Congratulations on the SNN you designed for defeating SwinTransformer with the same parameters! Looking forward to your code and pre-trained ckpts.

[jkhu29]

At the same time, I would like to kindly inquire about the training details, such as the dataset, pre-training, optimizer, and training time. Additionally, I am interested in the reasons behind the superior performance of this SNN network architecture compared to ANN. Thank you for your valuable contribution to the field of SNN!

[ZhangHanN1]

Thank you for your congratulations. I am currently organizing the codes for SGLFormer and will open source pre-trained ckpts and training parameter settings together.

[jkhu29]

LFTI :)

[ldkong1205]

Hi @ZhangHanN1, thanks for sharing your work! Any update for the code release?