wassname/ETSformer
1
0
Fork 0
mirror of https://github.com/wassname/ETSformer.git synced 2026-06-27 16:59:02 +08:00
Code Issues Packages Projects Releases Wiki Activity

multivar 2 single, an plot

Browse Source
This commit is contained in:
wassname
2022-11-28 14:10:49 +08:00
parent 244766827e
commit 0bf1606713
8 changed files with 1509 additions and 229 deletions
Download Patch File Download Diff File Expand all files Collapse all files
data_provider/data_loader.py
+14 -1
View File
@@ -229,6 +229,7 @@ class Dataset_Custom(Dataset):
cols = list(df_raw.columns) cols = list(df_raw.columns)
cols.remove(self.target) cols.remove(self.target)
cols.remove('date') cols.remove('date')
self.cols = ['date'] + cols + [self.target]
df_raw = df_raw[['date'] + cols + [self.target]] df_raw = df_raw[['date'] + cols + [self.target]]
# print(cols) # print(cols)
num_train = int(len(df_raw) * 0.7) num_train = int(len(df_raw) * 0.7)
@@ -244,6 +245,11 @@ class Dataset_Custom(Dataset):
df_data = df_raw[cols_data] df_data = df_raw[cols_data]
elif self.features == 'S': elif self.features == 'S':
df_data = df_raw[[self.target]] df_data = df_raw[[self.target]]
elif self.features == 'M2S':
df_data = df_raw[cols + [self.target]]
self.n_dims = 1 # len(cols + [self.target])
else:
raise NotImplementedError(self.features)
if self.scale: if self.scale:
train_data = df_data[border1s[0]:border2s[0]] train_data = df_data[border1s[0]:border2s[0]]
@@ -265,8 +271,15 @@ class Dataset_Custom(Dataset):
data_stamp = data_stamp.transpose(1, 0) data_stamp = data_stamp.transpose(1, 0)
self.data_x = data[border1:border2] self.data_x = data[border1:border2]
self.data_y = data[border1:border2] self.dates = df_raw['date'][border1:border2]
# self.data_y = data[border1:border2]
# y is just the col we predict
self.data_y = data[border1:border2][:, [-1]]
self.data_stamp = data_stamp self.data_stamp = data_stamp
# TODO check this is right. For example check that it makes the df
o = self.seq_len - self.label_len # + self.label_len
self.index = self.dates.iloc[o:].iloc[:len(self)]
def __getitem__(self, index): def __getitem__(self, index):
s_begin = index s_begin = index
exp/exp_main.py
+52 -44
View File
@@ -5,6 +5,8 @@ from utils.tools import EarlyStopping, adjust_learning_rate
from utils.metrics import metric from utils.metrics import metric
from utils.Adam import Adam from utils.Adam import Adam
from tqdm.auto import tqdm
import numpy as np import numpy as np
import torch import torch
import torch.nn as nn import torch.nn as nn
@@ -121,58 +123,64 @@ class Exp_Main(Exp_Basic):
model_optim = self._select_optimizer() model_optim = self._select_optimizer()
criterion = self._select_criterion() criterion = self._select_criterion()
for epoch in range(self.args.train_epochs): with tqdm(unit='epoch', total=self.args.train_epochs) as p:
iter_count = 0 for epoch in range(self.args.train_epochs):
train_loss = [] iter_count = 0
train_loss = []
self.model.train() self.model.train()
epoch_time = time.time() epoch_time = time.time()
for i, (batch_x, batch_y, batch_x_mark, batch_y_mark) in enumerate(train_loader): for i, (batch_x, batch_y, batch_x_mark, batch_y_mark) in enumerate(tqdm(train_loader, leave=False, desc='train')):
iter_count += 1 iter_count += 1
model_optim.zero_grad() model_optim.zero_grad()
batch_x = batch_x.float().to(self.device) batch_x = batch_x.float().to(self.device)
batch_y = batch_y.float().to(self.device) batch_y = batch_y.float().to(self.device)
batch_x_mark = batch_x_mark.float().to(self.device) batch_x_mark = batch_x_mark.float().to(self.device)
batch_y_mark = batch_y_mark.float().to(self.device) batch_y_mark = batch_y_mark.float().to(self.device)
# decoder input # decoder input
dec_inp = torch.zeros_like(batch_y[:, -self.args.pred_len:, :]).float() dec_inp = torch.zeros_like(batch_y[:, -self.args.pred_len:, :]).float()
dec_inp = torch.cat([batch_y[:, :self.args.label_len, :], dec_inp], dim=1).float().to(self.device) dec_inp = torch.cat([batch_y[:, :self.args.label_len, :], dec_inp], dim=1).float().to(self.device)
# encoder - decoder # encoder - decoder
outputs = self.model(batch_x, batch_x_mark, dec_inp, batch_y_mark) outputs = self.model(batch_x, batch_x_mark, dec_inp, batch_y_mark)
f_dim = -1 if self.args.features == 'MS' else 0 f_dim = -1 if self.args.features == 'MS' else 0
batch_y = batch_y[:, -self.args.pred_len:, f_dim:].to(self.device) batch_y = batch_y[:, -self.args.pred_len:, f_dim:].to(self.device)
loss = criterion(outputs, batch_y) loss = criterion(outputs, batch_y)
train_loss.append(loss.item()) train_loss.append(loss.item())
if (i + 1) % 100 == 0: # if (i + 1) % 10 == 0:
print("\titers: {0}, epoch: {1} | loss: {2:.7f}".format(i + 1, epoch + 1, loss.item())) # print("\titers: {0}, epoch: {1} | loss: {2:.7f}".format(i + 1, epoch + 1, loss.item()))
speed = (time.time() - time_now) / iter_count # speed = (time.time() - time_now) / iter_count
left_time = speed * ((self.args.train_epochs - epoch) * train_steps - i) # left_time = speed * ((self.args.train_epochs - epoch) * train_steps - i)
print('\tspeed: {:.4f}s/iter; left time: {:.4f}s'.format(speed, left_time)) # print('\tspeed: {:.4f}s/iter; left time: {:.4f}s'.format(speed, left_time))
iter_count = 0 # iter_count = 0
time_now = time.time() # time_now = time.time()
# p.desc = f'loss: {loss.item():.7f}'
loss.backward() loss.backward()
torch.nn.utils.clip_grad_norm(self.model.parameters(), 1.0) torch.nn.utils.clip_grad_norm(self.model.parameters(), 1.0)
model_optim.step() model_optim.step()
print("Epoch: {} cost time: {}".format(epoch + 1, time.time() - epoch_time)) print("Epoch: {} cost time: {}".format(epoch + 1, time.time() - epoch_time))
train_loss = np.average(train_loss) train_loss = np.average(train_loss)
vali_loss = self.vali(vali_data, vali_loader, criterion) vali_loss = self.vali(vali_data, vali_loader, criterion)
test_loss = self.vali(test_data, test_loader, criterion) test_loss = self.vali(test_data, test_loader, criterion)
print("Epoch: {0}, Steps: {1} | Train Loss: {2:.7f} Vali Loss: {3:.7f} Test Loss: {4:.7f}".format( p.update()
epoch + 1, train_steps, train_loss, vali_loss, test_loss)) p.desc = f'Train Loss: {test_loss:.7f} Vali Loss: {vali_loss:.7f}'
early_stopping(vali_loss, self.model, path) print("Epoch: {0}, Steps: {1} | Train Loss: {2:.7f} Vali Loss: {3:.7f} Test Loss: {4:.7f}".format(
if early_stopping.early_stop: epoch + 1, train_steps, train_loss, vali_loss, test_loss))
print("Early stopping") early_stopping(vali_loss, self.model, path)
break if early_stopping.early_stop:
print("Early stopping")
break
adjust_learning_rate(model_optim, epoch + 1, self.args) adjust_learning_rate(model_optim, epoch + 1, self.args)
best_model_path = path + '/' + 'checkpoint.pth' best_model_path = path + '/' + 'checkpoint.pth'
self.model.load_state_dict(torch.load(best_model_path)) self.model.load_state_dict(torch.load(best_model_path))
@@ -191,7 +199,7 @@ class Exp_Main(Exp_Basic):
self.model.eval() self.model.eval()
with torch.no_grad(): with torch.no_grad():
for i, (batch_x, batch_y, batch_x_mark, batch_y_mark) in enumerate(test_loader): for i, (batch_x, batch_y, batch_x_mark, batch_y_mark) in enumerate(tqdm(test_loader, leave=False)):
batch_x = batch_x.float().to(self.device) batch_x = batch_x.float().to(self.device)
batch_y = batch_y.float().to(self.device) batch_y = batch_y.float().to(self.device)
@@ -237,4 +245,4 @@ class Exp_Main(Exp_Basic):
np.save(folder_path + 'pred.npy', preds) np.save(folder_path + 'pred.npy', preds)
np.save(folder_path + 'true.npy', trues) np.save(folder_path + 'true.npy', trues)
return return preds, trues
mjc_notes.md
+3 -3
View File
@@ -2,6 +2,6 @@
TODO: TODO:
- [ ] run on stocks data - [x] run on stocks data
- [ ] multivariate in, univariate out - [x] multivariate in, univariate out
- [ ] graph - [x] graph
models/etsformer/encoder.py
+4
View File
@@ -183,6 +183,10 @@ class EncoderLayer(nn.Module):
self.dropout2 = nn.Dropout(dropout) self.dropout2 = nn.Dropout(dropout)
def forward(self, res, level, attn_mask=None): def forward(self, res, level, attn_mask=None):
# when c_in!=c_out assume target columns are at end of channels
level = level[:, :, -self.c_out:]
season, season_attn = self._season_block(res) season, season_attn = self._season_block(res)
res = res - season[:, :-self.pred_len] res = res - season[:, :-self.pred_len]
growth, growth_attn = self._growth_block(res) growth, growth_attn = self._growth_block(res)
models/etsformer/model.py
+1
View File
@@ -73,6 +73,7 @@ class ETSformer(nn.Module):
if self.training: if self.training:
x_enc = self.transform.transform(x_enc) x_enc = self.transform.transform(x_enc)
res = self.enc_embedding(x_enc) res = self.enc_embedding(x_enc)
level, growths, seasons, season_attns, growth_attns = self.encoder(res, x_enc, attn_mask=enc_self_mask) level, growths, seasons, season_attns, growth_attns = self.encoder(res, x_enc, attn_mask=enc_self_mask)
growth, season, growth_dampings = self.decoder(growths, seasons) growth, season, growth_dampings = self.decoder(growths, seasons)
notebooks/001_001_run.ipynb
+1435 -181
View File
File diff suppressed because one or more lines are too long
notebooks/results/Exchange_ETSformer_custom_ftM_sl128_pl48_dm512_nh8_el2_dl2_df2048_K0_lr0.001_'Exp'_0/test_metrics.npy
BIN
View File
Binary file not shown.
notebooks/results/Exchange_ETSformer_custom_ftM_sl128_pl48_dm512_nh8_el2_dl2_df2048_K0_lr0.001_'Exp'_0/val_metrics.npy
BIN
View File
Binary file not shown.