Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 4 additions & 2 deletions built-in-nodes/LatentApplyOperationCFG.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -14,8 +14,10 @@ The LatentApplyOperationCFG node applies a latent operation inside the classifie
| --- | --- | --- | --- | --- |
| `model` | The model to which the CFG operation will be applied | MODEL | Yes | - |
| `operation` | The latent operation to apply during the CFG sampling process | LATENT_OPERATION | Yes | - |
| `start_percent` | Fraction of the denoising schedule at which the operation starts being applied; 0 is the beginning of the schedule (default: 0.0) | FLOAT | No | 0.0 to 1.0 (step 0.001) |
| `end_percent` | Fraction of the denoising schedule at which the operation stops being applied; 1 is the end of the schedule (default: 1.0) | FLOAT | No | 0.0 to 1.0 (step 0.001) |

Note: This node is marked as experimental. The operation is applied to the model's conditioning outputs during the CFG sampling process. When two conditioning outputs are present, the operation is applied to the difference between the first and second output, and the second output is added back to the result. When only one conditioning output is present, the operation is applied directly to it.
Note: This node is marked as experimental. The operation is applied to the model's conditioning outputs during the CFG sampling process. When two conditioning outputs are present, the operation is applied to the difference between the first and second output, and the second output is added back to the result. When only one conditioning output is present, the operation is applied directly to it. The operation only runs between the `start_percent` and `end_percent` points of the denoising schedule, so it can be limited to part of the schedule; outside that window the conditioning outputs are returned unchanged.

## Outputs

Expand All @@ -26,4 +28,4 @@ Note: This node is marked as experimental. The operation is applied to the model
> This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LatentApplyOperationCFG/en.md)

---
**Source fingerprint (SHA-256):** `e383684a785878bfa4004c2fac78ae562d8e035fdfe081f8e4ebbb2c50161987`
**Source fingerprint (SHA-256):** `6a5f59f02eaec38334c63d871e48e89aa983a5ac2ca10801161cdc9e13cacdf2`
27 changes: 27 additions & 0 deletions built-in-nodes/LatentOperationBlend.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
---
title: "LatentOperationBlend - ComfyUI Built-in Node Documentation"
description: "This node creates a latent operation that blends a latent toward a reference latent"
sidebarTitle: "LatentOperationBlend"
icon: "circle"
mode: wide
---

This node creates a latent operation that blends a latent toward a reference latent, and returns that operation so it can be plugged into nodes such as Latent Apply Operation or Latent Apply Operation CFG. When the reference is a different spatial size, it is resized to the target latent with nearest-neighbor interpolation, and a reference with a smaller batch is repeated to match the target batch size. A strength of 0 leaves the latent unchanged. This node is marked as experimental.

## Inputs

| Parameter | Description | Data Type | Required | Range |
| --- | --- | --- | --- | --- |
| `reference` | The latent to blend toward. Its samples are cast to the device and dtype of the latent being processed, then resized and repeated to match it. | LATENT | Yes | - |
| `strength` | How far to blend toward the reference: 0 leaves the latent unchanged, 1 matches the resized reference latent (default: 1.0). | FLOAT | Yes | 0.0 to 1.0 (step 0.0001) |

## Outputs

| Output Name | Description | Data Type |
| --- | --- | --- |
| `operation` | A blend operation that can be applied to latent samples. | LATENT_OPERATION |

> This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! [Edit on GitHub](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LatentOperationBlend/en.md)

---
**Source fingerprint (SHA-256):** `5890089afddf83ddd4edd992606509b118aac9ef13eb89589f73fa75e0b9dd5a`
4 changes: 4 additions & 0 deletions docs.json
Original file line number Diff line number Diff line change
Expand Up @@ -1885,6 +1885,7 @@
"pages": [
"built-in-nodes/LatentApplyOperation",
"built-in-nodes/LatentApplyOperationCFG",
"built-in-nodes/LatentOperationBlend",
"built-in-nodes/LatentOperationSharpen",
"built-in-nodes/LatentOperationTonemapReinhard"
]
Expand Down Expand Up @@ -5353,6 +5354,7 @@
"pages": [
"zh/built-in-nodes/LatentApplyOperation",
"zh/built-in-nodes/LatentApplyOperationCFG",
"zh/built-in-nodes/LatentOperationBlend",
"zh/built-in-nodes/LatentOperationSharpen",
"zh/built-in-nodes/LatentOperationTonemapReinhard"
]
Expand Down Expand Up @@ -9225,6 +9227,7 @@
"pages": [
"ja/built-in-nodes/LatentApplyOperation",
"ja/built-in-nodes/LatentApplyOperationCFG",
"ja/built-in-nodes/LatentOperationBlend",
"ja/built-in-nodes/LatentOperationSharpen",
"ja/built-in-nodes/LatentOperationTonemapReinhard"
]
Expand Down Expand Up @@ -13030,6 +13033,7 @@
"pages": [
"ko/built-in-nodes/LatentApplyOperation",
"ko/built-in-nodes/LatentApplyOperationCFG",
"ko/built-in-nodes/LatentOperationBlend",
"ko/built-in-nodes/LatentOperationSharpen",
"ko/built-in-nodes/LatentOperationTonemapReinhard"
]
Expand Down
6 changes: 4 additions & 2 deletions ja/built-in-nodes/LatentApplyOperationCFG.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -14,8 +14,10 @@ LatentApplyOperationCFG ノードは、モデルのサンプリングプロセ
| --- | --- | --- | --- | --- |
| `モデル` | CFG 操作を適用する対象のモデル | MODEL | はい | - |
| `操作` | CFG サンプリングプロセス中に適用する latent 操作 | LATENT_OPERATION | はい | - |
| `start_percent` | 操作の適用を開始するノイズ除去スケジュール上の割合。0 はスケジュールの開始位置です(デフォルト: 0.0) | FLOAT | いいえ | 0.0 ~ 1.0(ステップ 0.001) |
| `end_percent` | 操作の適用を停止するノイズ除去スケジュール上の割合。1 はスケジュールの終了位置です(デフォルト: 1.0) | FLOAT | いいえ | 0.0 ~ 1.0(ステップ 0.001) |

注:このノードは experimental としてマークされています。操作は、CFG サンプリングプロセス中にモデルの conditioning 出力に対して適用されます。conditioning 出力が 2 つ存在する場合、操作は 1 つ目と 2 つ目の出力の差分に適用され、2 つ目の出力が結果に加算して戻されます。conditioning 出力が 1 つだけの場合は、その出力に直接操作が適用されます。
注:このノードは experimental としてマークされています。操作は、CFG サンプリングプロセス中にモデルの conditioning 出力に対して適用されます。conditioning 出力が 2 つ存在する場合、操作は 1 つ目と 2 つ目の出力の差分に適用され、2 つ目の出力が結果に加算して戻されます。conditioning 出力が 1 つだけの場合は、その出力に直接操作が適用されます。 この操作は、ノイズ除去スケジュールの `start_percent` と `end_percent` の間でのみ実行されるため、スケジュールの一部に限定できます。その区間外では、conditioning 出力は変更されずに返されます。

## 出力

Expand All @@ -26,4 +28,4 @@ LatentApplyOperationCFG ノードは、モデルのサンプリングプロセ
> このドキュメントは AI によって生成されました。エラーを見つけた場合や改善のご提案がある場合は、ぜひ貢献してください! [GitHub で編集](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LatentApplyOperationCFG/ja.md)

---
**Source fingerprint (SHA-256):** `e383684a785878bfa4004c2fac78ae562d8e035fdfe081f8e4ebbb2c50161987`
**Source fingerprint (SHA-256):** `6a5f59f02eaec38334c63d871e48e89aa983a5ac2ca10801161cdc9e13cacdf2`
27 changes: 27 additions & 0 deletions ja/built-in-nodes/LatentOperationBlend.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
---
title: "LatentOperationBlend - ComfyUI Built-in Node Documentation"
Comment thread
coderabbitai[bot] marked this conversation as resolved.
description: "このノードは、latent を参照 latent に向けてブレンドする latent 操作を作成し、その操作を返します。"
sidebarTitle: "LatentOperationBlend"
icon: "circle"
mode: wide
---

このノードは、latent を参照 latent に向けてブレンドする latent 操作を作成し、その操作を返します。これにより、Latent Apply Operation や Latent Apply Operation CFG などのノードに接続できます。参照の空間サイズが異なる場合は、最近傍補間を用いて対象の latent にリサイズされ、バッチ数が少ない参照は対象のバッチサイズに合わせて繰り返されます。strength が 0 の場合、latent は変更されません。このノードは実験的としてマークされています。

## 入力

| パラメータ | 説明 | データ型 | 必須 | 範囲 |
| --- | --- | --- | --- | --- |
| `reference` | ブレンド先となる latent。そのサンプルは処理対象の latent のデバイスと dtype にキャストされた後、それに合わせてリサイズおよび繰り返されます。 | LATENT | はい | - |
| `strength` | 参照へどの程度ブレンドするか: 0 は latent を変更せず、1 はリサイズされた参照 latent と一致します(デフォルト: 1.0)。 | FLOAT | はい | 0.0 〜 1.0 (ステップ 0.0001) |

## 出力

| 出力名 | 説明 | データ型 |
| --- | --- | --- |
| `operation` | latent サンプルに適用できるブレンド操作。 | LATENT_OPERATION |

> このドキュメントは AI によって生成されました。エラーを見つけた場合や改善のご提案がある場合は、ぜひ貢献してください! [GitHub で編集](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LatentOperationBlend/ja.md)

---
**Source fingerprint (SHA-256):** `5890089afddf83ddd4edd992606509b118aac9ef13eb89589f73fa75e0b9dd5a`
6 changes: 4 additions & 2 deletions ko/built-in-nodes/LatentApplyOperationCFG.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -14,8 +14,10 @@ LatentApplyOperationCFG 노드는 모델 샘플링 과정의 classifier-free gui
| --- | --- | --- | --- | --- |
| `모델` | CFG 연산이 적용될 모델입니다 | MODEL | 예 | - |
| `연산` | CFG 샘플링 과정에서 적용할 잠재 연산입니다 | LATENT_OPERATION | 예 | - |
| `start_percent` | 연산 적용이 시작되는 디노이징 스케줄상의 비율입니다. 0은 스케줄의 시작입니다(기본값: 0.0) | FLOAT | 아니요 | 0.0 ~ 1.0 (단계 0.001) |
| `end_percent` | 연산 적용이 중지되는 디노이징 스케줄상의 비율입니다. 1은 스케줄의 끝입니다(기본값: 1.0) | FLOAT | 아니요 | 0.0 ~ 1.0 (단계 0.001) |

참고: 이 노드는 실험적인 것으로 표시되어 있습니다. 이 연산은 CFG 샘플링 과정에서 모델의 컨디셔닝 출력에 적용됩니다. 컨디셔닝 출력이 두 개 있는 경우, 연산은 첫 번째 출력과 두 번째 출력의 차이에 적용되고 두 번째 출력이 결과에 다시 더해집니다. 컨디셔닝 출력이 하나만 있는 경우에는 해당 출력에 연산이 직접 적용됩니다.
참고: 이 노드는 실험적인 것으로 표시되어 있습니다. 이 연산은 CFG 샘플링 과정에서 모델의 컨디셔닝 출력에 적용됩니다. 컨디셔닝 출력이 두 개 있는 경우, 연산은 첫 번째 출력과 두 번째 출력의 차이에 적용되고 두 번째 출력이 결과에 다시 더해집니다. 컨디셔닝 출력이 하나만 있는 경우에는 해당 출력에 연산이 직접 적용됩니다. 이 연산은 디노이징 스케줄의 `start_percent`와 `end_percent` 사이에서만 실행되므로 스케줄의 일부로 제한할 수 있습니다. 해당 구간 밖에서는 컨디셔닝 출력이 변경 없이 반환됩니다.

## 출력

Expand All @@ -26,4 +28,4 @@ LatentApplyOperationCFG 노드는 모델 샘플링 과정의 classifier-free gui
> 이 문서는 AI에 의해 생성되었습니다. 오류를 발견하거나 개선 제안이 있으시면 기여해 주세요! [GitHub에서 편집](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LatentApplyOperationCFG/ko.md)

---
**Source fingerprint (SHA-256):** `e383684a785878bfa4004c2fac78ae562d8e035fdfe081f8e4ebbb2c50161987`
**Source fingerprint (SHA-256):** `6a5f59f02eaec38334c63d871e48e89aa983a5ac2ca10801161cdc9e13cacdf2`
27 changes: 27 additions & 0 deletions ko/built-in-nodes/LatentOperationBlend.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
---
title: "LatentOperationBlend - ComfyUI Built-in Node Documentation"
description: "이 노드는 레이턴트를 참조 레이턴트 쪽으로 블렌딩하는 레이턴트 연산을 생성하고, 해당 연산을 반환하여 Latent Apply Operation 또는 Latent Apply Operation CFG와 같은 노드에 연결할 수 있게 합니다."
sidebarTitle: "LatentOperationBlend"
icon: "circle"
mode: wide
---

이 노드는 레이턴트를 참조 레이턴트 쪽으로 블렌딩하는 레이턴트 연산을 생성하고, 해당 연산을 반환하여 Latent Apply Operation 또는 Latent Apply Operation CFG와 같은 노드에 연결할 수 있게 합니다. 참조가 다른 공간 크기를 가질 경우, 대상 레이턴트에 맞게 최근접 이웃 보간으로 크기가 조정되며, 배치 항목이 더 적은 참조는 대상 배치 크기에 맞게 반복됩니다. `strength`가 0이면 레이턴트는 변경되지 않습니다. 이 노드는 실험용으로 표시되어 있습니다.

## 입력

| 매개변수 | 설명 | 데이터 타입 | 필수 | 범위 |
| --- | --- | --- | --- | --- |
| `reference` | 블렌딩 대상이 되는 레이턴트입니다. 그 샘플은 처리 중인 레이턴트의 device 및 dtype으로 캐스팅된 다음, 이에 맞게 크기가 조정되고 반복됩니다. | LATENT | 예 | - |
| `strength` | 참조 쪽으로 얼마나 블렌딩할지 지정합니다. 0이면 레이턴트가 변경되지 않고, 1이면 크기가 조정된 참조 레이턴트와 일치합니다(기본값: 1.0). | FLOAT | 예 | 0.0 ~ 1.0 (단계 0.0001) |

## 출력

| 출력 이름 | 설명 | 데이터 타입 |
| --- | --- | --- |
| `operation` | 레이턴트 샘플에 적용할 수 있는 블렌드 연산입니다. | LATENT_OPERATION |

> 이 문서는 AI에 의해 생성되었습니다. 오류를 발견하거나 개선 제안이 있으시면 기여해 주세요! [GitHub에서 편집](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LatentOperationBlend/ko.md)

---
**Source fingerprint (SHA-256):** `5890089afddf83ddd4edd992606509b118aac9ef13eb89589f73fa75e0b9dd5a`
6 changes: 4 additions & 2 deletions zh/built-in-nodes/LatentApplyOperationCFG.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -14,8 +14,10 @@ LatentApplyOperationCFG 节点在模型采样过程的无分类器引导(CFG
| --- | --- | --- | --- | --- |
| `模型` | 将应用 CFG 操作的模型 | MODEL | 是 | - |
| `操作` | 在 CFG 采样过程中应用的 latent 操作 | LATENT_OPERATION | 是 | - |
| `start_percent` | 操作开始应用时的去噪调度比例;0 表示调度的起点(默认值:0.0) | FLOAT | 否 | 0.0 到 1.0(步长 0.001) |
| `end_percent` | 操作停止应用时的去噪调度比例;1 表示调度的终点(默认值:1.0) | FLOAT | 否 | 0.0 到 1.0(步长 0.001) |

注意:此节点被标记为实验性。该操作会在 CFG 采样过程中应用于模型的条件输出。当存在两个条件输出时,该操作会应用于第一个和第二个输出之间的差值,并将第二个输出加回结果。当只存在一个条件输出时,该操作会直接应用于该输出。
注意:此节点被标记为实验性。该操作会在 CFG 采样过程中应用于模型的条件输出。当存在两个条件输出时,该操作会应用于第一个和第二个输出之间的差值,并将第二个输出加回结果。当只存在一个条件输出时,该操作会直接应用于该输出。 该操作仅在去噪调度中 `start_percent` 与 `end_percent` 之间的区间内运行,因此可以限定在调度的一部分;在该区间之外,条件输出会原样返回。

## 输出

Expand All @@ -26,4 +28,4 @@ LatentApplyOperationCFG 节点在模型采样过程的无分类器引导(CFG
> 本文档由 AI 生成。如果您发现任何错误或有改进建议,欢迎贡献! [在 GitHub 上编辑](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LatentApplyOperationCFG/zh.md)

---
**Source fingerprint (SHA-256):** `e383684a785878bfa4004c2fac78ae562d8e035fdfe081f8e4ebbb2c50161987`
**Source fingerprint (SHA-256):** `6a5f59f02eaec38334c63d871e48e89aa983a5ac2ca10801161cdc9e13cacdf2`
27 changes: 27 additions & 0 deletions zh/built-in-nodes/LatentOperationBlend.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
---
title: "LatentOperationBlend - ComfyUI Built-in Node Documentation"
description: "此节点创建一个 latent 操作,用于将 latent 向参考 latent 混合,并返回该操作,以便将其接入 Latent Apply Operation 或 Latent Apply Operation CFG 等节点。"
sidebarTitle: "LatentOperationBlend"
icon: "circle"
mode: wide
---

此节点创建一个 latent 操作,用于将 latent 向参考 latent 混合,并返回该操作,以便将其接入 Latent Apply Operation 或 Latent Apply Operation CFG 等节点。当参考 latent 的空间尺寸不同时,会使用最近邻插值将其调整为目标 latent 的尺寸;如果参考 latent 的批次较小,则会重复以匹配目标批次大小。`strength` 为 0 时,latent 保持不变。此节点被标记为实验性节点。

## 输入

| 参数 | 描述 | 数据类型 | 必填 | 范围 |
| --- | --- | --- | --- | --- |
| `reference` | 要向其混合的目标 latent。其样本会被转换到正在处理的 latent 的设备与 dtype 上,然后调整尺寸并重复以与其匹配。 | LATENT | 是 | - |
| `strength` | 向参考 latent 混合的程度:0 表示 latent 保持不变,1 表示与调整尺寸后的参考 latent 匹配(默认值:1.0)。 | FLOAT | 是 | 0.0 到 1.0 (步长:0.0001) |

## 输出

| 输出名称 | 描述 | 数据类型 |
| --- | --- | --- |
| `operation` | 一种可应用于 latent 样本的混合操作。 | LATENT_OPERATION |

> 本文档由 AI 生成。如果您发现任何错误或有改进建议,欢迎贡献! [在 GitHub 上编辑](https://github.com/Comfy-Org/embedded-docs/blob/main/comfyui_embedded_docs/docs/LatentOperationBlend/zh.md)

---
**Source fingerprint (SHA-256):** `5890089afddf83ddd4edd992606509b118aac9ef13eb89589f73fa75e0b9dd5a`
Loading