HLM-Spatial

HLM-Spatial applies polynomial Hopfield networks to 3D perception tasks. The same Energy Language operations that edit language models work on spatial models.

Status

HLM-Spatial is currently training at large scale. Join the waiting list for early access.

Domains

• LIDAR — point cloud processing for autonomous systems
• Medical3D — volumetric medical imaging analysis
• Industrial3D — equipment inspection and defect detection

Surgery on Spatial Models

Spatial models store geometric patterns as attractor basins. Surgery operations modify how the model recognizes and classifies 3D structures:

surgeon = BasinSurgeon.from_checkpoint("hlm-spatial.pt")

# Survey what the model has learned
survey = surgeon.survey(layer=0)

# Remove a misclassification pattern
surgeon.remove(layer=3, seed=17, strength=0.1)

# Inject a new geometric pattern
surgeon.inject(layer=3, seed=42, strength=0.1)

surgeon.apply(layer=3)

Custom Training & Pilots

For industrial, medical, or autonomous vehicle applications requiring custom spatial model training, contact us.