الانتقال إلى المحتوى الرئيسي

RL Playground

Reinforcement learning playground — train agents in grid worlds with policy visualization

App ID rl-playground Chapter AI · Signal · Quantum Live https://zeq.dev/apps/rl-playground/ Local https://zeq.dev/apps/rl-playground/ Precision ≤0.1%

What it does​

Reinforcement learning playground — train agents in grid worlds with policy visualization. Built on the Zeq OS kernel — every simulation step is phase-locked to the 1.287 Hz HulyaPulse and averaged across a 0.777 s Zeqond window, so results are reproducible to ≤0.1% error regardless of wall-clock jitter on the host.

Run it​

Open the live instance: https://zeq.dev/apps/rl-playground/

Or run locally from a cloned api-server:

cd app/artifacts/api-server
pnpm dev                          # starts on :3010
open https://zeq.dev/apps/rl-playground/

Integrate​

  • Embed — the app is a static bundle under public/apps/rl-playground/. Drop the folder behind any reverse proxy or iframe it directly.
  • Drive via API — the same compute calls the app makes are public. See the API surface and filter for operators this app uses.
  • Compose — chain this app's outputs into another app by forwarding the compute result's zeq_resonance_objects array through POST /api/playground/compute.

Reference​

Middleware active. Kernel on the 1.287 Hz HulyaPulse. Awaiting next Zeqond.