jiarenyf
jiarenyf
> Hi @dsp6414 MMDet's model shares a different architecture from MMOCR, and we are working hard to refactor MMOCR recently to bridge the gap. Soon MMOCR will be able to...
@gaotongxiao @Mountchicken What about using backbone in mm segmentation ...
> We have tested on Pytorch1.1/1.3. Try lower version Pytorch or see [here](https://github.com/sgrvinod/a-PyTorch-Tutorial-to-Object-Detection/issues/24) Using pytorch 1.3.1, met the same problem ...
看起来是读数据越来越慢了
@pookiefoof @claforte @SusungHong @guanyingc
> Currently we don't have any dataset that needs heavy computation or I/O, so I don't think this would speedup training. BTW, the reason to use `num_workers=0` is I found...
模型忘记设置evel()了吧。。。