-
Notifications
You must be signed in to change notification settings - Fork 22
/
kitti_submission.py
133 lines (109 loc) · 5.5 KB
/
kitti_submission.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
import os
import cv2
import utils
import hydra
import shutil
import logging
import torch
import torch.optim
import torch.utils.data
import torch.backends.cudnn as cudnn
import numpy as np
from tqdm import tqdm
from omegaconf import DictConfig
from factory import model_factory
from kitti import KITTITest
from models.utils import mesh_grid, knn_interpolation
from utils import copy_to_device, size_of_batch, save_flow_png, save_disp_png, load_disp_png, disp2pc
class Evaluator:
def __init__(self, device: torch.device, cfgs: DictConfig):
self.cfgs = cfgs
self.device = device
logging.info('Loading test set from %s' % self.cfgs.testset.root_dir)
logging.info('Dataset split: %s' % self.cfgs.testset.split)
self.test_dataset = KITTITest(self.cfgs.testset)
self.test_loader = utils.FastDataLoader(
dataset=self.test_dataset,
batch_size=self.cfgs.model.batch_size,
num_workers=self.cfgs.testset.n_workers,
pin_memory=True
)
logging.info('Creating model: %s' % self.cfgs.model.name)
self.model = model_factory(self.cfgs.model).to(device=self.device)
logging.info('Loading checkpoint from %s' % self.cfgs.ckpt.path)
checkpoint = torch.load(self.cfgs.ckpt.path, self.device)
self.model.load_state_dict(checkpoint['state_dict'], strict=self.cfgs.ckpt.strict)
@torch.no_grad()
def run(self):
logging.info('Generating outputs for KITTI submission...')
self.model.eval()
for inputs in tqdm(self.test_loader):
inputs = copy_to_device(inputs, self.device)
with torch.cuda.amp.autocast(enabled=False):
outputs = self.model.forward(inputs)
for batch_id in range(size_of_batch(inputs)):
flow_2d_pred = outputs['flow_2d'][batch_id]
flow_3d_pred = outputs['flow_3d'][batch_id]
test_id = inputs['index'][batch_id].item()
input_h = inputs['input_h'][batch_id].item()
input_w = inputs['input_w'][batch_id].item()
if 'training' in self.cfgs.testset.split:
out_dir = 'submission/training'
else:
out_dir = 'submission/testing'
f = inputs['intrinsics'][batch_id][0].item()
cx = inputs['intrinsics'][batch_id][1].item()
cy = inputs['intrinsics'][batch_id][2].item()
if self.test_dataset.cfgs.disp_provider == 'kitti':
disp1 = load_disp_png(os.path.join(
self.test_dataset.root_dir, 'disp_occ_0', '%06d_10.png' % test_id
))[0]
else:
disp1 = load_disp_png(os.path.join(
self.test_dataset.root_dir,
'disp_%s' % self.test_dataset.cfgs.disp_provider,
'%06d_10.png' % test_id
))[0]
os.makedirs('%s/disp_0' % out_dir, exist_ok=True)
save_disp_png('%s/disp_0/%06d_10.png' % (out_dir, test_id), disp1)
flow_2d_pred = flow_2d_pred.permute(1, 2, 0).clamp(-500, 500).cpu().numpy()
flow_2d_pred = flow_2d_pred[:input_h, :input_w]
os.makedirs('%s/flow_initial' % out_dir, exist_ok=True)
save_flow_png('%s/flow_initial/%06d_10.png' % (out_dir, test_id), flow_2d_pred)
# densification
pc1_dense = disp2pc(disp1, baseline=0.54, f=f, cx=cx, cy=cy)
pc1_dense = torch.from_numpy(pc1_dense.reshape([-1, 3]).transpose()).to(self.device)
pc1 = inputs['pcs'][batch_id, :3]
flow_3d_dense = knn_interpolation(pc1[None, ...], flow_3d_pred[None, ...], pc1_dense[None, ...])[0]
pc1_dense_warp = pc1_dense + flow_3d_dense
disp_c = 0.54 * f / pc1_dense_warp[2].cpu().numpy().reshape(input_h, input_w)
disp_c[disp1 < 0] = -1.0
os.makedirs('%s/disp_1_initial' % out_dir, exist_ok=True)
save_disp_png('%s/disp_1_initial/%06d_10.png' % (out_dir, test_id), disp_c)
# generate a coarse occlusion mask for rigid background refinement
grid = mesh_grid(1, input_h, input_w, device='cpu', channel_first=False)[0].numpy()
grid_warp = grid + flow_2d_pred
x_out = np.logical_or(grid_warp[..., 0] < 0, grid_warp[..., 0] > input_w)
y_out = np.logical_or(grid_warp[..., 1] < 0, grid_warp[..., 1] > input_h)
occ_mask1_2d = np.logical_or(x_out, y_out).astype(np.uint8) * 255
os.makedirs('%s/occ' % out_dir, exist_ok=True)
cv2.imwrite('%s/occ/%06d_10.png' % (out_dir, test_id), occ_mask1_2d)
@hydra.main(config_path='conf', config_name='evaluator')
def main(cfgs: DictConfig):
utils.init_logging()
# change working directory
shutil.rmtree(os.getcwd(), ignore_errors=True)
os.chdir(hydra.utils.get_original_cwd())
if torch.cuda.device_count() == 0:
device = torch.device('cpu')
logging.info('No CUDA device detected, using CPU for evaluation')
elif torch.cuda.device_count() == 1:
device = torch.device('cuda:0')
logging.info('Using GPU: %s' % torch.cuda.get_device_name(device))
cudnn.benchmark = True
else:
raise RuntimeError('Submission script does not support multi-GPU systems.')
evaluator = Evaluator(device, cfgs)
evaluator.run()
if __name__ == '__main__':
main()