[Bug] atari_wrapper - numpy problem (AttributeError: 'numpy.random._generator.Generator' object has no attribute 'randint')

[semihtasbas]

Hi,

I try to use your Atari example in doc, but I receive an error about numpy. I couldn't find a solution about that. I even delete my anaconda and setup environment from scratch. I shared details below.

from stable_baselines3.common.env_util import make_atari_env
from stable_baselines3.common.vec_env import VecFrameStack
from stable_baselines3 import A2C

# There already exists an environment generator
# that will make and wrap atari environments correctly.
# Here we are also multi-worker training (n_envs=4 => 4 environments)
env = make_atari_env('PongNoFrameskip-v4', n_envs=4, seed=0)
# Frame-stacking with 4 frames
env = VecFrameStack(env, n_stack=4)

model = A2C('CnnPolicy', env, verbose=1)
model.learn(total_timesteps=25_000)

obs = env.reset()
while True:
    action, _states = model.predict(obs)
    obs, rewards, dones, info = env.step(action)
    env.render()

Output:

A.L.E: Arcade Learning Environment (version 0.7.5+db37282)
[Powered by Stella]
Using cuda device
Wrapping the env in a VecTransposeImage.
Traceback (most recent call last):
  File "/home/x/PycharmProjects/tries/main.py", line 207, in <module>
    model.learn(total_timesteps=25_000)
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/a2c/a2c.py", line 200, in learn
    reset_num_timesteps=reset_num_timesteps,
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/common/on_policy_algorithm.py", line 243, in learn
    total_timesteps, eval_env, callback, eval_freq, n_eval_episodes, eval_log_path, reset_num_timesteps, tb_log_name
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/common/base_class.py", line 429, in _setup_learn
    self._last_obs = self.env.reset()  # pytype: disable=annotation-type-mismatch
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/common/vec_env/vec_transpose.py", line 110, in reset
    return self.transpose_observations(self.venv.reset())
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/common/vec_env/vec_frame_stack.py", line 58, in reset
    observation = self.venv.reset()  # pytype:disable=annotation-type-mismatch
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/common/vec_env/dummy_vec_env.py", line 61, in reset
    obs = self.envs[env_idx].reset()
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/gym/core.py", line 292, in reset
    return self.env.reset(**kwargs)
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/gym/core.py", line 333, in reset
    return self.env.reset(**kwargs)
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/gym/core.py", line 319, in reset
    observation = self.env.reset(**kwargs)
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/common/atari_wrappers.py", line 59, in reset
    self.env.reset(**kwargs)
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/common/atari_wrappers.py", line 106, in reset
    obs = self.env.reset(**kwargs)
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/common/atari_wrappers.py", line 154, in reset
    return self.env.reset(**kwargs)
  File "/home/x/anaconda3/envs/rl/lib/python3.7/site-packages/stable_baselines3/common/atari_wrappers.py", line 36, in reset
    noops = self.unwrapped.np_random.randint(1, self.noop_max + 1)
AttributeError: 'numpy.random._generator.Generator' object has no attribute 'randint'

Process finished with exit code 1

• I installed pytorch from official website, package:Conda
• Python = 3.7.13
• PyTorch = 1.11.0
• Gym version = 0.21.0
• Numpy= 1.21.5
• Stable Baseline = 1.5.0
  (from pip list)

[araffin]

Hello,
This is due to latest ale-py version, and fixed in 3c468ff or #780
This easiest fix for you is pip install ale-py==0.7.4 or pip install git+https://github.com/DLR-RM/stable-baselines3 (use master version of SB3)