fix RuntimeError: Input type (float) and bias type (c10::Half) should be the same

[mwkldeveloper]

What does this PR do?

fix RuntimeError: Input type (float) and bias type (c10::Half) should be the same
According to : #6208

Fixes #
By adding autocast block to

diffusers/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py

Line 1034 in ff43dba

image = self.vae.decode(latents / self.vae.config.scaling_factor, return_dict=False, generator=generator)[

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[sayakpaul]

We don't include autocast blocks within the pipeline. It's the responsibility of the caller to have the pipeline called under the autocast context. Here's an example:

diffusers/examples/dreambooth/train_dreambooth_lora_sdxl.py

Line 1568 in 0532cec

with torch.cuda.amp.autocast():

[mwkldeveloper]

ok got it