Sage Attention for diffuser library

**Is your feature request related to a problem? No

**Describe the solution you'd like.**
A clear and concise description of what you want to happen.
Incorporate a way to add sage attention to the diffusers library: Flux pipeline, Wan pipeline, etc.

**Describe alternatives you've considered.**
None

**Additional context.**
When I incorporated sage attention in the flux pipeline (text to image) I achieved a 16% speed advantage vs no sage attention.
My environment was the same save for including / excluding sage attention in my 4 image benchmark creation.

How to incorporate sage attention? We must consider that this only applies to the Transformer. With this in mind I did the following to the FluxPipeline. Obviously there must be a way to do this via a variable of sorts so that we may/may not run it:

Need some kind of indicator to decide whether to include or not! This must be done before the denoising step in the model pipeline.
` import torch.nn.functional as F
sage_function = False
try:
from sageattention import sageattn
self.transformer.scaled_dot_product_attention = F.scaled_dot_product_attention = sageattn
sage_function = True
except (ImportError):
pass

# 6. Denoising loop
with self.progress_bar(total=num_inference_steps) as progress_bar:
for i, t in enumerate(timesteps):
if self.interrupt:
continue
`
After the denoising step we must remove sage attention else we get a VAE error due to Sage Attn wanting only torch.float16 or torch.bfloat16 dtypes which the VAE doesn't want:

` if output_type == "latent":
image = latents
else:
if sage_function:
self.transformer.scaled_dot_product_attention = F.scaled_dot_product_attention = torch._C._nn.scaled_dot_product_attention
`
Hopefully this helps.

Start by reviewing the Flux pipeline example and the requested coverage for Flux and Wan pipelines. Examine how the transformer and VAE stages surround the denoising loop, and determine how optional Sage Attention could be enabled only for transformer execution and removed before VAE processing. Done means the design works across the requested pipelines without breaking non-Sage execution.

Scritto dal modello di indicizzazione a partire dal testo della issue.