random-network-distillation

Random Network Distillation pytorch

A pytorch tutorial for DRL(Deep Reinforcement Learning)

Official implementation for "Anti-Exploration by Random Network Distillation", ICML 2023

Deep Reinforcement Learning by using Proximal Policy Optimization and Random Network Distillation in Tensorflow 2 and Pytorch with some explanation

Random network distillation on Montezuma's Revenge and Super Mario Bros.

Deep Potential Evolution Accelerator

My reproduction of various reinforcement learning algorithms (DQN variants, A3C, DPPO, RND with PPO) in Tensorflow.

Deep Reinforcement Learning Algorithms Implementation in PyTorch

Using Proximal Policy Optimization and Random Network Distillation on Pommerman

Combining Experience Replay with Exploration by Random Network Distillation

Exploration methods in Deep Reinforcement Learning.

Algorithm RND-PPO is implemented in super mario bros.

A minimal, GitHub-ready example project that generates a random network (graph). Designed for quick demos, testing pipelines, or starter repositories.

Source code for Master's Thesis: Curiosity-driven Planning with Reinforcement Learning.

Code for an AIMS South Africa master's essay (2019-2020) on curiosity in reinforcement learning: a Deep Q-Network baseline and DQN combined with Random Network Distillation (RND) on OpenAI Gym's MountainCar-v0. Jupyter notebooks with TensorFlow/Keras and PyTorch.

Bachelor's thesis code: does a more accurate world model always help a curious RL agent explore? PPO with ICM and RND curiosity modules in PyTorch, tested on sparse-reward MiniGrid/BabyAI tasks over 59 runs. Result: more accuracy helps, but with diminishing returns.

Visual reinforcement learning on Super Mario Galaxy 1/2 in a patched Dolphin: IQN with an IMPALA encoder sees stacked 84x84 grayscale frames; RAM telemetry drives only reward, curriculum and evaluation. Backward human-anchor curriculum with savestate-search demonstrations; no learned full-start star yet.