-
Notifications
You must be signed in to change notification settings - Fork 19
/
Copy pathtrain_funcs.py
129 lines (104 loc) · 5.48 KB
/
train_funcs.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
import os
import torch
from tqdm import tqdm
import numpy as np
def train_autoencoder_dataloader(dataloader_train, dataloader_val,
device, model, optim, loss_fn,
bsize, start_epoch, n_epochs, eval_freq, scheduler = None,
writer=None, save_recons=True, shapedata = None,
metadata_dir=None, samples_dir = None, checkpoint_path = None):
if not shapedata.normalization:
shapedata_mean = torch.Tensor(shapedata.mean).to(device)
shapedata_std = torch.Tensor(shapedata.std).to(device)
total_steps = start_epoch*len(dataloader_train)
for epoch in range(start_epoch, n_epochs):
model.train()
tloss = []
for b, sample_dict in enumerate(tqdm(dataloader_train)):
optim.zero_grad()
tx = sample_dict['points'].to(device)
cur_bsize = tx.shape[0]
tx_hat = model(tx)
loss = loss_fn(tx, tx_hat)
loss.backward()
optim.step()
if shapedata.normalization:
tloss.append(cur_bsize * loss.item())
else:
with torch.no_grad():
if shapedata.mean.shape[0]!=tx.shape[1]:
tx_norm = tx[:,:-1,:]
tx_hat_norm = tx_hat[:,:-1,:]
else:
tx_norm = tx
tx_hat_norm = tx_hat
tx_norm = (tx_norm - shapedata_mean)/shapedata_std
tx_norm = torch.cat((tx_norm,torch.zeros(tx.shape[0],1,tx.shape[2]).to(device)),1)
tx_hat_norm = (tx_hat_norm -shapedata_mean)/shapedata_std
tx_hat_norm = torch.cat((tx_hat_norm,torch.zeros(tx.shape[0],1,tx.shape[2]).to(device)),1)
loss_norm = loss_fn(tx_norm, tx_hat_norm)
tloss.append(cur_bsize * loss_norm.item())
if writer and total_steps % eval_freq == 0:
writer.add_scalar('loss/loss/data_loss',loss.item(),total_steps)
writer.add_scalar('training/learning_rate', optim.param_groups[0]['lr'],total_steps)
total_steps += 1
# validate
model.eval()
vloss = []
with torch.no_grad():
for b, sample_dict in enumerate(tqdm(dataloader_val)):
tx = sample_dict['points'].to(device)
cur_bsize = tx.shape[0]
tx_hat = model(tx)
loss = loss_fn(tx, tx_hat)
if shapedata.normalization:
vloss.append(cur_bsize * loss.item())
else:
with torch.no_grad():
if shapedata.mean.shape[0]!=tx.shape[1]:
tx_norm = tx[:,:-1,:]
tx_hat_norm = tx_hat[:,:-1,:]
else:
tx_norm = tx
tx_hat_norm = tx_hat
tx_norm = (tx_norm - shapedata_mean)/shapedata_std
tx_norm = torch.cat((tx_norm,torch.zeros(tx.shape[0],1,tx.shape[2]).to(device)),1)
tx_hat_norm = (tx_hat_norm - shapedata_mean)/shapedata_std
tx_hat_norm = torch.cat((tx_hat_norm,torch.zeros(tx.shape[0],1,tx.shape[2]).to(device)),1)
loss_norm = loss_fn(tx_norm, tx_hat_norm)
vloss.append(cur_bsize * loss_norm.item())
if scheduler:
scheduler.step()
epoch_tloss = sum(tloss) / float(len(dataloader_train.dataset))
writer.add_scalar('avg_epoch_train_loss',epoch_tloss,epoch)
if len(dataloader_val.dataset) > 0:
epoch_vloss = sum(vloss) / float(len(dataloader_val.dataset))
writer.add_scalar('avg_epoch_valid_loss', epoch_vloss,epoch)
print('epoch {0} | tr {1} | val {2}'.format(epoch,epoch_tloss,epoch_vloss))
else:
print('epoch {0} | tr {1} '.format(epoch,epoch_tloss))
model = model.cpu()
torch.save({'epoch': epoch,
'autoencoder_state_dict': model.state_dict(),
'optimizer_state_dict' : optim.state_dict(),
'scheduler_state_dict': scheduler.state_dict(),
},os.path.join(metadata_dir, checkpoint_path+'.pth.tar'))
if epoch % 10 == 0:
torch.save({'epoch': epoch,
'autoencoder_state_dict': model.state_dict(),
'optimizer_state_dict' : optim.state_dict(),
'scheduler_state_dict': scheduler.state_dict(),
},os.path.join(metadata_dir, checkpoint_path+'%s.pth.tar'%(epoch)))
model = model.to(device)
if save_recons:
with torch.no_grad():
if epoch == 0:
mesh_ind = [0]
msh = tx[mesh_ind[0]:1,0:-1,:].detach().cpu().numpy()
shapedata.save_meshes(os.path.join(samples_dir,'input_epoch_{0}'.format(epoch)),
msh, mesh_ind)
mesh_ind = [0]
msh = tx_hat[mesh_ind[0]:1,0:-1,:].detach().cpu().numpy()
shapedata.save_meshes(os.path.join(samples_dir,'epoch_{0}'.format(epoch)),
msh, mesh_ind)
print('~FIN~')