def sample(self, obs):
means, logvars = self.policy(obs)
print(means, logvars)
sampled_act = means + (
layers.exp(logvars / 2.0) * # stddev
layers.gaussian_random(shape=(self.act_dim,), dtype='float32'))
return sampled_act
def policy(self, obs):
hid1 = self.fc1(obs)
hid2 = self.fc2(hid1)
mu = self.mean_linear(hid2)
log_std = self.log_std_linear(hid2)
log_std = LOG_STD_MIN + 0.5 * (LOG_STD_MAX - LOG_STD_MIN) * (log_std +
1)
log_std = layers.exp(log_std)
return mu, log_std