Hugging Face's logo

Buckets:
hf-doc-build
/
doc-dev

Files
xet
hf-doc-build/doc-dev / diffusers /pr_13881 /zh /_app /immutable /nodes /31.BdH_aBdU.js
HuggingFaceDocBuilder's picture
HuggingFaceDocBuilder
download
raw
48.2 kB
import"../chunks/DsnmJJEf.js";import{i as Q,h as v,C as A,H as h,b as m,a as i,E as F,s as E}from"../chunks/DdZvggmf.js";import{p as R,o as S,s as l,f as o,a as n,b as N,c,n as r}from"../chunks/BbekZcyp.js";import{H as J}from"../chunks/BcnRgdDK.js";const Y='{"title":"ParaAttention","local":"paraattention","sections":[{"title":"第一块缓存","local":"第一块缓存","sections":[],"depth":2},{"title":"fp8 量化","local":"fp8-量化","sections":[],"depth":2},{"title":"上下文并行性","local":"上下文并行性","sections":[],"depth":2},{"title":"基准测试","local":"基准测试","sections":[],"depth":2}],"depth":1}';var H=c('<meta name="hf:doc:metadata"/>'),z=c('<p>要在 FLUX.1-dev 上应用第一块缓存，请调用 <code>apply_cache_on_pipe</code>，如下所示。0.08 是 FLUX 模型的默认残差差异值。</p> <!> <table><thead><tr><th>优化</th><th>原始</th><th>FBCache rdt=0.06</th><th>FBCache rdt=0.08</th><th>FBCache rdt=0.10</th><th>FBCache rdt=0.12</th></tr></thead><tbody><tr><td>预览</td><td><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/flux-original.png" alt="Original"/></td><td><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/flux-fbc-0.06.png" alt="FBCache rdt=0.06"/></td><td><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/flux-fbc-0.08.png" alt="FBCache rdt=0.08"/></td><td><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/flux-fbc-0.10.png" alt="FBCache rdt=0.10"/></td><td><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/flux-fbc-0.12.png" alt="FBCache rdt=0.12"/></td></tr><tr><td>墙时间 (s)</td><td>26.36</td><td>21.83</td><td>17.01</td><td>16.00</td><td>13.78</td></tr></tbody></table> <p>First Block Cache 将推理速度降低到 17.01 秒，与基线相比，或快 1.55 倍，同时保持几乎零质量损失。</p>',1),x=c('<p>要在 HunyuanVideo 上应用 First Block Cache，请使用 <code>apply_cache_on_pipe</code>，如下所示。0.06 是 HunyuanVideo 模型的默认残差差值。</p> <!> <video controls=""><source src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/hunyuan-video-original.mp4" type="video/mp4"/> 您的浏览器不支持视频标签。</video> <small>HunyuanVideo 无 FBCache</small> <video controls=""><source src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/hunyuan-video-fbc.mp4" type="video/mp4"/> Your browser does not support the video tag.</video> <small>HunyuanVideo 与 FBCache</small> <p>First Block Cache 将推理速度降低至 2271.06 秒，相比基线快了 1.62 倍，同时保持了几乎为零的质量损失。</p>',3),j=c("<!> <!>",1),q=c("<!> <p>fp8 动态量化和 torch.compile 将推理速度降低至 7.56 秒，相比基线快了 3.48 倍。</p>",1),L=c("<!> <p>NVIDIA L20 GPU 仅有 48GB 内存，在编译后且如果未调用 <code>enable_model_cpu_offload</code> 时，可能会遇到内存不足（OOM）错误，因为 HunyuanVideo 在高分辨率和大量帧数运行时具有非常大的激活张量。对于内存少于 80GB 的 GPU，可以尝试降低分辨率和帧数来避免 OOM 错误。</p> <p>大型视频生成模型通常受注意力计算而非全连接层的瓶颈限制。这些模型不会从量化和 torch.compile 中显著受益。</p>",1),D=c('<p>以下代码示例结合了第一块缓存、fp8动态量化、torch.compile和上下文并行，以实现最快的推理速度。</p> <!> <p>保存到<code>run_flux.py</code>并使用<a href="https://pytorch.org/docs/stable/elastic/run.html" rel="nofollow">torchrun</a>启动。</p> <!> <p>推理速度降至8.20秒，相比基线快了3.21倍，使用2个NVIDIA L20 GPU。在4个L20上，推理速度为3.90秒，快了6.75倍。</p>',1),K=c('<p>以下代码示例结合了第一块缓存和上下文并行，以实现最快的推理速度。</p> <!> <p>保存到 <code>run_hunyuan_video.py</code> 并使用 <a href="https://pytorch.org/docs/stable/elastic/run.html" rel="nofollow">torchrun</a> 启动。</p> <!> <p>推理速度降低到 649.23 秒，相比基线快 5.66 倍，使用 8 个 NVIDIA L20 GPU。</p>',1),P=c("<table><thead><tr><th>GPU 类型</th><th>GPU 数量</th><th>优化</th><th>墙钟时间 (s)</th><th>加速比</th></tr></thead><tbody><tr><td>NVIDIA L20</td><td>1</td><td>基线</td><td>26.36</td><td>1.00x</td></tr><tr><td>NVIDIA L20</td><td>1</td><td>FBCache (rdt=0.08)</td><td>17.01</td><td>1.55x</td></tr><tr><td>NVIDIA L20</td><td>1</td><td>FP8 DQ</td><td>13.40</td><td>1.96x</td></tr><tr><td>NVIDIA L20</td><td>1</td><td>FBCache (rdt=0.12) + FP8 DQ</td><td>7.56</td><td>3.48x</td></tr><tr><td>NVIDIA L20</td><td>2</td><td>FBCache (rdt=0.12) + FP8 DQ + CP</td><td>4.92</td><td>5.35x</td></tr><tr><td>NVIDIA L20</td><td>4</td><td>FBCache (rdt=0.12) + FP8 DQ + CP</td><td>3.90</td><td>6.75x</td></tr></tbody></table>"),O=c(`<table><thead><tr><th>GPU 类型</th><th>GPU 数量</th><th>优化</th><th>墙钟时间 (s)</th><th>加速比</th></tr></thead><tbody><tr><td>NVIDIA L20</td><td>1</td><td>基线</td><td>3675.71</td><td>1.00x</td></tr></tbody></table> <p>| NVIDIA
L20 | 1 | FBCache | 2271.06 | 1.62x |
| NVIDIA L20 | 2 | FBCache + CP | 1132.90 | 3.24x |
| NVIDIA L20 | 4 | FBCache + CP | 718.15 | 5.12x |
| NVIDIA L20 | 8 | FBCache + CP | 649.23 | 5.66x |</p>`,1),$=c(`<p></p> <!> <!> <div class="flex justify-center"><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/flux-performance.png"/></div> <div class="flex justify-center"><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/hunyuan-video-performance.png"/></div> <p>大型图像和视频生成模型，如 <a href="https://huggingface.co/black-forest-labs/FLUX.1-dev" rel="nofollow">FLUX.1-dev</a> 和 <a href="https://huggingface.co/tencent/HunyuanVideo" rel="nofollow">HunyuanVideo</a>，由于其规模，可能对实时应用和部署构成推理挑战。</p> <p><a href="https://github.com/chengzeyi/ParaAttention" rel="nofollow">ParaAttention</a> 是一个实现了<strong>上下文并行</strong>和<strong>第一块缓存</strong>的库，可以与其他技术（如 torch.compile、fp8 动态量化）结合使用，以加速推理。</p> <p>本指南将展示如何在 NVIDIA L20 GPU 上对 FLUX.1-dev 和 HunyuanVideo 应用 ParaAttention。
在我们的基线基准测试中，除了 HunyuanVideo 为避免内存不足错误外，未应用任何优化。</p> <p>我们的基线基准测试显示，FLUX.1-dev 能够在 28 步中生成 1024x1024 分辨率图像，耗时 26.36 秒；HunyuanVideo 能够在 30 步中生成 129 帧 720p 分辨率视频，耗时 3675.71 秒。</p> <blockquote class="tip"><p>对于更快的上下文并行推理，请尝试使用支持 NVLink 的 NVIDIA A100 或 H100 GPU（如果可用），尤其是在 GPU 数量较多时。</p></blockquote> <!> <p>缓存模型中 transformer 块的输出并在后续推理步骤中重用它们，可以降低计算成本并加速推理。</p> <p>然而，很难决定何时重用缓存以确保生成图像或视频的质量。ParaAttention 直接使用<strong>第一个 transformer 块输出的残差差异</strong>来近似模型输出之间的差异。当差异足够小时，重用先前推理步骤的残差差异。换句话说，跳过去噪步骤。</p> <p>这在 FLUX.1-dev 和 HunyuanVideo 推理上实现了 2 倍加速，且质量非常好。</p> <figure><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/para-attn/ada-cache.png" alt="Cache in Diffusion Transformer"/> <figcaption>AdaCache 的工作原理，第一块缓存是其变体</figcaption></figure> <!> <!> <p>fp8 动态量化进一步加速推理并减少内存使用。为了使用 8 位 <a href="https://www.nvidia.com/en-us/data-center/tensor-cores/" rel="nofollow">NVIDIA Tensor Cores</a>，必须对激活和权重进行量化。</p> <p>使用 <code>float8_weight_only</code> 和 <code>float8_dynamic_activation_float8_weight</code> 来量化文本编码器和变换器模型。</p> <p>默认量化方法是逐张量量化，但如果您的 GPU 支持逐行量化，您也可以尝试它以获得更好的准确性。</p> <p>使用以下命令安装 <a href="https://github.com/pytorch/ao/tree/main" rel="nofollow">torchao</a>。</p> <!> <p><a href="https://pytorch.org/tutorials/intermediate/torch_compile_tutorial.html" rel="nofollow">torch.compile</a> 使用 <code>mode="max-autotune-no-cudagraphs"</code> 或 <code>mode="max-autotune"</code> 选择最佳内核以获得性能。如果是第一次调用模型，编译可能会花费很长时间，但一旦模型编译完成，这是值得的。</p> <p>此示例仅量化变换器模型，但您也可以量化文本编码器以进一步减少内存使用。</p> <blockquote class="tip"><p>动态量化可能会显著改变模型输出的分布，因此您需要将 <code>residual_diff_threshold</code> 设置为更大的值以使其生效。</p></blockquote> <!> <!> <p>上下文并行性并行化推理并随多个 GPU 扩展。ParaAttention 组合设计允许您将上下文并行性与第一块缓存和动态量化结合使用。</p> <blockquote class="tip"><p>请参考 <a href="https://github.com/chengzeyi/ParaAttention/tree/main" rel="nofollow">ParaAttention</a> 仓库获取详细说明和如何使用多个 GPU 扩展推理的示例。</p></blockquote> <p>如果推理过程需要持久化和可服务，建议使用 <a href="https://pytorch.org/docs/stable/multiprocessing.html" rel="nofollow">torch.multiprocessing</a> 编写您自己的推理处理器。这可以消除启动进程以及加载和重新编译模型的开销。</p> <!> <!> <!> <!> <p></p>`,1);function tl(C,X){R(X,!1),S(()=>{new URLSearchParams(window.location.search).get("fw")}),Q();var w=$();v("fka9nn",M=>{var U=H();E(U,"content",Y),n(M,U)});var T=l(o(w),2);A(T,{containerStyle:"float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"});var b=l(T,2);h(b,{title:"ParaAttention",local:"paraattention",headingTag:"h1"});var Z=l(b,16);h(Z,{title:"第一块缓存",local:"第一块缓存",headingTag:"h2"});var f=l(Z,10);m(f,{id:"first-block-cache",options:["FLUX-1.dev","HunyuanVideo"],children:(M,U)=>{var p=j(),e=o(p);J(e,{id:"first-block-cache",option:"FLUX-1.dev",children:(a,y)=>{var s=z(),t=l(o(s),2);i(t,{code:"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",highlighted:`<span class="hljs-keyword">import</span> time
<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> FluxPipeline
pipe = FluxPipeline.from_pretrained(
 <span class="hljs-string">&quot;black-forest-labs/FLUX.1-dev&quot;</span>,
 torch_dtype=torch.bfloat16,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)
<span class="hljs-keyword">from</span> para_attn.first_block_cache.diffusers_adapters <span class="hljs-keyword">import</span> apply_cache_on_pipe
apply_cache_on_pipe(pipe, residual_diff_thre
shold=<span class="hljs-number">0.08</span>)
<span class="hljs-comment"># 启用内存节省</span>
<span class="hljs-comment"># pipe.enable_model_cpu_offload()</span>
<span class="hljs-comment"># pipe.enable_sequential_cpu_offload()</span>
begin = time.time()
image = pipe(
 <span class="hljs-string">&quot;A cat holding a sign that says hello world&quot;</span>,
 num_inference_steps=<span class="hljs-number">28</span>,
).images[<span class="hljs-number">0</span>]
end = time.time()
<span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;Time: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
<span class="hljs-built_in">print</span>(<span class="hljs-string">&quot;Saving image to flux.png&quot;</span>)
image.save(<span class="hljs-string">&quot;flux.png&quot;</span>)`,lang:"python",wrap:!1}),r(4),n(a,s)},$$slots:{default:!0}});var d=l(e,2);J(d,{id:"first-block-cache",option:"HunyuanVideo",children:(a,y)=>{var s=x(),t=l(o(s),2);i(t,{code:"aW1wb3J0JTIwdGltZSUwQWltcG9ydCUyMHRvcmNoJTBBZnJvbSUyMGRpZmZ1c2VycyUyMGltcG9ydCUyMEh1bnl1YW5WaWRlb1BpcGVsaW5lJTJDJTIwSHVueXVhblZpZGVvVHJhbnNmb3JtZXIzRE1vZGVsJTBBZnJvbSUyMGRpZmZ1c2Vycy51dGlscyUyMGltcG9ydCUyMGV4cG9ydF90b192aWRlbyUwQSUwQW1vZGVsX2lkJTIwJTNEJTIwJTIydGVuY2VudCUyRkh1bnl1YW5WaWRlbyUyMiUwQXRyYW5zZm9ybWVyJTIwJTNEJTIwSHVueXVhblZpZGVvVHJhbnNmb3JtZXIzRE1vZGVsLmZyb21fcHJldHJhaW5lZCglMEElMjAlMjAlMjAlMjBtb2RlbF9pZCUyQyUwQSUyMCUyMCUyMCUyMHN1YmZvbGRlciUzRCUyMnRyYW5zZm9ybWVyJTIyJTJDJTBBJTIwJTIwJTIwJTIwdG9yY2hfZHR5cGUlM0R0b3JjaC5iZmxvYXQxNiUyQyUwQSUyMCUyMCUyMCUyMHJldmlzaW9uJTNEJTIycmVmcyUyRnByJTJGMTglMjIlMkMlMEEpJTBBcGlwZSUyMCUzRCUyMEh1bnl1YW5WaWRlb1BpcGVsaW5lLmZyb21fcHJldHJhaW5lZCglMEElMjAlMjAlMjAlMjBtb2RlbF9pZCUyQyUwQSUyMCUyMCUyMCUyMHRyYW5zZm9ybWVyJTNEdHJhbnNmb3JtZXIlMkMlMEElMjAlMjAlMjAlMjB0b3JjaF9kdHlwZSUzRHRvcmNoLmZsb2F0MTYlMkMlMEElMjAlMjAlMjAlMjByZXZpc2lvbiUzRCUyMnJlZnMlMkZwciUyRjE4JTIyJTJDJTBBKS50byglMjJjdWRhJTIyKSUwQSUwQWZyb20lMjBwYXJhX2F0dG4uZmlyc3RfYmxvY2tfY2FjaGUuZGlmZnVzZXJzX2FkYXB0ZXJzJTIwaW1wb3J0JTIwYXBwbHlfY2FjaGVfb25fcGlwZSUwQSUwQWFwcGx5X2NhY2hlX29uX3BpcGUocGlwZSUyQyUyMHJlc2lkdWFsX2RpZmZfdGhyZXNob2xkJTNEMC42KSUwQSUwQXBpcGUudmFlLmVuYWJsZV90aWxpbmcoKSUwQSUwQWJlZ2luJTIwJTNEJTIwdGltZS50aW1lKCklMEFvdXRwdXQlMjAlM0QlMjBwaXBlKCUwQSUyMCUyMCUyMCUyMHByb21wdCUzRCUyMkElMjBjYXQlMjB3YWxrcyUyMG9uJTIwdGhlJTIwZ3Jhc3MlMkMlMjByZWFsaXN0aWMlMjIlMkMlMEElMjAlMjAlMjAlMjBoZWlnaHQlM0Q3MjAlMkMlMEElMjAlMjAlMjAlMjB3aWR0aCUzRDEyODAlMkMlMEElMjAlMjAlMjAlMjBudW1fZnJhbWVzJTNEMTI5JTJDJTBBJTIwJTIwJTIwJTIwbnVtX2luZmVyZW5jZV9zdGVwcyUzRDMwJTJDJTBBKS5mcmFtZXMlNUIwJTVEJTBBZW5kJTIwJTNEJTIwdGltZS50aW1lKCklMEFwcmludChmJTIyVGltZSUzQSUyMCU3QmVuZCUyMC0lMjBiZWdpbiUzQS4yZiU3RHMlMjIpJTBBJTBBcHJpbnQoJTIyU2F2aW5nJTIwdmlkZW8lMjB0byUyMGh1bnl1YW5fdmlkZW8ubXA0JTIyKSUwQWV4cG9ydF90b192aWRlbyhvdXRwdXQlMkMlMjAlMjJodW55dWFuX3ZpZGVvLm1wNCUyMiUyQyUyMGZwcyUzRDE1KQ==",highlighted:`<span class="hljs-keyword">import</span> time
<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> HunyuanVideoPipeline, HunyuanVideoTransformer3DModel
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video
model_id = <span class="hljs-string">&quot;tencent/HunyuanVideo&quot;</span>
transformer = HunyuanVideoTransformer3DModel.from_pretrained(
 model_id,
 subfolder=<span class="hljs-string">&quot;transformer&quot;</span>,
 torch_dtype=torch.bfloat16,
 revision=<span class="hljs-string">&quot;refs/pr/18&quot;</span>,
)
pipe = HunyuanVideoPipeline.from_pretrained(
 model_id,
 transformer=transformer,
 torch_dtype=torch.float16,
 revision=<span class="hljs-string">&quot;refs/pr/18&quot;</span>,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)
<span class="hljs-keyword">from</span> para_attn.first_block_cache.diffusers_adapters <span class="hljs-keyword">import</span> apply_cache_on_pipe
apply_cache_on_pipe(pipe, residual_diff_threshold=<span class="hljs-number">0.6</span>)
pipe.vae.enable_tiling()
begin = time.time()
output = pipe(
 prompt=<span class="hljs-string">&quot;A cat walks on the grass, realistic&quot;</span>,
 height=<span class="hljs-number">720</span>,
 width=<span class="hljs-number">1280</span>,
 num_frames=<span class="hljs-number">129</span>,
 num_inference_steps=<span class="hljs-number">30</span>,
).frames[<span class="hljs-number">0</span>]
end = time.time()
<span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;Time: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
<span class="hljs-built_in">print</span>(<span class="hljs-string">&quot;Saving video to hunyuan_video.mp4&quot;</span>)
export_to_video(output, <span class="hljs-string">&quot;hunyuan_video.mp4&quot;</span>, fps=<span class="hljs-number">15</span>)`,lang:"python",wrap:!1}),r(10),n(a,s)},$$slots:{default:!0}}),n(M,p)},$$slots:{default:!0}});var B=l(f,2);h(B,{title:"fp8 量化",local:"fp8-量化",headingTag:"h2"});var I=l(B,10);i(I,{code:"cGlwMyUyMGluc3RhbGwlMjAtVSUyMHRvcmNoJTIwdG9yY2hhbw==",highlighted:"pip3 install -U torch torchao",lang:"bash",wrap:!1});var _=l(I,8);m(_,{id:"fp8-quantization",options:["FLUX-1.dev","HunyuanVideo"],children:(M,U)=>{var p=j(),e=o(p);J(e,{id:"fp8-quantization",option:"FLUX-1.dev",children:(a,y)=>{var s=q(),t=o(s);i(t,{code:"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",highlighted:`<span class="hljs-keyword">import</span> time
<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> FluxPipeline
pipe = FluxPipeline.from_pretrained(
 <span class="hljs-string">&quot;black-forest-labs/FLUX.1-dev&quot;</span>,
 torch_dtype=torch.bfloat16,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)
<span class="hljs-keyword">from</span> para_attn.first_block_cache.diffusers_adapters <span class="hljs-keyword">import</span> apply_cache_on_pipe
apply_cache_on_pipe(
 pipe,
 residual_diff_threshold=<span class="hljs-number">0.12</span>, <span class="hljs-comment"># 使用更大的值以使缓存生效</span>
)
<span class="hljs-keyword">from</span> torchao.quantization <span class="hljs-keyword">import</span> quantize_, float8_dynamic_activation_float8_weight, float8_weight_only
quantize_(pipe.text_encoder, float8_weight_only())
quantize_(pipe.transformer, float8_dynamic_activation_float8_weight())
pipe.transformer = torch.<span class="hljs-built_in">compile</span>(
 pipe.transformer, mode=<span class="hljs-string">&quot;max-autotune-no-cudagraphs&quot;</span>,
)
<span class="hljs-comment"># 启用内存节省</span>
<span class="hljs-comment"># pipe.enable_model_cpu_offload()</span>
<span class="hljs-comment"># pipe.enable_sequential_cpu_offload()</span>
<span class="hljs-keyword">for</span> i <span class="hljs-keyword">in</span> <span class="hljs-built_in">range</span>(<span class="hljs-number">2</span>):
 begin = time.time()
 image = pipe(
 <span class="hljs-string">&quot;A cat holding a sign that says hello world&quot;</span>,
 num_inference_steps=<span class="hljs-number">28</span>,
 ).images[<span class="hljs-number">0</span>]
 end = time.time()
 <span class="hljs-keyword">if</span> i == <span class="hljs-number">0</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;预热时间: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
 <span class="hljs-keyword">else</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;时间: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
<span class="hljs-built_in">print</span>(<span class="hljs-string">&quot;保存图像到 flux.png&quot;</span>)
image.save(<span class="hljs-string">&quot;flux.png&quot;</span>)`,lang:"python",wrap:!1}),r(2),n(a,s)},$$slots:{default:!0}});var d=l(e,2);J(d,{id:"fp8-quantization",option:"HunyuanVideo",children:(a,y)=>{var s=L(),t=o(s);i(t,{code:"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",highlighted:`<span class="hljs-keyword">import</span> time
<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> HunyuanVideoPipeline, HunyuanVideoTransformer3DModel
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video
model_id = <span class="hljs-string">&quot;tencent/HunyuanVideo&quot;</span>
transformer = HunyuanVideoTransformer3DModel.from_pretrained(
 model_id,
 subfolder=<span class="hljs-string">&quot;transformer&quot;</span>,
 torch_dtype=torch.bfloat16,
 revision=<span class="hljs-string">&quot;refs/pr/18&quot;</span>,
)
pipe = HunyuanVideoPipeline.from_pretrained(
 model_id,
 transformer=transformer,
 torch_dtype=torch.float16,
 revision=<span class="hljs-string">&quot;refs/pr/18&quot;</span>,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)
<span class="hljs-keyword">from</span> para_attn.first_block_cache.diffusers_adapters <span class="hljs-keyword">import</span> apply_cache_on_pipe
apply_cache_on_pipe(pipe)
<span class="hljs-keyword">from</span> torchao.quantization <span class="hljs-keyword">import</span> quantize_, float8_dynamic_activation_float8_weight, float8_weight_only
quantize_(pipe.text_encoder, float8_weight_only())
quantize_(pipe.transformer, float8_dynamic_activation_float8_weight())
pipe.transformer = torch.<span class="hljs-built_in">compile</span>(
 pipe.transformer, mode=<span class="hljs-string">&quot;max-autotune-no-cudagraphs&quot;</span>,
)
<span class="hljs-comment"># Enable memory savings</span>
pipe.vae.enable_tiling()
<span class="hljs-comment"># pipe.enable_model_cpu_offload()</span>
<span class="hljs-comment"># pipe.enable_sequential_cpu_offload()</span>
<span class="hljs-keyword">for</span> i <span class="hljs-keyword">in</span> <span class="hljs-built_in">range</span>(<span class="hljs-number">2</span>):
 begin = time.time()
 output = pipe(
 prompt=<span class="hljs-string">&quot;A cat walks on the grass, realistic&quot;</span>,
 height=<span class="hljs-number">720</span>,
 width=<span class="hljs-number">1280</span>,
 num_frames=<span class="hljs-number">129</span>,
 num_inference_steps=<span class="hljs-number">1</span> <span class="hljs-keyword">if</span> i == <span class="hljs-number">0</span> <span class="hljs-keyword">else</span> <span class="hljs-number">30</span>,
 ).frames[<span class="hljs-number">0</span>]
 end = time.time()
 <span class="hljs-keyword">if</span> i == <span class="hljs-number">0</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;Warm up time: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
 <span class="hljs-keyword">else</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;Time: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
<span class="hljs-built_in">print</span>(<span class="hljs-string">&quot;Saving video to hunyuan_video.mp4&quot;</span>)
export_to_video(output, <span class="hljs-string">&quot;hunyuan_video.mp4&quot;</span>, fps=<span class="hljs-number">15</span>)`,lang:"python",wrap:!1}),r(4),n(a,s)},$$slots:{default:!0}}),n(M,p)},$$slots:{default:!0}});var g=l(_,2);h(g,{title:"上下文并行性",local:"上下文并行性",headingTag:"h2"});var G=l(g,8);m(G,{id:"context-parallelism",options:["FLUX-1.dev","HunyuanVideo"],children:(M,U)=>{var p=j(),e=o(p);J(e,{id:"context-parallelism",option:"FLUX-1.dev",children:(a,y)=>{var s=D(),t=l(o(s),2);i(t,{code:"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",highlighted:`<span class="hljs-keyword">import</span> time
<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">import</span> torch.distributed <span class="hljs-keyword">as</span> dist
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> FluxPipeline
dist.init_process_group()
torch.cuda.set_device(dist.get_rank())
pipe = FluxPipeline.from_pretrained(
 <span class="hljs-string">&quot;black-forest-labs/FLUX.1-dev&quot;</span>,
 torch_dtype=torch.bfloat16,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)
<span class="hljs-keyword">from</span> para_attn.context_parallel <span class="hljs-keyword">import</span> init_context_parallel_mesh
<span class="hljs-keyword">from</span> para_attn.context_parallel.diffusers_adapters <span class="hljs-keyword">import</span> parallelize_pipe
<span class="hljs-keyword">from</span> para_attn.parallel_vae.diffusers_adapters <span class="hljs-keyword">import</span> parallelize_vae
mesh = init_context_parallel_mesh(
 pipe.device.<span class="hljs-built_in">type</span>,
 max_ring_dim_size=<span class="hljs-number">2</span>,
)
parallelize_pipe(
 pipe,
 mesh=mesh,
)
parallelize_vae(pipe.vae, mesh=mesh._flatten())
<span class="hljs-keyword">from</span> para_attn.first_block_cache.diffusers_adapters <span class="hljs-keyword">import</span> apply_cache_on_pipe
apply_cache_on_pipe(
 pipe,
 residual_diff_threshold=<span class="hljs-number">0.12</span>, <span class="hljs-comment"># 使用较大的值以使缓存生效</span>
)
<span class="hljs-keyword">from</span> torchao.quantization <span class="hljs-keyword">import</span> quantize_, float8_dynamic_activation_float8_weight, float8_weight_only
quantize_(pipe.text_encoder, float8_weight_only())
quantize_(pipe.transformer, float8_dynamic_activation_float8_weight())
torch._inductor.config.reorder_for_compute_comm_overlap = <span class="hljs-literal">True</span>
pipe.transformer = torch.<span class="hljs-built_in">compile</span>(
 pipe.transformer, mode=<span class="hljs-string">&quot;max-autotune-no-cudagraphs&quot;</span>,
)
<span class="hljs-comment"># 启用内存节省</span>
<span class="hljs-comment"># pipe.enable_model_cpu_offload(gpu_id=dist.get_rank())</span>
<span class="hljs-comment"># pipe.enable_sequential_cpu_offload(gpu_id=dist.get_rank())</span>
<span class="hljs-keyword">for</span> i <span class="hljs-keyword">in</span> <span class="hljs-built_in">range</span>(<span class="hljs-number">2</span>):
 begin = time.time()
 image = pipe(
 <span class="hljs-string">&quot;A cat holding a sign that says hello world&quot;</span>,
 num_inference_steps=<span class="hljs-number">28</span>,
 output_type=<span class="hljs-string">&quot;pil&quot;</span> <span class="hljs-keyword">if</span> dist.get_rank() == <span class="hljs-number">0</span> <span class="hljs-keyword">else</span> <span class="hljs-string">&quot;pt&quot;</span>,
 ).images[<span class="hljs-number">0</span>]
 end = time.time()
 <span class="hljs-keyword">if</span> dist.get_rank() == <span class="hljs-number">0</span>:
 <span class="hljs-keyword">if</span> i == <span class="hljs-number">0</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;预热时间: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
 <span class="hljs-keyword">else</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;时间: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
<span class="hljs-keyword">if</span> dist.get_rank() == <span class="hljs-number">0</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">&quot;将图像保存到flux.png&quot;</span>)
 image.save(<span class="hljs-string">&quot;flux.png&quot;</span>)
dist.destroy_process_group()`,lang:"python",wrap:!1});var u=l(t,4);i(u,{code:"JTIzJTIwJUU0JUJEJUJGJUU3JTk0JUE4LS1ucHJvY19wZXJfbm9kZSVFNiU4QyU4NyVFNSVBRSU5QUdQVSVFNiU5NSVCMCVFOSU4NyU4RiUwQXRvcmNocnVuJTIwLS1ucHJvY19wZXJfbm9kZSUzRDIlMjBydW5fZmx1eC5weQ==",highlighted:`<span class="hljs-comment"># 使用--nproc_per_node指定GPU数量</span>
torchrun --nproc_per_node=2 run_flux.py`,lang:"bash",wrap:!1}),r(2),n(a,s)},$$slots:{default:!0}});var d=l(e,2);J(d,{id:"context-parallelism",option:"HunyuanVideo",children:(a,y)=>{var s=K(),t=l(o(s),2);i(t,{code:"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",highlighted:`<span class="hljs-keyword">import</span> time
<span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">import</span> torch.distributed <span class="hljs-keyword">as</span> dist
<span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> HunyuanVideoPipeline, HunyuanVideoTransformer3DModel
<span class="hljs-keyword">from</span> diffusers.utils <span class="hljs-keyword">import</span> export_to_video
dist.init_process_group()
torch.cuda.set_device(dist.get_rank())
model_id = <span class="hljs-string">&quot;tencent/HunyuanVideo&quot;</span>
transformer = HunyuanVideoTransformer3DModel.from_pretrained(
 model_id,
 subfolder=<span class="hljs-string">&quot;transformer&quot;</span>,
 torch_dtype=torch.bfloat16,
 revision=<span class="hljs-string">&quot;refs/pr/18&quot;</span>,
)
pipe = HunyuanVideoPipeline.from_pretrained(
 model_id,
 transformer=transformer,
 torch_dtype=torch.float16,
 revision=<span class="hljs-string">&quot;refs/pr/18&quot;</span>,
).to(<span class="hljs-string">&quot;cuda&quot;</span>)
<span class="hljs-keyword">from</span> para_attn.context_parallel <span class="hljs-keyword">import</span> init_context_parallel_mesh
<span class="hljs-keyword">from</span> para_attn.context_parallel.diffusers_adapters <span class="hljs-keyword">import</span> parallelize_pipe
<span class="hljs-keyword">from</span> para_attn.parallel_vae.diffusers_adapters <span class="hljs-keyword">import</span> parallelize_vae
mesh = init_context_parallel_mesh(
 pipe.device.<span class="hljs-built_in">type</span>,
)
parallelize_pipe(
 pipe,
 mesh=mesh,
)
parallelize_vae(pipe.vae, mesh=mesh._flatten())
<span class="hljs-keyword">from</span> para_attn.first_block_cache.diffusers_adapters <span class="hljs-keyword">import</span> apply_cache_on_pipe
apply_cache_on_pipe(pipe)
<span class="hljs-comment"># from torchao.quantization import quantize_, float8_dynamic_activation_float8_weight, float8_weight_only</span>
<span class="hljs-comment">#</span>
<span class="hljs-comment"># torch._inductor.config.reorder_for_compute_comm_overlap = True</span>
<span class="hljs-comment">#</span>
<span class="hljs-comment"># quantize_(pipe.text_encoder, float8_weight_only())</span>
<span class="hljs-comment"># quantize_(pipe.transformer, float8_dynamic_activation_float8_weight())</span>
<span class="hljs-comment"># pipe.transformer = torch.compile(</span>
<span class="hljs-comment"># pipe.transformer, mode=&quot;max-autotune-no-cudagraphs&quot;,</span>
<span class="hljs-comment"># )</span>
<span class="hljs-comment"># 启用内存节省</span>
pipe.vae.enable_tiling()
<span class="hljs-comment"># pipe.enable_model_cpu_offload(gpu_id=dist.get_rank())</span>
<span class="hljs-comment"># pipe.enable_sequential_cpu_offload(gpu_id=dist.get_rank())</span>
<span class="hljs-keyword">for</span> i <span class="hljs-keyword">in</span> <span class="hljs-built_in">range</span>(<span class="hljs-number">2</span>):
 begin = time.time()
 output = pipe(
 prompt=<span class="hljs-string">&quot;A cat walks on the grass, realistic&quot;</span>,
 height=<span class="hljs-number">720</span>,
 width=<span class="hljs-number">1280</span>,
 num_frames=<span class="hljs-number">129</span>,
 num_inference_steps=<span class="hljs-number">1</span> <span class="hljs-keyword">if</span> i == <span class="hljs-number">0</span> <span class="hljs-keyword">else</span> <span class="hljs-number">30</span>,
 output_type=<span class="hljs-string">&quot;pil&quot;</span> <span class="hljs-keyword">if</span> dist.get_rank() == <span class="hljs-number">0</span> <span class="hljs-keyword">else</span> <span class="hljs-string">&quot;pt&quot;</span>,
 ).frames[<span class="hljs-number">0</span>]
 end = time.time()
 <span class="hljs-keyword">if</span> dist.get_rank() == <span class="hljs-number">0</span>:
 <span class="hljs-keyword">if</span> i == <span class="hljs-number">0</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;预热时间: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
 <span class="hljs-keyword">else</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">f&quot;时间: <span class="hljs-subst">{end - begin:<span class="hljs-number">.2</span>f}</span>s&quot;</span>)
<span class="hljs-keyword">if</span> dist.get_rank() == <span class="hljs-number">0</span>:
 <span class="hljs-built_in">print</span>(<span class="hljs-string">&quot;保存视频到 hunyuan_video.mp4&quot;</span>)
 export_to_video(output, <span class="hljs-string">&quot;hunyuan_video.mp4&quot;</span>, fps=<span class="hljs-number">15</span>)
dist.destroy_process_group()`,lang:"python",wrap:!1});var u=l(t,4);i(u,{code:"JTIzJTIwJUU0JUJEJUJGJUU3JTk0JUE4JTIwLS1ucHJvY19wZXJfbm9kZSUyMCVFNiU4QyU4NyVFNSVBRSU5QSUyMEdQVSUyMCVFNiU5NSVCMCVFOSU4NyU4RiUwQXRvcmNocnVuJTIwLS1ucHJvY19wZXJfbm9kZSUzRDglMjBydW5faHVueXVhbl92aWRlby5weQ==",highlighted:`<span class="hljs-comment"># 使用 --nproc_per_node 指定 GPU 数量</span>
torchrun --nproc_per_node=8 run_hunyuan_video.py`,lang:"bash",wrap:!1}),r(2),n(a,s)},$$slots:{default:!0}}),n(M,p)},$$slots:{default:!0}});var W=l(G,2);h(W,{title:"基准测试",local:"基准测试",headingTag:"h2"});var V=l(W,2);m(V,{id:"conclusion",options:["FLUX-1.dev","HunyuanVideo"],children:(M,U)=>{var p=j(),e=o(p);J(e,{id:"conclusion",option:"FLUX-1.dev",children:(a,y)=>{var s=P();n(a,s)},$$slots:{default:!0}});var d=l(e,2);J(d,{id:"conclusion",option:"HunyuanVideo",children:(a,y)=>{var s=O();r(2),n(a,s)},$$slots:{default:!0}}),n(M,p)},$$slots:{default:!0}});var k=l(V,2);F(k,{source:"https://github.com/huggingface/diffusers/blob/main/docs/source/zh/optimization/para_attn.md"}),r(2),n(C,w),N()}export{tl as component};

Xet Storage Details

Size:
48.2 kB
·
Xet hash:
839797bba3455a390ad45fb57be44d99c5166df2bffa1c26be7e74973cb1c09b

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.