortcut
This function appears in FANN >= 2.0.0.
FANN_EXTERNAL struct fann *FANN_API fann_create_standard_array( unsigned int num_layers, const unsigned int * layers )
Just like fann_create_standard, but with an array of layer sizes instead of individual parameters.
// Creating an ANN with 2 input neurons, 1 output neuron,
// and two hidden layers with 8 and 9 neurons
unsigned int layers[4] = {2, 8, 9, 1};
struct fann *ann = fann_create_standard_array(4, layers);fann_create_standard, fann_create_sparse, fann_create_shortcut
This function appears in FANN >= 2.0.0.
FANN_EXTERNAL struct fann *FANN_API fann_create_sparse( float connection_rate, unsigned int num_layers, ... )
Creates a standard backpropagation neural network, which is not fully connected.
| connection_rate | The connection rate controls how many connections there will be in the network. If the connection rate is set to 1, the network will be fully connected, but if it is set to 0.5 only half of the connections will be set. A connection rate of 1 will yield the same result as fann_create_standard |
| num_layers | The total number of layers including the input and the output layer. |
| ... | Integer values determining the number of neurons in each layer starting with the input layer and ending with the output layer. |
A pointer to the newly created struct fann.
fann_create_sparse_array, fann_create_standard, fann_create_shortcut
This function appears in FANN >= 2.0.0.
FANN_EXTERNAL struct fann *FANN_API fann_create_sparse_array( float connection_rate, unsigned int num_layers, const unsigned int * layers )
Just like fann_create_sparse, but with an array of layer sizes instead of individual parameters.
See fann_create_standard_array for a description of the parameters.
fann_create_sparse, fann_create_standard, fann_create_shortcut
This function appears in FANN >= 2.0.0.
FANN_EXTERNAL struct fann *FANN_API fann_create_shortcut( unsigned int num_layers, ... )
Creates a standard backpropagation neural network, which is fully connected and which also has shortcut connections.
Shortcut connections are connections that skip layers. A fully connected network with shortcut connections is a network where all neurons are connected to all neurons in later layers. Including direct connections from the input layer to the output layer.
See fann_create_standard for a description of the parameters.
fann_create_shortcut_array, fann_create_standard, fann_create_sparse,
This function appears in FANN >= 2.0.0.
FANN_EXTERNAL struct fann *FANN_API fann_create_shortcut_array( unsigned int num_layers, const unsigned int * layers )
Just like fann_create_shortcut, but with an array of layer sizes instead of individual parameters.
See fann_create_standard_array for a description of the parameters.
fann_create_shortcut, fann_create_standard, fann_create_sparse
This function appears in FANN >= 2.0.0.
FANN_EXTERNAL struct fann * FANN_API fann_copy( struct fann * ann )
Creates a copy of a fann structure.
Data in the user data fann_set_user_data is not copied, but the user data pointer is copied.
This function appears in FANN >= 2.2.0.
FANN_EXTERNAL void FANN_API fann_randomize_weights( struct fann * ann, fann_type min_weight, fann_type max_weight )
Give each connection a random weight between min_weight and max_weight
From the beginning the weights are random between -0.1 and 0.1.
This function appears in FANN >= 1.0.0.
FANN_EXTERNAL void FANN_API fann_init_weights( struct fann * ann, struct fann_train_data * train_data )
Initialize the weights using Widrow + Nguyen’s algorithm.
This function behaves similarly to fann_randomize_weights. It will use the algorithm developed by Derrick Nguyen and Bernard Widrow to set the weights in such a way as to speed up training. This technique is not always successful, and in some cases can be less efficient than a purely random initialization.
The algorithm requires access to the range of the input data (ie, largest and smallest input), and therefore accepts a second argument, data, which is the training data that will be used to train the network.
fann_randomize_weights, fann_read_train_from_file
This function appears in FANN >= 1.1.0.
FANN_EXTERNAL void FANN_API fann_print_connections( struct fann * ann )
Will print the connections of the ann in a compact matrix, for easy viewing of the internals of the ann.
The output from fann_print_connections on a small (2 2 1) network trained on the xor problem
Layer / Neuron 012345 L 1 / N 3 BBa... L 1 / N 4 BBA... L 1 / N 5 ...... L 2 / N 6 ...BBA L 2 / N 7 ......
This network has five real neurons and two bias neurons. This gives a total of seven neurons named from 0 to 6. The connections between these neurons can be seen in the matrix. “.” is a place where there is no connection, while a character tells how strong the connection is on a scale from a-z. The two real neurons in the hidden layer (neuron 3 and 4 in layer 1) have connections from the three neurons in the previous layer as is visible in the first two lines. The output neuron (6) has connections from the three neurons in the hidden layer 3 - 5 as is visible in the fourth line.
To simplify the matrix output neurons are not visible as neurons that connections can come from, and input and bias neurons are not visible as neurons that connections can go to.
This function appears in FANN >= 1.2.0.
FANN_EXTERNAL enum fann_nettype_enum FANN_API fann_get_network_type( struct fann * ann )
Get the type of neural network it was created as.
| ann | A previously created neural network structure of type struct fann pointer. |
The neural network type from enum fann_network_type_enum
This function appears in FANN >= 2.1.0
FANN_EXTERNAL float FANN_API fann_get_connection_rate( struct fann * ann )
Get the connection rate used when the network was created
| ann | A previously created neural network structure of type struct fann pointer. |
The connection rate
This function appears in FANN >= 2.1.0
FANN_EXTERNAL unsigned int FANN_API fann_get_num_layers( struct fann * ann )
Get the number of layers in the network
| ann | A previously created neural network structure of type struct fann pointer. |
The number of layers in the neural network
// Obtain the number of layers in a neural network struct fann *ann = fann_create_standard(4, 2, 8, 9, 1); unsigned int num_layers = fann_get_num_layers(ann);
This function appears in FANN >= 2.1.0
FANN_EXTERNAL void FANN_API fann_get_layer_array( struct fann * ann, unsigned int * layers )
Get the number of neurons in each layer in the network.
Bias is not included so the layers match the fann_create functions.
| ann | A previously created neural network structure of type struct fann pointer. |
The layers array must be preallocated to at least sizeof(unsigned int) * fann_num_layers() long.
This function appears in FANN >= 2.1.0
FANN_EXTERNAL void FANN_API fann_get_bias_array( struct fann * ann, unsigned int * bias )
Get the number of bias in each layer in the network.
| ann | A previously created neural network structure of type struct fann pointer. |
The bias array must be preallocated to at least sizeof(unsigned int) * fann_num_layers() long.
This function appears in FANN >= 2.1.0
FANN_EXTERNAL void FANN_API fann_get_connection_array( struct fann * ann, struct fann_connection * connections )
Get the connections in the network.
| ann | A previously created neural network structure of type struct fann pointer. |
The connections array must be preallocated to at least sizeof(struct fann_connection) * fann_get_total_connections() long.
This function appears in FANN >= 2.1.0
FANN_EXTERNAL void FANN_API fann_set_weight_array( struct fann * ann, struct fann_connection * connections, unsigned int num_connections )
Set connections in the network.
| ann | A previously created neural network structure of type struct fann pointer. |
Only the weights can be changed, connections and weights are ignored if they do not already exist in the network.
The array must have sizeof(struct fann_connection) * num_connections size.
This function appears in FANN >= 2.1.0
FANN_EXTERNAL void FANN_API fann_set_weight( struct fann * ann, unsigned int from_neuron, unsigned int to_neuron, fann_type weight )
Set a connection in the network.
| ann | A previously created neural network structure of type struct fann pointer. |
Only the weights can be changed. The connection/weight is ignored if it does not already exist in the network.
This function appears in FANN >= 2.1.0
FANN_EXTERNAL void FANN_API fann_get_weights( struct fann * ann, fann_type * weights )
Get all the network weights.
| ann | A previously created neural network structure of type struct fann pointer. |
| weights | A fann_type pointer to user data. It is the responsibility of the user to allocate sufficient space to store all the weights. |
This function appears in FANN >= x.y.z
FANN_EXTERNAL void FANN_API fann_set_weights( struct fann * ann, fann_type * weights )
Set network weights.
| ann | A previously created neural network structure of type struct fann pointer. |
| weights | A fann_type pointer to user data. It is the responsibility of the user to make the weights array sufficient long to store all the weights. |
This function appears in FANN >= x.y.z
FANN_EXTERNAL void FANN_API fann_set_user_data( struct fann * ann, void * user_data )
Store a pointer to user defined data. The pointer can be retrieved with fann_get_user_data for example in a callback. It is the user’s responsibility to allocate and deallocate any data that the pointer might point to.
| ann | A previously created neural network structure of type struct fann pointer. |
| user_data | A void pointer to user defined data. |
This function appears in FANN >= 2.1.0
FANN_EXTERNAL void * FANN_API fann_get_user_data( struct fann * ann )
Get a pointer to user defined data that was previously set with fann_set_user_data. It is the user’s responsibility to allocate and deallocate any data that the pointer might point to.
| ann | A previously created neural network structure of type struct fann pointer. |
A void pointer to user defined data.
This function appears in FANN >= 2.1.0
FANN_EXTERNAL void FANN_API fann_disable_seed_rand()
Disables the automatic random generator seeding that happens in FANN.
Per default FANN will always seed the random generator when creating a new network, unless FANN_NO_SEED is defined during compilation of the library. This method can disable this at runtime.
This function appears in FANN >= 2.3.0
FANN_EXTERNAL void FANN_API fann_enable_seed_rand()
Enables the automatic random generator seeding that happens in FANN.
Per default FANN will always seed the random generator when creating a new network, unless FANN_NO_SEED is defined during compilation of the library. This method can disable this at runtime.
This function appears in FANN >= 2.3.0
FANN_EXTERNAL unsigned int FANN_API fann_get_decimal_point( struct fann * ann )
Returns the position of the decimal point in the ann.
This function is only available when the ANN is in fixed point mode.
The decimal point is described in greater detail in the tutorial <Fixed Point Usage>.
<Fixed Point Usage>, fann_get_multiplier, fann_save_to_fixed, fann_save_train_to_fixed
This function appears in FANN >= 1.0.0.
FANN_EXTERNAL unsigned int FANN_API fann_get_multiplier( struct fann * ann )
returns the multiplier that fix point data is multiplied with.
This function is only available when the ANN is in fixed point mode.
The multiplier is the used to convert between floating point and fixed point notation. A floating point number is multiplied with the multiplier in order to get the fixed point number and visa versa.
The multiplier is described in greater detail in the tutorial <Fixed Point Usage>.
<Fixed Point Usage>, fann_get_decimal_point, fann_save_to_fixed, fann_save_train_to_fixed
This function appears in FANN >= 1.0.0.
Creates a standard fully connected backpropagation neural network.
FANN_EXTERNAL struct fann *FANN_API fann_create_standard( unsigned int num_layers, ... )
Just like fann_create_standard, but with an array of layer sizes instead of individual parameters.
FANN_EXTERNAL struct fann *FANN_API fann_create_standard_array( unsigned int num_layers, const unsigned int * layers )
Creates a standard backpropagation neural network, which is not fully connected.
FANN_EXTERNAL struct fann *FANN_API fann_create_sparse( float connection_rate, unsigned int num_layers, ... )
Just like fann_create_sparse, but with an array of layer sizes instead of individual parameters.
FANN_EXTERNAL struct fann *FANN_API fann_create_sparse_array( float connection_rate, unsigned int num_layers, const unsigned int * layers )
Creates a standard backpropagation neural network, which is fully connected and which also has shortcut connections.
FANN_EXTERNAL struct fann *FANN_API fann_create_shortcut( unsigned int num_layers, ... )
Just like fann_create_shortcut, but with an array of layer sizes instead of individual parameters.
FANN_EXTERNAL struct fann *FANN_API fann_create_shortcut_array( unsigned int num_layers, const unsigned int * layers )
Destroys the entire network, properly freeing all the associated memory.
FANN_EXTERNAL void FANN_API fann_destroy( struct fann * ann )
Creates a copy of a fann structure.
FANN_EXTERNAL struct fann * FANN_API fann_copy( struct fann * ann )
Will run input through the neural network, returning an array of outputs, the number of which being equal to the number of neurons in the output layer.
FANN_EXTERNAL fann_type * FANN_API fann_run( struct fann * ann, fann_type * input )
Give each connection a random weight between min_weight and max_weight
FANN_EXTERNAL void FANN_API fann_randomize_weights( struct fann * ann, fann_type min_weight, fann_type max_weight )
Initialize the weights using Widrow + Nguyen’s algorithm.
FANN_EXTERNAL void FANN_API fann_init_weights( struct fann * ann, struct fann_train_data * train_data )
Will print the connections of the ann in a compact matrix, for easy viewing of the internals of the ann.
FANN_EXTERNAL void FANN_API fann_print_connections( struct fann * ann )
Prints all of the parameters and options of the ANN
FANN_EXTERNAL void FANN_API fann_print_parameters( struct fann * ann )
Get the number of input neurons.
FANN_EXTERNAL unsigned int FANN_API fann_get_num_input( struct fann * ann )
Get the number of output neurons.
FANN_EXTERNAL unsigned int FANN_API fann_get_num_output( struct fann * ann )
Get the total number of neurons in the entire network.
FANN_EXTERNAL unsigned int FANN_API fann_get_total_neurons( struct fann * ann )
Get the total number of connections in the entire network.
FANN_EXTERNAL unsigned int FANN_API fann_get_total_connections( struct fann * ann )
Get the type of neural network it was created as.
FANN_EXTERNAL enum fann_nettype_enum FANN_API fann_get_network_type( struct fann * ann )
Get the connection rate used when the network was created
FANN_EXTERNAL float FANN_API fann_get_connection_rate( struct fann * ann )
Get the number of layers in the network
FANN_EXTERNAL unsigned int FANN_API fann_get_num_layers( struct fann * ann )
Get the number of neurons in each layer in the network.
FANN_EXTERNAL void FANN_API fann_get_layer_array( struct fann * ann, unsigned int * layers )
Get the number of bias in each layer in the network.
FANN_EXTERNAL void FANN_API fann_get_bias_array( struct fann * ann, unsigned int * bias )
Get the connections in the network.
FANN_EXTERNAL void FANN_API fann_get_connection_array( struct fann * ann, struct fann_connection * connections )
Set connections in the network.
FANN_EXTERNAL void FANN_API fann_set_weight_array( struct fann * ann, struct fann_connection * connections, unsigned int num_connections )
Set a connection in the network.
FANN_EXTERNAL void FANN_API fann_set_weight( struct fann * ann, unsigned int from_neuron, unsigned int to_neuron, fann_type weight )
Get all the network weights.
FANN_EXTERNAL void FANN_API fann_get_weights( struct fann * ann, fann_type * weights )
Set network weights.
FANN_EXTERNAL void FANN_API fann_set_weights( struct fann * ann, fann_type * weights )
Store a pointer to user defined data.
FANN_EXTERNAL void FANN_API fann_set_user_data( struct fann * ann, void * user_data )
Get a pointer to user defined data that was previously set with fann_set_user_data.
FANN_EXTERNAL void * FANN_API fann_get_user_data( struct fann * ann )
Disables the automatic random generator seeding that happens in FANN.
FANN_EXTERNAL void FANN_API fann_disable_seed_rand()
Enables the automatic random generator seeding that happens in FANN.
FANN_EXTERNAL void FANN_API fann_enable_seed_rand()
Returns the position of the decimal point in the ann.
FANN_EXTERNAL unsigned int FANN_API fann_get_decimal_point( struct fann * ann )
returns the multiplier that fix point data is multiplied with.
FANN_EXTERNAL unsigned int FANN_API fann_get_multiplier( struct fann * ann )
Does the same as fann_train_on_data, but reads the training data directly from a file.
FANN_EXTERNAL void FANN_API fann_train_on_file( struct fann * ann, const char * filename, unsigned int max_epochs, unsigned int epochs_between_reports, float desired_error )
The fast artificial neural network (fann) structure.
struct fann
Test with a set of inputs, and a set of desired outputs.
FANN_EXTERNAL fann_type * FANN_API fann_test( struct fann * ann, fann_type * input, fann_type * desired_output )
Reads a file that stores training data.
FANN_EXTERNAL struct fann_train_data *FANN_API fann_read_train_from_file( const char * filename )
Saves the entire network to a configuration file.
FANN_EXTERNAL int FANN_API fann_save_to_fixed( struct fann * ann, const char * configuration_file )
Saves the training structure to a fixed point data file.
FANN_EXTERNAL int FANN_API fann_save_train_to_fixed( struct fann_train_data * data, const char * filename, unsigned int decimal_point )