Policy Gradient Methods and Behaviour Cloning on CartPole-v1

HW1 — Research in LLMs, HSE × Central University

📄 Read the report

Reproduce

pip install -r requirements.txt

# Core experiments (5 seeds, ~7 min on M2)
cd src && python run_experiments.py --full

# Extended experiments (15 seeds + RLOO sweep + BC noise, ~20 min)
python additional_experiments.py --all

# Statistical analysis + extra figures
python statistical_analysis.py
python plotting.py
python plotting_extra.py