Deep Reinforcement Learning Library

A modular deep RL library written from scratch in PyTorch, with PPO, SAC, DQN and actor–critic methods validated on discrete and continuous-control tasks.

Date
2026 – present
Organisation
Personal project
Technologies
  • PyTorch
  • Gymnasium
  • PPO
  • SAC
  • DQN
  • Actor–Critic

Technical approach

  • Algorithms: PPO, SAC, DQN and actor–critic methods in PyTorch.
  • Infrastructure: reusable training, evaluation and logging pipelines shared across algorithms.
  • Validation: discrete and continuous-control Gymnasium environments.

Status

In progress. Code and learning curves will be added here.