cpsc2020/saved_models/__init__.py at 676b1ee616ae0c69522c841550c7a1a279793198 · DeepPSP/cpsc2020 · GitHub

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
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
""" """

import os
from typing import Optional, Tuple, Union

import torch
from torch import nn

from cfg import ModelCfg
from model import ECG_CRNN_CPSC2020, ECG_SEQ_LAB_NET_CPSC2020

_BASE_DIR = os.path.dirname(os.path.abspath(__file__))


__all__ = [
    "load_model",
]


def load_model(which: str = "both") -> Union[nn.Module, Tuple[nn.Module, ...]]:
    """finished, checked,

    Parameters:
    -----------
    which: str,
        choice of the models

    Returns:
    --------
    nn.Module, or sequence of nn.Module
    """
    if torch.cuda.is_available():
        device = torch.device("cuda")
    else:
        device = torch.device("cpu")
    _which = which.lower()
    if _which in ["both", "crnn"]:
        crnn_cfg = ModelCfg.crnn
        crnn_model = ECG_CRNN_CPSC2020(
            classes=crnn_cfg.classes,
            n_leads=crnn_cfg.n_leads,
            input_len=4000,
            config=crnn_cfg,
        )
        crnn_state_dict = torch.load(os.path.join(_BASE_DIR, "crnn_10s.pth"), map_location=device)
        crnn_state_dict["clf.lin_0.weight"] = crnn_state_dict.pop("clf.weight")
        crnn_state_dict["clf.lin_0.bias"] = crnn_state_dict.pop("clf.bias")
        crnn_model.load_state_dict(crnn_state_dict)
        crnn_model.eval()
        if _which == "crnn":
            return crnn_model
    if _which in ["both", "seq_lab"]:
        seq_lab_cfg = ModelCfg.seq_lab
        seq_lab_model = ECG_SEQ_LAB_NET_CPSC2020(
            classes=seq_lab_cfg.classes,
            n_leads=seq_lab_cfg.n_leads,
            input_len=4000,
            config=seq_lab_cfg,
        )
        seq_lab_state_dict = torch.load(os.path.join(_BASE_DIR, "seq_lab_10s.pth"), map_location=device)
        seq_lab_model.load_state_dict(seq_lab_state_dict)
        seq_lab_model.eval()
        if _which == "seq_lab":
            return seq_lab_model
    return crnn_model, seq_lab_model