This page is a lightweight preview for exploring QS before installing G2S locally. The local G2S server remains the recommended path for serious work and is typically 5–10× faster in the current setup.
Choose Stone or Strebelle and run an unconditional QS simulation inside a Web Worker. Every run uses a new random seed and produces a result with the same dimensions as the selected training image.
Stone · continuous · the simulation will use the original image dimensions
Checking browser thread support…
Open multithreaded version ↗Ready.
The full-size result will appear here after the simulation completes.
Leave this preview open and use -sa browser to test the interface without installing the local server. For repeated or large simulations, use the normal local-server address: it remains the faster and recommended route.
# Python
simulation, index_map, duration, meta = g2s(
"-a", "qs", "-sa", "browser", "-ti", ti,
"-di", np.full((200, 200), np.nan, dtype=np.float32),
"-dt", [0], "-k", 1.2, "-n", 50, "-j", 4,
"-TO", 30000, "-returnMeta"
)Bridge ready; waiting for a local QS command.
MATLAB uses the same QS flags and '-sa','browser'. The page’s maximum-thread spinner caps the effective -j value. GPU, distributed, and asynchronous server operations are intentionally not part of browser mode.
The browser preview is useful for trying QS before installation. The local server remains the faster, full-featured path for production simulations, larger grids, GPU/distributed workflows, and repeated runs.
Read the G2S installation overview ↗