-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmain.py
255 lines (198 loc) · 10.3 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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
import typing
from io import StringIO
from typing import Tuple
import json
import os
from tqdm import tqdm
import torch
from torch import nn
from torch import optim
from sklearn.preprocessing import StandardScaler
import joblib
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
import torch.utils.data as Data
import utils
# from modules import Encoder, Decoder
from modules_GRU import Encoder, Decoder
from custom_types import DaRnnNet, TrainData, TrainConfig
from utils import numpy_to_tvar
from constants import device
logger = utils.setup_log()
logger.info(f"Using computation device: {device}")
root_path = "/home/rr/Downloads/nsm_data/Train/"
total = 1753 # 1753
files_num = 50
def da_rnn(encoder_hidden_size=128, decoder_hidden_size=128, T=10, learning_rate=0.01, batch_size=128):
train_cfg = TrainConfig(T, int(files_num * 100 * 0.7), batch_size, nn.MSELoss())
logger.info(f"Training size: {int(total / files_num) * train_cfg.train_size:d}.")
enc_kwargs = {"input_size": 5307, "hidden_size": encoder_hidden_size, "T": T}
# enc_kwargs = {"input_size": 5307, "hidden_size": encoder_hidden_size, "T": T}
encoder = Encoder(**enc_kwargs).to(device)
with open(os.path.join("data", "enc_kwargs.json"), "w") as fi:
json.dump(enc_kwargs, fi, indent=4)
dec_kwargs = {"encoder_hidden_size": encoder_hidden_size,
"decoder_hidden_size": decoder_hidden_size, "T": T, "out_feats": 618}
decoder = Decoder(**dec_kwargs).to(device)
with open(os.path.join("data", "dec_kwargs.json"), "w") as fi:
json.dump(dec_kwargs, fi, indent=4)
encoder_optimizer = optim.Adam(
params=[p for p in encoder.parameters() if p.requires_grad],
lr=learning_rate,
weight_decay=0.001)
decoder_optimizer = optim.Adam(
params=[p for p in decoder.parameters() if p.requires_grad],
lr=learning_rate,
weight_decay=0.001)
da_rnn_net = DaRnnNet(encoder, decoder, encoder_optimizer, decoder_optimizer)
return train_cfg, da_rnn_net
def train(inputs_list, net: DaRnnNet, t_cfg: TrainConfig, n_epochs=10, save_plots=False):
# f = open(root_path + "record.txt", "w")
f = open("./record.txt", "w")
iter_per_epoch = int(np.ceil(int(total / files_num) * t_cfg.train_size * 1. / t_cfg.batch_size))
iter_losses = np.zeros((n_epochs + 1) * iter_per_epoch)
epoch_losses = np.zeros(n_epochs + 1)
logger.info(
f"Iterations per epoch: {int(total / files_num) * t_cfg.train_size * 1. / t_cfg.batch_size:3.3f} ~ {iter_per_epoch:d}.")
n_iter = 0
scale = StandardScaler()
for e_i in range(n_epochs):
train_input_data = pd.DataFrame()
train_label_data = pd.DataFrame()
all_val_loss = list()
all_test_loss = list()
n_iter_per_epoche = 0
for i, file in enumerate(tqdm(inputs_list)):
if i % files_num: # 50
single_input_data = pd.read_csv(root_path + "Input/" + file, sep=' ', header=None, dtype=float)
single_label_data = pd.read_csv(root_path + "Label/" + file, sep=' ', header=None, dtype=float)
train_input_data = train_input_data.append(single_input_data, ignore_index=True)
train_label_data = train_label_data.append(single_label_data, ignore_index=True)
elif i != 0 and i % files_num == 0:
scale = scale.fit(train_input_data)
t_input_data = scale.transform(train_input_data)
# t_input_data = np.array(train_input_data)
t_label_data = np.array(train_label_data)
perm_idx = np.random.permutation(t_cfg.train_size - t_cfg.T)
for t_i in range(0, t_cfg.train_size, t_cfg.batch_size):
batch_idx = perm_idx[t_i:(t_i + t_cfg.batch_size)]
feats, y_history, y_target = prep_train_data(batch_idx, t_cfg,
t_input_data[:t_cfg.train_size],
t_label_data[:t_cfg.train_size])
loss = train_iteration(net, t_cfg.loss_func, feats, y_history, y_target)
iter_losses[e_i * iter_per_epoch + n_iter_per_epoche] = loss
# print('itr_loss:', loss)
n_iter += 1
n_iter_per_epoche += 1
adjust_learning_rate(net, n_iter)
y_val_pred = predict(net, t_input_data[:t_cfg.train_size], t_label_data[:t_cfg.train_size],
t_cfg.train_size, t_cfg.batch_size, t_cfg.T,
on_train=True)
y_test_pred = predict(net, t_input_data, t_label_data,
t_cfg.train_size, t_cfg.batch_size, t_cfg.T,
on_train=False)
# TODO: make this MSE and make it work for multiple inputs
val_loss = [x - y for x, y in zip(y_val_pred, t_label_data[t_cfg.T - 1:]) if x.all() != 0]
test_loss = [x - y for x, y in zip(y_test_pred, t_label_data[t_cfg.train_size:]) if x.all() != 0]
all_val_loss = all_val_loss + val_loss
all_test_loss = all_test_loss + test_loss
train_input_data = train_input_data.drop(train_input_data.index, inplace=False)
train_label_data = train_label_data.drop(train_label_data.index, inplace=False)
epoch_losses[e_i] = np.mean(iter_losses[range(e_i * iter_per_epoch, (e_i + 1) * iter_per_epoch)])
if e_i % 1 == 0:
item = str(e_i) + ' ' + str(epoch_losses[e_i]) + ' ' + str(np.mean(np.abs(all_val_loss))) \
+ ' ' + str(np.mean(np.abs(all_test_loss))) + '\n'
f.write(item)
f.flush()
logger.info(
f"Epoch {e_i:d}, train loss: {epoch_losses[e_i]:3.5f}, val loss: {np.mean(np.abs(all_val_loss))}."
f"test loss: {np.mean(np.abs(all_test_loss))}.")
if e_i % 20 == 0 and e_i != 0:
torch.save(net.encoder.state_dict(), os.path.join("data", "encoder" + str(e_i) + ".torch"))
torch.save(net.decoder.state_dict(), os.path.join("data", "decoder" + str(e_i) + ".torch"))
f.close()
return iter_losses, epoch_losses
def prep_train_data(batch_idx: np.ndarray, t_cfg: TrainConfig, input_data, label_data):
feats = np.zeros((len(batch_idx), t_cfg.T - 1, input_data.shape[1]))
y_history = np.zeros((len(batch_idx), t_cfg.T - 1, label_data.shape[1]))
y_target = label_data[batch_idx + t_cfg.T]
# y_target = label_data.iloc[batch_idx + t_cfg.T]
for b_i, b_idx in enumerate(batch_idx):
b_slc = slice(b_idx, b_idx + t_cfg.T - 1)
feats[b_i, :, :] = input_data[b_slc, :]
y_history[b_i, :] = label_data[b_slc]
return feats, y_history, y_target
def adjust_learning_rate(net: DaRnnNet, n_iter: int):
# TODO: Where did this Learning Rate adjustment schedule come from?
# Should be modified to use Cosine Annealing with warm restarts https://www.jeremyjordan.me/nn-learning-rate/
if n_iter % 10000 == 0 and n_iter > 0:
for enc_params, dec_params in zip(net.enc_opt.param_groups, net.dec_opt.param_groups):
enc_params['lr'] = enc_params['lr'] * 0.9
dec_params['lr'] = dec_params['lr'] * 0.9
def train_iteration(t_net: DaRnnNet, loss_func: typing.Callable, X, y_history, y_target):
t_net.enc_opt.zero_grad()
t_net.dec_opt.zero_grad()
input_weighted, input_encoded = t_net.encoder(numpy_to_tvar(X))
y_pred = t_net.decoder(input_encoded, numpy_to_tvar(y_history))
y_true = numpy_to_tvar(y_target)
loss = loss_func(y_pred, y_true)
# regularization_loss = 0
# for param in t_net.decoder.parameters():
# regularization_loss += torch.sum(abs(param))
#
# classify_loss = loss_func(y_pred, y_true)
# loss = classify_loss + 0.5 * regularization_loss
loss.backward()
t_net.enc_opt.step()
t_net.dec_opt.step()
return loss.item()
def predict(t_net: DaRnnNet, input_data, label_data, train_size: int, batch_size: int, T: int, on_train=False):
out_size = label_data.shape[1]
if on_train:
y_pred = np.zeros((train_size - T + 1, out_size))
else:
y_pred = np.zeros((input_data.shape[0] - train_size, out_size))
for y_i in range(0, len(y_pred), batch_size):
y_slc = slice(y_i, y_i + batch_size)
batch_idx = range(len(y_pred))[y_slc]
b_len = len(batch_idx)
X = np.zeros((b_len, T - 1, input_data.shape[1]))
y_history = np.zeros((b_len, T - 1, label_data.shape[1]))
for b_i, b_idx in enumerate(batch_idx):
if on_train:
idx = range(b_idx, b_idx + T - 1)
else:
idx = range(b_idx + train_size - T, b_idx + train_size - 1)
X[b_i, :, :] = input_data[idx, :]
y_history[b_i, :] = label_data[idx]
y_history = numpy_to_tvar(y_history)
_, input_encoded = t_net.encoder(numpy_to_tvar(X))
y_pred[y_slc] = t_net.decoder(input_encoded, y_history).cpu().data.numpy()
return y_pred
def main():
save_plots = True
inputs_list = os.listdir(root_path + "Input/")
inputs_list.sort(key=lambda x: int(x[:-4]))
da_rnn_kwargs = {"batch_size": 64, "T": 15}
config, model = da_rnn(learning_rate=.001, **da_rnn_kwargs)
iter_loss, epoch_loss = train(inputs_list, model, config, n_epochs=100, save_plots=save_plots)
# final_y_pred = predict(model, data, config.train_size, config.batch_size, config.T)
plt.figure()
plt.semilogy(range(len(iter_loss)), iter_loss)
utils.save_or_show_plot("iter_loss.png", save_plots)
plt.figure()
plt.semilogy(range(len(epoch_loss)), epoch_loss)
utils.save_or_show_plot("epoch_loss.png", save_plots)
# plt.figure()
# plt.plot(final_y_pred, label='Predicted')
# plt.plot(data.targs[config.train_size:], label="True")
# plt.legend(loc='upper left')
# utils.save_or_show_plot("final_predicted.png", save_plots)
with open(os.path.join("data", "da_rnn_kwargs.json"), "w") as fi:
json.dump(da_rnn_kwargs, fi, indent=4)
# joblib.dump(scaler, os.path.join("data", "scaler.pkl"))
torch.save(model.encoder.state_dict(), os.path.join("data", "encoder.torch"))
torch.save(model.decoder.state_dict(), os.path.join("data", "decoder.torch"))
if __name__ == '__main__':
main()