Buckets:
| <meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"DeepCache","local":"deepcache","sections":[{"title":"Benchmark","local":"benchmark","sections":[],"depth":2}],"depth":1}"> | |
| <link href="/docs/diffusers/main/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/entry/start.21e27d66.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/scheduler.8c3d61f6.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/singletons.1db06f6d.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/index.0997d446.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/paths.085e8bc8.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/entry/app.de4fb612.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/index.da70eac4.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/nodes/0.f6117ae5.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/each.e59479a4.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/nodes/167.eab69b71.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/CodeBlock.00a903b3.js"> | |
| <link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/EditOnGithub.1e64e623.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"DeepCache","local":"deepcache","sections":[{"title":"Benchmark","local":"benchmark","sections":[],"depth":2}],"depth":1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <h1 class="relative group"><a id="deepcache" 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="#deepcache"><span><svg class="" 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>DeepCache</span></h1> <p data-svelte-h="svelte-93kxkg"><a href="https://huggingface.co/papers/2312.00858" rel="nofollow">DeepCache</a> accelerates <a href="/docs/diffusers/main/en/api/pipelines/stable_diffusion/text2img#diffusers.StableDiffusionPipeline">StableDiffusionPipeline</a> and <a href="/docs/diffusers/main/en/api/pipelines/stable_diffusion/stable_diffusion_xl#diffusers.StableDiffusionXLPipeline">StableDiffusionXLPipeline</a> by strategically caching and reusing high-level features while efficiently updating low-level features by taking advantage of the U-Net architecture.</p> <p data-svelte-h="svelte-yv0ij">Start by installing <a href="https://github.com/horseee/DeepCache" rel="nofollow">DeepCache</a>:</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 class="" 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> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->pip install DeepCache<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-sy0xvd">Then load and enable the <a href="https://github.com/horseee/DeepCache#usage" rel="nofollow"><code>DeepCacheSDHelper</code></a>:</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 class="" 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> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --> import torch | |
| from diffusers import StableDiffusionPipeline | |
| pipe = StableDiffusionPipeline.from_pretrained('runwayml/stable-diffusion-v1-5', torch_dtype=torch.float16).to("cuda") | |
| <span class="hljs-addition">+ from DeepCache import DeepCacheSDHelper</span> | |
| <span class="hljs-addition">+ helper = DeepCacheSDHelper(pipe=pipe)</span> | |
| <span class="hljs-addition">+ helper.set_params(</span> | |
| <span class="hljs-addition">+ cache_interval=3,</span> | |
| <span class="hljs-addition">+ cache_branch_id=0,</span> | |
| <span class="hljs-addition">+ )</span> | |
| <span class="hljs-addition">+ helper.enable()</span> | |
| image = pipe("a photo of an astronaut on a moon").images[0]<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1k27afg">The <code>set_params</code> method accepts two arguments: <code>cache_interval</code> and <code>cache_branch_id</code>. <code>cache_interval</code> means the frequency of feature caching, specified as the number of steps between each cache operation. <code>cache_branch_id</code> identifies which branch of the network (ordered from the shallowest to the deepest layer) is responsible for executing the caching processes. | |
| Opting for a lower <code>cache_branch_id</code> or a larger <code>cache_interval</code> can lead to faster inference speed at the expense of reduced image quality (ablation experiments of these two hyperparameters can be found in the <a href="https://arxiv.org/abs/2312.00858" rel="nofollow">paper</a>). Once those arguments are set, use the <code>enable</code> or <code>disable</code> methods to activate or deactivate the <code>DeepCacheSDHelper</code>.</p> <div class="flex justify-center" data-svelte-h="svelte-12c1pq6"><img src="https://github.com/horseee/Diffusion_DeepCache/raw/master/static/images/example.png"></div> <p data-svelte-h="svelte-a518zi">You can find more generated samples (original pipeline vs DeepCache) and the corresponding inference latency in the <a href="https://wandb.ai/horseee/DeepCache/runs/jwlsqqgt?workspace=user-horseee" rel="nofollow">WandB report</a>. The prompts are randomly selected from the <a href="https://cocodataset.org/#home" rel="nofollow">MS-COCO 2017</a> dataset.</p> <h2 class="relative group"><a id="benchmark" 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="#benchmark"><span><svg class="" 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>Benchmark</span></h2> <p data-svelte-h="svelte-18sczus">We tested how much faster DeepCache accelerates <a href="https://huggingface.co/stabilityai/stable-diffusion-2-1" rel="nofollow">Stable Diffusion v2.1</a> with 50 inference steps on an NVIDIA RTX A5000, using different configurations for resolution, batch size, cache interval (I), and cache branch (B).</p> <table data-svelte-h="svelte-whx3f0"><thead><tr><th><strong>Resolution</strong></th> <th><strong>Batch size</strong></th> <th><strong>Original</strong></th> <th><strong>DeepCache(I=3, B=0)</strong></th> <th><strong>DeepCache(I=5, B=0)</strong></th> <th><strong>DeepCache(I=5, B=1)</strong></th></tr></thead> <tbody><tr><td>512</td> <td>8</td> <td>15.96</td> <td>6.88(2.32x)</td> <td>5.03(3.18x)</td> <td>7.27(2.20x)</td></tr> <tr><td></td> <td>4</td> <td>8.39</td> <td>3.60(2.33x)</td> <td>2.62(3.21x)</td> <td>3.75(2.24x)</td></tr> <tr><td></td> <td>1</td> <td>2.61</td> <td>1.12(2.33x)</td> <td>0.81(3.24x)</td> <td>1.11(2.35x)</td></tr> <tr><td>768</td> <td>8</td> <td>43.58</td> <td>18.99(2.29x)</td> <td>13.96(3.12x)</td> <td>21.27(2.05x)</td></tr> <tr><td></td> <td>4</td> <td>22.24</td> <td>9.67(2.30x)</td> <td>7.10(3.13x)</td> <td>10.74(2.07x)</td></tr> <tr><td></td> <td>1</td> <td>6.33</td> <td>2.72(2.33x)</td> <td>1.97(3.21x)</td> <td>2.98(2.12x)</td></tr> <tr><td>1024</td> <td>8</td> <td>101.95</td> <td>45.57(2.24x)</td> <td>33.72(3.02x)</td> <td>53.00(1.92x)</td></tr> <tr><td></td> <td>4</td> <td>49.25</td> <td>21.86(2.25x)</td> <td>16.19(3.04x)</td> <td>25.78(1.91x)</td></tr> <tr><td></td> <td>1</td> <td>13.83</td> <td>6.07(2.28x)</td> <td>4.43(3.12x)</td> <td>7.15(1.93x)</td></tr></tbody></table> <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/optimization/deepcache.md" target="_blank"><span data-svelte-h="svelte-1kd6by1"><</span> <span data-svelte-h="svelte-x0xyl0">></span> <span data-svelte-h="svelte-1dajgef"><span class="underline ml-1.5">Update</span> on GitHub</span></a> <p></p> | |
| <script> | |
| { | |
| __sveltekit_1eu7tzz = { | |
| assets: "/docs/diffusers/main/en", | |
| base: "/docs/diffusers/main/en", | |
| env: {} | |
| }; | |
| const element = document.currentScript.parentElement; | |
| const data = [null,null]; | |
| Promise.all([ | |
| import("/docs/diffusers/main/en/_app/immutable/entry/start.21e27d66.js"), | |
| import("/docs/diffusers/main/en/_app/immutable/entry/app.de4fb612.js") | |
| ]).then(([kit, app]) => { | |
| kit.start(app, element, { | |
| node_ids: [0, 167], | |
| data, | |
| form: null, | |
| error: null | |
| }); | |
| }); | |
| } | |
| </script> | |
Xet Storage Details
- Size:
- 12.3 kB
- Xet hash:
- 838dc5d7e2ea74afb00f4fe1e9cac14368a34f75be393b8388f85ba2cefe730d
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.