We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
See https://github.com/DLR-RM/stable-baselines3/pull/53/checks?check_run_id=758580860 and https://gitlab.com/araffin/stable-baselines3/-/jobs/589877156
Since #50 was merged.
Traceback:
2020-06-10T16:29:13.7480849Z =================================== FAILURES =================================== 2020-06-10T16:29:13.7482359Z ___________________________ test_sync_vec_normalize ____________________________ 2020-06-10T16:29:13.7482761Z 2020-06-10T16:29:13.7484480Z def test_sync_vec_normalize(): 2020-06-10T16:29:13.7484937Z env = DummyVecEnv([make_env]) 2020-06-10T16:29:13.7485318Z 2020-06-10T16:29:13.7485746Z assert unwrap_vec_normalize(env) is None 2020-06-10T16:29:13.7486142Z 2020-06-10T16:29:13.7486588Z env = VecNormalize(env, norm_obs=True, norm_reward=True, clip_obs=10., clip_reward=10.) 2020-06-10T16:29:13.7487019Z 2020-06-10T16:29:13.7487477Z assert isinstance(unwrap_vec_normalize(env), VecNormalize) 2020-06-10T16:29:13.7488198Z 2020-06-10T16:29:13.7488624Z env = VecFrameStack(env, 1) 2020-06-10T16:29:13.7489719Z 2020-06-10T16:29:13.7490133Z assert isinstance(unwrap_vec_normalize(env), VecNormalize) 2020-06-10T16:29:13.7490566Z 2020-06-10T16:29:13.7490939Z eval_env = DummyVecEnv([make_env]) 2020-06-10T16:29:13.7491374Z eval_env = VecNormalize(eval_env, training=False, norm_obs=True, norm_reward=True, clip_obs=10., clip_reward=10.) 2020-06-10T16:29:13.7491818Z eval_env = VecFrameStack(eval_env, 1) 2020-06-10T16:29:13.7492237Z 2020-06-10T16:29:13.7494143Z env.reset() 2020-06-10T16:29:13.7494637Z # Initialize running mean 2020-06-10T16:29:13.7495040Z latest_reward = None 2020-06-10T16:29:13.7495440Z for _ in range(100): 2020-06-10T16:29:13.7495931Z _, latest_reward, _, _ = env.step([env.action_space.sample()]) 2020-06-10T16:29:13.7496345Z 2020-06-10T16:29:13.7496769Z # Check that unnormalized reward is same as original reward 2020-06-10T16:29:13.7497213Z original_latest_reward = env.get_original_reward() 2020-06-10T16:29:13.7497681Z assert np.allclose(original_latest_reward, env.unnormalize_reward(latest_reward)) 2020-06-10T16:29:13.7498090Z 2020-06-10T16:29:13.7498462Z obs = env.reset() 2020-06-10T16:29:13.7498870Z dummy_rewards = np.random.rand(10) 2020-06-10T16:29:13.7499270Z original_obs = env.get_original_obs() 2020-06-10T16:29:13.7499712Z # Check that unnormalization works 2020-06-10T16:29:13.7500113Z > assert np.allclose(original_obs, env.unnormalize_obs(obs)) 2020-06-10T16:29:13.7500514Z E assert False 2020-06-10T16:29:13.7501878Z E + where False = <function allclose at 0x7fc839db05f0>(array([[ 0.24142723, 0.97041893, -0.96644473]], dtype=float32), array([[-0.48370542, 0.9704189 , -0.96644476]])) 2020-06-10T16:29:13.7502405Z E + where <function allclose at 0x7fc839db05f0> = np.allclose 2020-06-10T16:29:13.7503978Z E + and array([[-0.48370542, 0.9704189 , -0.96644476]]) = <bound method VecNormalize.unnormalize_obs of <stable_baselines3.common.vec_env.vec_normalize.VecNormalize object at 0x7fc814804d50>>(array([[10. , 3.0962322 , -0.74936515]], dtype=float32)) 2020-06-10T16:29:13.7505725Z E + where <bound method VecNormalize.unnormalize_obs of <stable_baselines3.common.vec_env.vec_normalize.VecNormalize object at 0x7fc814804d50>> = <stable_baselines3.common.vec_env.vec_frame_stack.VecFrameStack object at 0x7fc814804450>.unnormalize_obs 2020-06-10T16:29:13.7506161Z 2020-06-10T16:29:13.7506578Z tests/test_vec_normalize.py:167: AssertionError
The text was updated successfully, but these errors were encountered:
Fixes #54
3bdfdb3
unnormalization may not always work because of the clipping, fixed in #53
Sorry, something went wrong.
Hotfix PPO + gSDE (#53)
494ebfd
* Fix variable being passed with gradients * Update changelog * Bump version * Fixes #54
Miffyli
Successfully merging a pull request may close this issue.
See https://github.com/DLR-RM/stable-baselines3/pull/53/checks?check_run_id=758580860
and https://gitlab.com/araffin/stable-baselines3/-/jobs/589877156
Since #50 was merged.
Traceback:
The text was updated successfully, but these errors were encountered: