Hu Di

correct cfg_name in demo/ipu_train_example.sh

1

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand...

add mixup to dataset.pipeline

1

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand...

Different samples for eval train test

1

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand...

[Bug]在cpu环境中，dataloder会出现段错误

1

推荐使用英语模板 Bug report，以便你的问题帮助更多人。 ### 描述 bug 如题 ### 复现流程 模块版本： torch==1.10.0+cpu torchvision==0.11.1 numpy==1.19.5 ```shell python3 dataloader_problem.py example.py ``` ### 相关信息 1. `pip list | grep "mmcv\|mmcls\|^torch"` 命令的输出 Package Version Editable...

could you please provide your training code, I want to learn it

It seems like there were a few wrong annos in the file.

I have tried training directly with 800x800, the result is not as well as training with 512x512 and fine tuning with 800x800. why this happened?

It seems like you will work out a new set of scales for activations on every batch, and use the new set of scales to quantize and dequantize activations?

lol