Panxjia
Panxjia
Thanks for your attention. This version can only set the batch to 1 for inference due to the generation of HSM. We will update a version supporting batchsize > 1...
Thanks for your attention, we did not use CRF and its variants.
around 23k images
please refer to [link](https://drive.google.com/drive/folders/1rd3iV9Xif2tRgofQWrL3qH1_lIFkicdI)
@chuangchuangtan I checked our code and models. The main differences compared with your configuration are: we used 8V100 GPUs and set bacth_size=1024, lr=0.005, lr_decay=10,15, total_epoch=20, cutmix_beta=1.0, cutmix_prob=0.5
Sorry for the confusing description. It means that we add SEM and SPA on the same model re-implemented by us, respectively. Since both methods work in the testing phase, it...