Fix guess_mode in StableDiffusionXLControlNetPipeline

[YashalKanungo]

Fix guess_mode in StableDiffusionXLControlNetPipeline

What does this PR do?

It fixes a bug due to which guess mode was not usable in StableDiffusionXLControlNetPipeline.

Reproducing the original bug

Following the example at https://huggingface.co/diffusers/controlnet-canny-sdxl-1.0
if we set guess_mode=True

images = pipe(
  prompt,
  negative_prompt=negative_prompt,
  image=image,
  controlnet_conditioning_scale=controlnet_conditioning_scale,
  guess_mode=True
).images

it throws the error,

RuntimeError                              Traceback (most recent call last)
Cell In[57], line 24
     21 image = np.concatenate([image, image, image], axis=2)
     22 image = Image.fromarray(image)
---> 24 images = pipe(
     25     prompt, negative_prompt=negative_prompt, image=image, controlnet_conditioning_scale=controlnet_conditioning_scale, guess_mode=True
     26     ).images

RuntimeError: The size of tensor a (8192) must match the size of tensor b (4096) at non-singleton dimension 1

Fixes

During ControlNet inference for only the conditional batch, also only use the corresponding add_text_embeds and add_time_ids.

Tests

1. Able to generate some images with guess_mode=True
2. Ran test_controlnet_sdxl

pytest tests/pipelines/controlnet/test_controlnet_sdxl.py
...
============================================================= 62 passed, 3 skipped, 165 warnings in 97.70s (0:01:37) ==============================================================

Before submitting

• [] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• [Y] Did you read the contributor guideline?
• [Y] Did you read our philosophy doc (important for complex PRs)?
• [] Was this discussed/approved via a Github issue or the forum? Please add a link to it if that's the case.
• [] Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• [] Did you write any new necessary tests?

Who can review?

• Pipelines: @patrickvonplaten and @sayakpaul

[sayakpaul]

Being addressed in #4155. So, closing this one.

Thank you for your contribution!