Cal-QL
Cal-QL copied to clipboard
nakamotoo
official implementation for our paper Cal-QL: Calibrated Offline RL Pre-Training for Efficient Online Fine-Tuning
Hi, thanks for your work! When I try cql in the pen binary environment, I find that for cql's value function always tend to diverge (tried mixing ratio 0.0 and...
Hi, Thanks for providing your code. My question is regarding the results in Appendix D. What hyperparameter configuration is used for these tasks? Thanks.
Hi @nakamotoo , Thank you for your amazing works. I am running your code. I had finished installing the environment based on the instruction in README. And when I run...
As inspecting through your codes, I found there is a function `cal_return_to_go` which requires a config dictionary for the high/low reward values for each env. What is its purpose and...
Hi Mitsuhiko, I have a few questions about replicating the repo as I am on MacOS. Not sure what to do with the "nvidia" or "cuda-nvcc". Add following environment variables...
