search

rll / rllab

rllab is a framework for developing and evaluating reinforcement learning algorithms, fully compatible with OpenAI Gym.
Other
2.91k stars 799 forks source link

example trpo_cartpole_recurrent.py does not work #189

Open lchenat opened 7 years ago

lchenat commented 7 years ago

when I ran examples/trpo_cartpole_recurrent.py I encountered the following errors:

Traceback (most recent call last): File "/home//Documents/rllab-master/scripts/run_experiment_lite.py", line 137, in run_experiment(sys.argv) File "/home//Documents/rllab-master/scripts/run_experiment_lite.py", line 121, in run_experiment method_call(variant_data) File "trpo_cartpole_recurrent.py", line 14, in run_task env_spec=env.spec, File "/home//Documents/rllab-master/rllab/policies/gaussian_gru_policy.py", line 50, in init output_nonlinearity=output_nonlinearity, File "/home//Documents/rllab-master/rllab/core/network.py", line 228, in init b=l_output_flat.b, File "/home/*/anaconda2/envs/py36/lib/python3.6/site-packages/lasagne/layers/dense.py", line 78, in init super(DenseLayer, self).init(incoming, *kwargs) File "/home//anaconda2/envs/py36/lib/python3.6/site-packages/lasagne/layers/base.py", line 44, in init if any(d is not None and d <= 0 for d in self.input_shape): File "/home/***/anaconda2/envs/py36/lib/python3.6/site-packages/lasagne/layers/base.py", line 44, in if any(d is not None and d <= 0 for d in self.input_shape): TypeError: '<=' not supported between instances of 'tuple' and 'int'

ryanjulian commented 6 years ago

Did you install the conda env directly from the environment.yml? It seems that your env uses python 3.6 (environment.yml specifies Python 3.5).