File size: 1,146 Bytes
04ae9bb | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 | """Registry for pluggable model architectures."""
from __future__ import annotations
from copy import deepcopy
from typing import Any, Dict, Type
import torch.nn as nn
MODEL_REGISTRY: Dict[str, Type[nn.Module]] = {}
def register_model(name: str):
def deco(cls: Type[nn.Module]):
if name in MODEL_REGISTRY:
raise ValueError(f"Model already registered: {name}")
MODEL_REGISTRY[name] = cls
cls.model_name = name
return cls
return deco
def build_model(cfg: Dict[str, Any], input_dim: int, output_dim: int) -> nn.Module:
name = cfg["name"]
if name not in MODEL_REGISTRY:
raise KeyError(f"Unknown model '{name}'. Available: {sorted(MODEL_REGISTRY)}")
params = deepcopy(cfg.get("params", {}))
params.pop("input_dim", None)
params.pop("output_dim", None)
params.pop("output_type", None)
output_type = cfg.get("output_type", "regression")
return MODEL_REGISTRY[name](
input_dim=input_dim,
output_dim=output_dim,
output_type=output_type,
**params,
)
def list_models() -> list[str]:
return sorted(MODEL_REGISTRY)
|