fhong-jpg

Can I skip trainning and test with the model you provided in FewX?

1

I download the file provide in the page about [FewX](https://github.com/fanq15/FewX):  will that file work in this situation?

Can anyone share me the trained model? Thanks.

2

The trained model is .pth file, like "Outputs/fsod_save_dir/ckpt/model_step59999.pth". Hope the repository owner can @fanq15

请问使用DistributedSampler，各个GPU的数据是如何分配的？是连续(互不相同)的还是相同的？

1

我仿照了您的方法实现了一次分布式训练：发现单机单卡和多机多卡完成相同次数epoch的时间差不多，遂有所问。