flush logger

async_worker/dpg.py

@@ -135,4 +135,6 @@ class DeterministicPolicyGradient:
         self.shared_reward_normalizer.offline_stats.merge(self.reward_normalizer.online_stats)
         self.reward_normalizer.online_stats.zero()
 
+        config.logger.writer.file_writer.flush()
+
         return steps, total_reward

async_worker/ppo.py

@@ -167,4 +167,6 @@ class ProximalPolicyOptimization:
                 self.critic_opt.step()
                 config.total_epochs.value += 1
 
+        config.logger.writer.file_writer.flush()
+
         return batched_steps, batched_rewards