Hzzone

You should have a desktop environment first... You can NOT display an INTERFACE with command line...sigh...

After a long time to try, I found a fast way to combine ddp with wandb, that is re-write `torch.distributed.launch` like below: ```python import sys import argparse import os import...

> @cavalleria my results wouldn't be compatible as I use a different face detector and dataset(s). > > I start with the fully trained baseline backbone and retrain with LR=0.01...

I have tried to fine-tune the models with motion blur or gaussian blur, but I failed to get better accuracy. I will try them from scratch with p=0.1, any update...

Unfortunately, it has not improved my performance:(

img_fake_att is the attention mask which is first proposed in [GANimation: Anatomically-aware Facial Animation from a Single Image, eccv 2018]. I am confused about why the code in https://github.com/ZhaoJ9014/High-Performance-Face-Recognition/blob/a0b6b6b269092051afb7a09a0100409cbccfa493/src/Look%20Across%20Elapse-%20Disentangled%20Representation%20Learning%20and%20Photorealistic%20Cross-Age%20Face%20Synthesis%20for%20Age-Invariant%20Face%20Recognition.TensorFlow/network.py#L150 has...

Besides, I have not found $L_{ip}$ in this code.....😥😥😥😥

I have no idea about your problem, but when I trained, it took me about 1s for 1 iteration with SSD and HKO data.

You should add the project path to your python environment.

detailly ``` docker run -it -p 8888:8888 -p 6006:6006 dl-docker:cpu sh -c 'jupyter notebook --ip=0.0.0.0 --no-browser' ``` Thanks a lot:)