Add support for Wan2.X I2V models

[avjves]

What?

Adds support for Wan2.X (Wan2.1 and Wan2.2) I2V models.

How?

Ports the model implementation from diffusers and adds SP support. Both Ulysses and Ring attention work. Support is gated behind diffusers version 0.35.2. For now it only supports the I2V models / task, but future tasks will be added in the future. The settings are set to be as close as possible to the original Wan2.2/Wan2.1 repo.

Tests

Wan2.2 with 8 MI300Xs, running:

torchrun --nproc_per_node=8 examples/wan_i2v_example.py --height 720 --width 1280 --num_frames 81 --model Wan-AI/Wan2.2-I2V-A14B-Diffusers --ulysses_degree 8 --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." --num_inference_steps 40 --use_torch_compile --seed 42 --img_file_path https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/wan_i2v_input.JPG

output:
https://github.com/user-attachments/assets/d5ac29f8-a54d-43b6-95ae-67d431c42263

Wan2.1 with 8 MI300Xs, running:

torchrun --nproc_per_node=8 examples/wan_i2v_example.py --height 720 --width 1280 --num_frames 81 --model Wan-AI/Wan2.1-I2V-14B-720P-Diffusers --ulysses_degree 8 --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." --num_inference_steps 40 --use_torch_compile --seed 42 --img_file_path https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/wan_i2v_input.JPG

output:
https://github.com/user-attachments/assets/2ed172b6-ba0a-4b5a-8802-dc5372350969

[jcaraban]

Looks more straightforward than I expected ✅
Would be great to test other vendor hardware before merge, to be extra sure.

[avjves]

Here's a sample from Wan 2.1 with 8xH100s:

wan_2_1_h100.mp4

[nuclearwu]

@avjves I ran it on the npu and it reported insufficient memory