diffusers/pr_11335/en/_app/immutable/nodes/10.09cdbc8e.js

import{s as Kp,o as Qp,n as D}from"../chunks/scheduler.8c3d61f6.js";import{S as em,i as om,g as a,s as t,r as m,A as tm,h as n,f as s,c as r,j as w,u as h,x as f,k as v,y as o,a as x,v as _,d as u,t as g,w as L}from"../chunks/index.da70eac4.js";import{T as C}from"../chunks/Tip.1d9b8c37.js";import{D as b}from"../chunks/Docstring.567bc132.js";import{C as rm}from"../chunks/CodeBlock.a9c4becf.js";import{E as am}from"../chunks/ExampleCodeBlock.15b54358.js";import{H as O,E as nm}from"../chunks/index.5d4ab994.js";function im(M){let i,$='To learn more about how to load LoRA weights, see the <a href="../../using-diffusers/loading_adapters#lora">LoRA</a> loading guide.';return{c(){i=a("p"),i.innerHTML=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-1fw6lx1"&&(i.innerHTML=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function sm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function dm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function lm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function cm(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function fm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function pm(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function mm(M){let i,$="Examples:",l,c,T;return c=new rm({props:{code:"JTIzJTIwQXNzdW1pbmclMjAlNjBwaXBlbGluZSU2MCUyMGlzJTIwYWxyZWFkeSUyMGxvYWRlZCUyMHdpdGglMjB0aGUlMjBMb1JBJTIwcGFyYW1ldGVycy4lMEFwaXBlbGluZS51bmxvYWRfbG9yYV93ZWlnaHRzKCklMEEuLi4=",highlighted:'<span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-comment"># Assuming `pipeline` is already loaded with the LoRA parameters.</span>\n<span class="hljs-meta">&gt;&gt;&gt; </span>pipeline.unload_lora_weights()\n<span class="hljs-meta">&gt;&gt;&gt; </span>...',wrap:!1}}),{c(){i=a("p"),i.textContent=$,l=t(),m(c.$$.fragment)},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-kvfsh7"&&(i.textContent=$),l=r(d),h(c.$$.fragment,d)},m(d,y){x(d,i,y),x(d,l,y),_(c,d,y),T=!0},p:D,i(d){T||(u(c.$$.fragment,d),T=!0)},o(d){g(c.$$.fragment,d),T=!1},d(d){d&&(s(i),s(l)),L(c,d)}}}function hm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function _m(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function um(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function gm(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function Lm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function xm(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function wm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function vm(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function bm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function $m(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function ym(M){let i,$="We support loading original format HunyuanVideo LoRA checkpoints.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-gyrs6h"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function Mm(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function Tm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function Dm(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function Am(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function Cm(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function Sm(M){let i,$="We support loading A1111 formatted LoRA checkpoints in a limited capacity.",l,c,T="This function is experimental and might change in the future.";return{c(){i=a("p"),i.textContent=$,l=t(),c=a("p"),c.textContent=T},l(d){i=n(d,"P",{"data-svelte-h":!0}),f(i)!=="svelte-15l1sdn"&&(i.textContent=$),l=r(d),c=n(d,"P",{"data-svelte-h":!0}),f(c)!=="svelte-3fufvn"&&(c.textContent=T)},m(d,y){x(d,i,y),x(d,l,y),x(d,c,y)},p:D,d(d){d&&(s(i),s(l),s(c))}}}function km(M){let i,$="This is an experimental API.";return{c(){i=a("p"),i.textContent=$},l(l){i=n(l,"P",{"data-svelte-h":!0}),f(i)!=="svelte-8w79b9"&&(i.textContent=$)},m(l,c){x(l,i,c)},p:D,d(l){l&&s(i)}}}function Rm(M){let i,$,l,c,T,d,y,af='LoRA is a fast and lightweight training method that inserts and trains a significantly smaller number of parameters instead of all the model parameters. This produces a smaller file (~100 MBs) and makes it easier to quickly train a model to learn a new concept. LoRA weights are typically loaded into the denoiser, text encoder or both. The denoiser usually corresponds to a UNet (<a href="/docs/diffusers/pr_11335/en/api/models/unet2d-cond#diffusers.UNet2DConditionModel">UNet2DConditionModel</a>, for example) or a Transformer (<a href="/docs/diffusers/pr_11335/en/api/models/sd3_transformer2d#diffusers.SD3Transformer2DModel">SD3Transformer2DModel</a>, for example). There are several classes for loading LoRA weights:',Li,ot,nf='<li><code>StableDiffusionLoraLoaderMixin</code> provides functions for loading and unloading, fusing and unfusing, enabling and disabling, and more functions for managing LoRA weights. This class can be used with any model.</li> <li><code>StableDiffusionXLLoraLoaderMixin</code> is a <a href="../../api/pipelines/stable_diffusion/stable_diffusion_xl">Stable Diffusion (SDXL)</a> version of the <code>StableDiffusionLoraLoaderMixin</code> class for loading and saving LoRA weights. It can only be used with the SDXL model.</li> <li><code>SD3LoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/blog/sd3" rel="nofollow">Stable Diffusion 3</a>.</li> <li><code>FluxLoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/flux" rel="nofollow">Flux</a>.</li> <li><code>CogVideoXLoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/cogvideox" rel="nofollow">CogVideoX</a>.</li> <li><code>Mochi1LoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/mochi" rel="nofollow">Mochi</a>.</li> <li><code>AuraFlowLoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/fal/AuraFlow" rel="nofollow">AuraFlow</a>.</li> <li><code>LTXVideoLoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/ltx_video" rel="nofollow">LTX-Video</a>.</li> <li><code>SanaLoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/sana" rel="nofollow">Sana</a>.</li> <li><code>HunyuanVideoLoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/hunyuan_video" rel="nofollow">HunyuanVideo</a>.</li> <li><code>Lumina2LoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/lumina2" rel="nofollow">Lumina2</a>.</li> <li><code>WanLoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/wan" rel="nofollow">Lumina2</a>.</li> <li><code>CogView4LoraLoaderMixin</code> provides similar functions for <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/cogview4" rel="nofollow">CogView3</a>.</li> <li><code>AmusedLoraLoaderMixin</code> is for the <a href="/docs/diffusers/pr_11335/en/api/pipelines/amused#diffusers.AmusedPipeline">AmusedPipeline</a>.</li> <li><code>LoraBaseMixin</code> provides a base class with several utility methods to fuse, unfuse, unload, LoRAs and more.</li>',xi,Xe,wi,tt,vi,P,rt,As,sa,sf=`Load LoRA layers into Stable Diffusion <a href="/docs/diffusers/pr_11335/en/api/models/unet2d-cond#diffusers.UNet2DConditionModel">UNet2DConditionModel</a> and
<a href="https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel" rel="nofollow"><code>CLIPTextModel</code></a>.`,Cs,qe,at,Ss,da,df="This will load the LoRA layers specified in <code>state_dict</code> into <code>text_encoder</code>",ks,ze,nt,Rs,la,lf="This will load the LoRA layers specified in <code>state_dict</code> into <code>unet</code>.",Ps,B,it,Hs,ca,cf=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.unet</code> and
<code>self.text_encoder</code>.`,Vs,fa,ff="All kwargs are forwarded to <code>self.lora_state_dict</code>.",Is,pa,pf=`See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is
loaded.`,Fs,ma,mf=`See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_unet">load_lora_into_unet()</a> for more details on how the state dict is
loaded into <code>self.unet</code>.`,Ws,ha,hf=`See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder">load_lora_into_text_encoder()</a> for more details on how the state
dict is loaded into <code>self.text_encoder</code>.`,Es,ce,st,Us,_a,_f="Return state dict for lora weights and the network alphas.",Ns,je,Xs,Be,dt,qs,ua,uf="Save the LoRA parameters corresponding to the UNet and text encoder.",bi,lt,$i,H,ct,zs,ga,gf=`Load LoRA layers into Stable Diffusion XL <a href="/docs/diffusers/pr_11335/en/api/models/unet2d-cond#diffusers.UNet2DConditionModel">UNet2DConditionModel</a>,
<a href="https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel" rel="nofollow"><code>CLIPTextModel</code></a>, and
<a href="https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModelWithProjection" rel="nofollow"><code>CLIPTextModelWithProjection</code></a>.`,js,Ge,ft,Bs,La,Lf="This will load the LoRA layers specified in <code>state_dict</code> into <code>text_encoder</code>",Gs,Oe,pt,Os,xa,xf="This will load the LoRA layers specified in <code>state_dict</code> into <code>unet</code>.",Js,G,mt,Ys,wa,wf=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.unet</code> and
<code>self.text_encoder</code>.`,Zs,va,vf="All kwargs are forwarded to <code>self.lora_state_dict</code>.",Ks,ba,bf=`See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is
loaded.`,Qs,$a,$f=`See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_unet">load_lora_into_unet()</a> for more details on how the state dict is
loaded into <code>self.unet</code>.`,ed,ya,yf=`See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder">load_lora_into_text_encoder()</a> for more details on how the state
dict is loaded into <code>self.text_encoder</code>.`,od,fe,ht,td,Ma,Mf="Return state dict for lora weights and the network alphas.",rd,Je,ad,Ye,_t,nd,Ta,Tf="Save the LoRA parameters corresponding to the UNet and text encoder.",yi,ut,Mi,S,gt,id,Da,Df=`Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/sd3_transformer2d#diffusers.SD3Transformer2DModel">SD3Transformer2DModel</a>,
<a href="https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel" rel="nofollow"><code>CLIPTextModel</code></a>, and
<a href="https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModelWithProjection" rel="nofollow"><code>CLIPTextModelWithProjection</code></a>.`,sd,Aa,Af='Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/stable_diffusion/stable_diffusion_3#diffusers.StableDiffusion3Pipeline">StableDiffusion3Pipeline</a>.',dd,Ze,Lt,ld,Ca,Cf="This will load the LoRA layers specified in <code>state_dict</code> into <code>text_encoder</code>",cd,Ke,xt,fd,Sa,Sf="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",pd,Z,wt,md,ka,kf=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.unet</code> and
<code>self.text_encoder</code>.`,hd,Ra,Rf="All kwargs are forwarded to <code>self.lora_state_dict</code>.",_d,Pa,Pf=`See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is
loaded.`,ud,Ha,Hf=`See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,gd,pe,vt,Ld,Va,Vf="Return state dict for lora weights and the network alphas.",xd,Qe,wd,eo,bt,vd,Ia,If="Save the LoRA parameters corresponding to the UNet and text encoder.",bd,me,$t,$d,Fa,Ff=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,yd,oo,Ti,yt,Di,A,Mt,Md,Wa,Wf=`Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/flux_transformer#diffusers.FluxTransformer2DModel">FluxTransformer2DModel</a>,
<a href="https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPTextModel" rel="nofollow"><code>CLIPTextModel</code></a>.`,Td,Ea,Ef='Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/stable_diffusion/stable_diffusion_3#diffusers.StableDiffusion3Pipeline">StableDiffusion3Pipeline</a>.',Dd,to,Tt,Ad,Ua,Uf="This will load the LoRA layers specified in <code>state_dict</code> into <code>text_encoder</code>",Cd,ro,Dt,Sd,Na,Nf="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",kd,K,At,Rd,Xa,Xf=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>.`,Pd,qa,qf="All kwargs are forwarded to <code>self.lora_state_dict</code>.",Hd,za,zf=`See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is
loaded.`,Vd,ja,jf=`See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,Id,he,Ct,Fd,Ba,Bf="Return state dict for lora weights and the network alphas.",Wd,ao,Ed,no,St,Ud,Ga,Gf="Save the LoRA parameters corresponding to the UNet and text encoder.",Nd,_e,kt,Xd,Oa,Of=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,qd,io,zd,ue,Rt,jd,Ja,Jf="Unloads the LoRA parameters.",Bd,so,Ai,Pt,Ci,V,Ht,Gd,Ya,Yf='Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/cogvideox_transformer3d#diffusers.CogVideoXTransformer3DModel">CogVideoXTransformer3DModel</a>. Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/cogvideox#diffusers.CogVideoXPipeline">CogVideoXPipeline</a>.',Od,lo,Vt,Jd,Za,Zf="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",Yd,co,It,Zd,Ka,Kf=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>. All kwargs are forwarded to <code>self.lora_state_dict</code>. See
<a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is loaded.
See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,Kd,ge,Ft,Qd,Qa,Qf="Return state dict for lora weights and the network alphas.",el,fo,ol,po,Wt,tl,en,ep="Save the LoRA parameters corresponding to the UNet and text encoder.",rl,Le,Et,al,on,op=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,nl,mo,Si,Ut,ki,I,Nt,il,tn,tp='Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/mochi_transformer3d#diffusers.MochiTransformer3DModel">MochiTransformer3DModel</a>. Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/mochi#diffusers.MochiPipeline">MochiPipeline</a>.',sl,ho,Xt,dl,rn,rp="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",ll,_o,qt,cl,an,ap=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>. All kwargs are forwarded to <code>self.lora_state_dict</code>. See
<a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is loaded.
See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,fl,xe,zt,pl,nn,np="Return state dict for lora weights and the network alphas.",ml,uo,hl,go,jt,_l,sn,ip="Save the LoRA parameters corresponding to the UNet and text encoder.",ul,we,Bt,gl,dn,sp=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,Ll,Lo,Ri,Gt,Pi,F,Ot,xl,ln,dp='Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/aura_flow_transformer2d#diffusers.AuraFlowTransformer2DModel">AuraFlowTransformer2DModel</a> Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/aura_flow#diffusers.AuraFlowPipeline">AuraFlowPipeline</a>.',wl,xo,Jt,vl,cn,lp="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",bl,wo,Yt,$l,fn,cp=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>. All kwargs are forwarded to <code>self.lora_state_dict</code>. See
<a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is loaded.
See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,yl,ve,Zt,Ml,pn,fp="Return state dict for lora weights and the network alphas.",Tl,vo,Dl,bo,Kt,Al,mn,pp="Save the LoRA parameters corresponding to the UNet and text encoder.",Cl,be,Qt,Sl,hn,mp=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,kl,$o,Hi,er,Vi,W,or,Rl,_n,hp='Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/ltx_video_transformer3d#diffusers.LTXVideoTransformer3DModel">LTXVideoTransformer3DModel</a>. Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/ltx_video#diffusers.LTXPipeline">LTXPipeline</a>.',Pl,yo,tr,Hl,un,_p="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",Vl,Mo,rr,Il,gn,up=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>. All kwargs are forwarded to <code>self.lora_state_dict</code>. See
<a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is loaded.
See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,Fl,$e,ar,Wl,Ln,gp="Return state dict for lora weights and the network alphas.",El,To,Ul,Do,nr,Nl,xn,Lp="Save the LoRA parameters corresponding to the UNet and text encoder.",Xl,ye,ir,ql,wn,xp=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,zl,Ao,Ii,sr,Fi,E,dr,jl,vn,wp='Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/sana_transformer2d#diffusers.SanaTransformer2DModel">SanaTransformer2DModel</a>. Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/sana#diffusers.SanaPipeline">SanaPipeline</a>.',Bl,Co,lr,Gl,bn,vp="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",Ol,So,cr,Jl,$n,bp=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>. All kwargs are forwarded to <code>self.lora_state_dict</code>. See
<a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is loaded.
See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,Yl,Me,fr,Zl,yn,$p="Return state dict for lora weights and the network alphas.",Kl,ko,Ql,Ro,pr,ec,Mn,yp="Save the LoRA parameters corresponding to the UNet and text encoder.",oc,Te,mr,tc,Tn,Mp=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,rc,Po,Wi,hr,Ei,U,_r,ac,Dn,Tp='Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/hunyuan_video_transformer_3d#diffusers.HunyuanVideoTransformer3DModel">HunyuanVideoTransformer3DModel</a>. Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/hunyuan_video#diffusers.HunyuanVideoPipeline">HunyuanVideoPipeline</a>.',nc,Ho,ur,ic,An,Dp="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",sc,Vo,gr,dc,Cn,Ap=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>. All kwargs are forwarded to <code>self.lora_state_dict</code>. See
<a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is loaded.
See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,lc,De,Lr,cc,Sn,Cp="Return state dict for lora weights and the network alphas.",fc,Io,pc,Fo,xr,mc,kn,Sp="Save the LoRA parameters corresponding to the UNet and text encoder.",hc,Ae,wr,_c,Rn,kp=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,uc,Wo,Ui,vr,Ni,N,br,gc,Pn,Rp='Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/lumina2_transformer2d#diffusers.Lumina2Transformer2DModel">Lumina2Transformer2DModel</a>. Specific to <code>Lumina2Text2ImgPipeline</code>.',Lc,Eo,$r,xc,Hn,Pp="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",wc,Uo,yr,vc,Vn,Hp=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>. All kwargs are forwarded to <code>self.lora_state_dict</code>. See
<a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is loaded.
See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,bc,Ce,Mr,$c,In,Vp="Return state dict for lora weights and the network alphas.",yc,No,Mc,Xo,Tr,Tc,Fn,Ip="Save the LoRA parameters corresponding to the UNet and text encoder.",Dc,Se,Dr,Ac,Wn,Fp=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,Cc,qo,Xi,Ar,qi,X,Cr,Sc,En,Wp='Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/wan_transformer_3d#diffusers.WanTransformer3DModel">WanTransformer3DModel</a>. Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/wan#diffusers.WanPipeline">WanPipeline</a> and <code>[WanImageToVideoPipeline</code>].',kc,zo,Sr,Rc,Un,Ep="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",Pc,jo,kr,Hc,Nn,Up=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>. All kwargs are forwarded to <code>self.lora_state_dict</code>. See
<a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is loaded.
See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,Vc,ke,Rr,Ic,Xn,Np="Return state dict for lora weights and the network alphas.",Fc,Bo,Wc,Go,Pr,Ec,qn,Xp="Save the LoRA parameters corresponding to the UNet and text encoder.",Uc,Re,Hr,Nc,zn,qp=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,Xc,Oo,zi,Vr,ji,q,Ir,qc,jn,zp='Load LoRA layers into <a href="/docs/diffusers/pr_11335/en/api/models/wan_transformer_3d#diffusers.WanTransformer3DModel">WanTransformer3DModel</a>. Specific to <a href="/docs/diffusers/pr_11335/en/api/pipelines/cogview4#diffusers.CogView4Pipeline">CogView4Pipeline</a>.',zc,Jo,Fr,jc,Bn,jp="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",Bc,Yo,Wr,Gc,Gn,Bp=`Load LoRA weights specified in <code>pretrained_model_name_or_path_or_dict</code> into <code>self.transformer</code> and
<code>self.text_encoder</code>. All kwargs are forwarded to <code>self.lora_state_dict</code>. See
<a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a> for more details on how the state dict is loaded.
See <code>~loaders.StableDiffusionLoraLoaderMixin.load_lora_into_transformer</code> for more details on how the state
dict is loaded into <code>self.transformer</code>.`,Oc,Pe,Er,Jc,On,Gp="Return state dict for lora weights and the network alphas.",Yc,Zo,Zc,Ko,Ur,Kc,Jn,Op="Save the LoRA parameters corresponding to the UNet and text encoder.",Qc,He,Nr,ef,Yn,Jp=`Reverses the effect of
<a href="https://huggingface.co/docs/diffusers/main/en/api/loaders#diffusers.loaders.LoraBaseMixin.fuse_lora" rel="nofollow"><code>pipe.fuse_lora()</code></a>.`,of,Qo,Bi,Xr,Gi,Ve,qr,tf,et,zr,rf,Zn,Yp="This will load the LoRA layers specified in <code>state_dict</code> into <code>transformer</code>.",Oi,jr,Ji,Br,Gr,Yi,Or,Zi,gi,Ki;return T=new O({props:{title:"LoRA",local:"lora",headingTag:"h1"}}),Xe=new C({props:{$$slots:{default:[im]},$$scope:{ctx:M}}}),tt=new O({props:{title:"StableDiffusionLoraLoaderMixin",local:"diffusers.loaders.StableDiffusionLoraLoaderMixin",headingTag:"h2"}}),rt=new b({props:{name:"class diffusers.loaders.StableDiffusionLoraLoaderMixin",anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L117"}}),at=new b({props:{name:"load_lora_into_text_encoder",anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder",parameters:[{name:"state_dict",val:""},{name:"network_alphas",val:""},{name:"text_encoder",val:""},{name:"prefix",val:" = None"},{name:"lora_scale",val:" = 1.0"},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The key should be prefixed with an
additional <code>text_encoder</code> to distinguish between unet lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder.network_alphas",description:`<strong>network_alphas</strong> (<code>Dict[str, float]</code>) &#x2014;
The value of the network alpha used for stable learning and preventing underflow. This value has the
same meaning as the <code>--network_alpha</code> option in the kohya-ss trainer script. Refer to <a href="https://github.com/darkstorm2150/sd-scripts/blob/main/docs/train_network_README-en.md#execute-learning" rel="nofollow">this
link</a>.`,name:"network_alphas"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder.text_encoder",description:`<strong>text_encoder</strong> (<code>CLIPTextModel</code>) &#x2014;
The text encoder model to load the LoRA layers into.`,name:"text_encoder"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder.prefix",description:`<strong>prefix</strong> (<code>str</code>) &#x2014;
Expected prefix of the <code>text_encoder</code> in the <code>state_dict</code>.`,name:"prefix"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder.lora_scale",description:`<strong>lora_scale</strong> (<code>float</code>) &#x2014;
How much to scale the output of the lora linear layer before it is added with the output of the regular
lora layer.`,name:"lora_scale"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_text_encoder.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L417"}}),nt=new b({props:{name:"load_lora_into_unet",anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_unet",parameters:[{name:"state_dict",val:""},{name:"network_alphas",val:""},{name:"unet",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_unet.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_unet.network_alphas",description:`<strong>network_alphas</strong> (<code>Dict[str, float]</code>) &#x2014;
The value of the network alpha used for stable learning and preventing underflow. This value has the
same meaning as the <code>--network_alpha</code> option in the kohya-ss trainer script. Refer to <a href="https://github.com/darkstorm2150/sd-scripts/blob/main/docs/train_network_README-en.md#execute-learning" rel="nofollow">this
link</a>.`,name:"network_alphas"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_unet.unet",description:`<strong>unet</strong> (<code>UNet2DConditionModel</code>) &#x2014;
The UNet model to load the LoRA layers into.`,name:"unet"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_unet.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_unet.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_into_unet.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L340"}}),it=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"hotswap",val:": bool = False"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.load_lora_weights.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L127"}}),st=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict.weight_name",description:`<strong>weight_name</strong> (<code>str</code>, <em>optional</em>, defaults to None) &#x2014;
Name of the serialized state dict file.`,name:"weight_name"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L225"}}),je=new C({props:{warning:!0,$$slots:{default:[sm]},$$scope:{ctx:M}}}),dt=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"unet_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"text_encoder_lora_layers",val:": typing.Dict[str, torch.nn.modules.module.Module] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.save_lora_weights.unet_lora_layers",description:`<strong>unet_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>unet</code>.`,name:"unet_lora_layers"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.save_lora_weights.text_encoder_lora_layers",description:`<strong>text_encoder_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>text_encoder</code>. Must explicitly pass the text
encoder LoRA state dict because it comes from &#x1F917; Transformers.`,name:"text_encoder_lora_layers"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.StableDiffusionLoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L491"}}),lt=new O({props:{title:"StableDiffusionXLLoraLoaderMixin",local:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin",headingTag:"h2"}}),ct=new b({props:{name:"class diffusers.loaders.StableDiffusionXLLoraLoaderMixin",anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L613"}}),ft=new b({props:{name:"load_lora_into_text_encoder",anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_text_encoder",parameters:[{name:"state_dict",val:""},{name:"network_alphas",val:""},{name:"text_encoder",val:""},{name:"prefix",val:" = None"},{name:"lora_scale",val:" = 1.0"},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_text_encoder.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The key should be prefixed with an
additional <code>text_encoder</code> to distinguish between unet lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_text_encoder.network_alphas",description:`<strong>network_alphas</strong> (<code>Dict[str, float]</code>) &#x2014;
The value of the network alpha used for stable learning and preventing underflow. This value has the
same meaning as the <code>--network_alpha</code> option in the kohya-ss trainer script. Refer to <a href="https://github.com/darkstorm2150/sd-scripts/blob/main/docs/train_network_README-en.md#execute-learning" rel="nofollow">this
link</a>.`,name:"network_alphas"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_text_encoder.text_encoder",description:`<strong>text_encoder</strong> (<code>CLIPTextModel</code>) &#x2014;
The text encoder model to load the LoRA layers into.`,name:"text_encoder"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_text_encoder.prefix",description:`<strong>prefix</strong> (<code>str</code>) &#x2014;
Expected prefix of the <code>text_encoder</code> in the <code>state_dict</code>.`,name:"prefix"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_text_encoder.lora_scale",description:`<strong>lora_scale</strong> (<code>float</code>) &#x2014;
How much to scale the output of the lora linear layer before it is added with the output of the regular
lora layer.`,name:"lora_scale"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_text_encoder.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_text_encoder.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_text_encoder.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L908"}}),pt=new b({props:{name:"load_lora_into_unet",anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_unet",parameters:[{name:"state_dict",val:""},{name:"network_alphas",val:""},{name:"unet",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_unet.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_unet.network_alphas",description:`<strong>network_alphas</strong> (<code>Dict[str, float]</code>) &#x2014;
The value of the network alpha used for stable learning and preventing underflow. This value has the
same meaning as the <code>--network_alpha</code> option in the kohya-ss trainer script. Refer to <a href="https://github.com/darkstorm2150/sd-scripts/blob/main/docs/train_network_README-en.md#execute-learning" rel="nofollow">this
link</a>.`,name:"network_alphas"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_unet.unet",description:`<strong>unet</strong> (<code>UNet2DConditionModel</code>) &#x2014;
The UNet model to load the LoRA layers into.`,name:"unet"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_unet.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_unet.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_into_unet.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L830"}}),mt=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:": typing.Optional[str] = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L624"}}),ht=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.lora_state_dict.weight_name",description:`<strong>weight_name</strong> (<code>str</code>, <em>optional</em>, defaults to None) &#x2014;
Name of the serialized state dict file.`,name:"weight_name"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L714"}}),Je=new C({props:{warning:!0,$$slots:{default:[dm]},$$scope:{ctx:M}}}),_t=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"unet_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"text_encoder_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"text_encoder_2_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.save_lora_weights.unet_lora_layers",description:`<strong>unet_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>unet</code>.`,name:"unet_lora_layers"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.save_lora_weights.text_encoder_lora_layers",description:`<strong>text_encoder_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>text_encoder</code>. Must explicitly pass the text
encoder LoRA state dict because it comes from &#x1F917; Transformers.`,name:"text_encoder_lora_layers"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.save_lora_weights.text_encoder_2_lora_layers",description:`<strong>text_encoder_2_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>text_encoder_2</code>. Must explicitly pass the text
encoder LoRA state dict because it comes from &#x1F917; Transformers.`,name:"text_encoder_2_lora_layers"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.StableDiffusionXLLoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L983"}}),ut=new O({props:{title:"SD3LoraLoaderMixin",local:"diffusers.loaders.SD3LoraLoaderMixin",headingTag:"h2"}}),gt=new b({props:{name:"class diffusers.loaders.SD3LoraLoaderMixin",anchor:"diffusers.loaders.SD3LoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1113"}}),Lt=new b({props:{name:"load_lora_into_text_encoder",anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_text_encoder",parameters:[{name:"state_dict",val:""},{name:"network_alphas",val:""},{name:"text_encoder",val:""},{name:"prefix",val:" = None"},{name:"lora_scale",val:" = 1.0"},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_text_encoder.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The key should be prefixed with an
additional <code>text_encoder</code> to distinguish between unet lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_text_encoder.network_alphas",description:`<strong>network_alphas</strong> (<code>Dict[str, float]</code>) &#x2014;
The value of the network alpha used for stable learning and preventing underflow. This value has the
same meaning as the <code>--network_alpha</code> option in the kohya-ss trainer script. Refer to <a href="https://github.com/darkstorm2150/sd-scripts/blob/main/docs/train_network_README-en.md#execute-learning" rel="nofollow">this
link</a>.`,name:"network_alphas"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_text_encoder.text_encoder",description:`<strong>text_encoder</strong> (<code>CLIPTextModel</code>) &#x2014;
The text encoder model to load the LoRA layers into.`,name:"text_encoder"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_text_encoder.prefix",description:`<strong>prefix</strong> (<code>str</code>) &#x2014;
Expected prefix of the <code>text_encoder</code> in the <code>state_dict</code>.`,name:"prefix"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_text_encoder.lora_scale",description:`<strong>lora_scale</strong> (<code>float</code>) &#x2014;
How much to scale the output of the lora linear layer before it is added with the output of the regular
lora layer.`,name:"lora_scale"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_text_encoder.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_text_encoder.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_text_encoder.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1388"}}),xt=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>SD3Transformer2DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1328"}}),wt=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"hotswap",val:": bool = False"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.load_lora_weights.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1223"}}),vt=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.SD3LoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.SD3LoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1126"}}),Qe=new C({props:{warning:!0,$$slots:{default:[lm]},$$scope:{ctx:M}}}),bt=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.SD3LoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"text_encoder_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"text_encoder_2_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.SD3LoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.save_lora_weights.text_encoder_lora_layers",description:`<strong>text_encoder_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>text_encoder</code>. Must explicitly pass the text
encoder LoRA state dict because it comes from &#x1F917; Transformers.`,name:"text_encoder_lora_layers"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.save_lora_weights.text_encoder_2_lora_layers",description:`<strong>text_encoder_2_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>text_encoder_2</code>. Must explicitly pass the text
encoder LoRA state dict because it comes from &#x1F917; Transformers.`,name:"text_encoder_2_lora_layers"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1463"}}),$t=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.SD3LoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer', 'text_encoder', 'text_encoder_2']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.SD3LoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"},{anchor:"diffusers.loaders.SD3LoraLoaderMixin.unfuse_lora.unfuse_text_encoder",description:`<strong>unfuse_text_encoder</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014;
Whether to unfuse the text encoder LoRA parameters. If the text encoder wasn&#x2019;t monkey-patched with the
LoRA parameters then it won&#x2019;t have any effect.`,name:"unfuse_text_encoder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1575"}}),oo=new C({props:{warning:!0,$$slots:{default:[cm]},$$scope:{ctx:M}}}),yt=new O({props:{title:"FluxLoraLoaderMixin",local:"diffusers.loaders.FluxLoraLoaderMixin",headingTag:"h2"}}),Mt=new b({props:{name:"class diffusers.loaders.FluxLoraLoaderMixin",anchor:"diffusers.loaders.FluxLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1929"}}),Tt=new b({props:{name:"load_lora_into_text_encoder",anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_text_encoder",parameters:[{name:"state_dict",val:""},{name:"network_alphas",val:""},{name:"text_encoder",val:""},{name:"prefix",val:" = None"},{name:"lora_scale",val:" = 1.0"},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_text_encoder.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The key should be prefixed with an
additional <code>text_encoder</code> to distinguish between unet lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_text_encoder.network_alphas",description:`<strong>network_alphas</strong> (<code>Dict[str, float]</code>) &#x2014;
The value of the network alpha used for stable learning and preventing underflow. This value has the
same meaning as the <code>--network_alpha</code> option in the kohya-ss trainer script. Refer to <a href="https://github.com/darkstorm2150/sd-scripts/blob/main/docs/train_network_README-en.md#execute-learning" rel="nofollow">this
link</a>.`,name:"network_alphas"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_text_encoder.text_encoder",description:`<strong>text_encoder</strong> (<code>CLIPTextModel</code>) &#x2014;
The text encoder model to load the LoRA layers into.`,name:"text_encoder"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_text_encoder.prefix",description:`<strong>prefix</strong> (<code>str</code>) &#x2014;
Expected prefix of the <code>text_encoder</code> in the <code>state_dict</code>.`,name:"prefix"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_text_encoder.lora_scale",description:`<strong>lora_scale</strong> (<code>float</code>) &#x2014;
How much to scale the output of the lora linear layer before it is added with the output of the regular
lora layer.`,name:"lora_scale"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_text_encoder.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_text_encoder.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_text_encoder.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L2341"}}),Dt=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"network_alphas",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_transformer.network_alphas",description:`<strong>network_alphas</strong> (<code>Dict[str, float]</code>) &#x2014;
The value of the network alpha used for stable learning and preventing underflow. This value has the
same meaning as the <code>--network_alpha</code> option in the kohya-ss trainer script. Refer to <a href="https://github.com/darkstorm2150/sd-scripts/blob/main/docs/train_network_README-en.md#execute-learning" rel="nofollow">this
link</a>.`,name:"network_alphas"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>FluxTransformer2DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L2216"}}),At=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"hotswap",val:": bool = False"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
\`Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.load_lora_weights.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter. If the new
adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need to call an
additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L2076"}}),Ct=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.FluxLoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"return_alphas",val:": bool = False"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.FluxLoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1942"}}),ao=new C({props:{warning:!0,$$slots:{default:[fm]},$$scope:{ctx:M}}}),St=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.FluxLoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"text_encoder_lora_layers",val:": typing.Dict[str, torch.nn.modules.module.Module] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.FluxLoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.save_lora_weights.text_encoder_lora_layers",description:`<strong>text_encoder_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>text_encoder</code>. Must explicitly pass the text
encoder LoRA state dict because it comes from &#x1F917; Transformers.`,name:"text_encoder_lora_layers"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.FluxLoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L2416"}}),kt=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.FluxLoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer', 'text_encoder']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.FluxLoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L2531"}}),io=new C({props:{warning:!0,$$slots:{default:[pm]},$$scope:{ctx:M}}}),Rt=new b({props:{name:"unload_lora_weights",anchor:"diffusers.loaders.FluxLoraLoaderMixin.unload_lora_weights",parameters:[{name:"reset_to_overwritten_params",val:" = False"}],parametersDescription:[{anchor:"diffusers.loaders.FluxLoraLoaderMixin.unload_lora_weights.reset_to_overwritten_params",description:`<strong>reset_to_overwritten_params</strong> (<code>bool</code>, defaults to <code>False</code>) &#x2014; Whether to reset the LoRA-loaded modules
to their original params. Refer to the <a href="https://huggingface.co/docs/diffusers/main/en/api/pipelines/flux" rel="nofollow">Flux
documentation</a> to learn more.`,name:"reset_to_overwritten_params"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L2552"}}),so=new am({props:{anchor:"diffusers.loaders.FluxLoraLoaderMixin.unload_lora_weights.example",$$slots:{default:[mm]},$$scope:{ctx:M}}}),Pt=new O({props:{title:"CogVideoXLoraLoaderMixin",local:"diffusers.loaders.CogVideoXLoraLoaderMixin",headingTag:"h2"}}),Ht=new b({props:{name:"class diffusers.loaders.CogVideoXLoraLoaderMixin",anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3031"}}),Vt=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>CogVideoXTransformer3DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3187"}}),It=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3137"}}),Ft=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3039"}}),fo=new C({props:{warning:!0,$$slots:{default:[hm]},$$scope:{ctx:M}}}),Wt=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3248"}}),Et=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.CogVideoXLoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3343"}}),mo=new C({props:{warning:!0,$$slots:{default:[_m]},$$scope:{ctx:M}}}),Ut=new O({props:{title:"Mochi1LoraLoaderMixin",local:"diffusers.loaders.Mochi1LoraLoaderMixin",headingTag:"h2"}}),Nt=new b({props:{name:"class diffusers.loaders.Mochi1LoraLoaderMixin",anchor:"diffusers.loaders.Mochi1LoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3361"}}),Xt=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>MochiTransformer3DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3518"}}),qt=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3468"}}),zt=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3369"}}),uo=new C({props:{warning:!0,$$slots:{default:[um]},$$scope:{ctx:M}}}),jt=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3579"}}),Bt=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.Mochi1LoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3676"}}),Lo=new C({props:{warning:!0,$$slots:{default:[gm]},$$scope:{ctx:M}}}),Gt=new O({props:{title:"AuraFlowLoraLoaderMixin",local:"diffusers.loaders.AuraFlowLoraLoaderMixin",headingTag:"h2"}}),Ot=new b({props:{name:"class diffusers.loaders.AuraFlowLoraLoaderMixin",anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1596"}}),Jt=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>AuraFlowTransformer2DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1753"}}),Yt=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1703"}}),Zt=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1604"}}),vo=new C({props:{warning:!0,$$slots:{default:[Lm]},$$scope:{ctx:M}}}),Kt=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1814"}}),Qt=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer', 'text_encoder']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.AuraFlowLoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L1911"}}),$o=new C({props:{warning:!0,$$slots:{default:[xm]},$$scope:{ctx:M}}}),er=new O({props:{title:"LTXVideoLoraLoaderMixin",local:"diffusers.loaders.LTXVideoLoraLoaderMixin",headingTag:"h2"}}),or=new b({props:{name:"class diffusers.loaders.LTXVideoLoraLoaderMixin",anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3694"}}),tr=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>LTXVideoTransformer3DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3851"}}),rr=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3801"}}),ar=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3702"}}),To=new C({props:{warning:!0,$$slots:{default:[wm]},$$scope:{ctx:M}}}),nr=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L3912"}}),ir=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.LTXVideoLoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4009"}}),Ao=new C({props:{warning:!0,$$slots:{default:[vm]},$$scope:{ctx:M}}}),sr=new O({props:{title:"SanaLoraLoaderMixin",local:"diffusers.loaders.SanaLoraLoaderMixin",headingTag:"h2"}}),dr=new b({props:{name:"class diffusers.loaders.SanaLoraLoaderMixin",anchor:"diffusers.loaders.SanaLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4027"}}),lr=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>SanaTransformer2DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4184"}}),cr=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4134"}}),fr=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.SanaLoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.SanaLoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4035"}}),ko=new C({props:{warning:!0,$$slots:{default:[bm]},$$scope:{ctx:M}}}),pr=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.SanaLoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.SanaLoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4245"}}),mr=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.SanaLoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.SanaLoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.SanaLoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4342"}}),Po=new C({props:{warning:!0,$$slots:{default:[$m]},$$scope:{ctx:M}}}),hr=new O({props:{title:"HunyuanVideoLoraLoaderMixin",local:"diffusers.loaders.HunyuanVideoLoraLoaderMixin",headingTag:"h2"}}),_r=new b({props:{name:"class diffusers.loaders.HunyuanVideoLoraLoaderMixin",anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4360"}}),ur=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>HunyuanVideoTransformer3DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4520"}}),gr=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4470"}}),Lr=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4368"}}),Io=new C({props:{warning:!0,$$slots:{default:[ym]},$$scope:{ctx:M}}}),xr=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4581"}}),wr=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.HunyuanVideoLoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4678"}}),Wo=new C({props:{warning:!0,$$slots:{default:[Mm]},$$scope:{ctx:M}}}),vr=new O({props:{title:"Lumina2LoraLoaderMixin",local:"diffusers.loaders.Lumina2LoraLoaderMixin",headingTag:"h2"}}),br=new b({props:{name:"class diffusers.loaders.Lumina2LoraLoaderMixin",anchor:"diffusers.loaders.Lumina2LoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4696"}}),$r=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>Lumina2Transformer2DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4857"}}),yr=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4807"}}),Mr=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4704"}}),No=new C({props:{warning:!0,$$slots:{default:[Tm]},$$scope:{ctx:M}}}),Tr=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L4918"}}),Dr=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.Lumina2LoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5015"}}),qo=new C({props:{warning:!0,$$slots:{default:[Dm]},$$scope:{ctx:M}}}),Ar=new O({props:{title:"WanLoraLoaderMixin",local:"diffusers.loaders.WanLoraLoaderMixin",headingTag:"h2"}}),Cr=new b({props:{name:"class diffusers.loaders.WanLoraLoaderMixin",anchor:"diffusers.loaders.WanLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5033"}}),Sr=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>WanTransformer3DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5223"}}),kr=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5169"}}),Rr=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.WanLoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.WanLoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5041"}}),Bo=new C({props:{warning:!0,$$slots:{default:[Am]},$$scope:{ctx:M}}}),Pr=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.WanLoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.WanLoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5284"}}),Hr=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.WanLoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.WanLoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.WanLoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5381"}}),Oo=new C({props:{warning:!0,$$slots:{default:[Cm]},$$scope:{ctx:M}}}),Vr=new O({props:{title:"CogView4LoraLoaderMixin",local:"diffusers.loaders.CogView4LoraLoaderMixin",headingTag:"h2"}}),Ir=new b({props:{name:"class diffusers.loaders.CogView4LoraLoaderMixin",anchor:"diffusers.loaders.CogView4LoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5399"}}),Fr=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>CogView4Transformer2DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5556"}}),Wr=new b({props:{name:"load_lora_weights",anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_weights",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"adapter_name",val:" = None"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_weights.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_weights.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_weights.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.load_lora_weights.kwargs",description:`<strong>kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
See <a href="/docs/diffusers/pr_11335/en/api/loaders/lora#diffusers.loaders.StableDiffusionLoraLoaderMixin.lora_state_dict">lora_state_dict()</a>.`,name:"kwargs"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5506"}}),Er=new b({props:{name:"lora_state_dict",anchor:"diffusers.loaders.CogView4LoraLoaderMixin.lora_state_dict",parameters:[{name:"pretrained_model_name_or_path_or_dict",val:": typing.Union[str, typing.Dict[str, torch.Tensor]]"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.lora_state_dict.pretrained_model_name_or_path_or_dict",description:`<strong>pretrained_model_name_or_path_or_dict</strong> (<code>str</code> or <code>os.PathLike</code> or <code>dict</code>) &#x2014;
Can be either:</p>
<ul>
<li>A string, the <em>model id</em> (for example <code>google/ddpm-celebahq-256</code>) of a pretrained model hosted on
the Hub.</li>
<li>A path to a <em>directory</em> (for example <code>./my_model_directory</code>) containing the model weights saved
with <a href="/docs/diffusers/pr_11335/en/api/models/overview#diffusers.ModelMixin.save_pretrained">ModelMixin.save_pretrained()</a>.</li>
<li>A <a href="https://pytorch.org/tutorials/beginner/saving_loading_models.html#what-is-a-state-dict" rel="nofollow">torch state
dict</a>.</li>
</ul>`,name:"pretrained_model_name_or_path_or_dict"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.lora_state_dict.cache_dir",description:`<strong>cache_dir</strong> (<code>Union[str, os.PathLike]</code>, <em>optional</em>) &#x2014;
Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
is not used.`,name:"cache_dir"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.lora_state_dict.force_download",description:`<strong>force_download</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether or not to force the (re-)download of the model weights and configuration files, overriding the
cached versions if they exist.`,name:"force_download"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.lora_state_dict.proxies",description:`<strong>proxies</strong> (<code>Dict[str, str]</code>, <em>optional</em>) &#x2014;
A dictionary of proxy servers to use by protocol or endpoint, for example, <code>{&apos;http&apos;: &apos;foo.bar:3128&apos;, &apos;http://hostname&apos;: &apos;foo.bar:4012&apos;}</code>. The proxies are used on each request.`,name:"proxies"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.lora_state_dict.local_files_only",description:`<strong>local_files_only</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to only load local model weights and configuration files or not. If set to <code>True</code>, the model
won&#x2019;t be downloaded from the Hub.`,name:"local_files_only"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.lora_state_dict.token",description:`<strong>token</strong> (<code>str</code> or <em>bool</em>, <em>optional</em>) &#x2014;
The token to use as HTTP bearer authorization for remote files. If <code>True</code>, the token generated from
<code>diffusers-cli login</code> (stored in <code>~/.huggingface</code>) is used.`,name:"token"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.lora_state_dict.revision",description:`<strong>revision</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;main&quot;</code>) &#x2014;
The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
allowed by Git.`,name:"revision"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.lora_state_dict.subfolder",description:`<strong>subfolder</strong> (<code>str</code>, <em>optional</em>, defaults to <code>&quot;&quot;</code>) &#x2014;
The subfolder location of a model file within a larger model repository on the Hub or locally.`,name:"subfolder"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5407"}}),Zo=new C({props:{warning:!0,$$slots:{default:[Sm]},$$scope:{ctx:M}}}),Ur=new b({props:{name:"save_lora_weights",anchor:"diffusers.loaders.CogView4LoraLoaderMixin.save_lora_weights",parameters:[{name:"save_directory",val:": typing.Union[str, os.PathLike]"},{name:"transformer_lora_layers",val:": typing.Dict[str, typing.Union[torch.nn.modules.module.Module, torch.Tensor]] = None"},{name:"is_main_process",val:": bool = True"},{name:"weight_name",val:": str = None"},{name:"save_function",val:": typing.Callable = None"},{name:"safe_serialization",val:": bool = True"}],parametersDescription:[{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.save_lora_weights.save_directory",description:`<strong>save_directory</strong> (<code>str</code> or <code>os.PathLike</code>) &#x2014;
Directory to save LoRA parameters to. Will be created if it doesn&#x2019;t exist.`,name:"save_directory"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.save_lora_weights.transformer_lora_layers",description:`<strong>transformer_lora_layers</strong> (<code>Dict[str, torch.nn.Module]</code> or <code>Dict[str, torch.Tensor]</code>) &#x2014;
State dict of the LoRA layers corresponding to the <code>transformer</code>.`,name:"transformer_lora_layers"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.save_lora_weights.is_main_process",description:`<strong>is_main_process</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether the process calling this is the main process or not. Useful during distributed training and you
need to call this function on all processes. In this case, set <code>is_main_process=True</code> only on the main
process to avoid race conditions.`,name:"is_main_process"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.save_lora_weights.save_function",description:`<strong>save_function</strong> (<code>Callable</code>) &#x2014;
The function to use to save the state dictionary. Useful during distributed training when you need to
replace <code>torch.save</code> with another method. Can be configured with the environment variable
<code>DIFFUSERS_SAVE_MODE</code>.`,name:"save_function"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.save_lora_weights.safe_serialization",description:`<strong>safe_serialization</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether to save the model using <code>safetensors</code> or the traditional PyTorch way with <code>pickle</code>.`,name:"safe_serialization"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5617"}}),Nr=new b({props:{name:"unfuse_lora",anchor:"diffusers.loaders.CogView4LoraLoaderMixin.unfuse_lora",parameters:[{name:"components",val:": typing.List[str] = ['transformer']"},{name:"**kwargs",val:""}],parametersDescription:[{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.unfuse_lora.components",description:"<strong>components</strong> (<code>List[str]</code>) &#x2014; List of LoRA-injectable components to unfuse LoRA from.",name:"components"},{anchor:"diffusers.loaders.CogView4LoraLoaderMixin.unfuse_lora.unfuse_transformer",description:"<strong>unfuse_transformer</strong> (<code>bool</code>, defaults to <code>True</code>) &#x2014; Whether to unfuse the UNet LoRA parameters.",name:"unfuse_transformer"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L5714"}}),Qo=new C({props:{warning:!0,$$slots:{default:[km]},$$scope:{ctx:M}}}),Xr=new O({props:{title:"AmusedLoraLoaderMixin",local:"diffusers.loaders.AmusedLoraLoaderMixin",headingTag:"h2"}}),qr=new b({props:{name:"class diffusers.loaders.AmusedLoraLoaderMixin",anchor:"diffusers.loaders.AmusedLoraLoaderMixin",parameters:[],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L2824"}}),zr=new b({props:{name:"load_lora_into_transformer",anchor:"diffusers.loaders.AmusedLoraLoaderMixin.load_lora_into_transformer",parameters:[{name:"state_dict",val:""},{name:"network_alphas",val:""},{name:"transformer",val:""},{name:"adapter_name",val:" = None"},{name:"_pipeline",val:" = None"},{name:"low_cpu_mem_usage",val:" = False"},{name:"hotswap",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.loaders.AmusedLoraLoaderMixin.load_lora_into_transformer.state_dict",description:`<strong>state_dict</strong> (<code>dict</code>) &#x2014;
A standard state dict containing the lora layer parameters. The keys can either be indexed directly
into the unet or prefixed with an additional <code>unet</code> which can be used to distinguish between text
encoder lora layers.`,name:"state_dict"},{anchor:"diffusers.loaders.AmusedLoraLoaderMixin.load_lora_into_transformer.network_alphas",description:`<strong>network_alphas</strong> (<code>Dict[str, float]</code>) &#x2014;
The value of the network alpha used for stable learning and preventing underflow. This value has the
same meaning as the <code>--network_alpha</code> option in the kohya-ss trainer script. Refer to <a href="https://github.com/darkstorm2150/sd-scripts/blob/main/docs/train_network_README-en.md#execute-learning" rel="nofollow">this
link</a>.`,name:"network_alphas"},{anchor:"diffusers.loaders.AmusedLoraLoaderMixin.load_lora_into_transformer.transformer",description:`<strong>transformer</strong> (<code>UVit2DModel</code>) &#x2014;
The Transformer model to load the LoRA layers into.`,name:"transformer"},{anchor:"diffusers.loaders.AmusedLoraLoaderMixin.load_lora_into_transformer.adapter_name",description:`<strong>adapter_name</strong> (<code>str</code>, <em>optional</em>) &#x2014;
Adapter name to be used for referencing the loaded adapter model. If not specified, it will use
<code>default_{i}</code> where i is the total number of adapters being loaded.`,name:"adapter_name"},{anchor:"diffusers.loaders.AmusedLoraLoaderMixin.load_lora_into_transformer.low_cpu_mem_usage",description:`<strong>low_cpu_mem_usage</strong> (<code>bool</code>, <em>optional</em>) &#x2014;
Speed up model loading by only loading the pretrained LoRA weights and not initializing the random
weights.`,name:"low_cpu_mem_usage"},{anchor:"diffusers.loaders.AmusedLoraLoaderMixin.load_lora_into_transformer.hotswap",description:`<strong>hotswap</strong>  &#x2014; (<code>bool</code>, <em>optional</em>)
Defaults to <code>False</code>. Whether to substitute an existing (LoRA) adapter with the newly loaded adapter
in-place. This means that, instead of loading an additional adapter, this will take the existing
adapter weights and replace them with the weights of the new adapter. This can be faster and more
memory efficient. However, the main advantage of hotswapping is that when the model is compiled with
torch.compile, loading the new adapter does not require recompilation of the model. When using
hotswapping, the passed <code>adapter_name</code> should be the name of an already loaded adapter.</p>
<p>If the new adapter and the old adapter have different ranks and/or LoRA alphas (i.e. scaling), you need
to call an additional method before loading the adapter:`,name:"hotswap"}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora/lora_pipeline.py#L2829"}}),jr=new O({props:{title:"LoraBaseMixin",local:"diffusers.loaders.LoraBaseMixin",headingTag:"h2"}}),Gr=new b({props:{name:"class diffusers.loaders.LoraBaseMixin",anchor:"diffusers.loaders.LoraBaseMixin",parameters:[{name:"*args",val:""},{name:"**kwargs",val:""}],source:"https://github.com/huggingface/diffusers/blob/vr_11335/src/diffusers/loaders/lora_base.py#L73"}}),Or=new nm({props:{source:"https://github.com/huggingface/diffusers/blob/main/docs/source/en/api/loaders/lora.md"}}),{c(){i=a("meta"),$=t(),l=a("p"),c=t(),m(T.$$.fragment),d=t(),y=a("p"),y.innerHTML=af,Li=t(),ot=a("ul"),ot.innerHTML=nf,xi=t(),m(Xe.$$.fragment),wi=t(),m(tt.$$.fragment),vi=t(),P=a("div"),m(rt.$$.fragment),As=t(),sa=a("p"),sa.innerHTML=sf,Cs=t(),qe=a("div"),m(at.$$.fragment),Ss=t(),da=a("p"),da.innerHTML=df,ks=t(),ze=a("div"),m(nt.$$.fragment),Rs=t(),la=a("p"),la.innerHTML=lf,Ps=t(),B=a("div"),m(it.$$.fragment),Hs=t(),ca=a("p"),ca.innerHTML=cf,Vs=t(),fa=a("p"),fa.innerHTML=ff,Is=t(),pa=a("p"),pa.innerHTML=pf,Fs=t(),ma=a("p"),ma.innerHTML=mf,Ws=t(),ha=a("p"),ha.innerHTML=hf,Es=t(),ce=a("div"),m(st.$$.fragment),Us=t(),_a=a("p"),_a.textContent=_f,Ns=t(),m(je.$$.fragment),Xs=t(),Be=a("div"),m(dt.$$.fragment),qs=t(),ua=a("p"),ua.textContent=uf,bi=t(),m(lt.$$.fragment),$i=t(),H=a("div"),m(ct.$$.fragment),zs=t(),ga=a("p"),ga.innerHTML=gf,js=t(),Ge=a("div"),m(ft.$$.fragment),Bs=t(),La=a("p"),La.innerHTML=Lf,Gs=t(),Oe=a("div"),m(pt.$$.fragment),Os=t(),xa=a("p"),xa.innerHTML=xf,Js=t(),G=a("div"),m(mt.$$.fragment),Ys=t(),wa=a("p"),wa.innerHTML=wf,Zs=t(),va=a("p"),va.innerHTML=vf,Ks=t(),ba=a("p"),ba.innerHTML=bf,Qs=t(),$a=a("p"),$a.innerHTML=$f,ed=t(),ya=a("p"),ya.innerHTML=yf,od=t(),fe=a("div"),m(ht.$$.fragment),td=t(),Ma=a("p"),Ma.textContent=Mf,rd=t(),m(Je.$$.fragment),ad=t(),Ye=a("div"),m(_t.$$.fragment),nd=t(),Ta=a("p"),Ta.textContent=Tf,yi=t(),m(ut.$$.fragment),Mi=t(),S=a("div"),m(gt.$$.fragment),id=t(),Da=a("p"),Da.innerHTML=Df,sd=t(),Aa=a("p"),Aa.innerHTML=Af,dd=t(),Ze=a("div"),m(Lt.$$.fragment),ld=t(),Ca=a("p"),Ca.innerHTML=Cf,cd=t(),Ke=a("div"),m(xt.$$.fragment),fd=t(),Sa=a("p"),Sa.innerHTML=Sf,pd=t(),Z=a("div"),m(wt.$$.fragment),md=t(),ka=a("p"),ka.innerHTML=kf,hd=t(),Ra=a("p"),Ra.innerHTML=Rf,_d=t(),Pa=a("p"),Pa.innerHTML=Pf,ud=t(),Ha=a("p"),Ha.innerHTML=Hf,gd=t(),pe=a("div"),m(vt.$$.fragment),Ld=t(),Va=a("p"),Va.textContent=Vf,xd=t(),m(Qe.$$.fragment),wd=t(),eo=a("div"),m(bt.$$.fragment),vd=t(),Ia=a("p"),Ia.textContent=If,bd=t(),me=a("div"),m($t.$$.fragment),$d=t(),Fa=a("p"),Fa.innerHTML=Ff,yd=t(),m(oo.$$.fragment),Ti=t(),m(yt.$$.fragment),Di=t(),A=a("div"),m(Mt.$$.fragment),Md=t(),Wa=a("p"),Wa.innerHTML=Wf,Td=t(),Ea=a("p"),Ea.innerHTML=Ef,Dd=t(),to=a("div"),m(Tt.$$.fragment),Ad=t(),Ua=a("p"),Ua.innerHTML=Uf,Cd=t(),ro=a("div"),m(Dt.$$.fragment),Sd=t(),Na=a("p"),Na.innerHTML=Nf,kd=t(),K=a("div"),m(At.$$.fragment),Rd=t(),Xa=a("p"),Xa.innerHTML=Xf,Pd=t(),qa=a("p"),qa.innerHTML=qf,Hd=t(),za=a("p"),za.innerHTML=zf,Vd=t(),ja=a("p"),ja.innerHTML=jf,Id=t(),he=a("div"),m(Ct.$$.fragment),Fd=t(),Ba=a("p"),Ba.textContent=Bf,Wd=t(),m(ao.$$.fragment),Ed=t(),no=a("div"),m(St.$$.fragment),Ud=t(),Ga=a("p"),Ga.textContent=Gf,Nd=t(),_e=a("div"),m(kt.$$.fragment),Xd=t(),Oa=a("p"),Oa.innerHTML=Of,qd=t(),m(io.$$.fragment),zd=t(),ue=a("div"),m(Rt.$$.fragment),jd=t(),Ja=a("p"),Ja.textContent=Jf,Bd=t(),m(so.$$.fragment),Ai=t(),m(Pt.$$.fragment),Ci=t(),V=a("div"),m(Ht.$$.fragment),Gd=t(),Ya=a("p"),Ya.innerHTML=Yf,Od=t(),lo=a("div"),m(Vt.$$.fragment),Jd=t(),Za=a("p"),Za.innerHTML=Zf,Yd=t(),co=a("div"),m(It.$$.fragment),Zd=t(),Ka=a("p"),Ka.innerHTML=Kf,Kd=t(),ge=a("div"),m(Ft.$$.fragment),Qd=t(),Qa=a("p"),Qa.textContent=Qf,el=t(),m(fo.$$.fragment),ol=t(),po=a("div"),m(Wt.$$.fragment),tl=t(),en=a("p"),en.textContent=ep,rl=t(),Le=a("div"),m(Et.$$.fragment),al=t(),on=a("p"),on.innerHTML=op,nl=t(),m(mo.$$.fragment),Si=t(),m(Ut.$$.fragment),ki=t(),I=a("div"),m(Nt.$$.fragment),il=t(),tn=a("p"),tn.innerHTML=tp,sl=t(),ho=a("div"),m(Xt.$$.fragment),dl=t(),rn=a("p"),rn.innerHTML=rp,ll=t(),_o=a("div"),m(qt.$$.fragment),cl=t(),an=a("p"),an.innerHTML=ap,fl=t(),xe=a("div"),m(zt.$$.fragment),pl=t(),nn=a("p"),nn.textContent=np,ml=t(),m(uo.$$.fragment),hl=t(),go=a("div"),m(jt.$$.fragment),_l=t(),sn=a("p"),sn.textContent=ip,ul=t(),we=a("div"),m(Bt.$$.fragment),gl=t(),dn=a("p"),dn.innerHTML=sp,Ll=t(),m(Lo.$$.fragment),Ri=t(),m(Gt.$$.fragment),Pi=t(),F=a("div"),m(Ot.$$.fragment),xl=t(),ln=a("p"),ln.innerHTML=dp,wl=t(),xo=a("div"),m(Jt.$$.fragment),vl=t(),cn=a("p"),cn.innerHTML=lp,bl=t(),wo=a("div"),m(Yt.$$.fragment),$l=t(),fn=a("p"),fn.innerHTML=cp,yl=t(),ve=a("div"),m(Zt.$$.fragment),Ml=t(),pn=a("p"),pn.textContent=fp,Tl=t(),m(vo.$$.fragment),Dl=t(),bo=a("div"),m(Kt.$$.fragment),Al=t(),mn=a("p"),mn.textContent=pp,Cl=t(),be=a("div"),m(Qt.$$.fragment),Sl=t(),hn=a("p"),hn.innerHTML=mp,kl=t(),m($o.$$.fragment),Hi=t(),m(er.$$.fragment),Vi=t(),W=a("div"),m(or.$$.fragment),Rl=t(),_n=a("p"),_n.innerHTML=hp,Pl=t(),yo=a("div"),m(tr.$$.fragment),Hl=t(),un=a("p"),un.innerHTML=_p,Vl=t(),Mo=a("div"),m(rr.$$.fragment),Il=t(),gn=a("p"),gn.innerHTML=up,Fl=t(),$e=a("div"),m(ar.$$.fragment),Wl=t(),Ln=a("p"),Ln.textContent=gp,El=t(),m(To.$$.fragment),Ul=t(),Do=a("div"),m(nr.$$.fragment),Nl=t(),xn=a("p"),xn.textContent=Lp,Xl=t(),ye=a("div"),m(ir.$$.fragment),ql=t(),wn=a("p"),wn.innerHTML=xp,zl=t(),m(Ao.$$.fragment),Ii=t(),m(sr.$$.fragment),Fi=t(),E=a("div"),m(dr.$$.fragment),jl=t(),vn=a("p"),vn.innerHTML=wp,Bl=t(),Co=a("div"),m(lr.$$.fragment),Gl=t(),bn=a("p"),bn.innerHTML=vp,Ol=t(),So=a("div"),m(cr.$$.fragment),Jl=t(),$n=a("p"),$n.innerHTML=bp,Yl=t(),Me=a("div"),m(fr.$$.fragment),Zl=t(),yn=a("p"),yn.textContent=$p,Kl=t(),m(ko.$$.fragment),Ql=t(),Ro=a("div"),m(pr.$$.fragment),ec=t(),Mn=a("p"),Mn.textContent=yp,oc=t(),Te=a("div"),m(mr.$$.fragment),tc=t(),Tn=a("p"),Tn.innerHTML=Mp,rc=t(),m(Po.$$.fragment),Wi=t(),m(hr.$$.fragment),Ei=t(),U=a("div"),m(_r.$$.fragment),ac=t(),Dn=a("p"),Dn.innerHTML=Tp,nc=t(),Ho=a("div"),m(ur.$$.fragment),ic=t(),An=a("p"),An.innerHTML=Dp,sc=t(),Vo=a("div"),m(gr.$$.fragment),dc=t(),Cn=a("p"),Cn.innerHTML=Ap,lc=t(),De=a("div"),m(Lr.$$.fragment),cc=t(),Sn=a("p"),Sn.textContent=Cp,fc=t(),m(Io.$$.fragment),pc=t(),Fo=a("div"),m(xr.$$.fragment),mc=t(),kn=a("p"),kn.textContent=Sp,hc=t(),Ae=a("div"),m(wr.$$.fragment),_c=t(),Rn=a("p"),Rn.innerHTML=kp,uc=t(),m(Wo.$$.fragment),Ui=t(),m(vr.$$.fragment),Ni=t(),N=a("div"),m(br.$$.fragment),gc=t(),Pn=a("p"),Pn.innerHTML=Rp,Lc=t(),Eo=a("div"),m($r.$$.fragment),xc=t(),Hn=a("p"),Hn.innerHTML=Pp,wc=t(),Uo=a("div"),m(yr.$$.fragment),vc=t(),Vn=a("p"),Vn.innerHTML=Hp,bc=t(),Ce=a("div"),m(Mr.$$.fragment),$c=t(),In=a("p"),In.textContent=Vp,yc=t(),m(No.$$.fragment),Mc=t(),Xo=a("div"),m(Tr.$$.fragment),Tc=t(),Fn=a("p"),Fn.textContent=Ip,Dc=t(),Se=a("div"),m(Dr.$$.fragment),Ac=t(),Wn=a("p"),Wn.innerHTML=Fp,Cc=t(),m(qo.$$.fragment),Xi=t(),m(Ar.$$.fragment),qi=t(),X=a("div"),m(Cr.$$.fragment),Sc=t(),En=a("p"),En.innerHTML=Wp,kc=t(),zo=a("div"),m(Sr.$$.fragment),Rc=t(),Un=a("p"),Un.innerHTML=Ep,Pc=t(),jo=a("div"),m(kr.$$.fragment),Hc=t(),Nn=a("p"),Nn.innerHTML=Up,Vc=t(),ke=a("div"),m(Rr.$$.fragment),Ic=t(),Xn=a("p"),Xn.textContent=Np,Fc=t(),m(Bo.$$.fragment),Wc=t(),Go=a("div"),m(Pr.$$.fragment),Ec=t(),qn=a("p"),qn.textContent=Xp,Uc=t(),Re=a("div"),m(Hr.$$.fragment),Nc=t(),zn=a("p"),zn.innerHTML=qp,Xc=t(),m(Oo.$$.fragment),zi=t(),m(Vr.$$.fragment),ji=t(),q=a("div"),m(Ir.$$.fragment),qc=t(),jn=a("p"),jn.innerHTML=zp,zc=t(),Jo=a("div"),m(Fr.$$.fragment),jc=t(),Bn=a("p"),Bn.innerHTML=jp,Bc=t(),Yo=a("div"),m(Wr.$$.fragment),Gc=t(),Gn=a("p"),Gn.innerHTML=Bp,Oc=t(),Pe=a("div"),m(Er.$$.fragment),Jc=t(),On=a("p"),On.textContent=Gp,Yc=t(),m(Zo.$$.fragment),Zc=t(),Ko=a("div"),m(Ur.$$.fragment),Kc=t(),Jn=a("p"),Jn.textContent=Op,Qc=t(),He=a("div"),m(Nr.$$.fragment),ef=t(),Yn=a("p"),Yn.innerHTML=Jp,of=t(),m(Qo.$$.fragment),Bi=t(),m(Xr.$$.fragment),Gi=t(),Ve=a("div"),m(qr.$$.fragment),tf=t(),et=a("div"),m(zr.$$.fragment),rf=t(),Zn=a("p"),Zn.innerHTML=Yp,Oi=t(),m(jr.$$.fragment),Ji=t(),Br=a("div"),m(Gr.$$.fragment),Yi=t(),m(Or.$$.fragment),Zi=t(),gi=a("p"),this.h()},l(e){const p=tm("svelte-u9bgzb",document.head);i=n(p,"META",{name:!0,content:!0}),p.forEach(s),$=r(e),l=n(e,"P",{}),w(l).forEach(s),c=r(e),h(T.$$.fragment,e),d=r(e),y=n(e,"P",{"data-svelte-h":!0}),f(y)!=="svelte-zw3qap"&&(y.innerHTML=af),Li=r(e),ot=n(e,"UL",{"data-svelte-h":!0}),f(ot)!=="svelte-1bkyah4"&&(ot.innerHTML=nf),xi=r(e),h(Xe.$$.fragment,e),wi=r(e),h(tt.$$.fragment,e),vi=r(e),P=n(e,"DIV",{class:!0});var z=w(P);h(rt.$$.fragment,z),As=r(z),sa=n(z,"P",{"data-svelte-h":!0}),f(sa)!=="svelte-1btiu64"&&(sa.innerHTML=sf),Cs=r(z),qe=n(z,"DIV",{class:!0});var Jr=w(qe);h(at.$$.fragment,Jr),Ss=r(Jr),da=n(Jr,"P",{"data-svelte-h":!0}),f(da)!=="svelte-1062ci4"&&(da.innerHTML=df),Jr.forEach(s),ks=r(z),ze=n(z,"DIV",{class:!0});var Yr=w(ze);h(nt.$$.fragment,Yr),Rs=r(Yr),la=n(Yr,"P",{"data-svelte-h":!0}),f(la)!=="svelte-u3q4so"&&(la.innerHTML=lf),Yr.forEach(s),Ps=r(z),B=n(z,"DIV",{class:!0});var J=w(B);h(it.$$.fragment,J),Hs=r(J),ca=n(J,"P",{"data-svelte-h":!0}),f(ca)!=="svelte-vs7s0z"&&(ca.innerHTML=cf),Vs=r(J),fa=n(J,"P",{"data-svelte-h":!0}),f(fa)!=="svelte-15b960v"&&(fa.innerHTML=ff),Is=r(J),pa=n(J,"P",{"data-svelte-h":!0}),f(pa)!=="svelte-1k09o7t"&&(pa.innerHTML=pf),Fs=r(J),ma=n(J,"P",{"data-svelte-h":!0}),f(ma)!=="svelte-1791xru"&&(ma.innerHTML=mf),Ws=r(J),ha=n(J,"P",{"data-svelte-h":!0}),f(ha)!=="svelte-1u1p9ls"&&(ha.innerHTML=hf),J.forEach(s),Es=r(z),ce=n(z,"DIV",{class:!0});var Ie=w(ce);h(st.$$.fragment,Ie),Us=r(Ie),_a=n(Ie,"P",{"data-svelte-h":!0}),f(_a)!=="svelte-flusvq"&&(_a.textContent=_f),Ns=r(Ie),h(je.$$.fragment,Ie),Ie.forEach(s),Xs=r(z),Be=n(z,"DIV",{class:!0});var Zr=w(Be);h(dt.$$.fragment,Zr),qs=r(Zr),ua=n(Zr,"P",{"data-svelte-h":!0}),f(ua)!=="svelte-1ufq5ot"&&(ua.textContent=uf),Zr.forEach(s),z.forEach(s),bi=r(e),h(lt.$$.fragment,e),$i=r(e),H=n(e,"DIV",{class:!0});var j=w(H);h(ct.$$.fragment,j),zs=r(j),ga=n(j,"P",{"data-svelte-h":!0}),f(ga)!=="svelte-ruj0o9"&&(ga.innerHTML=gf),js=r(j),Ge=n(j,"DIV",{class:!0});var Kr=w(Ge);h(ft.$$.fragment,Kr),Bs=r(Kr),La=n(Kr,"P",{"data-svelte-h":!0}),f(La)!=="svelte-1062ci4"&&(La.innerHTML=Lf),Kr.forEach(s),Gs=r(j),Oe=n(j,"DIV",{class:!0});var Qr=w(Oe);h(pt.$$.fragment,Qr),Os=r(Qr),xa=n(Qr,"P",{"data-svelte-h":!0}),f(xa)!=="svelte-u3q4so"&&(xa.innerHTML=xf),Qr.forEach(s),Js=r(j),G=n(j,"DIV",{class:!0});var Y=w(G);h(mt.$$.fragment,Y),Ys=r(Y),wa=n(Y,"P",{"data-svelte-h":!0}),f(wa)!=="svelte-vs7s0z"&&(wa.innerHTML=wf),Zs=r(Y),va=n(Y,"P",{"data-svelte-h":!0}),f(va)!=="svelte-15b960v"&&(va.innerHTML=vf),Ks=r(Y),ba=n(Y,"P",{"data-svelte-h":!0}),f(ba)!=="svelte-1k09o7t"&&(ba.innerHTML=bf),Qs=r(Y),$a=n(Y,"P",{"data-svelte-h":!0}),f($a)!=="svelte-1791xru"&&($a.innerHTML=$f),ed=r(Y),ya=n(Y,"P",{"data-svelte-h":!0}),f(ya)!=="svelte-1u1p9ls"&&(ya.innerHTML=yf),Y.forEach(s),od=r(j),fe=n(j,"DIV",{class:!0});var Fe=w(fe);h(ht.$$.fragment,Fe),td=r(Fe),Ma=n(Fe,"P",{"data-svelte-h":!0}),f(Ma)!=="svelte-flusvq"&&(Ma.textContent=Mf),rd=r(Fe),h(Je.$$.fragment,Fe),Fe.forEach(s),ad=r(j),Ye=n(j,"DIV",{class:!0});var ea=w(Ye);h(_t.$$.fragment,ea),nd=r(ea),Ta=n(ea,"P",{"data-svelte-h":!0}),f(Ta)!=="svelte-1ufq5ot"&&(Ta.textContent=Tf),ea.forEach(s),j.forEach(s),yi=r(e),h(ut.$$.fragment,e),Mi=r(e),S=n(e,"DIV",{class:!0});var R=w(S);h(gt.$$.fragment,R),id=r(R),Da=n(R,"P",{"data-svelte-h":!0}),f(Da)!=="svelte-1jjo27s"&&(Da.innerHTML=Df),sd=r(R),Aa=n(R,"P",{"data-svelte-h":!0}),f(Aa)!=="svelte-10llbq"&&(Aa.innerHTML=Af),dd=r(R),Ze=n(R,"DIV",{class:!0});var oa=w(Ze);h(Lt.$$.fragment,oa),ld=r(oa),Ca=n(oa,"P",{"data-svelte-h":!0}),f(Ca)!=="svelte-1062ci4"&&(Ca.innerHTML=Cf),oa.forEach(s),cd=r(R),Ke=n(R,"DIV",{class:!0});var ta=w(Ke);h(xt.$$.fragment,ta),fd=r(ta),Sa=n(ta,"P",{"data-svelte-h":!0}),f(Sa)!=="svelte-1lgbsz7"&&(Sa.innerHTML=Sf),ta.forEach(s),pd=r(R),Z=n(R,"DIV",{class:!0});var de=w(Z);h(wt.$$.fragment,de),md=r(de),ka=n(de,"P",{"data-svelte-h":!0}),f(ka)!=="svelte-vs7s0z"&&(ka.innerHTML=kf),hd=r(de),Ra=n(de,"P",{"data-svelte-h":!0}),f(Ra)!=="svelte-15b960v"&&(Ra.innerHTML=Rf),_d=r(de),Pa=n(de,"P",{"data-svelte-h":!0}),f(Pa)!=="svelte-1k09o7t"&&(Pa.innerHTML=Pf),ud=r(de),Ha=n(de,"P",{"data-svelte-h":!0}),f(Ha)!=="svelte-1ukghd4"&&(Ha.innerHTML=Hf),de.forEach(s),gd=r(R),pe=n(R,"DIV",{class:!0});var We=w(pe);h(vt.$$.fragment,We),Ld=r(We),Va=n(We,"P",{"data-svelte-h":!0}),f(Va)!=="svelte-flusvq"&&(Va.textContent=Vf),xd=r(We),h(Qe.$$.fragment,We),We.forEach(s),wd=r(R),eo=n(R,"DIV",{class:!0});var ra=w(eo);h(bt.$$.fragment,ra),vd=r(ra),Ia=n(ra,"P",{"data-svelte-h":!0}),f(Ia)!=="svelte-1ufq5ot"&&(Ia.textContent=If),ra.forEach(s),bd=r(R),me=n(R,"DIV",{class:!0});var Ee=w(me);h($t.$$.fragment,Ee),$d=r(Ee),Fa=n(Ee,"P",{"data-svelte-h":!0}),f(Fa)!=="svelte-ioswce"&&(Fa.innerHTML=Ff),yd=r(Ee),h(oo.$$.fragment,Ee),Ee.forEach(s),R.forEach(s),Ti=r(e),h(yt.$$.fragment,e),Di=r(e),A=n(e,"DIV",{class:!0});var k=w(A);h(Mt.$$.fragment,k),Md=r(k),Wa=n(k,"P",{"data-svelte-h":!0}),f(Wa)!=="svelte-ylropp"&&(Wa.innerHTML=Wf),Td=r(k),Ea=n(k,"P",{"data-svelte-h":!0}),f(Ea)!=="svelte-10llbq"&&(Ea.innerHTML=Ef),Dd=r(k),to=n(k,"DIV",{class:!0});var aa=w(to);h(Tt.$$.fragment,aa),Ad=r(aa),Ua=n(aa,"P",{"data-svelte-h":!0}),f(Ua)!=="svelte-1062ci4"&&(Ua.innerHTML=Uf),aa.forEach(s),Cd=r(k),ro=n(k,"DIV",{class:!0});var na=w(ro);h(Dt.$$.fragment,na),Sd=r(na),Na=n(na,"P",{"data-svelte-h":!0}),f(Na)!=="svelte-1lgbsz7"&&(Na.innerHTML=Nf),na.forEach(s),kd=r(k),K=n(k,"DIV",{class:!0});var le=w(K);h(At.$$.fragment,le),Rd=r(le),Xa=n(le,"P",{"data-svelte-h":!0}),f(Xa)!=="svelte-178gcly"&&(Xa.innerHTML=Xf),Pd=r(le),qa=n(le,"P",{"data-svelte-h":!0}),f(qa)!=="svelte-15b960v"&&(qa.innerHTML=qf),Hd=r(le),za=n(le,"P",{"data-svelte-h":!0}),f(za)!=="svelte-1k09o7t"&&(za.innerHTML=zf),Vd=r(le),ja=n(le,"P",{"data-svelte-h":!0}),f(ja)!=="svelte-1ukghd4"&&(ja.innerHTML=jf),le.forEach(s),Id=r(k),he=n(k,"DIV",{class:!0});var Ue=w(he);h(Ct.$$.fragment,Ue),Fd=r(Ue),Ba=n(Ue,"P",{"data-svelte-h":!0}),f(Ba)!=="svelte-flusvq"&&(Ba.textContent=Bf),Wd=r(Ue),h(ao.$$.fragment,Ue),Ue.forEach(s),Ed=r(k),no=n(k,"DIV",{class:!0});var ia=w(no);h(St.$$.fragment,ia),Ud=r(ia),Ga=n(ia,"P",{"data-svelte-h":!0}),f(Ga)!=="svelte-1ufq5ot"&&(Ga.textContent=Gf),ia.forEach(s),Nd=r(k),_e=n(k,"DIV",{class:!0});var Ne=w(_e);h(kt.$$.fragment,Ne),Xd=r(Ne),Oa=n(Ne,"P",{"data-svelte-h":!0}),f(Oa)!=="svelte-ioswce"&&(Oa.innerHTML=Of),qd=r(Ne),h(io.$$.fragment,Ne),Ne.forEach(s),zd=r(k),ue=n(k,"DIV",{class:!0});var Kn=w(ue);h(Rt.$$.fragment,Kn),jd=r(Kn),Ja=n(Kn,"P",{"data-svelte-h":!0}),f(Ja)!=="svelte-119cgd9"&&(Ja.textContent=Jf),Bd=r(Kn),h(so.$$.fragment,Kn),Kn.forEach(s),k.forEach(s),Ai=r(e),h(Pt.$$.fragment,e),Ci=r(e),V=n(e,"DIV",{class:!0});var Q=w(V);h(Ht.$$.fragment,Q),Gd=r(Q),Ya=n(Q,"P",{"data-svelte-h":!0}),f(Ya)!=="svelte-1b3j6g9"&&(Ya.innerHTML=Yf),Od=r(Q),lo=n(Q,"DIV",{class:!0});var Qi=w(lo);h(Vt.$$.fragment,Qi),Jd=r(Qi),Za=n(Qi,"P",{"data-svelte-h":!0}),f(Za)!=="svelte-1lgbsz7"&&(Za.innerHTML=Zf),Qi.forEach(s),Yd=r(Q),co=n(Q,"DIV",{class:!0});var es=w(co);h(It.$$.fragment,es),Zd=r(es),Ka=n(es,"P",{"data-svelte-h":!0}),f(Ka)!=="svelte-1ug3yl0"&&(Ka.innerHTML=Kf),es.forEach(s),Kd=r(Q),ge=n(Q,"DIV",{class:!0});var Qn=w(ge);h(Ft.$$.fragment,Qn),Qd=r(Qn),Qa=n(Qn,"P",{"data-svelte-h":!0}),f(Qa)!=="svelte-flusvq"&&(Qa.textContent=Qf),el=r(Qn),h(fo.$$.fragment,Qn),Qn.forEach(s),ol=r(Q),po=n(Q,"DIV",{class:!0});var os=w(po);h(Wt.$$.fragment,os),tl=r(os),en=n(os,"P",{"data-svelte-h":!0}),f(en)!=="svelte-1ufq5ot"&&(en.textContent=ep),os.forEach(s),rl=r(Q),Le=n(Q,"DIV",{class:!0});var ei=w(Le);h(Et.$$.fragment,ei),al=r(ei),on=n(ei,"P",{"data-svelte-h":!0}),f(on)!=="svelte-ioswce"&&(on.innerHTML=op),nl=r(ei),h(mo.$$.fragment,ei),ei.forEach(s),Q.forEach(s),Si=r(e),h(Ut.$$.fragment,e),ki=r(e),I=n(e,"DIV",{class:!0});var ee=w(I);h(Nt.$$.fragment,ee),il=r(ee),tn=n(ee,"P",{"data-svelte-h":!0}),f(tn)!=="svelte-1o1jg8l"&&(tn.innerHTML=tp),sl=r(ee),ho=n(ee,"DIV",{class:!0});var ts=w(ho);h(Xt.$$.fragment,ts),dl=r(ts),rn=n(ts,"P",{"data-svelte-h":!0}),f(rn)!=="svelte-1lgbsz7"&&(rn.innerHTML=rp),ts.forEach(s),ll=r(ee),_o=n(ee,"DIV",{class:!0});var rs=w(_o);h(qt.$$.fragment,rs),cl=r(rs),an=n(rs,"P",{"data-svelte-h":!0}),f(an)!=="svelte-1ug3yl0"&&(an.innerHTML=ap),rs.forEach(s),fl=r(ee),xe=n(ee,"DIV",{class:!0});var oi=w(xe);h(zt.$$.fragment,oi),pl=r(oi),nn=n(oi,"P",{"data-svelte-h":!0}),f(nn)!=="svelte-flusvq"&&(nn.textContent=np),ml=r(oi),h(uo.$$.fragment,oi),oi.forEach(s),hl=r(ee),go=n(ee,"DIV",{class:!0});var as=w(go);h(jt.$$.fragment,as),_l=r(as),sn=n(as,"P",{"data-svelte-h":!0}),f(sn)!=="svelte-1ufq5ot"&&(sn.textContent=ip),as.forEach(s),ul=r(ee),we=n(ee,"DIV",{class:!0});var ti=w(we);h(Bt.$$.fragment,ti),gl=r(ti),dn=n(ti,"P",{"data-svelte-h":!0}),f(dn)!=="svelte-ioswce"&&(dn.innerHTML=sp),Ll=r(ti),h(Lo.$$.fragment,ti),ti.forEach(s),ee.forEach(s),Ri=r(e),h(Gt.$$.fragment,e),Pi=r(e),F=n(e,"DIV",{class:!0});var oe=w(F);h(Ot.$$.fragment,oe),xl=r(oe),ln=n(oe,"P",{"data-svelte-h":!0}),f(ln)!=="svelte-1l8ah08"&&(ln.innerHTML=dp),wl=r(oe),xo=n(oe,"DIV",{class:!0});var ns=w(xo);h(Jt.$$.fragment,ns),vl=r(ns),cn=n(ns,"P",{"data-svelte-h":!0}),f(cn)!=="svelte-1lgbsz7"&&(cn.innerHTML=lp),ns.forEach(s),bl=r(oe),wo=n(oe,"DIV",{class:!0});var is=w(wo);h(Yt.$$.fragment,is),$l=r(is),fn=n(is,"P",{"data-svelte-h":!0}),f(fn)!=="svelte-1ug3yl0"&&(fn.innerHTML=cp),is.forEach(s),yl=r(oe),ve=n(oe,"DIV",{class:!0});var ri=w(ve);h(Zt.$$.fragment,ri),Ml=r(ri),pn=n(ri,"P",{"data-svelte-h":!0}),f(pn)!=="svelte-flusvq"&&(pn.textContent=fp),Tl=r(ri),h(vo.$$.fragment,ri),ri.forEach(s),Dl=r(oe),bo=n(oe,"DIV",{class:!0});var ss=w(bo);h(Kt.$$.fragment,ss),Al=r(ss),mn=n(ss,"P",{"data-svelte-h":!0}),f(mn)!=="svelte-1ufq5ot"&&(mn.textContent=pp),ss.forEach(s),Cl=r(oe),be=n(oe,"DIV",{class:!0});var ai=w(be);h(Qt.$$.fragment,ai),Sl=r(ai),hn=n(ai,"P",{"data-svelte-h":!0}),f(hn)!=="svelte-ioswce"&&(hn.innerHTML=mp),kl=r(ai),h($o.$$.fragment,ai),ai.forEach(s),oe.forEach(s),Hi=r(e),h(er.$$.fragment,e),Vi=r(e),W=n(e,"DIV",{class:!0});var te=w(W);h(or.$$.fragment,te),Rl=r(te),_n=n(te,"P",{"data-svelte-h":!0}),f(_n)!=="svelte-1l515ep"&&(_n.innerHTML=hp),Pl=r(te),yo=n(te,"DIV",{class:!0});var ds=w(yo);h(tr.$$.fragment,ds),Hl=r(ds),un=n(ds,"P",{"data-svelte-h":!0}),f(un)!=="svelte-1lgbsz7"&&(un.innerHTML=_p),ds.forEach(s),Vl=r(te),Mo=n(te,"DIV",{class:!0});var ls=w(Mo);h(rr.$$.fragment,ls),Il=r(ls),gn=n(ls,"P",{"data-svelte-h":!0}),f(gn)!=="svelte-1ug3yl0"&&(gn.innerHTML=up),ls.forEach(s),Fl=r(te),$e=n(te,"DIV",{class:!0});var ni=w($e);h(ar.$$.fragment,ni),Wl=r(ni),Ln=n(ni,"P",{"data-svelte-h":!0}),f(Ln)!=="svelte-flusvq"&&(Ln.textContent=gp),El=r(ni),h(To.$$.fragment,ni),ni.forEach(s),Ul=r(te),Do=n(te,"DIV",{class:!0});var cs=w(Do);h(nr.$$.fragment,cs),Nl=r(cs),xn=n(cs,"P",{"data-svelte-h":!0}),f(xn)!=="svelte-1ufq5ot"&&(xn.textContent=Lp),cs.forEach(s),Xl=r(te),ye=n(te,"DIV",{class:!0});var ii=w(ye);h(ir.$$.fragment,ii),ql=r(ii),wn=n(ii,"P",{"data-svelte-h":!0}),f(wn)!=="svelte-ioswce"&&(wn.innerHTML=xp),zl=r(ii),h(Ao.$$.fragment,ii),ii.forEach(s),te.forEach(s),Ii=r(e),h(sr.$$.fragment,e),Fi=r(e),E=n(e,"DIV",{class:!0});var re=w(E);h(dr.$$.fragment,re),jl=r(re),vn=n(re,"P",{"data-svelte-h":!0}),f(vn)!=="svelte-tre9co"&&(vn.innerHTML=wp),Bl=r(re),Co=n(re,"DIV",{class:!0});var fs=w(Co);h(lr.$$.fragment,fs),Gl=r(fs),bn=n(fs,"P",{"data-svelte-h":!0}),f(bn)!=="svelte-1lgbsz7"&&(bn.innerHTML=vp),fs.forEach(s),Ol=r(re),So=n(re,"DIV",{class:!0});var ps=w(So);h(cr.$$.fragment,ps),Jl=r(ps),$n=n(ps,"P",{"data-svelte-h":!0}),f($n)!=="svelte-1ug3yl0"&&($n.innerHTML=bp),ps.forEach(s),Yl=r(re),Me=n(re,"DIV",{class:!0});var si=w(Me);h(fr.$$.fragment,si),Zl=r(si),yn=n(si,"P",{"data-svelte-h":!0}),f(yn)!=="svelte-flusvq"&&(yn.textContent=$p),Kl=r(si),h(ko.$$.fragment,si),si.forEach(s),Ql=r(re),Ro=n(re,"DIV",{class:!0});var ms=w(Ro);h(pr.$$.fragment,ms),ec=r(ms),Mn=n(ms,"P",{"data-svelte-h":!0}),f(Mn)!=="svelte-1ufq5ot"&&(Mn.textContent=yp),ms.forEach(s),oc=r(re),Te=n(re,"DIV",{class:!0});var di=w(Te);h(mr.$$.fragment,di),tc=r(di),Tn=n(di,"P",{"data-svelte-h":!0}),f(Tn)!=="svelte-ioswce"&&(Tn.innerHTML=Mp),rc=r(di),h(Po.$$.fragment,di),di.forEach(s),re.forEach(s),Wi=r(e),h(hr.$$.fragment,e),Ei=r(e),U=n(e,"DIV",{class:!0});var ae=w(U);h(_r.$$.fragment,ae),ac=r(ae),Dn=n(ae,"P",{"data-svelte-h":!0}),f(Dn)!=="svelte-xw6y5o"&&(Dn.innerHTML=Tp),nc=r(ae),Ho=n(ae,"DIV",{class:!0});var hs=w(Ho);h(ur.$$.fragment,hs),ic=r(hs),An=n(hs,"P",{"data-svelte-h":!0}),f(An)!=="svelte-1lgbsz7"&&(An.innerHTML=Dp),hs.forEach(s),sc=r(ae),Vo=n(ae,"DIV",{class:!0});var _s=w(Vo);h(gr.$$.fragment,_s),dc=r(_s),Cn=n(_s,"P",{"data-svelte-h":!0}),f(Cn)!=="svelte-1ug3yl0"&&(Cn.innerHTML=Ap),_s.forEach(s),lc=r(ae),De=n(ae,"DIV",{class:!0});var li=w(De);h(Lr.$$.fragment,li),cc=r(li),Sn=n(li,"P",{"data-svelte-h":!0}),f(Sn)!=="svelte-flusvq"&&(Sn.textContent=Cp),fc=r(li),h(Io.$$.fragment,li),li.forEach(s),pc=r(ae),Fo=n(ae,"DIV",{class:!0});var us=w(Fo);h(xr.$$.fragment,us),mc=r(us),kn=n(us,"P",{"data-svelte-h":!0}),f(kn)!=="svelte-1ufq5ot"&&(kn.textContent=Sp),us.forEach(s),hc=r(ae),Ae=n(ae,"DIV",{class:!0});var ci=w(Ae);h(wr.$$.fragment,ci),_c=r(ci),Rn=n(ci,"P",{"data-svelte-h":!0}),f(Rn)!=="svelte-ioswce"&&(Rn.innerHTML=kp),uc=r(ci),h(Wo.$$.fragment,ci),ci.forEach(s),ae.forEach(s),Ui=r(e),h(vr.$$.fragment,e),Ni=r(e),N=n(e,"DIV",{class:!0});var ne=w(N);h(br.$$.fragment,ne),gc=r(ne),Pn=n(ne,"P",{"data-svelte-h":!0}),f(Pn)!=="svelte-13h7gvv"&&(Pn.innerHTML=Rp),Lc=r(ne),Eo=n(ne,"DIV",{class:!0});var gs=w(Eo);h($r.$$.fragment,gs),xc=r(gs),Hn=n(gs,"P",{"data-svelte-h":!0}),f(Hn)!=="svelte-1lgbsz7"&&(Hn.innerHTML=Pp),gs.forEach(s),wc=r(ne),Uo=n(ne,"DIV",{class:!0});var Ls=w(Uo);h(yr.$$.fragment,Ls),vc=r(Ls),Vn=n(Ls,"P",{"data-svelte-h":!0}),f(Vn)!=="svelte-1ug3yl0"&&(Vn.innerHTML=Hp),Ls.forEach(s),bc=r(ne),Ce=n(ne,"DIV",{class:!0});var fi=w(Ce);h(Mr.$$.fragment,fi),$c=r(fi),In=n(fi,"P",{"data-svelte-h":!0}),f(In)!=="svelte-flusvq"&&(In.textContent=Vp),yc=r(fi),h(No.$$.fragment,fi),fi.forEach(s),Mc=r(ne),Xo=n(ne,"DIV",{class:!0});var xs=w(Xo);h(Tr.$$.fragment,xs),Tc=r(xs),Fn=n(xs,"P",{"data-svelte-h":!0}),f(Fn)!=="svelte-1ufq5ot"&&(Fn.textContent=Ip),xs.forEach(s),Dc=r(ne),Se=n(ne,"DIV",{class:!0});var pi=w(Se);h(Dr.$$.fragment,pi),Ac=r(pi),Wn=n(pi,"P",{"data-svelte-h":!0}),f(Wn)!=="svelte-ioswce"&&(Wn.innerHTML=Fp),Cc=r(pi),h(qo.$$.fragment,pi),pi.forEach(s),ne.forEach(s),Xi=r(e),h(Ar.$$.fragment,e),qi=r(e),X=n(e,"DIV",{class:!0});var ie=w(X);h(Cr.$$.fragment,ie),Sc=r(ie),En=n(ie,"P",{"data-svelte-h":!0}),f(En)!=="svelte-kthno9"&&(En.innerHTML=Wp),kc=r(ie),zo=n(ie,"DIV",{class:!0});var ws=w(zo);h(Sr.$$.fragment,ws),Rc=r(ws),Un=n(ws,"P",{"data-svelte-h":!0}),f(Un)!=="svelte-1lgbsz7"&&(Un.innerHTML=Ep),ws.forEach(s),Pc=r(ie),jo=n(ie,"DIV",{class:!0});var vs=w(jo);h(kr.$$.fragment,vs),Hc=r(vs),Nn=n(vs,"P",{"data-svelte-h":!0}),f(Nn)!=="svelte-1ug3yl0"&&(Nn.innerHTML=Up),vs.forEach(s),Vc=r(ie),ke=n(ie,"DIV",{class:!0});var mi=w(ke);h(Rr.$$.fragment,mi),Ic=r(mi),Xn=n(mi,"P",{"data-svelte-h":!0}),f(Xn)!=="svelte-flusvq"&&(Xn.textContent=Np),Fc=r(mi),h(Bo.$$.fragment,mi),mi.forEach(s),Wc=r(ie),Go=n(ie,"DIV",{class:!0});var bs=w(Go);h(Pr.$$.fragment,bs),Ec=r(bs),qn=n(bs,"P",{"data-svelte-h":!0}),f(qn)!=="svelte-1ufq5ot"&&(qn.textContent=Xp),bs.forEach(s),Uc=r(ie),Re=n(ie,"DIV",{class:!0});var hi=w(Re);h(Hr.$$.fragment,hi),Nc=r(hi),zn=n(hi,"P",{"data-svelte-h":!0}),f(zn)!=="svelte-ioswce"&&(zn.innerHTML=qp),Xc=r(hi),h(Oo.$$.fragment,hi),hi.forEach(s),ie.forEach(s),zi=r(e),h(Vr.$$.fragment,e),ji=r(e),q=n(e,"DIV",{class:!0});var se=w(q);h(Ir.$$.fragment,se),qc=r(se),jn=n(se,"P",{"data-svelte-h":!0}),f(jn)!=="svelte-1dry0uy"&&(jn.innerHTML=zp),zc=r(se),Jo=n(se,"DIV",{class:!0});var $s=w(Jo);h(Fr.$$.fragment,$s),jc=r($s),Bn=n($s,"P",{"data-svelte-h":!0}),f(Bn)!=="svelte-1lgbsz7"&&(Bn.innerHTML=jp),$s.forEach(s),Bc=r(se),Yo=n(se,"DIV",{class:!0});var ys=w(Yo);h(Wr.$$.fragment,ys),Gc=r(ys),Gn=n(ys,"P",{"data-svelte-h":!0}),f(Gn)!=="svelte-1ug3yl0"&&(Gn.innerHTML=Bp),ys.forEach(s),Oc=r(se),Pe=n(se,"DIV",{class:!0});var _i=w(Pe);h(Er.$$.fragment,_i),Jc=r(_i),On=n(_i,"P",{"data-svelte-h":!0}),f(On)!=="svelte-flusvq"&&(On.textContent=Gp),Yc=r(_i),h(Zo.$$.fragment,_i),_i.forEach(s),Zc=r(se),Ko=n(se,"DIV",{class:!0});var Ms=w(Ko);h(Ur.$$.fragment,Ms),Kc=r(Ms),Jn=n(Ms,"P",{"data-svelte-h":!0}),f(Jn)!=="svelte-1ufq5ot"&&(Jn.textContent=Op),Ms.forEach(s),Qc=r(se),He=n(se,"DIV",{class:!0});var ui=w(He);h(Nr.$$.fragment,ui),ef=r(ui),Yn=n(ui,"P",{"data-svelte-h":!0}),f(Yn)!=="svelte-ioswce"&&(Yn.innerHTML=Jp),of=r(ui),h(Qo.$$.fragment,ui),ui.forEach(s),se.forEach(s),Bi=r(e),h(Xr.$$.fragment,e),Gi=r(e),Ve=n(e,"DIV",{class:!0});var Ts=w(Ve);h(qr.$$.fragment,Ts),tf=r(Ts),et=n(Ts,"DIV",{class:!0});var Ds=w(et);h(zr.$$.fragment,Ds),rf=r(Ds),Zn=n(Ds,"P",{"data-svelte-h":!0}),f(Zn)!=="svelte-1lgbsz7"&&(Zn.innerHTML=Yp),Ds.forEach(s),Ts.forEach(s),Oi=r(e),h(jr.$$.fragment,e),Ji=r(e),Br=n(e,"DIV",{class:!0});var Zp=w(Br);h(Gr.$$.fragment,Zp),Zp.forEach(s),Yi=r(e),h(Or.$$.fragment,e),Zi=r(e),gi=n(e,"P",{}),w(gi).forEach(s),this.h()},h(){v(i,"name","hf:doc:metadata"),v(i,"content",Pm),v(qe,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(ze,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(B,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(ce,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Be,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(P,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ge,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Oe,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(G,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(fe,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ye,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(H,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ze,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ke,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Z,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(pe,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(eo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(me,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(S,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(to,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(ro,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(K,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(he,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(no,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(_e,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(ue,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(A,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(lo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(co,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(ge,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(po,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Le,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(V,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(ho,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(_o,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(xe,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(go,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(we,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(I,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(xo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(wo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(ve,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(bo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(be,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(F,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(yo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Mo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v($e,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Do,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(ye,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(W,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Co,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(So,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Me,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ro,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Te,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(E,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ho,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Vo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(De,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Fo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ae,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(U,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Eo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Uo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ce,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Xo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Se,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(N,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(zo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(jo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(ke,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Go,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Re,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(X,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Jo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Yo,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Pe,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ko,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(He,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(q,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(et,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Ve,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),v(Br,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8")},m(e,p){o(document.head,i),x(e,$,p),x(e,l,p),x(e,c,p),_(T,e,p),x(e,d,p),x(e,y,p),x(e,Li,p),x(e,ot,p),x(e,xi,p),_(Xe,e,p),x(e,wi,p),_(tt,e,p),x(e,vi,p),x(e,P,p),_(rt,P,null),o(P,As),o(P,sa),o(P,Cs),o(P,qe),_(at,qe,null),o(qe,Ss),o(qe,da),o(P,ks),o(P,ze),_(nt,ze,null),o(ze,Rs),o(ze,la),o(P,Ps),o(P,B),_(it,B,null),o(B,Hs),o(B,ca),o(B,Vs),o(B,fa),o(B,Is),o(B,pa),o(B,Fs),o(B,ma),o(B,Ws),o(B,ha),o(P,Es),o(P,ce),_(st,ce,null),o(ce,Us),o(ce,_a),o(ce,Ns),_(je,ce,null),o(P,Xs),o(P,Be),_(dt,Be,null),o(Be,qs),o(Be,ua),x(e,bi,p),_(lt,e,p),x(e,$i,p),x(e,H,p),_(ct,H,null),o(H,zs),o(H,ga),o(H,js),o(H,Ge),_(ft,Ge,null),o(Ge,Bs),o(Ge,La),o(H,Gs),o(H,Oe),_(pt,Oe,null),o(Oe,Os),o(Oe,xa),o(H,Js),o(H,G),_(mt,G,null),o(G,Ys),o(G,wa),o(G,Zs),o(G,va),o(G,Ks),o(G,ba),o(G,Qs),o(G,$a),o(G,ed),o(G,ya),o(H,od),o(H,fe),_(ht,fe,null),o(fe,td),o(fe,Ma),o(fe,rd),_(Je,fe,null),o(H,ad),o(H,Ye),_(_t,Ye,null),o(Ye,nd),o(Ye,Ta),x(e,yi,p),_(ut,e,p),x(e,Mi,p),x(e,S,p),_(gt,S,null),o(S,id),o(S,Da),o(S,sd),o(S,Aa),o(S,dd),o(S,Ze),_(Lt,Ze,null),o(Ze,ld),o(Ze,Ca),o(S,cd),o(S,Ke),_(xt,Ke,null),o(Ke,fd),o(Ke,Sa),o(S,pd),o(S,Z),_(wt,Z,null),o(Z,md),o(Z,ka),o(Z,hd),o(Z,Ra),o(Z,_d),o(Z,Pa),o(Z,ud),o(Z,Ha),o(S,gd),o(S,pe),_(vt,pe,null),o(pe,Ld),o(pe,Va),o(pe,xd),_(Qe,pe,null),o(S,wd),o(S,eo),_(bt,eo,null),o(eo,vd),o(eo,Ia),o(S,bd),o(S,me),_($t,me,null),o(me,$d),o(me,Fa),o(me,yd),_(oo,me,null),x(e,Ti,p),_(yt,e,p),x(e,Di,p),x(e,A,p),_(Mt,A,null),o(A,Md),o(A,Wa),o(A,Td),o(A,Ea),o(A,Dd),o(A,to),_(Tt,to,null),o(to,Ad),o(to,Ua),o(A,Cd),o(A,ro),_(Dt,ro,null),o(ro,Sd),o(ro,Na),o(A,kd),o(A,K),_(At,K,null),o(K,Rd),o(K,Xa),o(K,Pd),o(K,qa),o(K,Hd),o(K,za),o(K,Vd),o(K,ja),o(A,Id),o(A,he),_(Ct,he,null),o(he,Fd),o(he,Ba),o(he,Wd),_(ao,he,null),o(A,Ed),o(A,no),_(St,no,null),o(no,Ud),o(no,Ga),o(A,Nd),o(A,_e),_(kt,_e,null),o(_e,Xd),o(_e,Oa),o(_e,qd),_(io,_e,null),o(A,zd),o(A,ue),_(Rt,ue,null),o(ue,jd),o(ue,Ja),o(ue,Bd),_(so,ue,null),x(e,Ai,p),_(Pt,e,p),x(e,Ci,p),x(e,V,p),_(Ht,V,null),o(V,Gd),o(V,Ya),o(V,Od),o(V,lo),_(Vt,lo,null),o(lo,Jd),o(lo,Za),o(V,Yd),o(V,co),_(It,co,null),o(co,Zd),o(co,Ka),o(V,Kd),o(V,ge),_(Ft,ge,null),o(ge,Qd),o(ge,Qa),o(ge,el),_(fo,ge,null),o(V,ol),o(V,po),_(Wt,po,null),o(po,tl),o(po,en),o(V,rl),o(V,Le),_(Et,Le,null),o(Le,al),o(Le,on),o(Le,nl),_(mo,Le,null),x(e,Si,p),_(Ut,e,p),x(e,ki,p),x(e,I,p),_(Nt,I,null),o(I,il),o(I,tn),o(I,sl),o(I,ho),_(Xt,ho,null),o(ho,dl),o(ho,rn),o(I,ll),o(I,_o),_(qt,_o,null),o(_o,cl),o(_o,an),o(I,fl),o(I,xe),_(zt,xe,null),o(xe,pl),o(xe,nn),o(xe,ml),_(uo,xe,null),o(I,hl),o(I,go),_(jt,go,null),o(go,_l),o(go,sn),o(I,ul),o(I,we),_(Bt,we,null),o(we,gl),o(we,dn),o(we,Ll),_(Lo,we,null),x(e,Ri,p),_(Gt,e,p),x(e,Pi,p),x(e,F,p),_(Ot,F,null),o(F,xl),o(F,ln),o(F,wl),o(F,xo),_(Jt,xo,null),o(xo,vl),o(xo,cn),o(F,bl),o(F,wo),_(Yt,wo,null),o(wo,$l),o(wo,fn),o(F,yl),o(F,ve),_(Zt,ve,null),o(ve,Ml),o(ve,pn),o(ve,Tl),_(vo,ve,null),o(F,Dl),o(F,bo),_(Kt,bo,null),o(bo,Al),o(bo,mn),o(F,Cl),o(F,be),_(Qt,be,null),o(be,Sl),o(be,hn),o(be,kl),_($o,be,null),x(e,Hi,p),_(er,e,p),x(e,Vi,p),x(e,W,p),_(or,W,null),o(W,Rl),o(W,_n),o(W,Pl),o(W,yo),_(tr,yo,null),o(yo,Hl),o(yo,un),o(W,Vl),o(W,Mo),_(rr,Mo,null),o(Mo,Il),o(Mo,gn),o(W,Fl),o(W,$e),_(ar,$e,null),o($e,Wl),o($e,Ln),o($e,El),_(To,$e,null),o(W,Ul),o(W,Do),_(nr,Do,null),o(Do,Nl),o(Do,xn),o(W,Xl),o(W,ye),_(ir,ye,null),o(ye,ql),o(ye,wn),o(ye,zl),_(Ao,ye,null),x(e,Ii,p),_(sr,e,p),x(e,Fi,p),x(e,E,p),_(dr,E,null),o(E,jl),o(E,vn),o(E,Bl),o(E,Co),_(lr,Co,null),o(Co,Gl),o(Co,bn),o(E,Ol),o(E,So),_(cr,So,null),o(So,Jl),o(So,$n),o(E,Yl),o(E,Me),_(fr,Me,null),o(Me,Zl),o(Me,yn),o(Me,Kl),_(ko,Me,null),o(E,Ql),o(E,Ro),_(pr,Ro,null),o(Ro,ec),o(Ro,Mn),o(E,oc),o(E,Te),_(mr,Te,null),o(Te,tc),o(Te,Tn),o(Te,rc),_(Po,Te,null),x(e,Wi,p),_(hr,e,p),x(e,Ei,p),x(e,U,p),_(_r,U,null),o(U,ac),o(U,Dn),o(U,nc),o(U,Ho),_(ur,Ho,null),o(Ho,ic),o(Ho,An),o(U,sc),o(U,Vo),_(gr,Vo,null),o(Vo,dc),o(Vo,Cn),o(U,lc),o(U,De),_(Lr,De,null),o(De,cc),o(De,Sn),o(De,fc),_(Io,De,null),o(U,pc),o(U,Fo),_(xr,Fo,null),o(Fo,mc),o(Fo,kn),o(U,hc),o(U,Ae),_(wr,Ae,null),o(Ae,_c),o(Ae,Rn),o(Ae,uc),_(Wo,Ae,null),x(e,Ui,p),_(vr,e,p),x(e,Ni,p),x(e,N,p),_(br,N,null),o(N,gc),o(N,Pn),o(N,Lc),o(N,Eo),_($r,Eo,null),o(Eo,xc),o(Eo,Hn),o(N,wc),o(N,Uo),_(yr,Uo,null),o(Uo,vc),o(Uo,Vn),o(N,bc),o(N,Ce),_(Mr,Ce,null),o(Ce,$c),o(Ce,In),o(Ce,yc),_(No,Ce,null),o(N,Mc),o(N,Xo),_(Tr,Xo,null),o(Xo,Tc),o(Xo,Fn),o(N,Dc),o(N,Se),_(Dr,Se,null),o(Se,Ac),o(Se,Wn),o(Se,Cc),_(qo,Se,null),x(e,Xi,p),_(Ar,e,p),x(e,qi,p),x(e,X,p),_(Cr,X,null),o(X,Sc),o(X,En),o(X,kc),o(X,zo),_(Sr,zo,null),o(zo,Rc),o(zo,Un),o(X,Pc),o(X,jo),_(kr,jo,null),o(jo,Hc),o(jo,Nn),o(X,Vc),o(X,ke),_(Rr,ke,null),o(ke,Ic),o(ke,Xn),o(ke,Fc),_(Bo,ke,null),o(X,Wc),o(X,Go),_(Pr,Go,null),o(Go,Ec),o(Go,qn),o(X,Uc),o(X,Re),_(Hr,Re,null),o(Re,Nc),o(Re,zn),o(Re,Xc),_(Oo,Re,null),x(e,zi,p),_(Vr,e,p),x(e,ji,p),x(e,q,p),_(Ir,q,null),o(q,qc),o(q,jn),o(q,zc),o(q,Jo),_(Fr,Jo,null),o(Jo,jc),o(Jo,Bn),o(q,Bc),o(q,Yo),_(Wr,Yo,null),o(Yo,Gc),o(Yo,Gn),o(q,Oc),o(q,Pe),_(Er,Pe,null),o(Pe,Jc),o(Pe,On),o(Pe,Yc),_(Zo,Pe,null),o(q,Zc),o(q,Ko),_(Ur,Ko,null),o(Ko,Kc),o(Ko,Jn),o(q,Qc),o(q,He),_(Nr,He,null),o(He,ef),o(He,Yn),o(He,of),_(Qo,He,null),x(e,Bi,p),_(Xr,e,p),x(e,Gi,p),x(e,Ve,p),_(qr,Ve,null),o(Ve,tf),o(Ve,et),_(zr,et,null),o(et,rf),o(et,Zn),x(e,Oi,p),_(jr,e,p),x(e,Ji,p),x(e,Br,p),_(Gr,Br,null),x(e,Yi,p),_(Or,e,p),x(e,Zi,p),x(e,gi,p),Ki=!0},p(e,[p]){const z={};p&2&&(z.$$scope={dirty:p,ctx:e}),Xe.$set(z);const Jr={};p&2&&(Jr.$$scope={dirty:p,ctx:e}),je.$set(Jr);const Yr={};p&2&&(Yr.$$scope={dirty:p,ctx:e}),Je.$set(Yr);const J={};p&2&&(J.$$scope={dirty:p,ctx:e}),Qe.$set(J);const Ie={};p&2&&(Ie.$$scope={dirty:p,ctx:e}),oo.$set(Ie);const Zr={};p&2&&(Zr.$$scope={dirty:p,ctx:e}),ao.$set(Zr);const j={};p&2&&(j.$$scope={dirty:p,ctx:e}),io.$set(j);const Kr={};p&2&&(Kr.$$scope={dirty:p,ctx:e}),so.$set(Kr);const Qr={};p&2&&(Qr.$$scope={dirty:p,ctx:e}),fo.$set(Qr);const Y={};p&2&&(Y.$$scope={dirty:p,ctx:e}),mo.$set(Y);const Fe={};p&2&&(Fe.$$scope={dirty:p,ctx:e}),uo.$set(Fe);const ea={};p&2&&(ea.$$scope={dirty:p,ctx:e}),Lo.$set(ea);const R={};p&2&&(R.$$scope={dirty:p,ctx:e}),vo.$set(R);const oa={};p&2&&(oa.$$scope={dirty:p,ctx:e}),$o.$set(oa);const ta={};p&2&&(ta.$$scope={dirty:p,ctx:e}),To.$set(ta);const de={};p&2&&(de.$$scope={dirty:p,ctx:e}),Ao.$set(de);const We={};p&2&&(We.$$scope={dirty:p,ctx:e}),ko.$set(We);const ra={};p&2&&(ra.$$scope={dirty:p,ctx:e}),Po.$set(ra);const Ee={};p&2&&(Ee.$$scope={dirty:p,ctx:e}),Io.$set(Ee);const k={};p&2&&(k.$$scope={dirty:p,ctx:e}),Wo.$set(k);const aa={};p&2&&(aa.$$scope={dirty:p,ctx:e}),No.$set(aa);const na={};p&2&&(na.$$scope={dirty:p,ctx:e}),qo.$set(na);const le={};p&2&&(le.$$scope={dirty:p,ctx:e}),Bo.$set(le);const Ue={};p&2&&(Ue.$$scope={dirty:p,ctx:e}),Oo.$set(Ue);const ia={};p&2&&(ia.$$scope={dirty:p,ctx:e}),Zo.$set(ia);const Ne={};p&2&&(Ne.$$scope={dirty:p,ctx:e}),Qo.$set(Ne)},i(e){Ki||(u(T.$$.fragment,e),u(Xe.$$.fragment,e),u(tt.$$.fragment,e),u(rt.$$.fragment,e),u(at.$$.fragment,e),u(nt.$$.fragment,e),u(it.$$.fragment,e),u(st.$$.fragment,e),u(je.$$.fragment,e),u(dt.$$.fragment,e),u(lt.$$.fragment,e),u(ct.$$.fragment,e),u(ft.$$.fragment,e),u(pt.$$.fragment,e),u(mt.$$.fragment,e),u(ht.$$.fragment,e),u(Je.$$.fragment,e),u(_t.$$.fragment,e),u(ut.$$.fragment,e),u(gt.$$.fragment,e),u(Lt.$$.fragment,e),u(xt.$$.fragment,e),u(wt.$$.fragment,e),u(vt.$$.fragment,e),u(Qe.$$.fragment,e),u(bt.$$.fragment,e),u($t.$$.fragment,e),u(oo.$$.fragment,e),u(yt.$$.fragment,e),u(Mt.$$.fragment,e),u(Tt.$$.fragment,e),u(Dt.$$.fragment,e),u(At.$$.fragment,e),u(Ct.$$.fragment,e),u(ao.$$.fragment,e),u(St.$$.fragment,e),u(kt.$$.fragment,e),u(io.$$.fragment,e),u(Rt.$$.fragment,e),u(so.$$.fragment,e),u(Pt.$$.fragment,e),u(Ht.$$.fragment,e),u(Vt.$$.fragment,e),u(It.$$.fragment,e),u(Ft.$$.fragment,e),u(fo.$$.fragment,e),u(Wt.$$.fragment,e),u(Et.$$.fragment,e),u(mo.$$.fragment,e),u(Ut.$$.fragment,e),u(Nt.$$.fragment,e),u(Xt.$$.fragment,e),u(qt.$$.fragment,e),u(zt.$$.fragment,e),u(uo.$$.fragment,e),u(jt.$$.fragment,e),u(Bt.$$.fragment,e),u(Lo.$$.fragment,e),u(Gt.$$.fragment,e),u(Ot.$$.fragment,e),u(Jt.$$.fragment,e),u(Yt.$$.fragment,e),u(Zt.$$.fragment,e),u(vo.$$.fragment,e),u(Kt.$$.fragment,e),u(Qt.$$.fragment,e),u($o.$$.fragment,e),u(er.$$.fragment,e),u(or.$$.fragment,e),u(tr.$$.fragment,e),u(rr.$$.fragment,e),u(ar.$$.fragment,e),u(To.$$.fragment,e),u(nr.$$.fragment,e),u(ir.$$.fragment,e),u(Ao.$$.fragment,e),u(sr.$$.fragment,e),u(dr.$$.fragment,e),u(lr.$$.fragment,e),u(cr.$$.fragment,e),u(fr.$$.fragment,e),u(ko.$$.fragment,e),u(pr.$$.fragment,e),u(mr.$$.fragment,e),u(Po.$$.fragment,e),u(hr.$$.fragment,e),u(_r.$$.fragment,e),u(ur.$$.fragment,e),u(gr.$$.fragment,e),u(Lr.$$.fragment,e),u(Io.$$.fragment,e),u(xr.$$.fragment,e),u(wr.$$.fragment,e),u(Wo.$$.fragment,e),u(vr.$$.fragment,e),u(br.$$.fragment,e),u($r.$$.fragment,e),u(yr.$$.fragment,e),u(Mr.$$.fragment,e),u(No.$$.fragment,e),u(Tr.$$.fragment,e),u(Dr.$$.fragment,e),u(qo.$$.fragment,e),u(Ar.$$.fragment,e),u(Cr.$$.fragment,e),u(Sr.$$.fragment,e),u(kr.$$.fragment,e),u(Rr.$$.fragment,e),u(Bo.$$.fragment,e),u(Pr.$$.fragment,e),u(Hr.$$.fragment,e),u(Oo.$$.fragment,e),u(Vr.$$.fragment,e),u(Ir.$$.fragment,e),u(Fr.$$.fragment,e),u(Wr.$$.fragment,e),u(Er.$$.fragment,e),u(Zo.$$.fragment,e),u(Ur.$$.fragment,e),u(Nr.$$.fragment,e),u(Qo.$$.fragment,e),u(Xr.$$.fragment,e),u(qr.$$.fragment,e),u(zr.$$.fragment,e),u(jr.$$.fragment,e),u(Gr.$$.fragment,e),u(Or.$$.fragment,e),Ki=!0)},o(e){g(T.$$.fragment,e),g(Xe.$$.fragment,e),g(tt.$$.fragment,e),g(rt.$$.fragment,e),g(at.$$.fragment,e),g(nt.$$.fragment,e),g(it.$$.fragment,e),g(st.$$.fragment,e),g(je.$$.fragment,e),g(dt.$$.fragment,e),g(lt.$$.fragment,e),g(ct.$$.fragment,e),g(ft.$$.fragment,e),g(pt.$$.fragment,e),g(mt.$$.fragment,e),g(ht.$$.fragment,e),g(Je.$$.fragment,e),g(_t.$$.fragment,e),g(ut.$$.fragment,e),g(gt.$$.fragment,e),g(Lt.$$.fragment,e),g(xt.$$.fragment,e),g(wt.$$.fragment,e),g(vt.$$.fragment,e),g(Qe.$$.fragment,e),g(bt.$$.fragment,e),g($t.$$.fragment,e),g(oo.$$.fragment,e),g(yt.$$.fragment,e),g(Mt.$$.fragment,e),g(Tt.$$.fragment,e),g(Dt.$$.fragment,e),g(At.$$.fragment,e),g(Ct.$$.fragment,e),g(ao.$$.fragment,e),g(St.$$.fragment,e),g(kt.$$.fragment,e),g(io.$$.fragment,e),g(Rt.$$.fragment,e),g(so.$$.fragment,e),g(Pt.$$.fragment,e),g(Ht.$$.fragment,e),g(Vt.$$.fragment,e),g(It.$$.fragment,e),g(Ft.$$.fragment,e),g(fo.$$.fragment,e),g(Wt.$$.fragment,e),g(Et.$$.fragment,e),g(mo.$$.fragment,e),g(Ut.$$.fragment,e),g(Nt.$$.fragment,e),g(Xt.$$.fragment,e),g(qt.$$.fragment,e),g(zt.$$.fragment,e),g(uo.$$.fragment,e),g(jt.$$.fragment,e),g(Bt.$$.fragment,e),g(Lo.$$.fragment,e),g(Gt.$$.fragment,e),g(Ot.$$.fragment,e),g(Jt.$$.fragment,e),g(Yt.$$.fragment,e),g(Zt.$$.fragment,e),g(vo.$$.fragment,e),g(Kt.$$.fragment,e),g(Qt.$$.fragment,e),g($o.$$.fragment,e),g(er.$$.fragment,e),g(or.$$.fragment,e),g(tr.$$.fragment,e),g(rr.$$.fragment,e),g(ar.$$.fragment,e),g(To.$$.fragment,e),g(nr.$$.fragment,e),g(ir.$$.fragment,e),g(Ao.$$.fragment,e),g(sr.$$.fragment,e),g(dr.$$.fragment,e),g(lr.$$.fragment,e),g(cr.$$.fragment,e),g(fr.$$.fragment,e),g(ko.$$.fragment,e),g(pr.$$.fragment,e),g(mr.$$.fragment,e),g(Po.$$.fragment,e),g(hr.$$.fragment,e),g(_r.$$.fragment,e),g(ur.$$.fragment,e),g(gr.$$.fragment,e),g(Lr.$$.fragment,e),g(Io.$$.fragment,e),g(xr.$$.fragment,e),g(wr.$$.fragment,e),g(Wo.$$.fragment,e),g(vr.$$.fragment,e),g(br.$$.fragment,e),g($r.$$.fragment,e),g(yr.$$.fragment,e),g(Mr.$$.fragment,e),g(No.$$.fragment,e),g(Tr.$$.fragment,e),g(Dr.$$.fragment,e),g(qo.$$.fragment,e),g(Ar.$$.fragment,e),g(Cr.$$.fragment,e),g(Sr.$$.fragment,e),g(kr.$$.fragment,e),g(Rr.$$.fragment,e),g(Bo.$$.fragment,e),g(Pr.$$.fragment,e),g(Hr.$$.fragment,e),g(Oo.$$.fragment,e),g(Vr.$$.fragment,e),g(Ir.$$.fragment,e),g(Fr.$$.fragment,e),g(Wr.$$.fragment,e),g(Er.$$.fragment,e),g(Zo.$$.fragment,e),g(Ur.$$.fragment,e),g(Nr.$$.fragment,e),g(Qo.$$.fragment,e),g(Xr.$$.fragment,e),g(qr.$$.fragment,e),g(zr.$$.fragment,e),g(jr.$$.fragment,e),g(Gr.$$.fragment,e),g(Or.$$.fragment,e),Ki=!1},d(e){e&&(s($),s(l),s(c),s(d),s(y),s(Li),s(ot),s(xi),s(wi),s(vi),s(P),s(bi),s($i),s(H),s(yi),s(Mi),s(S),s(Ti),s(Di),s(A),s(Ai),s(Ci),s(V),s(Si),s(ki),s(I),s(Ri),s(Pi),s(F),s(Hi),s(Vi),s(W),s(Ii),s(Fi),s(E),s(Wi),s(Ei),s(U),s(Ui),s(Ni),s(N),s(Xi),s(qi),s(X),s(zi),s(ji),s(q),s(Bi),s(Gi),s(Ve),s(Oi),s(Ji),s(Br),s(Yi),s(Zi),s(gi)),s(i),L(T,e),L(Xe,e),L(tt,e),L(rt),L(at),L(nt),L(it),L(st),L(je),L(dt),L(lt,e),L(ct),L(ft),L(pt),L(mt),L(ht),L(Je),L(_t),L(ut,e),L(gt),L(Lt),L(xt),L(wt),L(vt),L(Qe),L(bt),L($t),L(oo),L(yt,e),L(Mt),L(Tt),L(Dt),L(At),L(Ct),L(ao),L(St),L(kt),L(io),L(Rt),L(so),L(Pt,e),L(Ht),L(Vt),L(It),L(Ft),L(fo),L(Wt),L(Et),L(mo),L(Ut,e),L(Nt),L(Xt),L(qt),L(zt),L(uo),L(jt),L(Bt),L(Lo),L(Gt,e),L(Ot),L(Jt),L(Yt),L(Zt),L(vo),L(Kt),L(Qt),L($o),L(er,e),L(or),L(tr),L(rr),L(ar),L(To),L(nr),L(ir),L(Ao),L(sr,e),L(dr),L(lr),L(cr),L(fr),L(ko),L(pr),L(mr),L(Po),L(hr,e),L(_r),L(ur),L(gr),L(Lr),L(Io),L(xr),L(wr),L(Wo),L(vr,e),L(br),L($r),L(yr),L(Mr),L(No),L(Tr),L(Dr),L(qo),L(Ar,e),L(Cr),L(Sr),L(kr),L(Rr),L(Bo),L(Pr),L(Hr),L(Oo),L(Vr,e),L(Ir),L(Fr),L(Wr),L(Er),L(Zo),L(Ur),L(Nr),L(Qo),L(Xr,e),L(qr),L(zr),L(jr,e),L(Gr),L(Or,e)}}}const Pm='{"title":"LoRA","local":"lora","sections":[{"title":"StableDiffusionLoraLoaderMixin","local":"diffusers.loaders.StableDiffusionLoraLoaderMixin","sections":[],"depth":2},{"title":"StableDiffusionXLLoraLoaderMixin","local":"diffusers.loaders.StableDiffusionXLLoraLoaderMixin","sections":[],"depth":2},{"title":"SD3LoraLoaderMixin","local":"diffusers.loaders.SD3LoraLoaderMixin","sections":[],"depth":2},{"title":"FluxLoraLoaderMixin","local":"diffusers.loaders.FluxLoraLoaderMixin","sections":[],"depth":2},{"title":"CogVideoXLoraLoaderMixin","local":"diffusers.loaders.CogVideoXLoraLoaderMixin","sections":[],"depth":2},{"title":"Mochi1LoraLoaderMixin","local":"diffusers.loaders.Mochi1LoraLoaderMixin","sections":[],"depth":2},{"title":"AuraFlowLoraLoaderMixin","local":"diffusers.loaders.AuraFlowLoraLoaderMixin","sections":[],"depth":2},{"title":"LTXVideoLoraLoaderMixin","local":"diffusers.loaders.LTXVideoLoraLoaderMixin","sections":[],"depth":2},{"title":"SanaLoraLoaderMixin","local":"diffusers.loaders.SanaLoraLoaderMixin","sections":[],"depth":2},{"title":"HunyuanVideoLoraLoaderMixin","local":"diffusers.loaders.HunyuanVideoLoraLoaderMixin","sections":[],"depth":2},{"title":"Lumina2LoraLoaderMixin","local":"diffusers.loaders.Lumina2LoraLoaderMixin","sections":[],"depth":2},{"title":"WanLoraLoaderMixin","local":"diffusers.loaders.WanLoraLoaderMixin","sections":[],"depth":2},{"title":"CogView4LoraLoaderMixin","local":"diffusers.loaders.CogView4LoraLoaderMixin","sections":[],"depth":2},{"title":"AmusedLoraLoaderMixin","local":"diffusers.loaders.AmusedLoraLoaderMixin","sections":[],"depth":2},{"title":"LoraBaseMixin","local":"diffusers.loaders.LoraBaseMixin","sections":[],"depth":2}],"depth":1}';function Hm(M){return Qp(()=>{new URLSearchParams(window.location.search).get("fw")}),[]}class Xm extends em{constructor(i){super(),om(this,i,Hm,Rm,Kp,{})}}export{Xm as component};