search

liuyuan-pal / Gen6D

[ECCV2022] Gen6D: Generalizable Model-Free 6-DoF Object Pose Estimation from RGB Images
GNU General Public License v3.0
606 stars 75 forks source link

训练中途报错ValueError: 'a' cannot be empty unless no samples are taken #69

Open baldman12 opened 1 year ago

baldman12 commented 1 year ago

在训练detector和selector时会报错以下是详细信息： File "/media/dj/F054ABE054ABA7B4/code/py/6D/Gen6D-main/train_model.py", line 11, in trainer.run() File "/media/dj/F054ABE054ABA7B4/code/py/6D/Gen6D-main/train/trainer.py", line 111, in run train_data = next(train_iter) File "/home/dj/conda/envs/pyotrch/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 628, in next data = self._next_data() File "/home/dj/conda/envs/pyotrch/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1313, in _next_data return self._process_data(data) File "/home/dj/conda/envs/pyotrch/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1359, in _process_data data.reraise() File "/home/dj/conda/envs/pyotrch/lib/python3.9/site-packages/torch/_utils.py", line 543, in reraise raise exception ValueError: Caught ValueError in DataLoader worker process 4. Original Traceback (most recent call last): File "/home/dj/conda/envs/pyotrch/lib/python3.9/site-packages/torch/utils/data/_utils/worker.py", line 302, in _worker_loop data = fetcher.fetch(index) File "/home/dj/conda/envs/pyotrch/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 58, in fetch data = [self.dataset[idx] for idx in possibly_batched_index] File "/home/dj/conda/envs/pyotrch/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 58, in data = [self.dataset[idx] for idx in possibly_batched_index] File "/media/dj/F054ABE054ABA7B4/code/py/6D/Gen6D-main/dataset/train_dataset.py", line 992, in getitem database, que_ids = self._select_query(index) File "/media/dj/F054ABE054ABA7B4/code/py/6D/Gen6D-main/dataset/train_dataset.py", line 247, in _select_query database = self.name2database[np.random.choice(names)] File "mtrand.pyx", line 915, in numpy.random.mtrand.RandomState.choice ValueError: 'a' cannot be empty unless no samples are taken

好像是集合为空的问题，请问应该如何解决。谢谢

liuyuan-pal commented 1 year ago

你要不看看是不是漏了其中哪个数据集，看看这个names是不是空的~

baldman12 commented 1 year ago

|-- data |-- GenMOP |-- chair ... |-- LINEMOD |-- cat ... |-- shapenet |-- shapenet_cache |-- shapenet_render |-- shapenet_render_v1.pkl |-- co3d_256_512 |-- apple ... |-- google_scanned_objects |-- 06K3jXvzqIM ... |-- coco |-- train2017 我是这样布置的，是否还需要其他数据集。并且这个报错是训练中途报错，有时是第3次有时是第15次

EvdoTheo commented 1 year ago

Hello @baldman12, I had the same problem while I was training the detector and I solved this issue by modifying the value use_database_sample_prob in the config file from true to false. @liuyuan-pal can you explain the use of this value? Thank you!

Oskris commented 6 days ago

@baldman12 您好，请问可以给我参考一下你的 co3d、google scanned objects和shapeNet数据集吗，我直接下载的数据集组织形式并没有像作者的这样