Tutorials

• Getting Started
  • Install
  • Prepare Dataset
  • Setup Algorithm
  • Setup Metrics
  • Start Training
  • Save and Load
• Data Collection
  • Prepare Environment
  • Data Collection with Random Policy
  • Data Collection with Trained Policy
  • Data Collection while Training Policy
• Create Your Dataset
  • Prepare Logged Data
  • Build MDPDataset
  • Set Timeout Flags
• Preprocess / Postprocess
  • Preprocess Observations
  • Preprocess / Postprocess Actions
  • Preprocess Rewards
• Customize Neural Network
  • Prepare PyTorch Model
  • Setup EncoderFactory
  • Support Q-function for Actor-Critic
  • Make your models loadable
• Online RL
  • Prepare Environment
  • Setup Algorithm
  • Setup Online RL Utilities
  • Start Training
  • Train with Stochastic Policy
• Finetuning
  • Prepare Dataset and Environment
  • Pretrain with Dataset
  • Finetune with Environment
  • Finetune with Saved Policy
  • Finetune with Different Algorithm
• Offline Policy Selection
  • Prepare trained policies
  • Train FQE with the trained policies
• Use Distributional Q-Function
• After Training Policies (Save and Load)
  • Prepare Pretrained Policies
  • Load Trained Policies
  • Inference
  • Export Policies as TorchScript
  • Export Policies as ONNX