Xingyuan Zhang

icon indicating a website link https://icaruswizard.github.io/ icon indicating an email [email protected]

icon company Machine Learning Research Lab, Volkswagen AG icon location Munich, Germany icon location An intelligent machine who wants to understand himself. Doing PhD in world models.

Results 4 comments of Xingyuan Zhang

Experimental results of MoREL for D4RL benchmarks

Hi guys. I just go through the code, and I find the setup of NumPy seed is missing in the `learn_model.py` script which will cause the results to be unreproducible.

Why MOPO is given access to terminal function in rollout generation?

I agree with your point. In real-world scenarios, the reward function and the terminal function are available in the most cases (MDP settings and some POMDP settings). I guess future...

[bug fix] make TanhNormal.mode a property

LGTM. cc @mzktbyjc2016 for merging.

Discrepancy between results reported in CQL and D4rl papers

Hi, I just cross check the CQL scores reported in D4RL (arXiv-v4) and CQL (NeurIPS) papers, there are few mismatches. |Task|D4RL (arXiv-v4)|CQL (NeurIPS)| |------|--------|--------------| |walker2d-medium|79.2|74.5| |hopper-medium|58.0|86.6| |walker2d-medium-replay|26.7|32.6| I hope you...