SURF-2025

Unsupervised Exploration

Classic Control

Run the whole script until you see title "Integrated", that's where the exploration and offline training begins
You can follow the interactive structure for parameter tuning in IQL which is not fully customizable now
The notebook should demonstrate the expected output already for you to follow, remember to operate on your own copy
Use wandb to monitor

code flow: preparation -> Exploration(parameter tuning)(out & few graphs) -> IQL(parameter tuning)(monitored by wandb)

Name		Name	Last commit message	Last commit date
Latest commit History 50 Commits
study-materials		study-materials
.DS_Store		.DS_Store
CCGymnax.ipynb		CCGymnax.ipynb
CartPole-v1_MountainCar_Heatmap_DisagreementExplorer_TOTAL_TIMESTEPS_16384_DEPTH_1_NUM_HIDDEN_128.pdf		CartPole-v1_MountainCar_Heatmap_DisagreementExplorer_TOTAL_TIMESTEPS_16384_DEPTH_1_NUM_HIDDEN_128.pdf
CheckList.md		CheckList.md
Dis.pdf		Dis.pdf
ME.pdf		ME.pdf
PPO.pdf		PPO.pdf
Pointax.ipynb		Pointax.ipynb
README.md		README.md
TQ.pdf		TQ.pdf
Xland.ipynb		Xland.ipynb
jax_tutorial.ipynb		jax_tutorial.ipynb
poster.md		poster.md