Kang Zhao

found some difference of dfl loss compared to the original Gfocal loss your dfl loss: original dfl loss: weight_right should be label minus dis_left rather than 1 minus dis_left. any...

reproducibility problems

1

Hi, Great works! I've tried to reproduce the results on imagenet data (10% labeled) with run_ddp_example.sh but got about 1% to 2% top-1 acc downgraded results. I saw that there...

Gradient check fails??

2

I am able to build setup.py and run test_GPUpool without errors. But gradient check results are always wrong. And I got the following warnings: UserWarning: At least one of the...