-
Thanks for sharing the code I was wondering what should I change to use this concept on my own data set and along side with the mixup concept presented in this reposotory
https://github.com/facebookr…
-
Hi, @iamhankai
I am trying to replicate the results of the TNT paper that have been fine-tuned to datasets such as CIFAR-10 and CIFAR-100. Could you give details about hyper-parameters used (like …
-
I use the deafult config and meet error, It seems like the parameters in config.py cannot be set to None，
I don't know why. please help me，many thanks
Traceback (most recent call last):
File "m…
-
Hi, I run following command to implement:
python -m torch.distributed.launch \
--nproc_per_node=8 \
--use_env main.py \
--model deit_base_patch16_224 \
--data-set INAT \
--batch-size 96 \
--l…
-
______________________________________________________________________________________________________________________________________________________
#####################
# TRAIN THE MODEL #
##…
-
Hi. In mean teacher, the consistency weight is 100. but in this work, all consistency weight is 1. Isn't this value too small?
Can you tell me the details of setting this parameter, because I have se…
-
Hello
Thank you for the nice work.
I have one question related to the loss.
When reading your paper, I could not figure out the difference between the 2 losses. Indeed, CutMix has the same loss th…
-
Can Augmix boost training on regular classification problems as Mixup or Cutmix do?
-
### 🚀 The feature
Currently they seem to only support old style `transform=`, `target_transform=`. Joint transforms are needed for augmentations. I guess these datasets just need to pass in `transfor…
-
您好，请问数据增强支持Cutmix、Grid Mask吗？