Separate Runtime + AIX Directory
Deep RL Lab
DRL stays a separate sister project with its own standalone runtime. AIX gives it a stable entry page, a readable table of contents, and one place to explain how the independent DRL site fits into the larger lab ecosystem.
Local repo: /drl (not present in this runtime)
Why DRL Stays Separate
- DRL already has its own app shape, runtime assumptions, and deployment scripts.
- AIX is App Engine multi-service right now; DRL is already shaped for a separate deployment path.
- Linking cleanly is cheaper and more reliable than trying to double-run DRL inside AIX.
Table Of Contents
This is the AIX-facing guide to the DRL app’s major pages and branches.
Core
| Preview | Route | Path | Launch |
|---|---|---|---|
| Overview Main landing page for the DRL sister app. | / |
Open in DRL | |
| Inventory Repository-wide inventory and recovery map. | /inventory |
Open in DRL |
Interactive
| Preview | Route | Path | Launch |
|---|---|---|---|
| Foundations Demo Interactive warm-up demo for the foundational RL arm. | /demos/foundations |
Open in DRL | |
| Finance Demo Interactive finance execution demo. | /demos/finance |
Open in DRL | |
| Lunar Lander Live Lunar runtime, checkpoints, and training jobs. | /lunar |
Open in DRL | |
| Grabber Live continuous-control arm lab with PPO playback and training snapshots. | /grabber |
Open in DRL |
Arms
| Preview | Route | Path | Launch |
|---|---|---|---|
| Foundations Tabular RL, dynamic programming, MC, TD, and early labs. | /sections/foundations |
Open in DRL | |
| Value-Based DQN lineage, navigation, and Lunar variants. | /sections/value-based |
Open in DRL | |
| Policy Gradients REINFORCE, PPO, actor-critic, and Pong-era material. | /sections/policy-gradients |
Open in DRL | |
| Continuous Control Reacher, DDPG, and Unity-heavy recovery work. | /sections/continuous-control |
Open in DRL | |
| Multi-Agent Tennis, Soccer, MARL notes, and theory material. | /sections/multi-agent |
Open in DRL | |
| Finance Domain application branch for RL-guided execution. | /sections/finance |
Open in DRL |
Archive
| Preview | Route | Path | Launch |
|---|---|---|---|
| Papers Reference shelf for the educational side of DRL. | /sections/papers |
Open in DRL | |
| Archive Legacy branches kept visible but not yet productized. | /sections/archive |
Open in DRL |
Recommended Entry Path
- Start with
/inventoryto see what material exists. - Use the two demos to understand the style of the app quickly.
- Then move into the long-form arms: foundations, value-based, continuous control, and multi-agent.
- Use Lunar only when you want the live runtime side rather than the archive/review side.