4G显存部署Flux，2分钟Wan2.1-14B视频生成，DiffSynth-Engine引擎开源！

pip3 install diffsynth-engine

from diffsynth_engine.pipelines import FluxImagePipeline, FluxModelConfig
from diffsynth_engine import fetch_model
model_path = fetch_model("muse/flux-with-vae", revision="20240902173035", path="flux1-dev-with-vae.safetensors")
config = FluxModelConfig(dit_path=model_path)
pipe = FluxImagePipeline.from_pretrained(config, offload_mode="cpu_offload").eval()
image = pipe(
  prompt="an astronaut on a horse, on the moon, black and white leica 1 9 7 0's photography, grainy, contrasty",
  width=1024,
  height=1024,
  num_inference_steps=30,
  seed=42,
)
image.save("flux_txt2img.png")

from diffsynth_engine.pipelines import FluxImagePipeline, FluxModelConfig
from diffsynth_engine import fetch_model
model_path = fetch_model("muse/flux-with-vae", revision="20240902173035", path="flux1-dev-with-vae.safetensors")
config = FluxModelConfig(dit_path=model_path)
pipe = FluxImagePipeline.from_pretrained(config, offload_mode="sequential_cpu_offload").eval()
image = pipe(
  prompt="an astronaut on a horse, on the moon, black and white leica 1 9 7 0's photography, grainy, contrasty",
  width=1024,
  height=1024,
  num_inference_steps=30,
  seed=42,
)
image.save("flux_txt2img.png")

from diffsynth_engine.pipelines import WanVideoPipeline, WanModelConfig
from diffsynth_engine.utils.download import fetch_model
from diffsynth_engine.utils.video import save_video
config = WanModelConfig(
    model_path=fetch_model("muse/wan2.1-14b-bf16", path="dit.safetensors"),
    t5_path=fetch_model("muse/wan2.1-umt5", path="umt5.safetensors"),
    vae_path=fetch_model("muse/wan2.1-vae", path="vae.safetensors"),
)
pipe = WanVideoPipeline.from_pretrained(config)
video = pipe(
    prompt="纪实摄影风格画面，一只活泼的小狗在绿茵茵的草地上迅速奔跑。小狗毛色棕黄，两只耳朵立起，神情专注而欢快。阳光洒在它身上，使得毛发看上去格外柔软而闪亮。背景是一片开阔的草地，偶尔点缀着几朵野花，远处隐约可见蓝天和几片白云。透视感鲜明，捕捉小狗奔跑时的动感和四周草地的生机。中景侧面移动视角。",
    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",
    num_frames=41,
    width=848,
    height=480,
    seed=42,
)
save_video(video, "wan_t2v.mp4", fps=15)

from diffsynth_engine.pipelines import WanVideoPipeline, WanModelConfig
from diffsynth_engine.utils.download import fetch_model
from diffsynth_engine.utils.video import save_video
config = WanModelConfig(
    model_path=fetch_model("muse/wan2.1-14b-bf16", path="dit.safetensors"),
    t5_path=fetch_model("muse/wan2.1-umt5", path="umt5.safetensors"),
    vae_path=fetch_model("muse/wan2.1-vae", path="vae.safetensors"),
)
pipe = WanVideoPipeline.from_pretrained(config, parallelism=4, use_cfg_parallel=True)
video = pipe(
    prompt="纪实摄影风格画面，一只活泼的小狗在绿茵茵的草地上迅速奔跑。小狗毛色棕黄，两只耳朵立起，神情专注而欢快。阳光洒在它身上，使得毛发看上去格外柔软而闪亮。背景是一片开阔的草地，偶尔点缀着几朵野花，远处隐约可见蓝天和几片白云。透视感鲜明，捕捉小狗奔跑时的动感和四周草地的生机。中景侧面移动视角。",
    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",
    num_frames=41,
    width=848,
    height=480,
    seed=42,
)
save_video(video, "wan_t2v_tp.mp4", fps=15)