Yuchao Gu
Yuchao Gu
> Yeah I also compared it with gfpgan and GPEN and the results are much better on GPEN and then gfpgan. I also don't know how they compare the results...
> Kind of agree with @lazy-nurd. I tried to test the code on one of the faces given as an example and it seems like it doesn't achieve to give...
> Yeah I also compared it with gfpgan and GPEN and the results are much better on GPEN and then gfpgan. I also don't know how they compare the results...
> Hey Thanks for the repository, But the results are not really good enough for blind face restoration. GFP-Gan had similar issue of high smoothing. GPEN works much better in...
> > @ersheng-ai > > > > * github.com/AlexeyAB/darknet - 47.1% AP - 71.0% AP50 > > * Tianxiaomo/pytorch-YOLOv4 - 46.6% AP (**-0.5**) - 70.3% AP50 (**-0.7**) > > >...
We fix the kernel search space in stage 1 and stage 2 because large kernels in the first two stages will largely add the computation cost in training.
The pretrain process contains the ImageNet-pretraining (if you modify the search space). But for downstream tasks, like SOD, semantic segmentation, you do not need to pre-train the search space repeatedlly....
How much GPU did you use? Single GPU or Multiple GPU?
The problem is because you run the training in single GPU. If you run the algorithm in multiple GPU(>=2), then the .module will be added by nn.distributed.DistributedDataParallel. I will fix...
you can try "pip install mtcnn"