This code is based on this repo, we add the splitting part and test it on HALF-Cheetah-v2 enviornment.
To run the small size baseline with seed i, run
bash run1.sh i
To run the splitting method with seed i, run
bash run2.sh i
To run the large size baseline with seed i, run
bash run3.sh i
To run the random splitting baseline with seed i, run
bash run44.sh i
i here could be any integer, the best result are seed = 1 in our report.