Merge pull request #166 from GFNOrg/turn_off_replay_buffer

turn off replay buffer while we determine issue with off policy sampling

tutorials/examples/train_hypergrid.py

@@ -292,7 +292,7 @@ def main(args):  # noqa: C901
     parser.add_argument(
         "--replay_buffer_size",
         type=int,
-        default=100,
+        default=0,
         help="If zero, no replay buffer is used. Otherwise, the replay buffer is used.",
     )