Spaces:
Running
on
Zero
Running
on
Zero
File size: 14,671 Bytes
9e15541 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 |
import os
from datasets.kitti_360.kitti_360_dataset import Kitti360Dataset
from datasets.kitti_odom.kitti_odometry_dataset import KittiOdometryDataset
from datasets.kitti_raw.kitti_raw_dataset import KittiRawDataset
from datasets.nyu_depth_v2.nyu_depth_v2_dataset import NYUDepthV2Dataset
from datasets.realestate10k.realestate10k_dataset import RealEstate10kDataset
from datasets.waymo.waymo_dataset import WaymoDataset
def make_datasets(config):
type = config.get("type", "KITTI_Raw")
if type == "KITTI_Odometry":
train_dataset = KittiOdometryDataset(
base_path=config["data_path"],
frame_count=config.get("data_fc", 1),
target_image_size=config.get("image_size", (128, 256)),
return_stereo=config.get("data_stereo", False),
sequences=config.get("train_sequences", ("00",)),
custom_pose_path=config.get("custom_pose_path", None),
keyframe_offset=0 #-(config.get("data_fc", 1) // 2)
)
test_dataset = KittiOdometryDataset(
base_path=config["data_path"],
frame_count=config.get("data_fc", 1),
target_image_size=config.get("image_size", (128, 256)),
return_stereo=config.get("data_stereo", False),
sequences=config.get("val_sequences", ("00",)),
custom_pose_path=config.get("custom_pose_path", None),
keyframe_offset=0 #-(config.get("data_fc", 1) // 2)
)
return train_dataset, test_dataset
elif type == "KITTI_Raw":
train_dataset = KittiRawDataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=os.path.join(config["split_path"], "train_files.txt"),
target_image_size=config.get("image_size", (192, 640)),
frame_count=config.get("data_fc", 1),
return_stereo=config.get("data_stereo", False),
keyframe_offset=config.get("keyframe_offset", 0),
dilation=config.get("dilation", 1),
color_aug=config.get("color_aug", False)
)
test_dataset = KittiRawDataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=os.path.join(config["split_path"], "val_files.txt"),
target_image_size=config.get("image_size", (192, 640)),
frame_count=config.get("data_fc", 1),
return_stereo=config.get("data_stereo", False),
keyframe_offset=config.get("keyframe_offset", 0),
dilation=config.get("dilation", 1),
)
return train_dataset, test_dataset
elif type == "KITTI_360":
if config.get("split_path", None) is None:
train_split_path = None
test_split_path = None
else:
train_split_path = os.path.join(config["split_path"], "train_files.txt")
test_split_path = os.path.join(config["split_path"], "val_files.txt")
train_dataset = Kitti360Dataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=train_split_path,
target_image_size=tuple(config.get("image_size", (192, 640))),
frame_count=config.get("data_fc", 3),
return_stereo=config.get("data_stereo", True),
return_fisheye=config.get("data_fisheye", True),
return_3d_bboxes=config.get("data_3d_bboxes", False),
return_segmentation=config.get("data_segmentation", False),
keyframe_offset=config.get("keyframe_offset", 0),
dilation=config.get("dilation", 1),
fisheye_rotation=config.get("fisheye_rotation", 0),
fisheye_offset=config.get("fisheye_offset", 1),
color_aug=config.get("color_aug", False),
is_preprocessed=config.get("is_preprocessed", False)
)
test_dataset = Kitti360Dataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=test_split_path,
target_image_size=tuple(config.get("image_size", (192, 640))),
frame_count=config.get("data_fc", 3),
return_stereo=config.get("data_stereo", True),
return_fisheye=config.get("data_fisheye", True),
return_3d_bboxes=config.get("data_3d_bboxes", False),
return_segmentation=config.get("data_segmentation", False),
keyframe_offset=config.get("keyframe_offset", 0),
fisheye_rotation=config.get("fisheye_rotation", 0),
fisheye_offset=config.get("fisheye_offset", 1),
dilation=config.get("dilation", 1),
is_preprocessed=config.get("is_preprocessed", False)
)
return train_dataset, test_dataset
elif type == "RealEstate10k":
train_dataset = RealEstate10kDataset(
data_path=config["data_path"],
split_path=None,
target_image_size=config.get("image_size", (256, 384)),
frame_count=config.get("data_fc", 2),
keyframe_offset=0, #-(config.get("data_fc", 1) // 2),
dilation=config.get("dilation", 10),
color_aug=config.get("color_aug", False)
)
test_dataset = RealEstate10kDataset(
data_path=config["data_path"],
split_path=os.path.join(config["split_path"], "val_files.txt"),
target_image_size=config.get("image_size", (256, 384)),
frame_count=config.get("data_fc", 2),
keyframe_offset=0, #-(config.get("data_fc", 1) // 2),
dilation=config.get("dilation", 10),
color_aug=False
)
return train_dataset, test_dataset
elif type == "Waymo":
if config.get("split_path", None) is None:
train_split_path = None
test_split_path = None
else:
train_split_path = os.path.join(config["split_path"], "train_files.txt")
test_split_path = os.path.join(config["split_path"], "val_files.txt")
train_dataset = WaymoDataset(
data_path=config["data_path"],
mode="training",
split_path=train_split_path,
target_image_size=tuple(config.get("image_size", (320, 480))),
frame_count=config.get("data_fc", 2),
keyframe_offset=config.get("keyframe_offset", 0),
return_45=config.get("return_45", True),
return_90=config.get("return_90", True),
offset_45=config.get("offset_45", 5),
offset_90=config.get("offset_90", 10),
dilation=config.get("dilation", 1),
color_aug=config.get("color_aug", True),
correct_exposure=config.get("correct_exposure", True),
)
test_dataset = WaymoDataset(
data_path=config["data_path"],
mode="validation",
split_path=test_split_path,
target_image_size=tuple(config.get("image_size", (320, 480))),
frame_count=config.get("data_fc", 2),
keyframe_offset=config.get("keyframe_offset", 0),
return_45=config.get("return_45", True),
return_90=config.get("return_90", True),
offset_45=config.get("offset_45", 5),
offset_90=config.get("offset_90", 10),
dilation=config.get("dilation", 1),
color_aug=False,
return_depth=True,
correct_exposure=config.get("correct_exposure", True),
)
return train_dataset, test_dataset
elif type == "KITTI_Raw_DFT":
train_dataset = KittiRawDataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=os.path.join(config["split_path"], "train_files.txt"),
target_image_size=config.get("image_size", (192, 640)),
frame_count=config.get("data_fc", 1),
return_stereo=config.get("data_stereo", False),
keyframe_offset=config.get("keyframe_offset", 0),
dilation=config.get("dilation", 1),
color_aug=config.get("color_aug", False)
)
test_dataset = KittiRawDataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=os.path.join(config["split_path"], "val_files.txt"),
target_image_size=config.get("image_size", (192, 640)),
frame_count=config.get("data_fc", 1),
return_stereo=config.get("data_stereo", False),
keyframe_offset=config.get("keyframe_offset", 0),
dilation=config.get("dilation", 1),
)
return train_dataset, test_dataset
elif type == "KITTI_360_DFT":
if config.get("split_path", None) is None:
train_split_path = None
test_split_path = None
else:
train_split_path = os.path.join(config["split_path"], "train_files.txt")
test_split_path = os.path.join(config["split_path"], "val_files.txt")
train_dataset = Kitti360Dataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=train_split_path,
target_image_size=tuple(config.get("image_size", (192, 640))),
frame_count=config.get("data_fc", 3),
return_stereo=config.get("data_stereo", True),
return_fisheye=config.get("data_fisheye", True),
return_3d_bboxes=config.get("data_3d_bboxes", False),
return_segmentation=config.get("data_segmentation", False),
keyframe_offset=config.get("keyframe_offset", 0),
dilation=config.get("dilation", 1),
fisheye_rotation=config.get("fisheye_rotation", 0),
fisheye_offset=config.get("fisheye_offset", 1),
stereo_offset=config.get("stereo_offset", 1),
color_aug=config.get("color_aug", False),
is_preprocessed=config.get("is_preprocessed", False)
)
test_dataset = Kitti360Dataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=test_split_path,
target_image_size=tuple(config.get("image_size", (192, 640))),
frame_count=config.get("data_fc", 3),
return_stereo=config.get("data_stereo", True),
return_fisheye=config.get("data_fisheye", True),
return_3d_bboxes=config.get("data_3d_bboxes", False),
return_segmentation=config.get("data_segmentation", False),
keyframe_offset=config.get("keyframe_offset", 0),
fisheye_rotation=config.get("fisheye_rotation", 0),
fisheye_offset=config.get("fisheye_offset", [10])[0], ## this modifies the offsets for all datasets including the training dataset
stereo_offset= config.get("stereo_offset", [1])[0], ## This is to set consistent evaluation with test and viz
dilation=config.get("dilation", 1),
is_preprocessed=config.get("is_preprocessed", False)
)
return train_dataset, test_dataset
else:
raise NotImplementedError(f"Unsupported dataset type: {type}")
def make_test_dataset(config):
type = config.get("type", "KITTI_Raw")
if type == "KITTI_Raw":
test_dataset = KittiRawDataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=os.path.join(config["split_path"], "test_files.txt"),
target_image_size=config.get("image_size", (192, 640)),
return_depth=True,
frame_count=config.get("data_fc", 1),
return_stereo=config.get("data_stereo", False),
keyframe_offset=0
)
return test_dataset
elif type == "KITTI_360":
test_dataset = Kitti360Dataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=os.path.join(config.get("split_path", None), "test_files.txt"),
target_image_size=tuple(config.get("image_size", (192, 640))),
frame_count=config.get("data_fc", 1),
return_stereo=config.get("data_stereo", False),
return_fisheye=config.get("data_fisheye", False),
return_3d_bboxes=config.get("data_3d_bboxes", False),
return_segmentation=config.get("data_segmentation", False),
keyframe_offset=0,
fisheye_rotation=config.get("fisheye_rotation", 0),
fisheye_offset=config.get("fisheye_offset", 1),
dilation=config.get("dilation", 1),
is_preprocessed=config.get("is_preprocessed", False)
)
return test_dataset
elif type == "KITTI_360_DFT":
test_dataset = Kitti360Dataset(
data_path=config["data_path"],
pose_path=config["pose_path"],
split_path=os.path.join(config.get("split_path", None), "test_files.txt"),
target_image_size=tuple(config.get("image_size", (192, 640))),
frame_count=config.get("data_fc", 1),
return_stereo=config.get("data_stereo", False),
return_fisheye=config.get("data_fisheye", False),
return_3d_bboxes=config.get("data_3d_bboxes", False),
return_segmentation=config.get("data_segmentation", False),
keyframe_offset=0,
fisheye_rotation=config.get("fisheye_rotation", 0),
fisheye_offset=config.get("fisheye_offset", [10])[0], ## this modifies the offsets for all datasets including the training dataset
stereo_offset= config.get("stereo_offset", [1])[0], ## This is to set consistent evaluation with test and viz
dilation=config.get("dilation", 1),
is_preprocessed=config.get("is_preprocessed", False),
return_depth=True
)
return test_dataset
elif type == "RealEstate10k":
test_dataset = RealEstate10kDataset(
data_path=config["data_path"],
split_path=os.path.join(config["split_path"], "test_files.txt"),
target_image_size=config.get("image_size", (256, 384)),
frame_count=config.get("data_fc", 2),
keyframe_offset=0,
dilation=config.get("dilation", 10),
color_aug=False
)
return test_dataset
elif type == "NYU_Depth_V2":
test_dataset = NYUDepthV2Dataset(
data_path=config["data_path"],
target_image_size=config.get("image_size", (256, 384)),
)
return test_dataset
else:
raise NotImplementedError(f"Unsupported dataset type: {type}") |