modifications for minecraft

configs.yaml

@@ -177,6 +177,9 @@ minecraft:
   task: minecraft_diamond
   step: 1e8
   envs: 16
+  # no eval
+  eval_episode_num: 0
+  eval_every: 1e8
   action_repeat: 1
   train_ratio: 16
   video_pred_log: true

dreamer.py

@@ -230,6 +230,7 @@ def make_env(config, mode):
     env = wrappers.TimeLimit(env, config.time_limit)
     env = wrappers.SelectAction(env, key="action")
     env = wrappers.UUID(env)
+    if suite == "minecraft":
         env = wrappers.RewardObs(env)
     return env
 
@@ -326,6 +327,7 @@ def main(config):
     # make sure eval will be executed once after config.steps
     while agent._step < config.steps + config.eval_every:
         logger.write()
+        if config.eval_episode_num > 0:
             print("Start evaluation.")
             eval_policy = functools.partial(agent, training=False)
             tools.simulate(

networks.py

@@ -415,7 +415,7 @@ class MultiDecoder(nn.Module):
         vector_dist,
     ):
         super(MultiDecoder, self).__init__()
-        excluded = ("is_first", "is_last", "is_terminal", "reward")
+        excluded = ("is_first", "is_last", "is_terminal")
         shapes = {k: v for k, v in shapes.items() if k not in excluded}
         self.cnn_shapes = {
             k: v for k, v in shapes.items() if len(v) == 3 and re.match(cnn_keys, k)