-
Notifications
You must be signed in to change notification settings - Fork 5
/
Copy pathkitti.yaml
105 lines (100 loc) · 4.35 KB
/
kitti.yaml
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
# Enable debug mode will disable multi-processor so that you can visualize the inputs in order
name: "kitti"
debug_mode: False
hardware_synced: True
sensors_info:
# http://www.cvlibs.net/datasets/kitti/setup.php
camera:
frequency: 10 # Hz
type: "Point Grey Flea 2 (FL2-14S3C-C)"
shutter: "global"
shape: [375, 1242] # h 375 , w 1242 # Original is 1382 x 512, the size is from processed
lidar:
frequency: 10 # Hz
type: "Velodyne HDL-64E"
beams: 64
range: 120 #80 # m
fov: [26.9, 360] # vertical, horizontal
raw_data:
name: kitti_2011_09_26
root_dir: "Replace this with your training / validation / testing raw data sub-folder"
camera_sub_folder: image_02
lidar_sub_folder: velodyne_points
training_data:
chunk_size: 5000
downsample_ratio: 0.5 # TODO: This depends on how big the machine is
skip_frames: 0
output_format: 'tfrecords'
# TODO: the file name shall be automatcially generated from the yaml configs
output_dir: "Replace this with where you would like to save the tfrecords"
name: "training" # ["training", "testing", "validation"]
sampling_window: 5 # by default, we use a single lidar frame
sampling_stride: 1 # skip every every two frames
z_norm_methods: 'lidar_range' #'lidar_range_and_lidar_depth' # 'group' #
crop_shape: [[32, 64], [27, 283]] # Cropping the H, W of X so that it's 2***
reduce_lidar_line_to: 64 # TODO: this needs to be moved to the args for the trigger
features:
X:
modality: 'image'
data_type: 'float32'
nhwc: True
H: 32 # original size is 93
W: 256 # origin size is 310
C: 24 # (sampling_window + 1) * 4
feature_name: 'X'
Y:
modality: 'scalar'
data_type: 'float32'
feature_name: 'Y'
# In my experiment, below is the training / testing data
# Testing Data
#{'num_data': 2161,
# 'num_data_stats': {'2011_09_26_drive_0005_sync': 154,
# '2011_09_26_drive_0015_sync': 297,
# '2011_09_26_drive_0023_sync': 474,
# '2011_09_26_drive_0036_sync': 803,
# '2011_09_26_drive_0093_sync': 433},
# 'num_drive': 5}
# Training Data
#{'num_data': 13719,
# 'num_data_stats': {'2011_09_26_drive_0001_sync': 108,
# '2011_09_26_drive_0002_sync': 77,
# '2011_09_26_drive_0009_sync': 443,
# '2011_09_26_drive_0011_sync': 233,
# '2011_09_26_drive_0013_sync': 144,
# '2011_09_26_drive_0014_sync': 314,
# '2011_09_26_drive_0017_sync': 114,
# '2011_09_26_drive_0018_sync': 270,
# '2011_09_26_drive_0019_sync': 481,
# '2011_09_26_drive_0020_sync': 86,
# '2011_09_26_drive_0022_sync': 800,
# '2011_09_26_drive_0027_sync': 188,
# '2011_09_26_drive_0028_sync': 430,
# '2011_09_26_drive_0029_sync': 430,
# '2011_09_26_drive_0032_sync': 390,
# '2011_09_26_drive_0035_sync': 131,
# '2011_09_26_drive_0039_sync': 395,
# '2011_09_26_drive_0046_sync': 125,
# '2011_09_26_drive_0048_sync': 22,
# '2011_09_26_drive_0051_sync': 438,
# '2011_09_26_drive_0052_sync': 78,
# '2011_09_26_drive_0056_sync': 294,
# '2011_09_26_drive_0057_sync': 361,
# '2011_09_26_drive_0059_sync': 373,
# '2011_09_26_drive_0060_sync': 78,
# '2011_09_26_drive_0061_sync': 703,
# '2011_09_26_drive_0064_sync': 570,
# '2011_09_26_drive_0070_sync': 420,
# '2011_09_26_drive_0079_sync': 100,
# '2011_09_26_drive_0084_sync': 383,
# '2011_09_26_drive_0086_sync': 706,
# '2011_09_26_drive_0087_sync': 729,
# '2011_09_26_drive_0091_sync': 340,
# '2011_09_26_drive_0095_sync': 268,
# '2011_09_26_drive_0096_sync': 475,
# '2011_09_26_drive_0101_sync': 936,
# '2011_09_26_drive_0104_sync': 312,
# '2011_09_26_drive_0106_sync': 227,
# '2011_09_26_drive_0113_sync': 87,
# '2011_09_26_drive_0117_sync': 660},
# 'num_drive': 40}