-
Notifications
You must be signed in to change notification settings - Fork 5
/
main.py
160 lines (143 loc) · 4.37 KB
/
main.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
"""
Ke Chen knutchen@ucsd.edu
Tone-Octave Network - main file
This file contains the main script
"""
import os
import random
import numpy as np
import argparse
import torch
from torch import nn
import torch.nn.functional as F
from torch.utils.data import DataLoader
import pytorch_lightning as pl
import config
from data_generator import TONetTrainDataset, TONetTestDataset
from model.msnet import MSnet
from model.tonet import TONet
from model.multi_dr import MLDRnet
from model.ftanet import FTAnet
from model.mcdnn import MCDNN
from util import tonpy_fn
def train():
train_dataset = TONetTrainDataset(
data_list = config.train_file,
config = config
)
train_dataloader = DataLoader(
dataset = train_dataset,
shuffle = True,
num_workers = config.n_workers,
batch_size = config.batch_size
)
test_datasets = [
TONetTestDataset(
data_list = d,
config = config
) for d in config.test_file
]
test_dataloaders = [
DataLoader(
dataset = d,
shuffle = False,
batch_size = 1,
collate_fn=tonpy_fn
) for d in test_datasets
]
loss_func = nn.BCELoss()
if config.model_type == "MCDNN":
me_model = MCDNN()
me_model_r = MCDNN()
elif config.model_type == "MLDRNet":
me_model = MLDRnet()
me_model_r = MLDRnet()
elif config.model_type == "FTANet":
me_model = FTAnet(freq_bin = config.freq_bin, time_segment=config.seg_frame)
me_model_r = FTAnet(freq_bin = config.freq_bin, time_segment=config.seg_frame)
else: # MSNet
me_model = MSnet()
me_model_r = MSnet()
if config.ablation_mode == "single" or config.ablation_mode == "spl" or config.ablation_mode == "spat":
me_model_r = None
model = TONet(
l_model = me_model,
r_model = me_model_r,
config = config,
loss_func = loss_func,
mode = config.ablation_mode
)
trainer = pl.Trainer(
deterministic = True,
gpus = 1,
checkpoint_callback = False,
max_epochs = config.max_epoch,
auto_lr_find = True,
sync_batchnorm=True,
# check_val_every_n_epoch = 1,
val_check_interval = 0.25,
# num_sanity_val_steps=0
)
trainer.fit(model, train_dataloader, test_dataloaders)
def test():
test_datasets = [
TONetTestDataset(
data_list = d,
config = config
) for d in config.test_file
]
test_dataloaders = [
DataLoader(
dataset = d,
shuffle = False,
batch_size = 1,
collate_fn=tonpy_fn
) for d in test_datasets
]
loss_func = nn.BCELoss()
if config.model_type == "MCDNN":
me_model = MCDNN()
me_model_r = MCDNN()
elif config.model_type == "MLDRNet":
me_model = MLDRnet()
me_model_r = MLDRnet()
elif config.model_type == "FTANet":
me_model = FTAnet(freq_bin = config.freq_bin, time_segment=config.seg_frame)
me_model_r = FTAnet(freq_bin = config.freq_bin, time_segment=config.seg_frame)
else: # MSNet
me_model = MSnet()
me_model_r = MSnet()
if config.ablation_mode == "single" or config.ablation_mode == "spl" or config.ablation_mode == "spat":
me_model_r = None
model = TONet(
l_model = me_model,
r_model = me_model_r,
config = config,
loss_func = loss_func,
mode = config.ablation_mode
)
trainer = pl.Trainer(
deterministic = True,
gpus = 1,
checkpoint_callback = False,
max_epochs = config.max_epoch,
auto_lr_find = True,
sync_batchnorm=True,
# check_val_every_n_epoch = 1,
val_check_interval = 0.25,
)
# load the checkpoint
ckpt = torch.load(config.resume_checkpoint, map_location="cpu")
model.load_state_dict(ckpt)
trainer.test(model, test_dataloaders)
if __name__ == "__main__":
parser = argparse.ArgumentParser(prog = "TONET for Singing Melody Extraction")
subparsers = parser.add_subparsers(dest = "mode")
parser_train = subparsers.add_parser("train")
parser_test = subparsers.add_parser("test")
args = parser.parse_args()
pl.seed_everything(config.random_seed)
if args.mode == "train":
train()
elif args.mode == "test":
test()