eiro-ecg-template/finetune_model.py at master · eiro-labs/eiro-ecg-template · 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
66
67
68
69
70
71
72
73
74
75
76
import numpy as np
import pandas as pd
import torch.nn as nn
import torch.nn.functional as F
import json
from net1d import Net1D

import torch.nn as nn
import torch


def ft_12lead_ECGFounder(device, pth, n_classes, linear_prob=False):
  model = Net1D(
      in_channels=12,
      base_filters=64, #32 64
      ratio=1,
      filter_list=[64,160,160,400,400,1024,1024],    #[16,32,32,80,80,256,256] [32,64,64,160,160,512,512] [64,160,160,400,400,1024,1024]
      m_blocks_list=[2,2,2,3,3,4,4],   #[2,2,2,2,2,2,2] [2,2,2,3,3,4,4]
      kernel_size=16,
      stride=2,
      groups_width=16,
      verbose=False,
      use_bn=False,
      use_do=False,
      n_classes=n_classes)

  checkpoint = torch.load(pth, map_location=device)
  state_dict = checkpoint['state_dict']

  state_dict = {k: v for k, v in state_dict.items() if not k.startswith('dense.')}

  model.load_state_dict(state_dict, strict=False)

  model.dense = nn.Linear(model.dense.in_features, n_classes).to(device)
  # freezing model
  if linear_prob == True:
    for name, param in model.named_parameters():
        if 'dense' not in name:  # no freezing last layer
            param.requires_grad = False

  model.to(device)

  return model


def ft_1lead_ECGFounder(device, pth, n_classes,linear_prob=False):
  model = Net1D(
      in_channels=1,
      base_filters=64, #32 64
      ratio=1,
      filter_list=[64,160,160,400,400,1024,1024],    #[16,32,32,80,80,256,256] [32,64,64,160,160,512,512] [64,160,160,400,400,1024,1024]
      m_blocks_list=[2,2,2,3,3,4,4],   #[2,2,2,2,2,2,2] [2,2,2,3,3,4,4]
      kernel_size=16,
      stride=2,
      groups_width=16,
      verbose=False,
      use_bn=False,
      use_do=False,
      n_classes=n_classes)

  checkpoint = torch.load(pth, map_location=device)
  state_dict = checkpoint['state_dict']

  state_dict = {k: v for k, v in state_dict.items() if not k.startswith('dense.')}

  model.load_state_dict(state_dict, strict=False)

  model.dense = nn.Linear(model.dense.in_features, n_classes).to(device)

  if linear_prob == True:
    for name, param in model.named_parameters():
        if 'dense' not in name:  # no freezing last layer
            param.requires_grad = False
  model.to(device)

  return model