support wan-series models

examples/wanvideo/model_inference/Wan2.2-TI2V-5B.py

@@ -0,0 +1,42 @@
+import torch
+from PIL import Image
+from diffsynth.utils.data import save_video
+from diffsynth.pipelines.wan_video import WanVideoPipeline, ModelConfig
+from modelscope import dataset_snapshot_download
+
+pipe = WanVideoPipeline.from_pretrained(
+    torch_dtype=torch.bfloat16,
+    device="cuda",
+    model_configs=[
+        ModelConfig(model_id="Wan-AI/Wan2.2-TI2V-5B", origin_file_pattern="models_t5_umt5-xxl-enc-bf16.pth"),
+        ModelConfig(model_id="Wan-AI/Wan2.2-TI2V-5B", origin_file_pattern="diffusion_pytorch_model*.safetensors"),
+        ModelConfig(model_id="Wan-AI/Wan2.2-TI2V-5B", origin_file_pattern="Wan2.2_VAE.pth"),
+    ],
+)
+
+# Text-to-video
+video = pipe(
+    prompt="两只可爱的橘猫戴上拳击手套，站在一个拳击台上搏斗。",
+    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",
+    seed=0, tiled=True,
+    height=704, width=1248,
+    num_frames=121,
+)
+save_video(video, "video_1_Wan2.2-TI2V-5B.mp4", fps=15, quality=5)
+
+# Image-to-video
+dataset_snapshot_download(
+    dataset_id="DiffSynth-Studio/examples_in_diffsynth",
+    local_dir="./",
+    allow_file_pattern=["data/examples/wan/cat_fightning.jpg"]
+)
+input_image = Image.open("data/examples/wan/cat_fightning.jpg").resize((1248, 704))
+video = pipe(
+    prompt="两只可爱的橘猫戴上拳击手套，站在一个拳击台上搏斗。",
+    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",
+    seed=0, tiled=True,
+    height=704, width=1248,
+    input_image=input_image,
+    num_frames=121,
+)
+save_video(video, "video_2_Wan2.2-TI2V-5B.mp4", fps=15, quality=5)