Skip to content

Latest commit

 

History

History
16 lines (15 loc) · 446 Bytes

README.md

File metadata and controls

16 lines (15 loc) · 446 Bytes
Raw

prompt_lstm-decision_transformer

Code for Offline Prompt Reinforcement Learning method Based on Feature Extraction

run

  1. Download D4RL datasets： http://rail.eecs.berkeley.edu/datasets/offline_rl/gym_mujoco_v2/ or run data/download_d4rl_datasets.py
  2. Set running tasks and related parameters for example:( --env=halfcheetah --dataset_mode=expert --prompt-length=5 --K=20 --no-propmt=False )
  3. run pldt.py