-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathtrain_basic.py
169 lines (128 loc) · 8.59 KB
/
train_basic.py
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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
"""
Main training script to train a Basic UNETR/nnUNet with MMWHS Challenge dataset for CT/MR and MSD-BraTS
"""
# pyright: reportPrivateImportUsage=false
from csv import writer
import logging, os, torch
from pyrsistent import b
from typing import Union
from monai.data.dataloader import DataLoader
from monai.metrics import DiceMetric, HausdorffDistanceMetric, SurfaceDistanceMetric
from monai.data.utils import pad_list_data_collate
from monai.losses.dice import DiceCELoss
import data
from configs import TrainingConfig
from torchmanager_monai import Manager, metrics
from networks import SelfDistillUNETRWithDictOutput as SelfDistilUNETR
from networks import SelfDistillnnUNetWithDictOutput as SelfDistilnnUNet
from networks import SelfDistillSwinUNETRWithDictOutput as SelfDistilSwinUNETR
from torchmanager import callbacks, losses
from torchmanager_core import random
from torch.backends import cudnn
from utils import count_parameters
# initialization
seed = 100
random.freeze_seed(seed)
cudnn.benchmark = False
cudnn.deterministic = True
if __name__ == "__main__":
# get configurations
config = TrainingConfig.from_arguments()
cudnn.benchmark = True
if config.show_verbose: config.show_settings()
# initialize checkpoint and data dirs
data_dir = os.path.join(config.experiment_dir, "data")
best_ckpt_dir = os.path.join(config.experiment_dir, "best.model")
last_ckpt_dir = os.path.join(config.experiment_dir, "last.model")
# # load dataset - Load MMWHS Challenge Data
# in_channels = 1
# training_dataset, validation_dataset, num_classes = data.load_challenge(config.data, config.img_size, train_split=config.training_split, show_verbose=config.show_verbose) # type:ignore
# load dataset - Load MSD-BraTS Data
in_channels = 4
training_dataset, validation_dataset, _, num_classes = data.load_msd(config.data, config.img_size, train_split=config.training_split, show_verbose=config.show_verbose)
training_dataset = DataLoader(training_dataset, batch_size=config.batch_size, shuffle=True, collate_fn=pad_list_data_collate)
validation_dataset = DataLoader(validation_dataset, batch_size=1, collate_fn=pad_list_data_collate)
##########################################################################################################
## Initialize the UNETR model
# model = SelfDistilUNETR(in_channels, num_classes, img_size=config.img_size, feature_size=16, hidden_size=768, mlp_dim=3072, num_heads=12, pos_embed="perceptron", norm_name="instance", res_block=True, dropout_rate=0.0) # for MSD-BraTS and MMWHS(MR/CT)
# model = SelfDistilUNETR(in_channels, num_classes, img_size=config.img_size, self_distillation=False, feature_size=32, hidden_size=768, mlp_dim=3072, num_heads=12, pos_embed="perceptron", norm_name="instance", res_block=True, dropout_rate=0.0) # MMWHS CT only
# model = SelfDistilUNETR(in_channels, num_classes, img_size=config.img_size, self_distillation=False, feature_size=16, hidden_size=768, mlp_dim=3072, num_heads=12, pos_embed="perceptron", norm_name="instance", res_block=True, dropout_rate=0.0) # MMWHS CT only Ablation and MSD-BraTS
##########################################################################################################
## Initialize the nnUNet model
# kernel_size = [[3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3], [3, 3, 3]] # input + 3 Enc-Dec Layers + Bottleneck
# strides = [[1, 1, 1], [2, 2, 2], [2, 2, 2], [2, 2, 2], [2, 2, 2]] # input + 3 Enc-Dec Layers + Bottleneck
# # filters = [32,64,128,256,320] # originally used for MMWHS
# filters = [16,32,64,128,256] # for MSD-BraTS due to memory limitations
# model = SelfDistilnnUNet(
# spatial_dims = 3,
# in_channels = in_channels,
# out_channels = num_classes,
# kernel_size = kernel_size,
# strides = strides,
# upsample_kernel_size = strides[1:],
# filters=filters,
# norm_name="instance",
# deep_supervision=False,
# deep_supr_num=3,
# self_distillation=False,
# self_distillation_num=4,
# res_block=True
# )
##########################################################################################################
## Initialize the SwinUNETR model
# model = SelfDistilSwinUNETR(img_size=config.img_size, in_channels=in_channels, out_channels=num_classes, feature_size=36, self_distillation=False) # MMWHS CT only
model = SelfDistilSwinUNETR(img_size=config.img_size, in_channels=in_channels, out_channels=num_classes, feature_size=12, self_distillation=False) # MSD-BraTS
##########################################################################################################
## Count model parameters
print(f'The total number of model parameter is: {count_parameters(model)}')
##########################################################################################################
# initialize optimizer, loss, metrics, and post processing
optimizer = torch.optim.AdamW(model.parameters(), lr=1e-3, weight_decay=1e-5) # lr used by MMWHS challenge winner/MSD-BraTS
# # initialize learning rate scheduler (lr used by MMWHS challenge winner)
lr_step = max(int(config.epochs / 6), 1) # for nnUNet and UNETR (MMWHS-CT)
lr_scheduler = torch.optim.lr_scheduler.StepLR(optimizer, lr_step, gamma=0.5) # used for MMWHS
# optimizer = torch.optim.AdamW(model.parameters(), lr=1e-4, weight_decay=1e-5) # lr=0.0001 # for MSD-BraTS
# Initilialize Loss Functions
loss_fn: Union[losses.Loss, dict[str, losses.Loss]]
## Basic UNETR/nnUNet ONLY
loss_dice = losses.Loss(DiceCELoss(include_background=True, to_onehot_y=True, softmax=True), target="out") # out_main and GT labels
loss_fn = {
"dice": loss_dice
}
# Initialize Metrics for evaluation
dice_fn = metrics.CumulativeIterationMetric(DiceMetric(include_background=False, reduction="none", get_not_nans=False), target="out")
hd_fn = metrics.CumulativeIterationMetric(HausdorffDistanceMetric(include_background=False, percentile=95.0, reduction="none", get_not_nans=False), target="out")
msd_fn = metrics.CumulativeIterationMetric(SurfaceDistanceMetric(include_background=False, reduction="none", get_not_nans=False), target="out")
metric_fns: dict[str, metrics.Metric] = {
"val_dice": dice_fn,
"val_hd": hd_fn,
"val_msd": msd_fn,
}
post_labels = data.transforms.AsDiscrete(to_onehot=num_classes)
post_predicts = data.transforms.AsDiscrete(argmax=True, to_onehot=num_classes)
# compile manager
manager = Manager(model, post_labels=post_labels, post_predicts=post_predicts, optimizer=optimizer, loss_fn=loss_fn, metrics=metric_fns, roi_size=config.img_size) # type: ignore
## All callbacks defined below
# initialize callbacks
tensorboard_callback = callbacks.TensorBoard(data_dir)
last_ckpt_callback = callbacks.LastCheckpoint(manager, last_ckpt_dir)
besti_ckpt_callback = callbacks.BestCheckpoint("dice", manager, best_ckpt_dir)
lr_scheduler_callback = callbacks.LrSchedueler(lr_scheduler, tf_board_writer=tensorboard_callback.writer)
# Final callbacks list
callbacks_list: list[callbacks.Callback] = [tensorboard_callback, besti_ckpt_callback, last_ckpt_callback, lr_scheduler_callback]
# callbacks_list: list[callbacks.Callback] = [tensorboard_callback, besti_ckpt_callback, last_ckpt_callback]
# train
manager.fit(training_dataset, config.epochs, val_dataset=validation_dataset, device=config.device, use_multi_gpus=config.use_multi_gpus, callbacks_list=callbacks_list, show_verbose=config.show_verbose)
# save and test with last model
model = manager.model
torch.save(model, config.output_model)
summary = manager.test(validation_dataset, device=config.device, use_multi_gpus=config.use_multi_gpus, show_verbose=config.show_verbose)
logging.info(summary)
# save and test with best model on validation dataset
manager = Manager.from_checkpoint("experiments/CT_MMWHS_UNETR_Basic_filters16_Ablation_Fold5.exp/best.model")
if isinstance(manager.model, torch.nn.parallel.DataParallel): model = manager.model.module
else: model = manager.model
manager.model = model
print(f'The best Dice score on validation set occurs at {manager.current_epoch + 1} epoch number')
summary = manager.test(validation_dataset, device=config.device, use_multi_gpus=config.use_multi_gpus, show_verbose=config.show_verbose)
logging.info(summary)