TODO

Add support for following functions:

* fann_get_bit_fail
* fann_reset_MSE
* fann_train_epoch
* fann_test_data
* fann_save_train

* fann_get/set_training_algorithm
* fann_get/set_learning_rate
* fann_get/set_learning_momentum
* fann_set_activation_function
* fann_get/set_activation_steepness
* fann_get/set_train_error_function
* fann_get/set_train_stop_function
* fann_get/set_bit_fail_limit
* fann_get/set_quickprop_decay
* fann_get/set_quickprop_mu
All below
fann_get_rprop_increase_factor
fann_set_rprop_increase_factor
fann_get_rprop_decrease_factor
fann_set_rprop_decrease_factor
fann_get_rprop_delta_min
fann_set_rprop_delta_min
fann_get_rprop_delta_max
fann_set_rprop_delta_max
fann_get_rprop_delta_zero
fann_set_rprop_delta_zero
fann_get_sarprop_weight_decay_shift
fann_set_sarprop_weight_decay_shift
fann_get_sarprop_step_error_threshold_factor
fann_set_sarprop_step_error_threshold_factor
fann_get_sarprop_step_error_shift
fann_set_sarprop_step_error_shift
fann_get_sarprop_temperature
fann_set_sarprop_temperature


Make all fann calls thread safe by using mutex locks

make training asynchronous