xnn.common

Shared abstractions used across all model families (cnn / dnn / gnn).

Modules

benchmark	Benchmark several pre-trained models on one dataset and tabulate their errors.
cli
config
data
deploy
distributed	Process-group helpers shared by training and benchmarking.
featurizers	Featurizer base + basis functions shared across families.
finetune	Fine-tuning pretrained potentials: multi-head replay, LoRA, references.
models
train