chanlilong / 4d_net_pytorch Goto Github PK

Hi,how can I get the model for "model_KITTI_exp_CP2.pth"?

Thank you for your excellent work, but I have some questions about some code. Can you help me try to answer them?
(1) Is the selection of label ranges from cameras to LiDAR coordinate systems too rough?
label.columns = self.col_names df = label[["type","z","x","y","l","w","h","yaw"]] #Camera Coord #从camera视角转化为lidar视角， df.columns = ["type","x","y","z","l","w","h","yaw"] #LiDAR Coord df["y"] = (-1*df["y"]).copy(deep=True) df = df[df["type"]!="DontCare"] #排除某些类别 xy_filter = (df["x"].values <= self.xyz_range[3]) & (df["x"].values >= self.xyz_range[0]) & (df["y"].values <= self.xyz_range[4]) & (df["y"].values >= self.xyz_range[1]) df = df[xy_filter]
(2) After setting 'max_voxels', should we randomly select or design a selector?