Minigrid --" Kernel size can't be greater than actual input size" for DQN

Getting Error: RuntimeError: Calculated padded input size per channel: (3 x 3). Kernel size: (4 x 4). Kernel size can’t be greater than actual input size

`import gym import gym_minigrid from stable_baselines3.common.env_checker import check_env from stable_baselines3 import PPO, A2C from stable_baselines3 import DQN

env = gym.make(‘MiniGrid-Empty-16x16-v0’) check_env(env, warn=True) env = make_vec_env(lambda: env, n_envs=1) model = DQN(“MultiInputPolicy”, env) `

Slight modifications from the original minigrid.py: changed the observation to only the image. tldr; obs = dict(‘image’: array(16,16,3))

I am not quite sure how I should fix this if it’s the layer of NN within the baseline model? Thanks

[Miffyli]

Next time, please fill in the issue template 😃.

Problem is that with conv layers the output size, which gets smaller on every conv layer, becomes negative. You need to either increase the input shape size or use a different feature extractor (see doc on making custom policies). Same applies to your SB2 issue (please close that one, we can answer questions here).

[dyllanwli]

Solved by using env_wrapper.