Buckets:

hf-doc-build
/

doc

Files
xet
hf-doc-build/doc / diffusers /v0.43.0 /zh /using-diffusers /text-img2vid.html
mishig
2 days ago
download
raw
53.9 kB
<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;视频生成&quot;,&quot;local&quot;:&quot;video-generation&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Pipeline 参数&quot;,&quot;local&quot;:&quot;pipeline-parameters&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;num_frames&quot;,&quot;local&quot;:&quot;numframes&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;guidance_scale&quot;,&quot;local&quot;:&quot;guidancescale&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;negative_prompt&quot;,&quot;local&quot;:&quot;negativeprompt&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;减少内存使用&quot;,&quot;local&quot;:&quot;reduce-memory-usage&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;推理速度&quot;,&quot;local&quot;:&quot;inference-speed&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/>
		<link href="/docs/diffusers/main/en/_app/immutable/entry/start.0S0_iEoG.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/I9eooN77.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/Bupkjr9f.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/entry/app.CPdOxCbw.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/CyvZ3FS5.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/l0QnnOED.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/DnVIJfC5.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/DsnmJJEf.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/nodes/0.DteTvmQZ.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/CoSrdQax.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/nodes/407.BqN650YR.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/DSPDl1gz.js" rel="modulepreload">
		<link href="/docs/diffusers/main/en/_app/immutable/chunks/D5eijdnr.js" rel="modulepreload">
		<!--g33fzy--><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Video generation&quot;,&quot;local&quot;:&quot;video-generation&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Pipeline parameters&quot;,&quot;local&quot;:&quot;pipeline-parameters&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;num_frames&quot;,&quot;local&quot;:&quot;numframes&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;guidance_scale&quot;,&quot;local&quot;:&quot;guidancescale&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3},{&quot;title&quot;:&quot;negative_prompt&quot;,&quot;local&quot;:&quot;negativeprompt&quot;,&quot;sections&quot;:[],&quot;depth&quot;:3}],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Reduce memory usage&quot;,&quot;local&quot;:&quot;reduce-memory-usage&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Inference speed&quot;,&quot;local&quot;:&quot;inference-speed&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/><!---->
		<link href="/docs/diffusers/main/en/_app/immutable/assets/0.tn0RQdqM.css" rel="modulepreload"> <!--[--><!--[0--><!--[--><!--[0--><!--[--><p></p> <div class="items-center shrink-0 min-w-[100px] max-sm:min-w-[50px] justify-end ml-auto flex" style="float: inline-end; margin-inline-start: 10px; display: inline-flex; position: relative; z-index: 10;"><div dir="ltr" class="inline-flex rounded-md max-sm:rounded-sm"><button class="inline-flex items-center gap-1 h-7 max-sm:h-7 px-2 max-sm:px-1.5 text-sm font-medium text-gray-800 border border-r-0 rounded-l-md max-sm:rounded-l-sm border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-live="polite"><span class="inline-flex items-center justify-center rounded-md p-0.5 max-sm:p-0 hover:text-gray-800 dark:hover:text-gray-200"><svg class="sm:size-3.5 size-3" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----></span> <span dir="auto">Copy page</span></button> <button class="inline-flex items-center justify-center w-6 max-sm:w-5 h-7 max-sm:h-7 disabled:pointer-events-none text-sm text-gray-500 hover:text-gray-700 dark:hover:text-white rounded-r-md max-sm:rounded-r-sm border border-l transition border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-haspopup="menu" aria-expanded="false" aria-label="Open copy menu"><svg class="transition-transform text-gray-400 overflow-visible sm:size-3.5 size-3 rotate-0" width="1em" height="1em" viewBox="0 0 12 7" fill="none" xmlns="http://www.w3.org/2000/svg"><path d="M1 1L6 6L11 1" stroke="currentColor"></path></svg><!----></button></div> <!--[-1--><!--]--></div><!----> <!--[0--><h1 class="relative group"><a id="video-generation" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#video-generation"><span><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg><!----></span></a> <span><!---->视频生成<!----></span></h1><!--]--><!----> <p>视频生成模型将图像生成（可视为单帧视频）扩展至同时处理与空间和时间相关的数据。确保所有这些文本、空间和时间数据在帧与帧之间保持一致且对齐，是生成长时长、高分辨率视频的一大挑战。</p> <p>现代视频模型采用扩散 Transformer（DiT）架构来应对这一挑战。该架构降低了计算成本，并能够更高效地扩展至更大规模、更高质量的图像和视频数据。</p> <p>请查看下方这些视频模型的部分能力展示。</p> <div class="flex space-x-2 items-center my-1.5 mr-8 h-7 !pl-0 -mx-3 md:mx-0"><!--[--><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd border-gray-800 bg-black dark:bg-gray-700 text-white">Wan2.1</div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">HunyuanVideo</div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">LTX-Video</div><!--]--></div> <div class="language-select"><div><div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600  " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> 已复制</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-comment"># pip install ftfy</span>
<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">import</span> numpy <span class="hljs-keyword">as</span> np
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> AutoModel, WanPipeline
<span class="hljs-keyword">from</span> diffusers.hooks.group_offloading <span class="hljs-keyword">import</span> apply_group_offloading
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video, load_image
<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> UMT5EncoderModel

text_encoder = UMT5EncoderModel.from_pretrained(<span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, subfolder=<span class="hljs-string">&quot;text_encoder&quot;</span>, dtype=torch.bfloat16)
vae = AutoModel.from_pretrained(<span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, subfolder=<span class="hljs-string">&quot;vae&quot;</span>, dtype=torch.float32)
transformer = AutoModel.from_pretrained(<span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, subfolder=<span class="hljs-string">&quot;transformer&quot;</span>, dtype=torch.bfloat16)

<span class="hljs-comment"># group-offloading</span>
onload_device = torch.device(<span class="hljs-string">&quot;cuda&quot;</span>)
offload_device = torch.device(<span class="hljs-string">&quot;cpu&quot;</span>)
apply_group_offloading(text_encoder,
    onload_device=onload_device,
    offload_device=offload_device,
    offload_type=<span class="hljs-string">&quot;block_level&quot;</span>,
    num_blocks_per_group=<span class="hljs-number">4</span>
)
transformer.enable_group_offload(
    onload_device=onload_device,
    offload_device=offload_device,
    offload_type=<span class="hljs-string">&quot;leaf_level&quot;</span>,
    use_stream=<span class="hljs-literal">True</span>
)

pipeline = WanPipeline.from_pretrained(
    <span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>,
    vae=vae,
    transformer=transformer,
    text_encoder=text_encoder,
    dtype=torch.bfloat16
)
pipeline.to(<span class="hljs-string">&quot;cuda&quot;</span>)  <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span>

prompt = <span class="hljs-string">&quot;&quot;&quot;
The camera rushes from far to near in a low-angle shot, 
revealing a white ferret on a log. It plays, leaps into the water, and emerges, as the camera zooms in 
for a close-up. Water splashes berry bushes nearby, while moss, snow, and leaves blanket the ground. 
Birch trees and a light blue sky frame the scene, with ferns in the foreground. Side lighting casts dynamic 
shadows and warm highlights. Medium composition, front view, low angle, with depth of field.
&quot;&quot;&quot;</span>
negative_prompt = <span class="hljs-string">&quot;&quot;&quot;
Bright tones, overexposed, static, blurred details, subtitles, style, works, paintings, images, static, overall gray, worst quality, 
low quality, JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn faces, deformed, disfigured, 
misshapen limbs, fused fingers, still picture, messy background, three legs, many people in the background, walking backwards
&quot;&quot;&quot;</span>

output = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    num_frames=<span class="hljs-number">81</span>,
    guidance_scale=<span class="hljs-number">5.0</span>,
).frames[<span class="hljs-number">0</span>]
export_to_video(output, <span class="hljs-string">&quot;output.mp4&quot;</span>, fps=<span class="hljs-number">16</span>)<!----></pre></div><!----></div><!----> <div hidden=""><div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600  " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> 已复制</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers importAutoModel, HunyuanVideoPipeline
<span class="hljs-keyword">from</span> diffusers.quantizers <span class="hljs-keyword">import</span> PipelineQuantizationConfig
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video

<span class="hljs-comment"># quantize weights to int4 with bitsandbytes</span>
pipeline_quant_config = PipelineQuantizationConfig(
  quant_backend=<span class="hljs-string">&quot;bitsandbytes_4bit&quot;</span>,
  quant_kwargs={
    <span class="hljs-string">&quot;load_in_4bit&quot;</span>: <span class="hljs-literal">True</span>,
    <span class="hljs-string">&quot;bnb_4bit_quant_type&quot;</span>: <span class="hljs-string">&quot;nf4&quot;</span>,
    <span class="hljs-string">&quot;bnb_4bit_compute_dtype&quot;</span>: torch.bfloat16
    },
  components_to_quantize=<span class="hljs-string">&quot;transformer&quot;</span>
)

pipeline = HunyuanVideoPipeline.from_pretrained(
    <span class="hljs-string">&quot;hunyuanvideo-community/HunyuanVideo&quot;</span>,
    quantization_config=pipeline_quant_config,
    dtype=torch.bfloat16,
)

<span class="hljs-comment"># model-offloading and tiling</span>
pipeline.enable_model_cpu_offload()
pipeline.vae.enable_tiling()

prompt = <span class="hljs-string">&quot;A fluffy teddy bear sits on a bed of soft pillows surrounded by children&#x27;s toys.&quot;</span>
video = pipeline(prompt=prompt, num_frames=<span class="hljs-number">61</span>, num_inference_steps=<span class="hljs-number">30</span>).frames[<span class="hljs-number">0</span>]
export_to_video(video, <span class="hljs-string">&quot;output.mp4&quot;</span>, fps=<span class="hljs-number">15</span>)<!----></pre></div><!----></div><!----> <div hidden=""><div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600  " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> 已复制</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> LTXPipeline, AutoModel
<span class="hljs-keyword">from</span> diffusers.hooks <span class="hljs-keyword">import</span> apply_group_offloading
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video

<span class="hljs-comment"># fp8 layerwise weight-casting</span>
transformer = AutoModel.from_pretrained(
    <span class="hljs-string">&quot;Lightricks/LTX-Video&quot;</span>,
    subfolder=<span class="hljs-string">&quot;transformer&quot;</span>,
    dtype=torch.bfloat16
)
transformer.enable_layerwise_casting(
    storage_dtype=torch.float8_e4m3fn, compute_dtype=torch.bfloat16
)

pipeline = LTXPipeline.from_pretrained(<span class="hljs-string">&quot;Lightricks/LTX-Video&quot;</span>, transformer=transformer, dtype=torch.bfloat16)

<span class="hljs-comment"># group-offloading</span>
onload_device = torch.device(<span class="hljs-string">&quot;cuda&quot;</span>)
offload_device = torch.device(<span class="hljs-string">&quot;cpu&quot;</span>)
pipeline.transformer.enable_group_offload(onload_device=onload_device, offload_device=offload_device, offload_type=<span class="hljs-string">&quot;leaf_level&quot;</span>, use_stream=<span class="hljs-literal">True</span>)
apply_group_offloading(pipeline.text_encoder, onload_device=onload_device, offload_type=<span class="hljs-string">&quot;block_level&quot;</span>, num_blocks_per_group=<span class="hljs-number">2</span>)
apply_group_offloading(pipeline.vae, onload_device=onload_device, offload_type=<span class="hljs-string">&quot;leaf_level&quot;</span>)

prompt = <span class="hljs-string">&quot;&quot;&quot;
A woman with long brown hair and light skin smiles at another woman with long blonde hair. The woman with brown hair wears a black jacket and has a small, barely noticeable mole on her right cheek. The camera angle is a close-up, focused on the woman with brown hair&#x27;s face. The lighting is warm and natural, likely from the setting sun, casting a soft glow on the scene. The scene appears to be real-life footage
&quot;&quot;&quot;</span>
negative_prompt = <span class="hljs-string">&quot;worst quality, inconsistent motion, blurry, jittery, distorted&quot;</span>

video = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    width=<span class="hljs-number">768</span>,
    height=<span class="hljs-number">512</span>,
    num_frames=<span class="hljs-number">161</span>,
    decode_timestep=<span class="hljs-number">0.03</span>,
    decode_noise_scale=<span class="hljs-number">0.025</span>,
    num_inference_steps=<span class="hljs-number">50</span>,
).frames[<span class="hljs-number">0</span>]
export_to_video(video, <span class="hljs-string">&quot;output.mp4&quot;</span>, fps=<span class="hljs-number">24</span>)<!----></pre></div><!----></div><!----><!----></div><!----> <p>本指南将介绍视频生成的基础知识，例如应配置哪些参数以及如何降低其内存占用。</p> <blockquote class="tip"><p>如果您想进一步了解如何使用特定模型，请参阅其 pipeline API 模型卡。</p></blockquote> <!--[1--><h2 class="relative group"><a id="pipeline-parameters" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#pipeline-parameters"><span><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg><!----></span></a> <span><!---->Pipeline 参数<!----></span></h2><!--]--><!----> <p>在 pipeline 中，有多个参数会影响视频生成的质量或速度。尝试不同的参数值对于找到合适的质量与速度之间的平衡至关重要。</p> <!--[2--><h3 class="relative group"><a id="numframes" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#numframes"><span><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg><!----></span></a> <span><!---->num_frames<!----></span></h3><!--]--><!----> <p>帧是作为其他帧序列中的一部分进行播放的静态图像，用于产生运动效果或视频。使用 <code>num_frames</code> 控制每秒生成的帧数。增加 <code>num_frames</code> 可以提升感知到的运动流畅度和视觉连贯性，对于包含动态内容的视频尤为重要。较高的 <code>num_frames</code> 值还会增加视频时长。</p> <p>某些视频模型在推理时需要更具体的 <code>num_frames</code> 值。例如，<a href="/docs/diffusers/v0.43.0/zh/api/pipelines/hunyuan_video#diffusers.HunyuanVideoPipeline">HunyuanVideoPipeline</a> 建议使用 <code>(4 * num_frames) +1</code> 来计算 <code>num_frames</code>。请务必查阅 pipeline 的 API 模型卡，以确认是否有推荐的值。</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600  " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> 已复制</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> LTXPipeline
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video

pipeline = LTXPipeline.from_pretrained(
    <span class="hljs-string">&quot;Lightricks/LTX-Video&quot;</span>, dtype=torch.bfloat16
).to(<span class="hljs-string">&quot;cuda&quot;</span>)  <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span>

prompt = <span class="hljs-string">&quot;&quot;&quot;
A woman with long brown hair and light skin smiles at another woman with long blonde hair. The woman 
with brown hair wears a black jacket and has a small, barely noticeable mole on her right cheek. The 
camera angle is a close-up, focused on the woman with brown hair&#x27;s face. The lighting is warm and 
natural, likely from the setting sun, casting a soft glow on the scene. The scene appears to be 
real-life footage
&quot;&quot;&quot;</span>

video = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    width=<span class="hljs-number">768</span>,
    height=<span class="hljs-number">512</span>,
    num_frames=<span class="hljs-number">161</span>,
    decode_timestep=<span class="hljs-number">0.03</span>,
    decode_noise_scale=<span class="hljs-number">0.025</span>,
    num_inference_steps=<span class="hljs-number">50</span>,
).frames[<span class="hljs-number">0</span>]
export_to_video(video, <span class="hljs-string">&quot;output.mp4&quot;</span>, fps=<span class="hljs-number">24</span>)<!----></pre></div><!----> <!--[2--><h3 class="relative group"><a id="guidancescale" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#guidancescale"><span><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg><!----></span></a> <span><!---->guidance_scale<!----></span></h3><!--]--><!----> <p>引导尺度（或“cfg”）控制生成的帧在多大程度上遵循输入条件（文本、图像或两者兼有）。增大 <code>guidance_scale</code> 会生成与输入条件更相似、细节更丰富的帧，但可能会引入伪影并降低输出的多样性。较低的 <code>guidance_scale</code> 值会鼓励更宽松的提示词遵循，从而增加输出的多样性，但细节可能不够出色。如果该值过低，模型可能会完全忽略提示词，生成随机噪声。</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600  " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> 已复制</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> CogVideoXPipeline, CogVideoXTransformer3DModel
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video

pipeline = CogVideoXPipeline.from_pretrained(
  <span class="hljs-string">&quot;THUDM/CogVideoX-2b&quot;</span>,
  dtype=torch.float16
).to(<span class="hljs-string">&quot;cuda&quot;</span>)  <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span>

prompt = <span class="hljs-string">&quot;&quot;&quot;
A detailed wooden toy ship with intricately carved masts and sails is seen gliding smoothly over
a plush, blue carpet that mimics the waves of the sea. The ship&#x27;s hull is painted a rich brown, 
with tiny windows. The carpet, soft and textured, provides a perfect backdrop, resembling an 
oceanic expanse. Surrounding the ship are various other toys and children&#x27;s items, hinting at 
a playful environment. The scene captures the innocence and imagination of childhood, 
with the toy ship&#x27;s journey symbolizing endless adventures in a whimsical, indoor setting.
&quot;&quot;&quot;</span>

video = pipeline(
  prompt=prompt,
  guidance_scale=<span class="hljs-number">6</span>,
  num_inference_steps=<span class="hljs-number">50</span>
).frames[<span class="hljs-number">0</span>]
export_to_video(video, <span class="hljs-string">&quot;output.mp4&quot;</span>, fps=<span class="hljs-number">8</span>)<!----></pre></div><!----> <!--[2--><h3 class="relative group"><a id="negativeprompt" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#negativeprompt"><span><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg><!----></span></a> <span><!---->negative_prompt<!----></span></h3><!--]--><!----> <p>负向提示词（negative prompt）有助于排除你不想在生成视频中看到的内容。它通常用于提升生成视频的质量和一致性，通过将模型引导远离“模糊、扭曲、丑陋”等不良元素，从而生成更干净、更聚焦的视频。</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600  " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> 已复制</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-comment"># pip install ftfy</span>
<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> WanPipeline
<span class="hljs-keyword">from</span> diffusers.schedulers.scheduling_unipc_multistep <span class="hljs-keyword">import</span> UniPCMultistepScheduler
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video

vae = AutoencoderKLWan.from_pretrained(
  <span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, subfolder=<span class="hljs-string">&quot;vae&quot;</span>, dtype=torch.float32
)
pipeline = WanPipeline.from_pretrained(
  <span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, vae=vae, dtype=torch.bfloat16
)
pipeline.scheduler = UniPCMultistepScheduler.from_config(
  pipeline.scheduler.config, flow_shift=<span class="hljs-number">5.0</span>
)
pipeline.to(<span class="hljs-string">&quot;cuda&quot;</span>)  <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span>

pipeline.load_lora_weights(<span class="hljs-string">&quot;benjamin-paine/steamboat-willie-14b&quot;</span>, adapter_name=<span class="hljs-string">&quot;steamboat-willie&quot;</span>)
pipeline.set_adapters(<span class="hljs-string">&quot;steamboat-willie&quot;</span>)

pipeline.enable_model_cpu_offload()

<span class="hljs-comment"># use &quot;steamboat willie style&quot; to trigger the LoRA</span>
prompt = <span class="hljs-string">&quot;&quot;&quot;
steamboat willie style, golden era animation, The camera rushes from far to near in a low-angle shot, 
revealing a white ferret on a log. It plays, leaps into the water, and emerges, as the camera zooms in 
for a close-up. Water splashes berry bushes nearby, while moss, snow, and leaves blanket the ground. 
Birch trees and a light blue sky frame the scene, with ferns in the foreground. Side lighting casts 
dynamic shadows and warm highlights. Medium composition, front view, low angle, with depth of field.
&quot;&quot;&quot;</span>

output = pipeline(
  prompt=prompt,
  num_frames=<span class="hljs-number">81</span>,
  guidance_scale=<span class="hljs-number">5.0</span>,
).frames[<span class="hljs-number">0</span>]
export_to_video(output, <span class="hljs-string">&quot;output.mp4&quot;</span>, fps=<span class="hljs-number">16</span>)<!----></pre></div><!----> <!--[1--><h2 class="relative group"><a id="reduce-memory-usage" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#reduce-memory-usage"><span><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg><!----></span></a> <span><!---->减少内存使用<!----></span></h2><!--]--><!----> <p>像 <a href="/docs/diffusers/v0.43.0/zh/api/pipelines/hunyuan_video#diffusers.HunyuanVideoPipeline">HunyuanVideoPipeline</a> 和 <a href="/docs/diffusers/v0.43.0/zh/api/pipelines/wan#diffusers.WanPipeline">WanPipeline</a> 这样拥有 100 亿以上参数的最新视频模型需要大量内存，通常会超出消费级硬件的可用内存。Diffusers 提供了多种技术来降低这些大型模型的内存需求。</p> <blockquote class="tip"><p>有关其他节省内存技术的详细信息，请参阅 <a href="../optimization/memory">减少内存使用</a> 指南。</p></blockquote> <p>其中一种技术是<a href="../optimization/memory#group-offloading">group-offloading</a>，它会在模型内部层组（例如<code>torch.nn.Sequential</code>）未被使用时，将其卸载到 CPU。这些层仅在计算需要时才会被加载，从而避免在 GPU 上存储<strong>所有</strong>模型组件。对于像<a href="/docs/diffusers/v0.43.0/zh/api/pipelines/wan#diffusers.WanPipeline">WanPipeline</a>这样的 14B 参数模型，group-offloading 可将所需显存降低至约 13GB。</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600  " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> 已复制</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-comment"># pip install ftfy</span>
<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">import</span> numpy <span class="hljs-keyword">as</span> np
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> AutoModel, WanPipeline
<span class="hljs-keyword">from</span> diffusers.hooks.group_offloading <span class="hljs-keyword">import</span> apply_group_offloading
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video, load_image
<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> UMT5EncoderModel

text_encoder = UMT5EncoderModel.from_pretrained(<span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, subfolder=<span class="hljs-string">&quot;text_encoder&quot;</span>, dtype=torch.bfloat16)
vae = AutoModel.from_pretrained(<span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, subfolder=<span class="hljs-string">&quot;vae&quot;</span>, dtype=torch.float32)
transformer = AutoModel.from_pretrained(<span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, subfolder=<span class="hljs-string">&quot;transformer&quot;</span>, dtype=torch.bfloat16)

<span class="hljs-comment"># group-offloading</span>
onload_device = torch.device(<span class="hljs-string">&quot;cuda&quot;</span>)
offload_device = torch.device(<span class="hljs-string">&quot;cpu&quot;</span>)
apply_group_offloading(text_encoder,
    onload_device=onload_device,
    offload_device=offload_device,
    offload_type=<span class="hljs-string">&quot;block_level&quot;</span>,
    num_blocks_per_group=<span class="hljs-number">4</span>
)
transformer.enable_group_offload(
    onload_device=onload_device,
    offload_device=offload_device,
    offload_type=<span class="hljs-string">&quot;leaf_level&quot;</span>,
    use_stream=<span class="hljs-literal">True</span>
)

pipeline = WanPipeline.from_pretrained(
    <span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>,
    vae=vae,
    transformer=transformer,
    text_encoder=text_encoder,
    dtype=torch.bfloat16
)
pipeline.to(<span class="hljs-string">&quot;cuda&quot;</span>)  <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span>

prompt = <span class="hljs-string">&quot;&quot;&quot;
The camera rushes from far to near in a low-angle shot, 
revealing a white ferret on a log. It plays, leaps into the water, and emerges, as the camera zooms in 
for a close-up. Water splashes berry bushes nearby, while moss, snow, and leaves blanket the ground. 
Birch trees and a light blue sky frame the scene, with ferns in the foreground. Side lighting casts dynamic 
shadows and warm highlights. Medium composition, front view, low angle, with depth of field.
&quot;&quot;&quot;</span>
negative_prompt = <span class="hljs-string">&quot;&quot;&quot;
Bright tones, overexposed, static, blurred details, subtitles, style, works, paintings, images, static, overall gray, worst quality, 
low quality, JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn faces, deformed, disfigured, 
misshapen limbs, fused fingers, still picture, messy background, three legs, many people in the background, walking backwards
&quot;&quot;&quot;</span>

output = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    num_frames=<span class="hljs-number">81</span>,
    guidance_scale=<span class="hljs-number">5.0</span>,
).frames[<span class="hljs-number">0</span>]
export_to_video(output, <span class="hljs-string">&quot;output.mp4&quot;</span>, fps=<span class="hljs-number">16</span>)<!----></pre></div><!----> <p>另一种减少内存占用的方法是考虑对模型进行量化，即将模型权重存储为更低精度的数据类型。不过，量化可能会影响视频质量，具体取决于所使用的视频模型。有关不同支持的量化后端的详细信息，请参阅量化<a href="../quantization/overview">概述</a>。</p> <p>以下示例使用 <a href="../quantization/bitsandbytes">bitsandbytes</a> 对模型进行量化。</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600  " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> 已复制</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-comment"># pip install ftfy</span>

<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> WanPipeline
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> AutoModel, WanPipeline
<span class="hljs-keyword">from</span> diffusers.quantizers <span class="hljs-keyword">import</span> PipelineQuantizationConfig
<span class="hljs-keyword">from</span> diffusers.schedulers.scheduling_unipc_multistep <span class="hljs-keyword">import</span> UniPCMultistepScheduler
<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> UMT5EncoderModel
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video

<span class="hljs-comment"># quantize transformer and text encoder weights with bitsandbytes</span>
pipeline_quant_config = PipelineQuantizationConfig(
  quant_backend=<span class="hljs-string">&quot;bitsandbytes_4bit&quot;</span>,
  quant_kwargs={<span class="hljs-string">&quot;load_in_4bit&quot;</span>: <span class="hljs-literal">True</span>},
  components_to_quantize=[<span class="hljs-string">&quot;transformer&quot;</span>, <span class="hljs-string">&quot;text_encoder&quot;</span>]
)

vae = AutoModel.from_pretrained(
  <span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, subfolder=<span class="hljs-string">&quot;vae&quot;</span>, dtype=torch.float32
)
pipeline = WanPipeline.from_pretrained(
  <span class="hljs-string">&quot;Wan-AI/Wan2.1-T2V-14B-Diffusers&quot;</span>, vae=vae, quantization_config=pipeline_quant_config, dtype=torch.bfloat16
)
pipeline.scheduler = UniPCMultistepScheduler.from_config(
  pipeline.scheduler.config, flow_shift=<span class="hljs-number">5.0</span>
)
pipeline.to(<span class="hljs-string">&quot;cuda&quot;</span>)  <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span>

pipeline.load_lora_weights(<span class="hljs-string">&quot;benjamin-paine/steamboat-willie-14b&quot;</span>, adapter_name=<span class="hljs-string">&quot;steamboat-willie&quot;</span>)
pipeline.set_adapters(<span class="hljs-string">&quot;steamboat-willie&quot;</span>)

pipeline.enable_model_cpu_offload()

<span class="hljs-comment"># use &quot;steamboat willie style&quot; to trigger the LoRA</span>
prompt = <span class="hljs-string">&quot;&quot;&quot;
steamboat willie style, golden era animation, The camera rushes from far to near in a low-angle shot, 
revealing a white ferret on a log. It plays, leaps into the water, and emerges, as the camera zooms in 
for a close-up. Water splashes berry bushes nearby, while moss, snow, and leaves blanket the ground. 
Birch trees and a light blue sky frame the scene, with ferns in the foreground. Side lighting casts 
dynamic shadows and warm highlights. Medium composition, front view, low angle, with depth of field.
&quot;&quot;&quot;</span>

output = pipeline(
  prompt=prompt,
  num_frames=<span class="hljs-number">81</span>,
  guidance_scale=<span class="hljs-number">5.0</span>,
).frames[<span class="hljs-number">0</span>]
export_to_video(output, <span class="hljs-string">&quot;output.mp4&quot;</span>, fps=<span class="hljs-number">16</span>)<!----></pre></div><!----> <!--[1--><h2 class="relative group"><a id="inference-speed" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#inference-speed"><span><svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg><!----></span></a> <span><!---->推理速度<!----></span></h2><!--]--><!----> <p><a href="https://docs.pytorch.org/tutorials/intermediate/torch_compile_tutorial.html" rel="nofollow">torch.compile</a> 可以通过使用优化后的内核来加速推理。首次编译耗时较长，但编译完成后速度会显著提升。最佳实践是只编译一次 pipeline，然后在不更改任何设置的情况下多次使用该 pipeline。任何更改（例如图像尺寸的变化）都会触发重新编译。</p> <p>以下示例会编译流水线中的 transformer，并使用 <code>"max-autotune"</code> 模式以最大化性能。</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5   text-gray-600  " title="code excerpt" type="button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg><!----> <div class=" absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0 "><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent;"></div> 已复制</div><!----></button><!----></div> <pre class="language-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> CogVideoXPipeline, CogVideoXTransformer3DModel
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video

pipeline = CogVideoXPipeline.from_pretrained(
  <span class="hljs-string">&quot;THUDM/CogVideoX-2b&quot;</span>,
  dtype=torch.float16
).to(<span class="hljs-string">&quot;cuda&quot;</span>)  <span class="hljs-comment"># or &quot;mps&quot;, &quot;xpu&quot;, &quot;cpu&quot;</span>

<span class="hljs-comment"># torch.compile</span>
pipeline.transformer.to(memory_format=torch.channels_last)
pipeline.transformer = torch.<span class="hljs-built_in">compile</span>(
    pipeline.transformer, mode=<span class="hljs-string">&quot;max-autotune&quot;</span>, fullgraph=<span class="hljs-literal">True</span>
)

prompt = <span class="hljs-string">&quot;&quot;&quot;
A detailed wooden toy ship with intricately carved masts and sails is seen gliding smoothly over a plush, blue carpet that mimics the waves of the sea. 
The ship&#x27;s hull is painted a rich brown, with tiny windows. The carpet, soft and textured, provides a perfect backdrop, resembling an oceanic expanse. 
Surrounding the ship are various other toys and children&#x27;s items, hinting at a playful environment. The scene captures the innocence and imagination of childhood, 
with the toy ship&#x27;s journey symbolizing endless adventures in a whimsical, indoor setting.
&quot;&quot;&quot;</span>

video = pipeline(
  prompt=prompt,
  guidance_scale=<span class="hljs-number">6</span>,
  num_inference_steps=<span class="hljs-number">50</span>
).frames[<span class="hljs-number">0</span>]
export_to_video(video, <span class="hljs-string">&quot;output.mp4&quot;</span>, fps=<span class="hljs-number">8</span>)<!----></pre></div><!----> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/diffusers/blob/main/docs/source/en/using-diffusers/text-img2vid.md" target="_blank"><svg class="mr-1" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M31,16l-7,7l-1.41-1.41L28.17,16l-5.58-5.59L24,9l7,7z"></path><path d="M1,16l7-7l1.41,1.41L3.83,16l5.58,5.59L8,23l-7-7z"></path><path d="M12.419,25.484L17.639,6.552l1.932,0.518L14.351,26.002z"></path></svg><!----> <span><span class="underline">Update</span> on GitHub</span></a><!----> <p></p><!--]--><!----><!--]--><!--]--><!--]--> <!--[-1--><!--]--><!--]-->
			
			<script>
				{
					__sveltekit_1qsdbwf = {
						base: "/docs/diffusers/main/en",
						assets: "/docs/diffusers/main/en"
					};
					const element = document.currentScript.parentElement;
					Promise.all([
						import("/docs/diffusers/main/en/_app/immutable/entry/start.0S0_iEoG.js"),
						import("/docs/diffusers/main/en/_app/immutable/entry/app.CPdOxCbw.js")
					]).then(([kit, app]) => {
						kit.start(app, element, {
							node_ids: [0, 407],
							data: [null,null],
							form: null,
							error: null
						});
					});
				}
			</script>
		
Xet Storage Details

Size:: 53.9 kB
Xet hash:: d2ff424ff18ac252c76c320999341f9b4a98a92f5c382284f33038eb8f87ce1e
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.