Download code/ampd/models/registry.py from fryabcfry/cs4486-amp-pd-checkpoints: direct link, hf CLI and curl.
- Browser
- Download file 1.15 kB
-
https://huggingface.co/fryabcfry/cs4486-amp-pd-checkpoints/resolve/main/code/ampd/models/registry.py
- Command line
-
hf download hf://fryabcfry/cs4486-amp-pd-checkpoints/code/ampd/models/registry.py
-
curl -L -o registry.py https://huggingface.co/fryabcfry/cs4486-amp-pd-checkpoints/resolve/main/code/ampd/models/registry.py
1.15 kB
| """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) | |