/
train_AAI.py
163 lines (154 loc) · 7.1 KB
/
train_AAI.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
import warnings
import librosa
warnings.filterwarnings('ignore', category=DeprecationWarning)
warnings.filterwarnings('ignore', category=FutureWarning)#remove numpy warnings
import numpy as np
import scipy.io
import csv
import os, time
import keras
import torch
import torch.nn as nn
import torch.nn.functional as F
import sys
from keras.preprocessing.sequence import pad_sequences
import pandas as pd
from keras.utils import to_categorical
import random
import load_and_log
import model
import models
import argparse
from keras.preprocessing.sequence import pad_sequences
from tqdm import tqdm
import early_stopper
from keras.utils import to_categorical
import logger
from shutil import copyfile
def parse_args(parser):
training = parser.add_argument_group('training setup')
training.add_argument('--epochs', type=int, default=500)
training.add_argument('--device', default='cuda:0')
training.add_argument('--cudnn-benchmark', default=True)
training.add_argument('--desc', type='str', default='exp') #describe any details of exp
training.add_argument('--log_path', type='str', default='logs/exp1.csv')
training.add_argument('--exp_results', type='str', default='logs/exp1_all_values.csv')
optimization = parser.add_argument_group('optimization setup')
optimization.add_argument('--optimizer', type=str, default='adam')
optimization.add_argument('-lr', '--learning-rate', type=float, default=0.0001)
optimization.add_argument('--weight-decay', default=1e-6, type=float)
optimization.add_argument('-batch_size', '--h-size', type=int, default=4,
setup = parser.add_argument_group('experimental setup')
setup.add_argument('--setup', type=str, required=True) #can be ss, pooled or ft (ss:single subject, ft: fine tuned)
setup.add_argument('--checkpoint', type=str, required=True)
setup.add_argument('--pool_checkpoint', type=str, required=False)#trained pooled model for ft
setup.add_argument('--model_desc', type=str, default='Transformer_AAI')#desc name to select model
return parser
def iter_(dataloader, train=True):
if train: net.train();
else: net.eval();
step_loss = []
for ema, mfcc in tqdm(dataloader):
if train:
optimizer.zero_grad()
ema, mfcc = ema.to(args.device), mfcc.to(args.device)
ema_out, dec_mask = net(inputs=mfcc.float(), dur_tgt=None)
loss = loss_fn(ema_out.float(), ema.float())
if train:
loss.backward()
optimizer.step()
step_loss.append(loss.item())
return sum(step_loss)/len(step_loss), step_loss
def get_cc(loader):
net.eval()
ema_, pred_ = [], []
for ema, mfcc in tqdm(loader):
ema, mfcc = ema.to(args.device), mfcc.to(args.device)
pred = net(mfcc.float(), dur_tgt=None)[0].float().permute(0, 2, 1)
ema_.extend(ema.float().detach().cpu().permute(0, 2, 1).tolist())
pred_.extend(pred.detach().cpu().tolist())
ema_ = np.array(ema_, dtype=np.float)
pred_ = np.array(pred_, dtype=np.float)
m = []
for j in range(len(pred_)):
c = []
for k in range(12):
c.append(scipy.stats.pearsonr(ema_[j][k], pred_[j][k])[0])
m.append(sum(c)/len(c))
return round(sum(m)/len(m), 3)
def main():
sub_loss = {}
sub_cc = {}
subjects = []
parser = argparse.ArgumentParser(description='AAI Training', allow_abbrev=False)
parser = parse_args(parser)
args, _ = parser.parse_known_args()
torch.manual_seed(args.seed + args.local_rank)
np.random.seed(args.seed + args.local_rank)
torch.backends.cudnn.benchmark = args.cudnn_benchmark
desc = args.model_desc
parser = models.parse_model_args(desc, parser)
args, unk_args = parser.parse_known_args()
model_config = models.get_model_config(desc, args)
pool_model = args.pool_checkpoint
if args.setup not in ['ss', 'ft', 'pooled']:
raise Exception('experimental setup not found')
total_runs = 1 if args.setup == 'pooled' else len(subjects)
for iter_num in range(total_runs):
subject = 'all' if args.setup == 'pooled' else subjects[iter_num]
net = models.get_model(desc, model_config, args.device).float()
if args.setup = 'ft':
try:
net.load_state_dict(torch.load(pool_model))
except:
raise Exception('Failed to load pooled model')
kw = dict(lr=args.learning_rate, betas=(0.9, 0.90), eps=1e-9,
weight_decay=args.weight_decay)
if args.optimizer == 'adam':
optimizer = torch.optim.Adam(net.parameters(), **kw)
Youtval, X_valseq, val_lengths, Youttrain, X_trainseq, train_lens = load_and_log.load_data(mode=args.setup, iter_num, subjects)
loss_fn = nn.MSELoss()
dataset = torch.utils.data.TensorDataset(torch.from_numpy(Youttrain), torch.from_numpy(X_trainseq))
dataloader = torch.utils.data.DataLoader(dataset, batch_size=4, shuffle=True)
dataset = torch.utils.data.TensorDataset(torch.from_numpy(Youtval), torch.from_numpy(X_valseq))
valloader = torch.utils.data.DataLoader(dataset, batch_size=4, shuffle=True)
min_loss = 100
scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, 'min', patience = 3, factor=0.5)
modelname = args.checkpoint
early_stopping = early_stopper.EarlyStopping(patience=11, verbose=True, path=modelname, min_run=1, delta=0.001)
start_time = time.time()
train_step_loss_arr, val_step_loss_arr = [], []
train_epoch_loss_arr, val_epoch_loss_arr = [], []
pytorch_total_params_trainable = sum(p.numel() for p in net.parameters() if p.requires_grad)
print(pytorch_total_params_trainable)
for i in range(args.epochs):
a = time.time()
trainloss, train_step_loss = iter_(dataloader)
valloss, val_step_loss = iter_(valloader, train=False)
train_step_loss_arr.extend(train_step_loss)
val_step_loss_arr.extend(val_step_loss)
train_epoch_loss_arr.append(trainloss)
val_epoch_loss_arr.append(valloss)
if valloss<min_loss:
min_loss = valloss
if i>5:
get_cc(dataloader)
get_cc(valloader)
scheduler.step(valloss)
early_stopping(valloss, net, i)
print('min:', min_loss)
if early_stopping.early_stop:
print("Early stopping at epoch ",i)
print('loss:', min_loss)
end_time = time.time() - start_time
break
end_time = time.time() - start_time
net.load_state_dict(torch.load(modelname))
tr = get_cc(dataloader)
v = get_cc(valloader)
sub_cc[subject] = [tr, v]
sub_loss[subject] = min_loss
with open(args.exp_results, 'wb') as f:
np.save(f, [train_epoch_loss, val_epoch_loss, train_step_loss, val_step_loss, model_config], allow_pickle=True)
logger.save(iter_num, args.desc, subject, modelname, min_loss, v, i,args.log_path,end_time, pytorch_total_params_trainable, model_config, args.exp_results)
print(args.traindesc, sub_cc, sub_loss)