ST3D: Cannot find pseudo label for frame

I am getting an error when running train.py, it seems to have something to do with PSEUDO_LABEL not being updated. The Traceback repeats for multiple frames, not just 002080 as seen below. I’ve also put the full output on this gist, in case the information below is not enough. Am I missing something? Thanks for any help!

Commands Run

$ NUM_GPUS=8
$ CONFIG_FILE=cfgs/da-waymo-kitti_models/pvrcnn_st3d/pvrcnn_st3d.yaml
$ bash scripts/dist_train.sh ${NUM_GPUS} --cfg_file ${CONFIG_FILE}

Error

[2021-07-21 15:05:09,022  train.py 168  INFO]  **********************Start training da-waymo-kitti_models/pvrcnn_st3d/pvrcnn_st3d(default)**********************
generate_ps_e0: 100%|████████████████████| 232/232 [03:14<00:00,  1.19it/s, pos_ps_box=0.000(0.000), ign_ps_box=15.000(14.899)]
Traceback (most recent call last):                                                                                             
  File "train.py", line 199, in <module>
    main()
  File "train.py", line 191, in main
    ema_model=None
  File "/home/user5/open-mmlab/ST3D/tools/train_utils/train_st_utils.py", line 157, in train_model_st
    dataloader_iter=dataloader_iter, ema_model=ema_model
  File "/home/user5/open-mmlab/ST3D/tools/train_utils/train_st_utils.py", line 42, in train_one_epoch_st
    target_batch = next(dataloader_iter)
  File "/home/user5/anaconda3/envs/st3d7/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 582, in __next__
    return self._process_next_batch(batch)
  File "/home/user5/anaconda3/envs/st3d7/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 608, in _process_next_batch
    raise batch.exc_type(batch.exc_msg)
ValueError: Traceback (most recent call last):
  File "/home/user5/anaconda3/envs/st3d7/lib/python3.7/site-packages/torch/utils/data/_utils/worker.py", line 99, in _worker_loop
    samples = collate_fn([dataset[i] for i in batch_indices])
  File "/home/user5/anaconda3/envs/st3d7/lib/python3.7/site-packages/torch/utils/data/_utils/worker.py", line 99, in <listcomp>
    samples = collate_fn([dataset[i] for i in batch_indices])
  File "/home/user5/open-mmlab/ST3D/tools/../pcdet/datasets/kitti/kitti_dataset.py", line 413, in __getitem__
    self.fill_pseudo_labels(input_dict)
  File "/home/user5/open-mmlab/ST3D/tools/../pcdet/datasets/dataset.py", line 146, in fill_pseudo_labels
    gt_boxes = self_training_utils.load_ps_label(input_dict['frame_id'])
  File "/home/user5/open-mmlab/ST3D/tools/../pcdet/utils/self_training_utils.py", line 221, in load_ps_label
    raise ValueError('Cannot find pseudo label for frame: %s' % frame_id)
ValueError: Cannot find pseudo label for frame: 002080

epochs:   0%|                                                                                           | 0/30 [04:05<?, ?it/s]

Environment

Python 3.7 CUDA 10.0 PyTorch 1.1 spconv 1.0 (commit 8da6f96) pcdet 0.2.0+73dda8c

About this issue

Original URL
State: closed
Created 3 years ago
Comments: 17 (6 by maintainers)

Most upvoted comments

@Liz66666 I followed @AndyYuan96’s advice on using Manager().dict() for PSEUDO_LABEL in self_training_utils.py.

from multiprocessing import Manager

PSEUDO_LABEL = Manager().dict()

Actually before doing this, I managed to run the training by adding the pkl.load code from check_already_exsit_pseudo_label to load_ps_label, but changed to the above solution and reran the training as it should be better.

hughjazzman on Aug 5, 2021

Yes.

>>> len(data['002829']['gt_boxes'])
36
>>> data['002829']['gt_boxes'][0]
array([-12.67259693,   2.17882085,   0.4152911 ,   4.12861061,
         1.84807122,   1.557392  ,   3.2007041 ,   1.        ,
         0.81863701])
>>> data['002829']['cls_scores']
array([0.93731874, 0.92046964, 0.9030588 , 0.9078988 , 0.8054646 ,
       0.80939907, 0.88196886, 0.85936695, 0.86304134, 0.8979211 ,
       0.8379814 , 0.6361919 , 0.71232146, 0.8505274 , 0.70933247,
       0.7508229 , 0.6692671 , 0.4446245 , 0.46104938, 0.15859666,
       0.3290353 , 0.1491799 , 0.24889277, 0.13674273, 0.13740459,
       0.11474051, 0.14078389, 0.14664578, 0.11770795, 0.25883213,
       0.11457415, 0.12707922, 0.13901637, 0.12712085, 0.16028087,
       0.18077461], dtype=float32)
>>> data['002829']['iou_scores']
array([0.818637  , 0.8166671 , 0.8074603 , 0.8054869 , 0.79433286,
       0.7942598 , 0.7878299 , 0.78539443, 0.7752307 , 0.76748055,
       0.75878924, 0.7525175 , 0.7474871 , 0.74367374, 0.7227673 ,
       0.7154487 , 0.67805743, 0.64385206, 0.61908954, 0.55265254,
       0.5234711 , 0.4574411 , 0.34837985, 0.33632967, 0.32025087,
       0.2999313 , 0.26421076, 0.24447218, 0.24083728, 0.21878265,
       0.21566562, 0.19372286, 0.19200402, 0.16185848, 0.1382765 ,
       0.10244821], dtype=float32)
>>> data['002829']['memory_counter']
array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
       0., 0.])

use one gpu to train is normal, for 8 gpus, you need to change the code, use multiprocessing library, use Manager().dict() to replace dict(), and then carefully change the pseudo label save and load code.

AndyYuan96 on Aug 4, 2021