Dose the batch have to be set 1?

[5huanghuai]

I have a question: does b have to be set to 1?

[kkkls]

During training, the batch size can be values such as 2, 4, or 8. During testing, a batch size of 1 is typically used for evaluation.

[ahmad-573]

@kkkls Here

EVSSM/models/EVSSM.py

Line 286 in 5098a52

assert b == 1

the batch size is checked to be equal to 1. Can you explain why? That is why its not possible to train on b>1

[kkkls]

@kkkls Here   这里

EVSSM/models/EVSSM.py

Line 286 in 5098a52

assert b == 1
the batch size is checked to be equal to 1. Can you explain why? That is why its not possible to train on b>1
将检查批处理大小是否等于 1。您能解释一下原因吗？这就是为什么不可能在 b>1 上训练的原因

Hello, this version is for inference/testing, so the batch size is set to 1.
You can find the training code here:
EVSSM/basicsr/models/archs/EVSSM_arch.py