Tutorials#

Tutorials connect the data, model, and evaluation guides into complete research tasks. Start by exploring a released checkpoint on PILArNet-M-mini, or complete the first experiment before training.

Explore Panda

Inspect one real test event and run the released encoder, semantic, particle, and interaction models with interactive Plotly figures.

Explore Panda
Explore PoLAr-MAE

Inspect semantic predictions, masked reconstruction, and token embeddings.

Explore PoLAr-MAE
Train semantic segmentation

Adapt the verified tiny pipeline into a per-point classification experiment.

Semantic segmentation
Parameter-efficient fine-tuning

Freeze a base model, inject LoRA into selected projections, and verify the trainable set.

Parameter-efficient fine-tuning