Fix wan i2v condition shape mismatch

[DefTruth]

What does this PR do?

fixed #12499

Related issue: vipshop/cache-dit#291

Fix wan i2v condition shape. I think we should use num_latent_frames not the original num_frames after:

num_latent_frames = (num_frames - 1) // self.vae_scale_factor_temporal + 1

w/o this fix:

(cu129)➜ pipeline git:(dev) ✗ python3 run_wan_2.2_i2v.py
torch.Size([1, 1, 4, 90, 68])
torch.Size([1, 1, 84, 90, 68]) 4
torch.Size([1, 21, 4, 90, 68])
torch.Size([1, 4, 21, 90, 68]) torch.Size([1, 16, 81, 90, 68])
Traceback (most recent call last):
  File "/workspace/dev/vipshop/cache-dit/examples/pipeline/run_wan_2.2_i2v.py", line 154, in <module>
    video = run_pipe()
            ^^^^^^^^^^
  File "/workspace/dev/vipshop/cache-dit/examples/pipeline/run_wan_2.2_i2v.py", line 130, in run_pipe
    video = pipe(
            ^^^^^
  File "/usr/local/lib/python3.12/dist-packages/torch/utils/_contextlib.py", line 120, in decorate_context
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/workspace/dev/vipshop/diffusers/src/diffusers/pipelines/wan/pipeline_wan_i2v.py", line 705, in __call__
    latents_outputs = self.prepare_latents(
                      ^^^^^^^^^^^^^^^^^^^^^
  File "/workspace/dev/vipshop/diffusers/src/diffusers/pipelines/wan/pipeline_wan_i2v.py", line 487, in prepare_latents
    return latents, torch.concat([mask_lat_size, latent_condition], dim=1)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 21 but got size 81 for tensor number 1 in the list.

w/ this fix:

(cu129)➜ pipeline git:(dev) ✗ python3 run_wan_2.2_i2v.py
video_condition shape: torch.Size([1, 3, 21, 720, 544])
torch.Size([1, 1, 81, 90, 68])
torch.Size([1, 1, 1, 90, 68])
torch.Size([1, 1, 4, 90, 68])
torch.Size([1, 1, 84, 90, 68]) 4
torch.Size([1, 21, 4, 90, 68])
torch.Size([1, 4, 21, 90, 68]) torch.Size([1, 16, 21, 90, 68])
torch.Size([1, 16, 21, 90, 68]) torch.Size([1, 4, 21, 90, 68]) torch.Size([1, 16, 21, 90, 68])
 42%|████████████████████                                                 | 17/40 [10:17<13:55, 36.33s/it]

Reproduce

from: https://huggingface.co/Wan-AI/Wan2.2-I2V-A14B-Diffusers

import torch
import numpy as np
from diffusers import WanImageToVideoPipeline
from diffusers.utils import export_to_video, load_image

model_id = "Wan-AI/Wan2.2-I2V-A14B-Diffusers"
dtype = torch.bfloat16
device = "cuda"

pipe = WanImageToVideoPipeline.from_pretrained(model_id, torch_dtype=dtype)
pipe.to(device)


image = load_image(
    "https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/wan_i2v_input.JPG"
)
max_area = 480 * 832
aspect_ratio = image.height / image.width
mod_value = pipe.vae_scale_factor_spatial * pipe.transformer.config.patch_size[1]
height = round(np.sqrt(max_area * aspect_ratio)) // mod_value * mod_value
width = round(np.sqrt(max_area / aspect_ratio)) // mod_value * mod_value
image = image.resize((width, height))
prompt = "Summer beach vacation style, a white cat wearing sunglasses sits on a surfboard. The fluffy-furred feline gazes directly at the camera with a relaxed expression. Blurred beach scenery forms the background featuring crystal-clear waters, distant green hills, and a blue sky dotted with white clouds. The cat assumes a naturally relaxed posture, as if savoring the sea breeze and warm sunlight. A close-up shot highlights the feline's intricate details and the refreshing atmosphere of the seaside."

negative_prompt = "色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走"
generator = torch.Generator(device=device).manual_seed(0)
output = pipe(
    image=image,
    prompt=prompt,
    negative_prompt=negative_prompt,
    height=height,
    width=width,
    num_frames=81,
    guidance_scale=3.5,
    num_inference_steps=40,
    generator=generator,
).frames[0]
export_to_video(output, "i2v_output.mp4", fps=16)

[DefTruth]

@yiyixuxu hi~ can you take a look to this PR. I have found that your are the author of:

• [wan2.2] follow-up #12024