Use of different optimizers in SegFormer

Open anirbala98 opened this issue 2 years ago • 0 comments

Keeping other parameters unchanged, if I try to use SGD instead of AdamW optimizer for Segformer model(modified with a Swin backbone), the results are so bad. I do get decent results with AdamW though. Only AdamW works with Segformer?

Dec 19 '23 07:12 anirbala98