lyakaap
How about to change `while(indices is None or len(indices) == 0):` to `while(indices is None or indices.dim == 0):` I can fix this issue by this in pytorch==1.0.
"labels" is the category input images belong to.
Thanks for your interest to my work. You can download the checkpoint including optimizer state here: https://drive.google.com/file/d/1Z9G2yhYep0woJuKitaLJ2W06WUHWxbAv/view?usp=sharing
Then, this will do :) https://drive.google.com/file/d/1ySea-NJp_J0aWvma_WmVbc3Hnwf5LHUf/view
That's weird. It should match the performance if you execute inference code as described README. Please make sure that you evaluate with the private-set of phase-1.
Sorry, it seems that I have deleeted the weights of the models prior to stage2...
@GorillaSX You can check this branch for reproducing our results: https://github.com/lyakaap/ISC21-Descriptor-Track-1st/tree/reproduce I think you can reproduce by following: ``` python v107.py \ -a tf_efficientnetv2_m_in21ft1k --dist-url 'tcp://localhost:10001' --multiprocessing-distributed --world-size 1 --rank...
I remembered that I used 16 A100 GPUS.
Thanks for the answer! I found that some of the hyper-parameters don't exist in https://github.com/alipay/VCSL/blob/main/scripts/test_video_vta_tune.sh . Do you have any idea how to determine the search space of such parameters?