Z-Lee-corder

Thanks, author, I have used the learning rate 0.0025 and batch size 4, and the result of 3dssd_sasa has achieved 84%(AP10) and 85%(AP40) in kitti val set. But I found...

> Replace `if pkl_file.exists():` by if `os.path.exists(pkl_file)` Later on you might also have to comment the line `info_path = self.check_sequence_name_with_all_version(info_path)` and replace `sequence_file_tfrecord = sequence_file[:-9] + '_with_camera_labels.tfrecord` by `sequence_file_tfrecord =...

> Replace `if pkl_file.exists():` by if `os.path.exists(pkl_file)` Later on you might also have to comment the line `info_path = self.check_sequence_name_with_all_version(info_path)` and replace `sequence_file_tfrecord = sequence_file[:-9] + '_with_camera_labels.tfrecord` by `sequence_file_tfrecord =...

> Yes, I've found too that the waymo preprocessing cost a lot of memory. A partial solution would be to remove the multiprocessing by replacing > > ``` > with...

> I think maybe you gave the wrong link in that issue. Considering your codes from 51 to 61 in file centerpoint_MM_waymo.py, You should use the weight pre-trained on COCO...