Merge pull request #1 from thu-ml/master

fix readme (#273)

README.md

@@ -235,7 +235,7 @@ result = ts.trainer.offpolicy_trainer(
     train_fn=lambda epoch, env_step: policy.set_eps(eps_train),
     test_fn=lambda epoch, env_step: policy.set_eps(eps_test),
     stop_fn=lambda mean_rewards: mean_rewards >= env.spec.reward_threshold,
-    writer=writer, task=task)
+    writer=writer)
 print(f'Finished training! Use {result["duration"]}')
 ```