in fucntion training_step, why x_start = (x_start) * 2 - 1

[cqlouis]

def training_step(self, batch):
        image, label = self.get_input(batch)
        x_start = label

        x_start = (x_start) * 2 - 1

why x_start = （x_start）*2 -1 ？

Can I use x_start directly, instead of two times and minus1？

[920232796]

Yes, it's ok. This operator makes x_start in [-1, 1] range, bacause the diffusion model's input is usually in [-1, 1].

[cqlouis]

Yes, it's ok. This operator makes x_start in [-1, 1] range, bacause the diffusion model's input is usually in [-1, 1].

Thanks for your quick reply.

Another question:
I was wondering if there should be some modification in the rest of coding on the condition of making x_start in [-1, 1] or not.

In Chinese: 把x_start 转化为[-1,1]和不转换，直接使用其值在[0,1]，这两种情况，在代码上，是不是有什么地方需要做对应的修改？

[920232796]

No. You can train the model directly, and this operator only has a little change on the validation results.

[cqlouis]

No. You can train the model directly, and this operator only has a little change on the validation results.

Thanks！