|
发表于 2024-6-14 17:03:17
|
显示全部楼层
想使用deepspeech试试能否解决训练出的视频抖动问题,但是在python data_utils/process.py data/renquan1/renquan1.mp4 --asr deepspeech时,在数据集切分报错
1100 2.4177558422088623 -758.0857543945312
face tracking params saved
[INFO] ===== finished face tracking =====
[INFO] ===== extract flow =====
num data: 4012
load done
0%| | 0/4012 [00:00<?, ?it/s]D:\SyncTalk_240528\python\lib\site-packages\torch\functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at C:\actions-runner\_work\pytorch\pytorch\builder\windows\pytorch\aten\src\ATen\native\TensorShape.cpp:3484.)
return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
100%|████████████████████████████████████████████████████████████████████████████████████████████████████| 4012/4012 [07:35<00:00, 8.80it/s]
Traceback (most recent call last):
File "D:\SyncTalk_240528\data_utils\process.py", line 477, in <module>
extract_flow(base_dir, ori_imgs_dir, mask_imgs_dir, flow_dir)
File "D:\SyncTalk_240528\data_utils\process.py", line 340, in extract_flow
track_paths = sorted(glob.glob(os.path.join(track_dir, '*.npy')), key=lambda x: int(x.split('/')[-1].split('.')[0]))
File "D:\SyncTalk_240528\data_utils\process.py", line 340, in <lambda>
track_paths = sorted(glob.glob(os.path.join(track_dir, '*.npy')), key=lambda x: int(x.split('/')[-1].split('.')[0]))
ValueError: invalid literal for int() with base 10: 'renquan1\\flow_result\\2_0'
有大佬知道是什么问题吗
|
|