-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathvalidate_train.py
185 lines (150 loc) · 7.28 KB
/
validate_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
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
# Common
from torch.utils.data import DataLoader
import torch.nn.functional as F
import torch.nn as nn
import torch
import MinkowskiEngine as ME
import wandb
from tqdm import tqdm
import numpy as np
from utils.np_ioueval import iouEval
from utils.avgMeter import AverageMeter
def my_worker_init_fn(worker_id):
np.random.seed(np.random.get_state()[1][0] + worker_id)
def square_distance(src, dst):
"""
Calculate Euclid distance between each two points.
src^T * dst = xn * xm + yn * ym + zn * zm;
sum(src^2, dim=-1) = xn*xn + yn*yn + zn*zn;
sum(dst^2, dim=-1) = xm*xm + ym*ym + zm*zm;
dist = (xn - xm)^2 + (yn - ym)^2 + (zn - zm)^2
= sum(src**2,dim=-1)+sum(dst**2,dim=-1)-2*src^T*dst
Input:
src: source points, [N, C]
dst: target points, [M, C]
Output:
dist: per-point square distance, [N, M]
"""
N, _ = src.shape
M, _ = dst.shape
dist = -2 * torch.matmul(src, dst.permute(1, 0))
dist += torch.sum(src ** 2, -1).view(N, 1)
dist += torch.sum(dst ** 2, -1).view(1, M)
return dist
class validater:
def __init__(self,
cfg,
dataset_name,
domain,
criterion,
tf_writer, logger):
self.cfg = cfg
self.domain = domain
self.criterion = criterion
self.tf_writer = tf_writer
self.logger = logger
self.dataset_name = dataset_name
# init dataset for domain
if dataset_name == 'SemanticKITTI':
from dataset.semkitti_test_Sparse_Batch import SemanticKITTI_infer_B
self.test_dataset = SemanticKITTI_infer_B(self.cfg, 'test') # [gen_pselab | test]
elif dataset_name == 'SynLiDAR':
from dataset.synlidar_test_Sparse_Batch import SynLiDAR_infer_B
self.test_dataset = SynLiDAR_infer_B(self.cfg, 'test') # [gen_pselab | test]
elif dataset_name == "SemanticPOSS":
from dataset.SemanticPoss_test_Sparse_Batch import semPoss_infer_B
self.test_dataset = semPoss_infer_B(self.cfg, 'test') # [gen_pselab | test]
else:
raise NotImplementedError('The domain: ** {} ** is not implement now.'.format(dataset_name))
self.test_dataloader = DataLoader(self.test_dataset,
batch_size=self.cfg.DATALOADER.VAL_BATCH_SIZE,
num_workers=self.cfg.DATALOADER.NUM_WORKERS,
worker_init_fn=my_worker_init_fn,
collate_fn=self.test_dataset.collate_fn,
pin_memory=True,
shuffle=False,
drop_last=False
)
def rolling_predict(self, net_G, old_G, c_iter, src_centers=None, domain='src'):
torch.cuda.empty_cache()
self.logger.info('Current iter:{}'.format(c_iter))
if src_centers is not None:
centers_vectors = src_centers.Proto
net_G.eval() # set model to eval mode (for bn and dp)
if old_G is not None:
old_G.eval()
tea_iou_calc = iouEval(self.cfg.MODEL_G.NUM_CLASSES)
tea_iou_calc.reset()
t_dict = {} # record logs for wandb & tensorboard
iou_calc = iouEval(self.cfg.MODEL_G.NUM_CLASSES)
iou_calc.reset()
losses = AverageMeter()
# iter_loader = iter(self.test_dataloader)
tqdm_loader = tqdm(self.test_dataloader,
total=len(self.test_dataloader), ncols=50)
with torch.no_grad():
# for batch_data in self.tgt_train_loader:
for batch_idx, batch_data in enumerate(tqdm_loader):
if batch_idx % 500 == 0:
torch.cuda.empty_cache()
batch_data = self.send_data2GPU(batch_data)
cloud_inds = batch_data['cloud_inds']
val_G_in = ME.SparseTensor(batch_data['feats_mink'], batch_data['coords_mink'])
val_logits_1 = net_G(val_G_in, is_train=False)
val_logits_1 = val_logits_1.F
val_logits_1 = val_logits_1[batch_data['inverse_map']]
if old_G is not None:
val_logits_tea = old_G(val_G_in, is_train=False)
val_logits_tea = val_logits_tea.F
val_logits_tea = val_logits_tea[batch_data['inverse_map']]
val_labels = batch_data['pc_labs']
# Processing each scan. 对每一帧单独处理
left_ind = 0
for scan_i in range(len(cloud_inds)):
# get a single scan
pc_i_len = batch_data['s_lens'][scan_i]
vo_i_2_lab = val_labels[left_ind: left_ind+pc_i_len]
# net G
vo_i_2_pc = val_logits_1[left_ind: left_ind+pc_i_len, :]
loss = self.criterion(vo_i_2_pc, vo_i_2_lab)
losses.update(loss.item(), self.cfg.DATALOADER.VAL_BATCH_SIZE)
# cal IoU
iou_calc.addBatch(vo_i_2_pc.argmax(dim=1), vo_i_2_lab.long())
if old_G is not None:
vo_i_tea_pc = val_logits_tea[left_ind: left_ind+pc_i_len, :]
# cal IoU
tea_iou_calc.addBatch(vo_i_tea_pc.argmax(dim=1), vo_i_2_lab.long())
# udpate left index
left_ind += pc_i_len
t_dict['valid_{0}/all_loss_avg'.format(self.domain)] = losses.avg
sp_mean_iou1, sp_iou_list1 = iou_calc.getIoU()# sp1
self.logger.info('domain: {}, sp1 IoU:{:.1f}'.format(self.domain, sp_mean_iou1 * 100))
t_dict['valid_{0}/sp1_IoU'.format(self.domain)] = 100 * sp_mean_iou1
s_sp_1 = ' \n dec1 IoU: \n'
for ci, iou_tmp in enumerate(sp_iou_list1):
cn = self.test_dataset.label_name[ci]
if ci != 0:
s_sp_1 += '{}:{:5.2f}|'.format(cn, 100 * iou_tmp)
t_dict['{0}_sp_net_EC/{1}'.format(self.domain, cn)] = 100 * iou_tmp
self.logger.info(s_sp_1)
if old_G is not None:
sp_mean_iou_tea, sp_iou_list_tea = tea_iou_calc.getIoU()# sp1
self.logger.info('domain: {}, sp tea IoU:{:.1f}'.format(self.domain, sp_mean_iou_tea * 100))
t_dict['valid_{0}/sp_tea_IoU'.format(self.domain)] = 100 * sp_mean_iou_tea
s_sp_tea = ' \n dec tea IoU: \n'
for ci, iou_tmp in enumerate(sp_iou_list_tea):
cn = self.test_dataset.label_name[ci]
if ci != 0:
s_sp_tea += '{}:{:5.2f}|'.format(cn, 100 * iou_tmp)
t_dict['{0}_sp_net_tea_EC/{1}'.format(self.domain, cn)] = 100 * iou_tmp
self.logger.info(s_sp_tea)
for k, v in t_dict.items():
self.tf_writer.add_scalar(k, v, c_iter)
wandb.log({k: v}, step= c_iter)
torch.cuda.empty_cache()
return sp_mean_iou1
@staticmethod
def send_data2GPU(batch_data):
for key in batch_data: # Target data to gpu
batch_data[key] = batch_data[key].cuda(non_blocking=True)
return batch_data