def mlp_gaussian_policy()

in safe_rl/sac/sac.py [0:0]

• 10 lines of code
• 1 McCabe index (conditional complexity)

def mlp_gaussian_policy(x, a, hidden_sizes, activation, output_activation):
    act_dim = a.shape.as_list()[-1]
    net = mlp(x, list(hidden_sizes), activation, activation)
    mu = tf.layers.dense(net, act_dim, activation=output_activation)
    log_std = tf.layers.dense(net, act_dim, activation=None)
    log_std = tf.clip_by_value(log_std, LOG_STD_MIN, LOG_STD_MAX)

    std = tf.exp(log_std)
    pi = mu + tf.random_normal(tf.shape(mu)) * std
    logp_pi = gaussian_likelihood(pi, mu, log_std)
    return mu, pi, logp_pi