Fix FluxControlNet img2img/inpainting with num_images_per_prompt>1

[Ricardo-M-L]

What does this PR do?

FluxControlNetImg2ImgPipeline and FluxControlNetInpaintPipeline already pre-resize control_image to batch_size * num_images_per_prompt, but only reshape control_mode to [1, 1]:

if control_mode is not None:
    control_mode = torch.tensor(control_mode).to(device, dtype=torch.long)
    control_mode = control_mode.reshape([-1, 1])

Inside FluxControlNetModel the union branch feeds control_mode to nn.Embedding and concatenates the result to encoder_hidden_states:

controlnet_mode_emb = self.controlnet_mode_embedder(controlnet_mode)
encoder_hidden_states = torch.cat([controlnet_mode_emb, encoder_hidden_states], dim=1)

So when num_images_per_prompt > 1 the embedded mode stays at [1, 1, inner_dim] while encoder_hidden_states is [B*num_images_per_prompt, seq, inner_dim], and the concat fails:

RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 2 but got size 1 for tensor number 1 in the list.

This matches the behaviour already implemented in FluxControlNetPipeline (non-img2img), which validates that control_mode is a single int and broadcasts it to the full image batch:

if not isinstance(control_mode, int):
    raise ValueError("For `FluxControlNet`, `control_mode` should be an `int` or `None`")
control_mode = torch.tensor(control_mode).to(device, dtype=torch.long)
control_mode = control_mode.view(-1, 1).expand(control_image.shape[0], 1)

This PR brings the img2img and inpainting pipelines in line with the same validation + broadcast so num_images_per_prompt > 1 works end-to-end.

Fixes #10741

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline?
• 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?

@sayakpaul @yiyixuxu