def test_exclude_classes4(self):
'''
exclude Car Pedestrian
'''
data_cfg = Test_exclude_classes.TRAINDATA_cfg
data_cfg["prep"]["@classes_to_exclude"] = ["Car", "Pedestrian"]
voxelizer = voxelizer_builder.build(Test_exclude_classes.VOXELIZER_cfg)
target_assigner = target_assigner_builder.build(Test_exclude_classes.TARGETASSIGNER_cfg)
dataloader = build(data_cfg,
ext_dict={
"voxelizer": voxelizer,
"target_assigner": target_assigner,
"feature_map_size": [1, 200, 176]
})
has1 = False
has2 = False
for data in dataloader:
labels =data["labels"]
labels1 = labels[labels == 1]
labels2 = labels[labels == 2]
if labels1.shape[0] > 0:
has1 = True
if labels2.shape[0] > 0:
has2 = True
self.assertFalse(has1)
self.assertFalse(has2)
def __init__(self, *args, **kwargs):
super(Test_kittidata_general, self).__init__(*args, **kwargs)
data_cfg = Test_kittidata_general.TRAINDATA_cfg
voxelizer = voxelizer_builder.build(Test_kittidata_general.VOXELIZER_cfg)
target_assigner = target_assigner_builder.build(Test_kittidata_general.TARGETASSIGNER_cfg)
dataloader = build(data_cfg,
ext_dict={
"voxelizer": voxelizer,
"target_assigner": target_assigner,
"feature_map_size": [1, 200, 176]
})
self.dataloader = dataloader
self.data = None
self.box_coder = target_assigner.box_coder
for i, data in enumerate(self.dataloader):
if i == 1:
self.data = data
break
else:
continue
def test_prep_infos1(self):
'''
filt
'''
data_cfg = Test_prep_infos.TRAINDATA_cfg
data_cfg["prep_infos"] = {
"@valid_range": [0, -32, -3, 52.8, 32.0, 1],
"@target_classes": ["Pedestrian"]
}
voxelizer = voxelizer_builder.build(Test_exclude_classes.VOXELIZER_cfg)
target_assigner = target_assigner_builder.build(Test_exclude_classes.TARGETASSIGNER_cfg)
dataloader = build(data_cfg,
ext_dict={
"voxelizer": voxelizer,
"target_assigner": target_assigner,
"feature_map_size": [1, 200, 176]
})
for data in dataloader:
labels =data["labels"]
labels1 = labels[labels == 1]
labels2 = labels[labels == 2]
self.assertTrue(labels2.shape[0] > 0)
def build_dataloader():
from incdet3.builders.dataloader_builder import build
from incdet3.builders import voxelizer_builder, target_assigner_builder
VOXELIZER_cfg = {
"type": "VoxelizerV1",
"@voxel_size": [0.05, 0.05, 0.1],
"@point_cloud_range": [0, -32, -3, 52.8, 32.0, 1],
"@max_num_points": 5,
"@max_voxels": 20000
}
TARGETASSIGNER_cfg = {
"type": "TaskAssignerV1",
"@classes": ["Car", "Pedestrian", "Cyclist"],
"@feature_map_sizes": None,
"@positive_fraction": None,
"@sample_size": 512,
"@assign_per_class": True,
"box_coder": {
"type": "BoxCoderV1",
"@custom_ndim": 0
},
"class_settings_car": {
"AnchorGenerator": {
"type": "AnchorGeneratorBEV",
"@class_name": "Car",
"@anchor_ranges": [0, -32, 0, 52.8, 32.0,
0], # TBD in modify_cfg(cfg)
"@sizes": [1.6, 3.9, 1.56], # wlh
"@rotations": [0, 1.57],
"@match_threshold": 0.6,
"@unmatch_threshold": 0.45,
},
"SimilarityCalculator": {
"type": "NearestIoUSimilarity"
}
},
"class_settings_pedestrian": {
"AnchorGenerator": {
"type": "AnchorGeneratorBEV",
"@class_name": "Pedestrian",
"@anchor_ranges": [0, -32, 0, 52.8, 32.0,
0], # TBD in modify_cfg(cfg)
"@sizes": [0.6, 0.8, 1.73], # wlh
"@rotations": [0, 1.57],
"@match_threshold": 0.6,
"@unmatch_threshold": 0.45,
},
"SimilarityCalculator": {
"type": "NearestIoUSimilarity"
}
},
"class_settings_cyclist": {
"AnchorGenerator": {
"type": "AnchorGeneratorBEV",
"@class_name": "Cyclist",
"@anchor_ranges": [0, -32, 0, 52.8, 32.0,
0], # TBD in modify_cfg(cfg)
"@sizes": [0.6, 0.8, 1.73], # wlh
"@rotations": [0, 1.57],
"@match_threshold": 0.6,
"@unmatch_threshold": 0.45,
},
"SimilarityCalculator": {
"type": "NearestIoUSimilarity"
}
},
}
TRAINDATA_cfg = {
"dataset": "kitti", # carla
"training": False, # set this to false to avoid shuffle
"batch_size": 1,
"num_workers": 1,
"@root_path": "unit_tests/data/test_kittidata",
"@info_path": "unit_tests/data/test_kittidata/KITTI_infos_train.pkl",
"@class_names": ["Car", "Pedestrian", "Cyclist"],
"prep": {
"@training": True, # set this to True to return targets
"@augment_dict": None,
"@filter_label_dict": {
"keep_classes": ["Car", "Pedestrian", "Cyclist"],
"min_num_pts": -1,
"label_range": [0, -32, -3, 52.8, 32.0, 1],
# [min_x, min_y, min_z, max_x, max_y, max_z] FIMU
},
"@feature_map_size": [1, 200, 176] # TBD
}
}
data_cfg = TRAINDATA_cfg
voxelizer = voxelizer_builder.build(VOXELIZER_cfg)
target_assigner = target_assigner_builder.build(TARGETASSIGNER_cfg)
dataloader = build(data_cfg,
ext_dict={
"voxelizer": voxelizer,
"target_assigner": target_assigner,
"feature_map_size": [1, 200, 176]
})
return dataloader
shuffle=False,
num_workers=data_cfg["num_workers"],
pin_memory=True,
collate_fn=merge_batch,
worker_init_fn=_worker_init_fn,
drop_last=False)
return dataloader
if __name__ == "__main__":
from incdet3.data.carladataset import CarlaDataset
from incdet3.configs.dev_cfg import cfg
from incdet3.builders import voxelizer_builder, target_assigner_builder
from tqdm import tqdm
data_cfg = cfg.TRAINDATA
voxelizer = voxelizer_builder.build(cfg.VOXELIZER)
print(voxelizer.grid_size)
target_assigner = target_assigner_builder.build(cfg.TARGETASSIGNER)
dataloader = build(data_cfg,
ext_dict={
"voxelizer": voxelizer,
"target_assigner": target_assigner,
"feature_map_size": [1, 200, 176]
})
for data in dataloader:
data = example_convert_to_torch(data)
print(data.keys())
print(type(data["voxels"]))
# from det3.utils.utils import save_pickle
# save_path = "./unit_tests/data/test_build_model_and_init.pkl"
# save_pickle(data, save_path)
def test_filt_label_by_range(self):
def limit_period_torch(val, offset=0.5, period=np.pi):
return val - torch.floor(val / period + offset) * period
data_cfg = Test_prep_infos.TRAINDATA_cfg
voxelizer = voxelizer_builder.build(Test_exclude_classes.VOXELIZER_cfg)
target_assigner = target_assigner_builder.build(Test_exclude_classes.TARGETASSIGNER_cfg)
dataloader = build(data_cfg,
ext_dict={
"voxelizer": voxelizer,
"target_assigner": target_assigner,
"feature_map_size": [1, 200, 176]
})
box_coder = target_assigner.box_coder
import torch
import torch.nn as nn
from det3.methods.second.ops.torch_ops import rotate_nms
from det3.dataloader.carladata import CarlaObj, CarlaLabel
from incdet3.utils.utils import filt_label_by_range
for data in dataloader:
tag = data["metadata"][0]["tag"]
if tag != "000006":
continue
label = data["metadata"][0]["label"]
cls_pred = torch.from_numpy(data["labels"]).cuda().float()
cls_pred *= (cls_pred >= 0).float()
cls_pred = cls_pred.long()
cls_pred = nn.functional.one_hot(cls_pred, num_classes=2+1)
cls_pred = cls_pred[..., 1:]
anchors = torch.from_numpy(data["anchors"]).cuda().float()
box_pred = torch.from_numpy(data["reg_targets"]).cuda().float()
box_pred = box_coder.decode(box_pred, anchors)
for box_preds, cls_preds in zip(box_pred, cls_pred):
box_preds = box_preds.float()
cls_preds = cls_preds.float()
total_scores = cls_preds
nms_func = rotate_nms
top_scores, top_labels = torch.max(
total_scores, dim=-1)
top_scores_keep = top_scores >= 0.5
top_scores = top_scores.masked_select(top_scores_keep)
box_preds = box_preds[top_scores_keep]
top_labels = top_labels[top_scores_keep]
boxes_for_nms = box_preds[:, [0, 1, 3, 4, 6]]
selected = nms_func(
boxes_for_nms,
top_scores,
pre_max_size=1000,
post_max_size=1000,
iou_threshold=0.3,
)
selected_boxes = box_preds[selected]
selected_labels = top_labels[selected]
selected_scores = top_scores[selected]
box_preds = selected_boxes
scores = selected_scores
label_preds = selected_labels
final_box_preds = box_preds
final_scores = scores
final_labels = label_preds
predictions_dict = {
"box3d_lidar": final_box_preds,
"scores": final_scores,
"label_preds": label_preds,
}
label_est = KittiLabel()
calib = KittiCalib(f"unit_tests/data/test_kittidata/training/calib/{tag}.txt").read_calib_file()
for box3d_lidar, label_preds, score in zip(
predictions_dict["box3d_lidar"],
predictions_dict["label_preds"],
predictions_dict["scores"]):
obj = KittiObj()
obj.type = "Car" if label_preds == 0 else "Pedestrian"
xyzwlhry_Flidar = box3d_lidar.cpu().numpy().flatten()
bcenter_Flidar = xyzwlhry_Flidar[:3].reshape(-1, 3)
bcenter_Fcam = calib.lidar2leftcam(bcenter_Flidar)
obj.x, obj.y, obj.z = bcenter_Fcam.flatten()
obj.w, obj.l, obj.h, obj.ry = xyzwlhry_Flidar[3:]
obj.truncated = 0
obj.occluded = 0
obj.alpha = 0
obj.bbox_l = 0
obj.bbox_t = 0
obj.bbox_r = 0
obj.bbox_b = 0
label_est.add_obj(obj)
label_est.current_frame = "Cam2"
label_filt = filt_label_by_range(label_est, valid_range=[20, -35.2, -3, 52.8, 12.85, -0.26], calib=calib)
self.assertTrue(len(label_filt) == 1)