Evaluated return of the pre-trained model does not match the reported return

chenhongge / SA_DQN

[NeurIPS 2020, Spotlight] State-Adversarial DQN (SA-DQN) for robust deep reinforcement learning

https://arxiv.org/pdf/2003.08938.pdf

31 stars 11 forks source link

Evaluated return of the pre-trained model does not match the reported return #1

Open YaqiXie opened 2 years ago

YaqiXie commented 2 years ago

I'm trying to evaluate the pre-trained RoadRunner model. The return I got from running python test.py --config config/RoadRunner_cov.json test_config:load_model_path=models/RoadRunner-convex.model is around one thousand (as shown in the figure below), which is far from the reported 44638.0±7367.0. I'm wondering what are the causes. Is there any updated environment or package that affects the performance?

Screenshot 2022-09-01 at 9 32 47 PM

i-hu commented 1 year ago

hi，how did you solve it？