Jiang Ji

你可以在inference的代码里找找 里面肯定有提取提取输入图片的landmark的部分 我之前看到过 具体不愿找了。反正是用到了facexlib这个库里的函数

> @chhsieh0309 have you successfully trained the model with 2.5x scale factor? I was able to train with scale factor of integer values (e.g. 2x, 4x), but not sure how...

> Have you experimented with the idea, how did it go?

> Thanks for your interest. It depends on the dataset resolution. Larger input size will help with retraining details of those images with larger sizes. In our next paper, we...

> > > Thanks for your interest. It depends on the dataset resolution. Larger input size will help with retraining details of those images with larger sizes. In our next...

也许可以试试这个？ https://github.com/microsoft/genalog

应该要自己在纯文本的数据集上重新训练或者finetune吧

gt = ground truth

请问你有进一步研究吗 作者说他的训练用的ground truth 图片时经过锐化处理的 但是问他这个锐化时怎么做的 一直没有得到回复