msamogh/nonechucks
Deal with bad samples in your dataset dynamically, use Transforms as Filters, and more!
PythonMIT
Issues
- 4
ValueError: dataset attribute should not be set after SafeDataLoader is initialized
#28 opened by neeek2303 - 0
support newer versions of pytorch
#37 opened by yairVanti - 10
Memory leak
#13 opened by shrubb - 1
Generator keyword error
#35 opened by KyriaAnnwyn - 2
Colab issue
#33 opened by ValdemarQ - 0
- 2
Pytorch's IterableDataset
#24 opened by jungerm2 - 1
- 0
- 3
import nonechucks fails for torch1.4.0+cu100
#25 opened by Hvitgar - 0
Potential bug in `_reset_index`
#31 opened by mdfirman - 5
- 0
Skip filtering step for safe samples
#27 opened by despoisj - 2
Not compatible with PyTorch v1.2 any more
#17 opened by iyah4888 - 1
- 3
KeyError: (<function SafeDataset.__getitem__ at 0x0000025B61310B70>, (1527,), frozenset())
#20 opened by Baldins - 2
- 1
Not compatible with PyTorch 1.2.0
#16 opened by ezyang - 2
Overriding torch's classes
#12 opened by shigabeev - 9
batch size reduction
#15 opened by Baldins - 2
Override some functions of the original dataset
#14 opened by yanxu1per - 7
How to handle TypError using nonechucks
#9 opened by simongeek - 0
Write tests
#11 opened by msamogh - 0
- 6
- 1
Failed to shuffle data across epoches
#8 opened by chadHGY - 8
- 3
- 11
- 3
- 8
- 2
Unable to import nonechucks
#1 opened by sannawag