Skip to content
LayerDynamics — LatticeLabs Toolkit LayerDynamics — LatticeLabs Toolkit LatticeLabs Toolkit

ll_gen — Installation

Terminal window
pip install -e ./ll_gen

What each path needs

Section titled “What each path needs”

ll_gen has two proposal paths and a shared dispose stage:

  • Dispose (always) — executes proposals in a sandboxed CadQuery subprocess. CadQuery (and its OpenCASCADE backend) is needed to turn proposals into real geometry and validate them.
  • Neural path — needs PyTorch and the stepnet package, whose STEPVAE, StructuredDiffusion, VQVAEModel, and CADGenerationPipeline the neural generators drive. Install ll_stepnet too.
  • Code path — proposes CadQuery/OpenSCAD code; an LLM proposer is optional.

Verify

Section titled “Verify”
from ll_gen import GenerationOrchestrator


orch = GenerationOrchestrator()   # uses default LLGenConfig
print(type(orch).__name__)

Next

Section titled “Next”

Continue to Usage.