diff --git a/run_maddpg.py b/run_maddpg.py
index 59db98d..740318e 100644
--- a/run_maddpg.py
+++ b/run_maddpg.py
@@ -2,7 +2,7 @@
 from ray.tune import run_experiments
 from ray.tune.registry import register_trainable, register_env
 from env import MultiAgentParticleEnv
-import ray.rllib.agents.maddpg.maddpg as maddpg
+import ray.rllib.contrib.maddpg.maddpg as maddpg
 import argparse
 
 import os
@@ -116,7 +116,7 @@ def gen_policy(i):
 
     run_experiments({
         "MADDPG_RLLib": {
-            "run": "MADDPG",
+            "run": "contrib/MADDPG",
             "env": "mpe",
             "stop": {
                 "episodes_total": args.num_episodes,