-
Notifications
You must be signed in to change notification settings - Fork 1
/
train.py
120 lines (110 loc) · 5.14 KB
/
train.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
import numpy as np
import scipy.misc,scipy.io
import time,os,sys
import threading
import util
print(util.toYellow("======================================================="))
print(util.toYellow("train.py (train with joint 2D optimization with novel viewpoints)"))
print(util.toYellow("======================================================="))
import tensorflow as tf
import data,graph,transform
import options
print(util.toMagenta("setting configurations..."))
opt = options.set(training=True)
# create directories for model output
util.mkdir("models_{0}".format(opt.group))
print(util.toMagenta("building graph..."))
tf.reset_default_graph()
# build graph
with tf.device("/gpu:0"):
# ------ define input data ------
inputImage = tf.placeholder(tf.float32,shape=[opt.batchSize,opt.inH,opt.inW,3])
renderTrans = tf.placeholder(tf.float32,shape=[opt.batchSize,opt.novelN,4])
depthGT = tf.placeholder(tf.float32,shape=[opt.batchSize,opt.novelN,opt.H,opt.W,1])
maskGT = tf.placeholder(tf.float32,shape=[opt.batchSize,opt.novelN,opt.H,opt.W,1])
PH = [inputImage,renderTrans,depthGT,maskGT]
# ------ build encoder-decoder ------
encoder = graph.encoder if opt.arch=="original" else \
graph.encoder_resnet if opt.arch=="resnet" else None
decoder = graph.decoder if opt.arch=="original" else \
graph.decoder_resnet if opt.arch=="resnet" else None
latent = encoder(opt,inputImage)
XYZ,maskLogit = decoder(opt,latent) # [B,H,W,3V],[B,H,W,V]
mask = tf.to_float(maskLogit>0)
# ------ build transformer ------
fuseTrans = tf.nn.l2_normalize(opt.fuseTrans,dim=1)
XYZid,ML = transform.fuse3D(opt,XYZ,maskLogit,fuseTrans) # [B,1,VHW]
newDepth,newMaskLogit,collision = transform.render2D(opt,XYZid,ML,renderTrans) # [B,N,H,W,1]
# ------ define loss ------
loss_depth = graph.masked_l1_loss(newDepth-depthGT,tf.equal(collision,1))/(opt.batchSize*opt.novelN)
loss_mask = graph.cross_entropy_loss(newMaskLogit,maskGT)/(opt.batchSize*opt.novelN)
loss = loss_mask+opt.lambdaDepth*loss_depth
# ------ optimizer ------
lr_PH = tf.placeholder(tf.float32,shape=[])
optim = tf.train.AdamOptimizer(learning_rate=lr_PH).minimize(loss)
# ------ generate summaries ------
summaryImage = [util.imageSummary(opt,"image_RGB",inputImage,opt.inH,opt.inW),
util.imageSummary(opt,"image_depth/pred",((1-newDepth)*tf.to_float(tf.equal(collision,1)))[:,0,:,:,0:1],opt.H,opt.W),
util.imageSummary(opt,"image_depth/GT",(1-depthGT)[:,0,:,:,0:1],opt.H,opt.W),
util.imageSummary(opt,"image_mask/new",tf.sigmoid(newMaskLogit[:,0,:,:,0:1]),opt.H,opt.W),
util.imageSummary(opt,"image_mask",tf.sigmoid(maskLogit[:,:,:,0:1]),opt.outH,opt.outW),
util.imageSummary(opt,"image_mask/GT",maskGT[:,0,:,:,0:1],opt.H,opt.W)]
summaryImage = tf.summary.merge(summaryImage)
summaryLoss = [tf.summary.scalar("loss_total",loss),
tf.summary.scalar("loss_mask",loss_mask),
tf.summary.scalar("loss_depth",loss_depth)]
summaryLoss = tf.summary.merge(summaryLoss)
# load data
print(util.toMagenta("loading dataset..."))
dataloader = data.Loader(opt)
dataloader.loadChunk(opt)
# prepare model saver/summary writer
saver = tf.train.Saver()
summaryWriter = tf.summary.FileWriter("summary_{0}/{1}".format(opt.group,opt.model))
print(util.toYellow("======= TRAINING START ======="))
timeStart = time.time()
# start session
tfConfig = tf.ConfigProto(allow_soft_placement=True)
tfConfig.gpu_options.allow_growth = True
with tf.Session(config=tfConfig) as sess:
sess.run(tf.global_variables_initializer())
if opt.fromIt!=0:
util.restoreModelFromIt(opt,sess,saver,opt.fromIt)
print(util.toMagenta("resuming from iteration {0}...".format(opt.fromIt)))
elif opt.load:
util.restoreModel(opt,sess,saver)
print(util.toMagenta("loading pretrained ({0}) to fine-tune...".format(opt.load)))
summaryWriter.add_graph(sess.graph)
print(util.toMagenta("start training..."))
chunkResumeN,chunkMaxN = opt.fromIt//opt.itPerChunk,opt.toIt//opt.itPerChunk
# training loop
for c in range(chunkResumeN,chunkMaxN):
dataloader.shipChunk()
dataloader.thread = threading.Thread(target=dataloader.loadChunk,args=[opt])
dataloader.thread.start()
for i in range(c*opt.itPerChunk,(c+1)*opt.itPerChunk):
lr = opt.lr*opt.lrDecay**(i//opt.lrStep)
# make training batch
batch = data.makeBatch(opt,dataloader,PH)
batch[lr_PH] = lr
# run one step
runList = [optim,loss,loss_depth,loss_mask,maskLogit]
_,l,ld,lm,ml = sess.run(runList,feed_dict=batch)
if (i+1)%20==0:
print("it. {0}/{1}, lr={2}, loss={4} ({5},{6}), time={3}"
.format(util.toCyan("{0}".format(i+1)),
opt.toIt,
util.toYellow("{0:.0e}".format(lr)),
util.toGreen("{0:.2f}".format(time.time()-timeStart)),
util.toRed("{0:.2f}".format(l)),
util.toRed("{0:.2f}".format(ld)),
util.toRed("{0:.2f}".format(lm))))
if (i+1)%100==0:
summaryWriter.add_summary(sess.run(summaryLoss,feed_dict=batch),i+1)
if (i+1)%500==0:
summaryWriter.add_summary(sess.run(summaryImage,feed_dict=batch),i+1)
if (i+1)%2000==0:
util.saveModel(opt,sess,saver,i+1)
print(util.toGreen("model saved: {0}/{1}, it.{2}".format(opt.group,opt.model,i+1)))
dataloader.thread.join()
print(util.toYellow("======= TRAINING DONE ======="))