diffusers/pr_13921/en/_app/immutable/nodes/46.25123c5c.js

import{s as oe,n as te,o as ne}from"../chunks/scheduler.53228c21.js";import{S as re,i as se,e as i,s as r,c as D,h as ae,a as m,d as t,b as s,f as O,g as C,j as R,k as U,l,m as a,n as M,t as w,o as y,p as k}from"../chunks/index.cac5d66a.js";import{C as de}from"../chunks/CopyLLMTxtMenu.0ef49226.js";import{D as Z}from"../chunks/Docstring.9de32ff4.js";import{H as ee,E as ie}from"../chunks/MermaidChart.svelte_svelte_type_style_lang.48d5cb47.js";function me(J){let d,P,L,j,h,H,p,z,u,K='A modified flux Transformer model from <a href="https://huggingface.co/lodestones/Chroma1-HD" rel="nofollow">Chroma</a>',E,_,A,n,g,V,x,Q="The Transformer model introduced in Flux, modified for Chroma.",W,$,X='Reference: <a href="https://huggingface.co/lodestones/Chroma1-HD" rel="nofollow">https://huggingface.co/lodestones/Chroma1-HD</a>',G,c,T,B,v,Y='The <a href="/docs/diffusers/pr_13921/en/api/models/flux_transformer#diffusers.FluxTransformer2DModel">FluxTransformer2DModel</a> forward method.',I,b,F,N,q;return h=new de({props:{containerStyle:"float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"}}),p=new ee({props:{title:"ChromaTransformer2DModel",local:"chromatransformer2dmodel",headingTag:"h1"}}),_=new ee({props:{title:"ChromaTransformer2DModel",local:"diffusers.ChromaTransformer2DModel",headingTag:"h2"}}),g=new Z({props:{name:"class diffusers.ChromaTransformer2DModel",anchor:"diffusers.ChromaTransformer2DModel",parameters:[{name:"patch_size",val:": int = 1"},{name:"in_channels",val:": int = 64"},{name:"out_channels",val:": int | None = None"},{name:"num_layers",val:": int = 19"},{name:"num_single_layers",val:": int = 38"},{name:"attention_head_dim",val:": int = 128"},{name:"num_attention_heads",val:": int = 24"},{name:"joint_attention_dim",val:": int = 4096"},{name:"axes_dims_rope",val:": tuple = (16, 56, 56)"},{name:"approximator_num_channels",val:": int = 64"},{name:"approximator_hidden_dim",val:": int = 5120"},{name:"approximator_layers",val:": int = 5"}],parametersDescription:[{anchor:"diffusers.ChromaTransformer2DModel.patch_size",description:`<strong>patch_size</strong> (<code>int</code>, defaults to <code>1</code>) &#x2014;
Patch size to turn the input data into small patches.`,name:"patch_size"},{anchor:"diffusers.ChromaTransformer2DModel.in_channels",description:`<strong>in_channels</strong> (<code>int</code>, defaults to <code>64</code>) &#x2014;
The number of channels in the input.`,name:"in_channels"},{anchor:"diffusers.ChromaTransformer2DModel.out_channels",description:`<strong>out_channels</strong> (<code>int</code>, <em>optional</em>, defaults to <code>None</code>) &#x2014;
The number of channels in the output. If not specified, it defaults to <code>in_channels</code>.`,name:"out_channels"},{anchor:"diffusers.ChromaTransformer2DModel.num_layers",description:`<strong>num_layers</strong> (<code>int</code>, defaults to <code>19</code>) &#x2014;
The number of layers of dual stream DiT blocks to use.`,name:"num_layers"},{anchor:"diffusers.ChromaTransformer2DModel.num_single_layers",description:`<strong>num_single_layers</strong> (<code>int</code>, defaults to <code>38</code>) &#x2014;
The number of layers of single stream DiT blocks to use.`,name:"num_single_layers"},{anchor:"diffusers.ChromaTransformer2DModel.attention_head_dim",description:`<strong>attention_head_dim</strong> (<code>int</code>, defaults to <code>128</code>) &#x2014;
The number of dimensions to use for each attention head.`,name:"attention_head_dim"},{anchor:"diffusers.ChromaTransformer2DModel.num_attention_heads",description:`<strong>num_attention_heads</strong> (<code>int</code>, defaults to <code>24</code>) &#x2014;
The number of attention heads to use.`,name:"num_attention_heads"},{anchor:"diffusers.ChromaTransformer2DModel.joint_attention_dim",description:`<strong>joint_attention_dim</strong> (<code>int</code>, defaults to <code>4096</code>) &#x2014;
The number of dimensions to use for the joint attention (embedding/channel dimension of
<code>encoder_hidden_states</code>).`,name:"joint_attention_dim"},{anchor:"diffusers.ChromaTransformer2DModel.axes_dims_rope",description:`<strong>axes_dims_rope</strong> (<code>tuple[int]</code>, defaults to <code>(16, 56, 56)</code>) &#x2014;
The dimensions to use for the rotary positional embeddings.`,name:"axes_dims_rope"}],source:"https://github.com/huggingface/diffusers/blob/vr_13921/src/diffusers/models/transformers/transformer_chroma.py#L370"}}),T=new Z({props:{name:"forward",anchor:"diffusers.ChromaTransformer2DModel.forward",parameters:[{name:"hidden_states",val:": Tensor"},{name:"encoder_hidden_states",val:": Tensor = None"},{name:"timestep",val:": LongTensor = None"},{name:"img_ids",val:": Tensor = None"},{name:"txt_ids",val:": Tensor = None"},{name:"attention_mask",val:": Tensor = None"},{name:"joint_attention_kwargs",val:": dict[str, typing.Any] | None = None"},{name:"controlnet_block_samples",val:" = None"},{name:"controlnet_single_block_samples",val:" = None"},{name:"return_dict",val:": bool = True"},{name:"controlnet_blocks_repeat",val:": bool = False"}],parametersDescription:[{anchor:"diffusers.ChromaTransformer2DModel.forward.hidden_states",description:`<strong>hidden_states</strong> (<code>torch.Tensor</code> of shape <code>(batch_size, image_sequence_length, in_channels)</code>) &#x2014;
Input <code>hidden_states</code>.`,name:"hidden_states"},{anchor:"diffusers.ChromaTransformer2DModel.forward.encoder_hidden_states",description:`<strong>encoder_hidden_states</strong> (<code>torch.Tensor</code> of shape <code>(batch_size, text_sequence_length, joint_attention_dim)</code>) &#x2014;
Conditional embeddings (embeddings computed from the input conditions such as prompts) to use.`,name:"encoder_hidden_states"},{anchor:"diffusers.ChromaTransformer2DModel.forward.timestep",description:`<strong>timestep</strong> ( <code>torch.LongTensor</code>) &#x2014;
Used to indicate denoising step.`,name:"timestep"},{anchor:"diffusers.ChromaTransformer2DModel.forward.img_ids",description:`<strong>img_ids</strong> (<code>torch.Tensor</code>) &#x2014;
Image position ids used to compute the rotary positional embeddings.`,name:"img_ids"},{anchor:"diffusers.ChromaTransformer2DModel.forward.txt_ids",description:`<strong>txt_ids</strong> (<code>torch.Tensor</code>) &#x2014;
Text position ids used to compute the rotary positional embeddings.`,name:"txt_ids"},{anchor:"diffusers.ChromaTransformer2DModel.forward.attention_mask",description:`<strong>attention_mask</strong> (<code>torch.Tensor</code>, <em>optional</em>) &#x2014;
Mask applied to <code>encoder_hidden_states</code> during attention.`,name:"attention_mask"},{anchor:"diffusers.ChromaTransformer2DModel.forward.controlnet_block_samples",description:`<strong>controlnet_block_samples</strong> (<code>list</code> of <code>torch.Tensor</code>, <em>optional</em>) &#x2014;
A list of tensors that if specified are added to the residuals of transformer blocks.`,name:"controlnet_block_samples"},{anchor:"diffusers.ChromaTransformer2DModel.forward.controlnet_single_block_samples",description:`<strong>controlnet_single_block_samples</strong> (<code>list</code> of <code>torch.Tensor</code>, <em>optional</em>) &#x2014;
A list of tensors that if specified are added to the residuals of single transformer blocks.`,name:"controlnet_single_block_samples"},{anchor:"diffusers.ChromaTransformer2DModel.forward.controlnet_blocks_repeat",description:`<strong>controlnet_blocks_repeat</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) &#x2014;
Whether to repeat the controlnet block samples across all transformer blocks.`,name:"controlnet_blocks_repeat"},{anchor:"diffusers.ChromaTransformer2DModel.forward.joint_attention_kwargs",description:`<strong>joint_attention_kwargs</strong> (<code>dict</code>, <em>optional</em>) &#x2014;
A kwargs dictionary that if specified is passed along to the <code>AttentionProcessor</code> as defined under
<code>self.processor</code> in
<a href="https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py" rel="nofollow">diffusers.models.attention_processor</a>.`,name:"joint_attention_kwargs"},{anchor:"diffusers.ChromaTransformer2DModel.forward.return_dict",description:`<strong>return_dict</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) &#x2014;
Whether or not to return a <code>~models.transformer_2d.Transformer2DModelOutput</code> instead of a plain
tuple.`,name:"return_dict"}],source:"https://github.com/huggingface/diffusers/blob/vr_13921/src/diffusers/models/transformers/transformer_chroma.py#L476",returnDescription:`<script context="module">export const metadata = 'undefined';<\/script>


<p>If <code>return_dict</code> is True, an <code>~models.transformer_2d.Transformer2DModelOutput</code> is returned, otherwise a
<code>tuple</code> where the first element is the sample tensor.</p>
`}}),b=new ie({props:{source:"https://github.com/huggingface/diffusers/blob/main/docs/source/en/api/models/chroma_transformer.md"}}),{c(){d=i("meta"),P=r(),L=i("p"),j=r(),D(h.$$.fragment),H=r(),D(p.$$.fragment),z=r(),u=i("p"),u.innerHTML=K,E=r(),D(_.$$.fragment),A=r(),n=i("div"),D(g.$$.fragment),V=r(),x=i("p"),x.textContent=Q,W=r(),$=i("p"),$.innerHTML=X,G=r(),c=i("div"),D(T.$$.fragment),B=r(),v=i("p"),v.innerHTML=Y,I=r(),D(b.$$.fragment),F=r(),N=i("p"),this.h()},l(e){const o=ae("svelte-u9bgzb",document.head);d=m(o,"META",{name:!0,content:!0}),o.forEach(t),P=s(e),L=m(e,"P",{}),O(L).forEach(t),j=s(e),C(h.$$.fragment,e),H=s(e),C(p.$$.fragment,e),z=s(e),u=m(e,"P",{"data-svelte-h":!0}),R(u)!=="svelte-1ynrsh5"&&(u.innerHTML=K),E=s(e),C(_.$$.fragment,e),A=s(e),n=m(e,"DIV",{class:!0});var f=O(n);C(g.$$.fragment,f),V=s(f),x=m(f,"P",{"data-svelte-h":!0}),R(x)!=="svelte-x7vpty"&&(x.textContent=Q),W=s(f),$=m(f,"P",{"data-svelte-h":!0}),R($)!=="svelte-gq0pya"&&($.innerHTML=X),G=s(f),c=m(f,"DIV",{class:!0});var S=O(c);C(T.$$.fragment,S),B=s(S),v=m(S,"P",{"data-svelte-h":!0}),R(v)!=="svelte-zht8lg"&&(v.innerHTML=Y),S.forEach(t),f.forEach(t),I=s(e),C(b.$$.fragment,e),F=s(e),N=m(e,"P",{}),O(N).forEach(t),this.h()},h(){U(d,"name","hf:doc:metadata"),U(d,"content",le),U(c,"class","docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"),U(n,"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,o){l(document.head,d),a(e,P,o),a(e,L,o),a(e,j,o),M(h,e,o),a(e,H,o),M(p,e,o),a(e,z,o),a(e,u,o),a(e,E,o),M(_,e,o),a(e,A,o),a(e,n,o),M(g,n,null),l(n,V),l(n,x),l(n,W),l(n,$),l(n,G),l(n,c),M(T,c,null),l(c,B),l(c,v),a(e,I,o),M(b,e,o),a(e,F,o),a(e,N,o),q=!0},p:te,i(e){q||(w(h.$$.fragment,e),w(p.$$.fragment,e),w(_.$$.fragment,e),w(g.$$.fragment,e),w(T.$$.fragment,e),w(b.$$.fragment,e),q=!0)},o(e){y(h.$$.fragment,e),y(p.$$.fragment,e),y(_.$$.fragment,e),y(g.$$.fragment,e),y(T.$$.fragment,e),y(b.$$.fragment,e),q=!1},d(e){e&&(t(P),t(L),t(j),t(H),t(z),t(u),t(E),t(A),t(n),t(I),t(F),t(N)),t(d),k(h,e),k(p,e),k(_,e),k(g),k(T),k(b,e)}}}const le='{"title":"ChromaTransformer2DModel","local":"chromatransformer2dmodel","sections":[{"title":"ChromaTransformer2DModel","local":"diffusers.ChromaTransformer2DModel","sections":[],"depth":2}],"depth":1}';function ce(J){return ne(()=>{new URLSearchParams(window.location.search).get("fw")}),[]}class ge extends re{constructor(d){super(),se(this,d,ce,me,oe,{})}}export{ge as component};