LL-XSJ

Training related issues

1

Author, why am I training very slowly with 8 cards and batch size=8, and when I set batch size=16, it will report a graphics memory error? Why is this 

About Training Your Own Dataset

1

May I ask which configuration file in the model folder was used to train the weight of your DN-DAB Deformable DETR (Deformable Encoder Only)?

请问作者怎么用你的代码训练自己的数据集呢？

initialization error

2

AssertionError: vision_transformers not found

训练自己的数据集出现错误

16

self._epoch = checkpoint['meta']['epoch'] KeyError: 'meta'