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
8 changes: 4 additions & 4 deletions ko/changelog/index.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -108,7 +108,7 @@
<Update label="v0.27.1" description="2026년 7월 8일">

**파트너 노드 업데이트**
* [**Seedream 5 Pro**](https://links.comfy.org/3R12fXw): ByteDance Seedream 5 Pro 이미지 생성 모델 지원

Check warning on line 111 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L111

Did you really mean 'Seedream'?

Check warning on line 111 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L111

Did you really mean 'Seedream'?
* [**Seed Audio 1.0**](https://github.com/Comfy-Org/ComfyUI/pull/14731): ByteDance Seed Audio 1.0 - 음성, 음악, 효과음 및 다중 화자 대화 지원
* [**Ideogram V1/V2 지원 중단**](https://github.com/Comfy-Org/ComfyUI/pull/14712): 레거시 Ideogram V1 및 V2 파트너 노드 지원 중단
* [**StabilityAI 임시 제거**](https://github.com/Comfy-Org/ComfyUI/pull/14737): StabilityAI 파트너 노드 임시 제거
Expand All @@ -121,21 +121,21 @@
<Update label="v0.27.0" description="2026년 6월 30일">

**int8 모델 지원**
이번 릴리스는 **int8 convrot 모델**에 대한 지원을 추가하여, 양자화된 모델에 상당한 성능 향상을 가져옵니다. 이제 int8 지원은 Turing GPU 이상에서 작동하며, 최적화된 LoRA 적용 및 메모리 관리가 포함됩니다.

Check warning on line 124 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L124

Did you really mean 'convrot'?

**새로운 노드**
* [**Seed Node**](https://github.com/Comfy-Org/ComfyUI/pull/14616): 워크플로에서 시드 값을 관리하는 새로운 시드 노드 (CORE-295)
* [**Advanced Krea 2 Model Merging**](https://github.com/Comfy-Org/ComfyUI/pull/14621): Krea 2 모델 병합을 위한 고급 노드

Check warning on line 128 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L128

Did you really mean 'Krea'?

Check warning on line 128 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L128

Did you really mean 'Krea'?
* [**Bounding Box Canvas**](https://github.com/Comfy-Org/ComfyUI/pull/14537): 바운딩 박스 캔버스 노드 및 Ideogram JSON 프롬프트 지원 (CORE-292)
* [**ConditioningMultiply**](https://github.com/Comfy-Org/ComfyUI/pull/14686): 조건화 값을 곱하는 새로운 조건화 노드
* [**GLSL Update**](https://github.com/Comfy-Org/ComfyUI/pull/13195): GLSL 노드가 ANGLE 라이브러리를 사용하도록 업데이트됨 (CORE-162)

**파트너 노드 업데이트**
* [**Nano Banana 2 Lite**](https://github.com/Comfy-Org/ComfyUI/pull/14693): Google의 Nano Banana 2 Lite 모델 지원 추가

Check warning on line 134 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L134

Did you really mean 'Nano'?

Check warning on line 134 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L134

Did you really mean 'Nano'?
* [**Gemini Video Omni**](https://github.com/Comfy-Org/ComfyUI/pull/14695): 비디오 생성을 위한 새로운 Gemini Video Omni 노드 추가

Check warning on line 135 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L135

Did you really mean 'Omni'?

Check warning on line 135 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L135

Did you really mean 'Omni'?

**성능 및 안정성**
* [**int8 Models**](https://github.com/Comfy-Org/ComfyUI/pull/14636): 더 빠른 추론을 위한 int8 convrot 모델 지원 추가

Check warning on line 138 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L138

Did you really mean 'convrot'?
* [**Faster int8**](https://github.com/Comfy-Org/ComfyUI/pull/14641): int8 커널 성능 최적화
* [**Turing GPU Support**](https://github.com/Comfy-Org/ComfyUI/pull/14662): int8 지원이 Turing 아키텍처 GPU로 확장됨
* [**Better int8 LoRA**](https://github.com/Comfy-Org/ComfyUI/pull/14685): 더 나은 품질과 속도로 int8 LoRA 적용 개선
Expand All @@ -143,7 +143,7 @@
* [**Asset Hashing Opt-In**](https://github.com/Comfy-Org/ComfyUI/pull/14663): 에셋 해싱이 `--enable-asset-hashing` 플래그를 통해 옵트인으로 변경됨 (기본적으로 비활성화)

**버그 수정**
* [**int8 LoRA requant**](https://github.com/Comfy-Org/ComfyUI/pull/14650): int8 LoRA에 잘못된 설정 사용 시 낮은 품질의 재양자화 수정

Check warning on line 146 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L146

Did you really mean 'requant'?
* [**int8 mm offloaded**](https://github.com/Comfy-Org/ComfyUI/pull/14669): 오프로드된 LoRA 가중치에서 int8 행렬 곱셈이 건너뛰어지는 문제 수정
* [**Nodes categories**](https://github.com/Comfy-Org/ComfyUI/pull/14674): 노드 카테고리 업데이트 및 재구성
* [**Frontend**](https://github.com/Comfy-Org/ComfyUI/pull/14684): comfyui-frontend-package를 1.45.20으로 업데이트
Expand Down Expand Up @@ -172,18 +172,18 @@
* [**Krea2**](https://links.comfy.org/4eC8714): 폭넓은 미적 및 스타일 범위를 제공하는 Krea의 오픈소스 이미지 모델 — 오픈 가중치 **RAW** / **Turbo** 쌍; RAW에서 LoRA 학습, Turbo에서 8단계 추론 실행
* [**Boogu-Image**](https://github.com/Comfy-Org/ComfyUI/pull/14523): Qwen3-VL-8B 텍스트 인코더가 포함된 텍스트 기반 이미지 생성 모델; [**TextEncodeBooguEdit**](https://github.com/Comfy-Org/ComfyUI/pull/14529) 노드는 네거티브 프롬프트와 선택적 참조 이미지를 통해 이미지 편집 기능을 추가합니다
* [**Qwen3-VL Text Generation**](https://github.com/Comfy-Org/ComfyUI/pull/14298): Qwen3-VL 4B 및 8B 모델을 사용한 이미지 이해가 포함된 텍스트 생성
* [**SCAIL-2 Multireference**](https://github.com/Comfy-Org/ComfyUI/pull/14509): SCAIL-2 캐릭터 교체를 위한 다중 참조 지원

Check warning on line 175 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L175

Did you really mean 'Multireference'?

**새로운 노드**
* [**LTX2 Context Windows**](https://github.com/Comfy-Org/ComfyUI/pull/13325): IC-LoRA 가이드 스플라이싱과 함께 LTX-2 멀티모달 비디오+오디오 모델을 위한 컨텍스트 창 샘플링
* [**Load3DAdvanced**](https://github.com/Comfy-Org/ComfyUI/pull/14316): model_3d, camera_info, model_3d_info, 너비, 높이 출력이 있는 메시 전용 3D 로더

Check warning on line 179 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L179

Did you really mean 'camera_info'?
* [**Save Nodes Output Socket**](https://github.com/Comfy-Org/ComfyUI/pull/13866): 다운스트림 파이프라인 연결을 위해 저장 노드에 출력 소켓이 추가되었습니다

**파트너 노드 업데이트**
* [**Luma Rays 3.2**](https://github.com/Comfy-Org/ComfyUI/pull/14540): Luma Ray 3.2 비디오 생성을 위한 7개의 새로운 노드

Check warning on line 183 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L183

Did you really mean 'Luma'?

Check warning on line 183 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L183

Did you really mean 'Luma'?

**성능 및 안정성**
* [**Krea 2**](https://github.com/Comfy-Org/ComfyUI/pull/14594): 더 정확한 메모리 사용 계수

Check warning on line 186 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L186

Did you really mean 'Krea'?
* [**Dynamic VRAM**](https://github.com/Comfy-Org/ComfyUI/pull/14577): 동적 VRAM을 비활성화하는 대신 네이티브 형식을 시도하도록 사용자에게 프롬프트가 표시됩니다
* [**Anima**](https://github.com/Comfy-Org/ComfyUI/pull/14557): 작은 추론 최적화
* [**Jobs API**](https://github.com/Comfy-Org/ComfyUI/pull/14493): 취소 엔드포인트가 추가되었습니다 (`POST /api/jobs/{job_id}/cancel`, `POST /api/jobs/cancel`)
Expand All @@ -198,7 +198,7 @@
<Update label="v0.25.1" description="2026년 6월 16일">

**파트너 노드 업데이트**
* [**Kling V3-Turbo**](https://github.com/Comfy-Org/ComfyUI/pull/14528): Kling V3-Turbo 모델 지원 추가

Check warning on line 201 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L201

Did you really mean 'Kling'?

Check warning on line 201 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L201

Did you really mean 'Kling'?

</Update>

Expand All @@ -208,7 +208,7 @@
* [**Depth Anything 3**](https://github.com/Comfy-Org/ComfyUI/pull/13853): LiheYoung의 단안 깊이 추정 모델
* [**SCAIL-2**](https://github.com/Comfy-Org/ComfyUI/pull/14373): 향상된 비디오 편집을 위한 캐릭터 교체 모델
* [**Bernini-R**](https://github.com/Comfy-Org/ComfyUI/pull/14216): Wan 비디오 모델 지원
* **Ideogram 4**: 안정성 향상을 위해 fp8 dtype 문제 수정

Check warning on line 211 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L211

Did you really mean 'dtype'?
* **10비트 비디오 지원**: 10비트 비디오 입출력을 위한 코어 노드 지원 추가

**신규 노드**
Expand All @@ -217,21 +217,21 @@
* **개선된 ResolutionSelector** ([#14309](https://github.com/Comfy-Org/ComfyUI/pull/14309)): 더 나은 기본값으로 해상도 선택 기능 향상

**파트너 노드 업데이트**
* **Bria Transparent Video Background + Green Background + Replace Background** 노드

Check warning on line 220 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L220

Did you really mean 'Bria'?
* **Gemini Text Node** ([#14299](https://github.com/Comfy-Org/ComfyUI/pull/14299)): 새로운 LLM 파트너 노드
* **Runway Aleph2** ([#14306](https://github.com/Comfy-Org/ComfyUI/pull/14306)): 새로운 비디오 생성 노드
* **Krea 2 Medium Turbo** 모델 지원

Check warning on line 223 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L223

Did you really mean 'Krea'?
* **Tripo3D Import 3D** 노드
* **Flux Erase**: 시드 입력 추가
* **NanoBanana**: temperature 및 top_p 제어 추가

Check warning on line 226 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L226

Did you really mean 'top_p'?
* **Flux KV 캐시**: 분할 배치 관련 충돌 수정

**성능 및 안정성**
* **Aimdo 0.4.10** + `--reserve-vram` 및 `--vram-headroom` 옵션

Check warning on line 230 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L230

Did you really mean 'Aimdo'?
* 안정성 개선이 포함된 Comfy Aimdo 0.4.9

Check warning on line 231 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L231

Did you really mean 'Aimdo'?
* 미디어 로딩: `--high-ram` 옵션 일반화
* **에셋 API**: 커서 기반 페이지네이션, 수집 시 이미지 크기 추출, WebSocket 메시지 내 에셋 ID 포함
* CUDA 130+에서 항상 CUDA malloc 활성화

Check warning on line 234 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L234

Did you really mean 'malloc'?
* 결정적 결과를 위해 cudnn.benchmark를 false로 강제
* 정렬되지 않은 너비의 이미지/비디오 디코딩에서 홀수 높이 충돌 및 가장자리 번짐 수정
* 정렬되지 않은 너비에서 비결정적 비디오 디코딩 수정
Expand All @@ -243,20 +243,20 @@
<Update label="v0.24.1" description="2026년 6월 3일">

**파트너 노드 업데이트**
* Bria 초록색 배경 노드 추가

Check warning on line 246 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L246

Did you really mean 'Bria'?
* SaveWEBM에 알파 채널 지원이 포함된 Bria 투명 비디오 배경 노드 추가

Check warning on line 247 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L247

Did you really mean 'Bria'?
* Krea 2 Medium Turbo 모델 추가

Check warning on line 248 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L248

Did you really mean 'Krea'?
* Flux Erase 노드에 시드 입력 추가

**버그 수정**
* Seedance 2.0 1080p 첫 번째/마지막 프레임 늘어나는 현상 수정

Check warning on line 252 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L252

Did you really mean 'Seedance'?

</Update>

<Update label="v0.24.0" description="2026년 6월 3일">

**새로운 오픈소스 모델 지원**
* [**Ideogram 4**](https://github.com/Comfy-Org/ComfyUI/pull/14259): Ideogram의 오픈웨이트 텍스트-이미지 모델에 대한 당일 지원. 텍스트 인코더로 Qwen3-VL-8B(13계층 hidden-state 탭)를 사용하는 NextDiT/Lumina2 계열의 싱글 스트림 DiT를 기반으로 합니다. 새로운 **Ideogram4Scheduler**, **DualModelGuider** (별도 모델에서 조건부/비조건부), **CFGOverride** 노드를 포함합니다.

Check warning on line 259 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L259

Did you really mean 'CFGOverride'?
* [**PiD Models**](https://github.com/Comfy-Org/ComfyUI/pull/14240): PiD 컨디셔닝 노드에 SDXL 및 QwenImage 변형을 추가했습니다.
* [**Radiance**](https://github.com/Comfy-Org/ComfyUI/pull/14206): 0이 아닌 순차적 txt_ids가 있는 체크포인트에 대한 지원을 추가했습니다.

Expand All @@ -283,7 +283,7 @@
<Update label="v0.23.0" description="2026년 6월 1일">

**새로운 오픈소스 모델 지원**
* [**Microsoft Lens**](https://github.com/Comfy-Org/ComfyUI/pull/14077): Microsoft의 Lens 텍스트-이미지 모델 지원, GPT-OSS-20B MoE 텍스트 인코더(nvfp4로 변환) 포함. 토크나이저는 safetensors 파일에 직접 내장

Check warning on line 286 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L286

Did you really mean 'safetensors'?
* [**NVIDIA PixelDiT & PiD**](https://github.com/Comfy-Org/ComfyUI/pull/14103): NVIDIA의 PixelDiT 1300M 텍스트-이미지 모델과 PiD 모델군(SDXL, QwenImage 변형은 v0.24.0에 추가) 지원
* [**TripoSplat**](https://github.com/Comfy-Org/ComfyUI/pull/14210): VAST-AI의 이미지-3D 가우시안 스플랫 모델 지원 — 단일 이미지를 DINOv3 + Flux2 VAE로 인코딩 후 flow-matching 디노이저 + 옥트리 가우시안 디코더를 실행해 3D 가우시안 스플랫 생성
* [**MediaPipe Face Detection**](https://github.com/Comfy-Org/ComfyUI/pull/14009): 의존성 없는 Google MediaPipe 얼굴 감지 및 랜드마크 재구현, 가중치는 safetensors로 배포
Expand All @@ -291,7 +291,7 @@
**새로운 노드**
* [**Save Image Advanced**](https://github.com/Comfy-Org/ComfyUI/pull/13850): 파일 출력을 더 세밀하게 제어하는 고급 이미지 저장 노드
* [**Gaussian Splat Nodes**](https://github.com/Comfy-Org/ComfyUI/pull/14190): 외부 의존성 없이 3D 가우시안 스플랫을 다루는 새로운 `GAUSSIAN` 타입과 유틸리티 노드 (`.ply`, `.splat`, `.spz`, `.ksplat`로 직렬화)
* [**Logic Nodes**](https://github.com/Comfy-Org/ComfyUI/pull/14004): 새로운 And, Or, Not 유틸리티 노드 (utils 카테고리로 이동)

Check warning on line 294 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L294

Did you really mean 'utils'?
* [**Preview3DAdvanced**](https://github.com/Comfy-Org/ComfyUI/pull/14175): 고급 3D 미리보기 노드

**성능 개선**
Expand All @@ -307,7 +307,7 @@
* **OpenRouter LLM 노드:** OpenRouter 언어 모델 API를 위한 새 파트너 노드
* **Rodin2.5 노드:** Rodin의 새로운 3D 생성 노드
* **Krea2 이미지 노드:** Krea의 새로운 이미지 생성 파트너 노드
* **Beeble SwitchX 노드:** Beeble의 새로운 조명 재설정/배경 교체 노드

Check warning on line 310 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L310

Did you really mean 'Beeble'?
* **Tripo3D P1 모델:** Tripo3D의 P1 생성 모델을 위한 새 노드
* **Flux 가상 피팅 & 지우기:** 새로운 Flux 기반 가상 피팅 및 객체 지우기 노드
* **SeeDance 2:** 비디오 참조 업로드 개선; 런타임 입력 검증 추가
Expand All @@ -327,7 +327,7 @@
* 프론트엔드 경고 스팸 통합
* `LTXVCropGuides`: 가이드가 동일한 시작 위치를 공유할 때 잘못된 잠재 프레임 수정
* LTX AV 크로스 어텐션 AdaLN 변조 내 타임스텝 전달 수정
* 단일 채널 텐서에서의 Lanczos 차원 전치 수정

Check warning on line 330 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L330

Did you really mean 'Lanczos'?
* 배경 제거 마스크 출력 형태 수정
* LoRA 적용 시 리쉐이핑 수정
* Stable Audio DiT 및 VAE에 대해 Sage attention 비활성화
Expand All @@ -349,7 +349,7 @@
<Update label="v0.22.2" description="2026년 5월 20일">

**파트너 노드 업데이트**
* 품질 메시 옵션이 포함된 Rodin2.5 노드 추가 (usdz 내보내기 비활성화)

Check warning on line 352 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L352

Did you really mean 'usdz'?
* Rodin2.5에서 시드 파라미터가 항상 서버로 전달되던 문제 수정

</Update>
Expand All @@ -370,9 +370,9 @@
* [**Stable Audio 3.0**](https://blog.comfy.org/p/stable-audio-3-day-0-support): 향상된 오디오 생성 기능을 갖춘 최신 Stable Audio 3 모델 지원
* **MoGe:** 강화된 지오메트리 처리를 위한 MoGe 모델 지원 추가
* **HiDream-O1:** 보다 정밀한 이미지 생성 제어를 위한 영역 조건부 지원 추가
* **LTXV 개선사항:** 다운스케일 IC-LoRA 지원 및 선택적 attention_mask 입력 추가로 비디오 제어 개선

Check warning on line 373 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L373

Did you really mean 'attention_mask'?
* **시간적 처리:** 비디오 워크플로우에서 더 나은 시간적 제어를 위해 downscale_ratio_temporal 기능 구현

Check warning on line 374 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L374

Did you really mean 'downscale_ratio_temporal'?
* **LTX2.3 최적화:** guide_mask 사용 시 피크 VRAM 사용량 감소로 보다 효율적인 비디오 생성

Check warning on line 375 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L375

Did you really mean 'guide_mask'?

**파트너 노드**
* **Seed2.0:** ByteDance의 향상된 언어 모델 기능을 위한 새로운 노드 추가
Expand All @@ -385,7 +385,7 @@
* **음수 값 지원:** 특정 노드 파라미터에 음수 값을 허용해 창의적 제어 범위 확장

**기타**
* **모델 샘플링:** 체인된 model_sampling 패치 간에 noise_scale/shift 유지하여 일관된 결과 제공

Check warning on line 388 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L388

Did you really mean 'model_sampling'?
* **텍스트 생성:** Qwen3.5 다중 이미지 프롬프트 처리 개선으로 텍스트 생성 워크플로우 향상
* **동적 CLIP:** 적절한 동적 CLIP 저장 기능 구현
* **BiRefNet:** 배경 제거 성능 향상을 위한 호환성 문제 해결
Expand All @@ -404,16 +404,16 @@

**새로운 파트너 노드**
* **Flux2ImageNode 및 GrokImageEditNodeV2:** 이미지 생성 및 편집 워크플로우를 강화하기 위한 새로운 노드 추가
* **ByteDanceSeedreamNodeV2:** DynamicCombo 및 Autogrow 기능을 갖춘 새 버전으로, 파라미터 처리가 개선되었습니다.

Check warning on line 407 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L407

Did you really mean 'Autogrow'?
* **OpenAI 이미지 노드:** DynamicCombo 및 Autogrow 기능을 갖춘 새로운 통합

Check warning on line 408 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L408

Did you really mean 'Autogrow'?
* **Claude LLM 노드:** 텍스트 생성 워크플로우를 위한 Claude 언어 모델 지원 추가

**모델 지원 개선**
* **Anima TE LoRA:** Kohya 형식의 Anima TE LoRA 지원 추가

Check warning on line 412 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L412

Did you really mean 'Kohya'?
* **HiDream-O1-Image:** dtype 수정 및 비동적 VRAM 구성에 대한 메모리 사용량 개선과 함께 HiDream-O1-Image 모델 완벽 지원

Check warning on line 413 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L413

Did you really mean 'dtype'?

**기술적 개선**
* 모델 패처에서 fp8 형식의 safetensors 저장 문제 해결

Check warning on line 416 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L416

Did you really mean 'safetensors'?
* LTXV 미디어 중간 프레임 가이드 정렬을 수정하여 동영상 처리 향상
* RuntimeError 문제를 해결하여 VOID 모델 호환성 개선
* Save3D 노드를 개선하여 정점 색상 및 텍스처를 지원해 종합적인 3D 워크플로우 지원 가능하게 함
Expand All @@ -438,8 +438,8 @@
* [**Void**](https://github.com/Comfy-Org/ComfyUI/pull/13403): 패스 1 및 패스 2 모델 지원

**파트너 모델 지원**
* [**Tripo 3.1**](https://github.com/Comfy-Org/ComfyUI/pull/13788): 생산 준비 완료 3D 생성

Check warning on line 441 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L441

Did you really mean 'Tripo'?
* **NanoBanana2** 노드 v2: DynamicCombo 및 Autogrow UX 리프레시

Check warning on line 442 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L442

Did you really mean 'Autogrow'?

**노드 개선 사항**
* **수학 표현식:** 보다 복잡한 계산을 위한 부울 지원 추가
Expand All @@ -463,7 +463,7 @@
* 자기회귀 비디오; 인과 모델용 이미지-비디오 변환; 긴 시퀀스용 `causal_window_fix`; 프레임 보간 메모리 및 오버헤드 개선

**더 많은 파트너 통합**
* **Luma UNI-1**, **Topaz Astra 2**, **GPT-5.5** / **GPT-5.5-pro**

Check warning on line 466 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L466

Did you really mean 'Luma'?

Check warning on line 466 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L466

Did you really mean 'Astra'?

**더 나은 노드 수정**
* `--gpu-only`를 이용한 이미지 블렌딩 수정; JoinImageWithAlpha에서 배치 브로드캐스팅; SplitImageToTileList 및 ImageMergeTileList에서 타일 스트라이드 수정
Expand All @@ -476,28 +476,28 @@
<Update label="v0.20.3" description="2026년 4월 27일">

**파트너 노드 업데이트**
* Luma UNI-1 모델 추가

Check warning on line 479 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L479

Did you really mean 'Luma'?
* GPT 5.5 및 5.5-pro LLM 모델 추가
* grok-imagine-image-quality 모델 추가
* DynamicCombo/Autogrow를 지원하는 새로운 NanoBanana2 노드

**버그 수정**
* ImageBlend/ImageCompositeMasked에서 채널 수가 다른 이미지(알파 채널) 처리 문제 수정
* Motion Control 노드 내 Kling V3 가격 배지 수정

Check warning on line 486 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L486

Did you really mean 'Kling'?

</Update>

<Update label="v0.20.2" description="2026년 4월 27일">

**파트너 노드 업데이트**
* Topaz Astra 2 모델 추가 (이제 기본 업스케일러)

Check warning on line 493 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L493

Did you really mean 'Astra'?
* 일반 이미지용 ByteDance 가상 인물 라이브러리
* GPTImage2 노드에 사용자 지정 해상도 지원 추가
* Moonvalley API 노드 제거

Check warning on line 496 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L496

Did you really mean 'Moonvalley'?
* 파트너 API 노드 작업의 기본 타임아웃 증가

**모델 지원**
* SDPose 크기 조정 수정

Check warning on line 500 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L500

Did you really mean 'SDPose'?
* OneTainer ERNIE LoRA 지원

</Update>
Expand All @@ -511,8 +511,8 @@
* **LTXV 오디오 VAE:** 개선된 기본 통합 기능을 갖춘 독립형 LTXV 오디오 VAE 모델 지원

**파트너 노드 향상**
* **비디오 생성:** ByteDance 2, Veo 모델 및 Kling 노드에 4K 해상도 지원 추가

Check warning on line 514 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L514

Did you really mean 'Veo'?

Check warning on line 514 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L514

Did you really mean 'Kling'?
* **Veo 3 Lite:** 더 빠른 비디오 생성을 위한 Veo 모델의 새로운 경량 버전

Check warning on line 515 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L515

Did you really mean 'Veo'?

Check warning on line 515 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L515

Did you really mean 'Veo'?
* **SD2 리얼 휴먼:** 스테이블 디퓨전 2 리얼 휴먼 생성 지원 추가
* **GPT-Image 2:** 추가 해상도 지원과 고정 가격 배지가 포함된 새 버전 옵션
* **HappyHorse:** 이미지 생성 능력을 강화한 새로운 모델 통합
Expand All @@ -529,7 +529,7 @@
* ByteDance 2 노드의 자동 다운스케일링 기능 추가로 처리 최적화

**기술 업데이트**
* 데이터베이스 호환성을 높이기 위해 SQLAlchemy 버전 고정(>=2.0) 수정

Check warning on line 532 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L532

Did you really mean 'SQLAlchemy'?
* 디버깅을 더 잘하기 위한 ComfyUI 앱 파일 내 로깅 기능 강화
* 포괄적인 API 문서를 위한 OpenAPI 3.1 사양 추가
* 텍스처 크기를 이용한 흐림/선명 효과 처리를 위한 셰이더 처리 개선
Expand Down Expand Up @@ -560,9 +560,9 @@
* ByteDance 2 노드의 자동 다운스케일링 기능으로 대용량 동영상 파일 처리 가능

**파트너 노드 업데이트**
* **Veo 모델:** 4K 해상도 지원 및 Veo 3 Lite 모델 추가

Check warning on line 563 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L563

Did you really mean 'Veo'?

Check warning on line 563 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L563

Did you really mean 'Veo'?
* **Kling 노드:** 4K 해상도 기능 강화

Check warning on line 564 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L564

Did you really mean 'Kling'?
* **GPTImage:** gpt-image-2 버전 옵션 및 새로운 해상도 지원 추가

Check warning on line 565 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L565

Did you really mean 'GPTImage'?
* **HappyHorse 모델:** 새로운 파트너 모델 통합
* **SD2 실제 인간 지원:** 사실적인 인간 생성 기능 강화

Expand All @@ -572,8 +572,8 @@
* 모든 노드의 미리보기를 개선하여 디버깅을 위한 더 많은 텐서 값 표시

**버그 수정**
* SQLAlchemy 호환성 문제 해결: 버전 ≥2.0 고정

Check warning on line 575 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L575

Did you really mean 'SQLAlchemy'?
* Stable_Zero123 cc_projection 가중치의 매개변수 할당 문제 해결

Check warning on line 576 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L576

Did you really mean 'cc_projection'?
* LTXV 참조 오디오 노드 기능 수정
* veo-3.0 모델의 4K 해상도 제한 문제 해결

Expand All @@ -583,27 +583,27 @@

**파트너 노드 업데이트**
* Stable Diffusion 2용 SD2 Real Human 지원 추가
* Kling 노드에 4K 해상도 추가

Check warning on line 586 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L586

Did you really mean 'Kling'?
* GPTImage 가격 배지 수정 및 새 해상도 추가

Check warning on line 587 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L587

Did you really mean 'GPTImage'?

**버그 수정**
* veo-3.0 모델의 4K 해상도 제한 수정
* Windows(aimdo 활성화)에서 Stable_Zero123 cc_projection 가중치 할당 수정

Check warning on line 591 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L591

Did you really mean 'cc_projection'?

</Update>

<Update label="v0.19.4" description="2026년 4월 17일">

**파트너 노드 업데이트**
* Veo 모델의 4K 해상도 지원 및 Veo 3 Lite 모델 추가

Check warning on line 598 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L598

Did you really mean 'Veo'?

Check warning on line 598 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L598

Did you really mean 'Veo'?
* gpt-image-2를 버전 옵션으로 추가
* ByteDance 2 노드에 자동 비디오 다운스케일링 추가
* SD2 Real Human 지원 추가

**버그 수정**
* Windows에서 Stable_Zero123 cc_projection 가중치 할당 수정 (aimdo 활성화 시)

Check warning on line 604 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L604

Did you really mean 'cc_projection'?

Check warning on line 604 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L604

Did you really mean 'aimdo'?
* veo-3.0 모델의 4K 해상도 제한 수정
* GPTImage 가격 배지 수정

Check warning on line 606 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L606

Did you really mean 'GPTImage'?

</Update>

Expand Down Expand Up @@ -639,8 +639,8 @@
<Update label="v0.19.1" description="2026년 4월 16일">

**새로운 모델 지원**
* [Erine 이미지 텍스트-to-이미지](https://blog.comfy.org/p/ernie-image-day-0-support)
* [Erine 이미지 텍스트 기반 이미지 생성](https://blog.comfy.org/p/ernie-image-day-0-support)
* LLM: Ministral 모델

Check warning on line 643 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L643

Did you really mean 'Ministral'?

**파트너 노드**
* [Sonilo: 동영상에 맞는 음악 만들기](https://blog.comfy.org/p/comfyui-now-supports-sonilo-via-partner)
Expand All @@ -656,7 +656,7 @@
- LTX2 참조 오디오(ID-LoRA) 기능 지원 추가
- TextGenerate 노드를 위한 향상된 8B 모델 호환성을 갖춘 Qwen3.5 텍스트 생성 모델 구현
- 개선된 객체 탐지 기능을 제공하는 RT-DETRv4 탐지 모델 지원 추가
- 중간 dtype 호환성을 갖춘 Ace Step 1.5 XL 모델 지원 도입

Check warning on line 659 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L659

Did you really mean 'dtype'?
- 소형 flux.2 디코더 지원 추가 및 풀링 출력 없이 Flux 조정 가능하게 설정
- 확장된 생성 옵션을 위한 Ernie Image 모델 구현

Expand All @@ -666,7 +666,7 @@
- 이미지 분석 기능을 제공하는 Image Histogram 노드 추가
- 기본값이 0인 Color Adjustment 노드 구현
- 곡선 입력과 상향된 유니폼 한계를 갖춘 GLSL 셰이더 노드 강화
- 대화형 UI 요소를 가진 노드에 대해 has_intermediate_output 플래그 추가

Check warning on line 669 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L669

Did you really mean 'has_intermediate_output'?

**메모리 및 성능**
- 모델 RAM 관리를 통한 RAM 캐시 통합으로 효율성 향상
Expand All @@ -682,7 +682,7 @@
- 개선된 기능을 갖춘 Tencent3D 노드 업데이트

**버그 수정 및 안정성**
- training_dtype "none" 및 bfloat16 LoRA 가중치와 함께 발생하는 Train LoRA 크래시 수정

Check warning on line 685 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L685

Did you really mean 'training_dtype'?
- LTXAV 모델과의 텍스트 생성 회귀 문제 해결
- FP8 스케일링 체크포인트 호환성 문제 수정
- 보안성이 낮은 브라우저에 대한 브라우저 보안 처리 개선
Expand Down Expand Up @@ -755,15 +755,15 @@
**모델 및 하드웨어 지원**
- 향상된 모델 호환성을 위한 mxfp8 정밀도 형식 지원 추가
- AMD gfx1150(Strix Point) GPU용 PyTorch Attention 활성화
- FP4, 8, 16 기본 dtype 지원 및 양자화된 선형 자동차분 함수로 훈련 기능 향상

Check warning on line 758 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L758

Did you really mean 'dtype'?
- 텍스트 인코더 호환성 문제 수정 및 Hunyuan3D v2.1 DiT 모델의 SageAttention 비활성화

**노드 및 API 개선**
- Tencent TextToModel 및 ImageToModel API 노드 수정

Check warning on line 762 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L762

Did you really mean 'Tencent'?
- Nano Banana 2 'thought_image' 지원 및 Quiver SVG 노드 추가

Check warning on line 763 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L763

Did you really mean 'Nano'?

Check warning on line 763 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L763

Did you really mean 'thought_image'?
- seedream-3-0-t2i 및 seedance-1-0-lite 모델을 오래된 버전으로 표시
- slice_cond 및 모델별 컨텍스트 윈도우 조정 기능 추가

Check warning on line 765 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L765

Did you really mean 'slice_cond'?
- EmptyLatentImage 및 EmptyImage 노드를 중간 dtype 설정에 맞춰 개선

Check warning on line 766 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L766

Did you really mean 'dtype'?

**프론트엔드 및 인프라**
- 프론트엔드 패키지를 1.41.21 버전으로 업데이트, 캐싱 헤더 개선으로 오래된 청크 방지
Expand All @@ -774,7 +774,7 @@

**버그 수정 및 안정성**
- 픽셀 공간 VAE 호환성 문제 및 종료 시 모델 파이널라이저 실행 문제 수정
- Load Diffusion Model 노드에서 가중치 dtype 처리 개선(고급 입력으로 표시)

Check warning on line 777 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L777

Did you really mean 'dtype'?
- 서브클래스화된 모델의 지연 가중치 초기화 강화
- 드문 특수 상황에서의 다양한 메모리 누수 및 손상 문제 수정

Expand Down Expand Up @@ -803,13 +803,13 @@
**모델 지원 및 개선**
- FluxKVCache 노드를 통해 Flux 2 Klein KV 캐시 모델 지원을 추가했습니다.
- Flux 모델용 정리 함수와 사전 주의 패치를 포함한 모델 패칭 시스템을 개선했습니다.
- Qwen 이미지 모델에 사전 주의 및 입력 후 패치를 추가했습니다.

Check warning on line 806 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L806

Did you really mean 'Qwen'?
- 래핑된 모델의 텍스트 인코더 LoRA 로딩 문제를 수정했습니다.
- 다양한 모델에서 batch_size > 1 문제를 해결하여 배치 처리를 개선했습니다.

Check warning on line 808 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L808

Did you really mean 'batch_size'?

**새로운 노드 및 기능**
- 이미지 편집 기능을 강화한 Painter 노드를 추가했습니다.
- 확장된 기능을 제공하는 Reve Image API 노드를 도입했습니다.

Check warning on line 812 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L812

Did you really mean 'Reve'?
- 워크플로우 템플릿을 버전 0.9.21로 업데이트했습니다.
- 그래디언트 스톱 포맷을 개선하여 프론트엔드와의 호환성을 높였습니다.

Expand All @@ -828,7 +828,7 @@
<Update label="v0.16.4" description="2026년 3월 7일">

**새로운 노드 및 기능**
- 워크플로우 내 수학적 연산을 위한 simpleeval 평가를 지원하는 수학 표현식 노드 추가

Check warning on line 831 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L831

Did you really mean 'simpleeval'?
- 토폴로지 처리 기능을 강화한 TencentSmartTopology API 노드 도입
- LLM 노드에 Gemini 3.1 Flash-Lite 모델 지원 추가, AI 언어 모델 옵션 확장

Expand All @@ -847,8 +847,8 @@

**LTX 오디오 모델 개선 사항**
- LTX2 보코더의 conv_transpose1d를 수동 캐스팅으로 수정하여 오디오 처리 안정성을 향상했습니다.
- 오디오 생성 시 더 나은 메모리 관리를 위해 LTX 오디오 VAE의 novram 호환성 문제를 해결했습니다.

Check warning on line 850 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L850

Did you really mean 'novram'?
- causal_fix 매개변수를 add_keyframe_index 및 append_keyframe 함수에 추가하여 비디오 워크플로우에서 키프레임 처리 정밀도를 높였습니다.

Check warning on line 851 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L851

Did you really mean 'causal_fix'?

Check warning on line 851 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L851

Did you really mean 'add_keyframe_index'?

Check warning on line 851 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L851

Did you really mean 'append_keyframe'?

</Update>

Expand Down Expand Up @@ -884,7 +884,7 @@
- 이미지 처리 기능을 강화한 네이티브 LongCat-Image 구현

**모델 및 형식 지원**
- LoKR 호환성을 위해 ACE-Step 1.5 lycoris 키 별칭 매핑을 추가했습니다.

Check warning on line 887 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L887

Did you really mean 'lycoris'?
- SDPose-OOD 모델 지원
- SCAIL WanVideo 모델 지원
- zeta 크로마 가중치 로딩 지원
Expand All @@ -900,13 +900,13 @@

**API 및 UI 개선**
- NanoBanana2 API 노드에 "IMAGE+TEXT" 지원을 추가했습니다.
- Mahiro CFG를 명확성을 위해 Similarity-Adaptive Guidance로 이름을 변경했습니다.

Check warning on line 903 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L903

Did you really mean 'Mahiro'?
- GLSL 노드 입력 치수 요구사항을 수정했습니다.
- VAEDecodeAudioTiled를 개선하여 tile_size 입력을 올바르게 사용하도록 했습니다.

Check warning on line 905 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L905

Did you really mean 'tile_size'?

**버그 수정**
- feat_map 처리 시 WanVAE 인코더/디코더 혼동 문제를 수정했습니다.

Check warning on line 908 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L908

Did you really mean 'feat_map'?
- lm_metadata 누락으로 인한 ACE-1.5 메모리 추정 문제를 해결했습니다.

Check warning on line 909 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L909

Did you really mean 'lm_metadata'?
- 서브스텝 시그마의 컨텍스트 윈도우 단계 처리를 개선했습니다.
- BytesIO에 쓰는 동안 VideoFromComponents 충돌을 방지했습니다.
- 이미지 타일 오버랩을 제한하여 처리 문제를 예방했습니다.
Expand All @@ -919,8 +919,8 @@
<Update label="v0.15.1" description="2026년 2월 26일">

**버그 수정 및 안정성**
- 프롬프트 항목에 class_type 키가 없을 때 발생하는 KeyError를 수정하여 워크플로우 충돌 방지

Check warning on line 922 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L922

Did you really mean 'class_type'?
- pyopengl, accelerate, numpy 텍스처와의 GLSL 노드 호환성 문제 해결

Check warning on line 923 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L923

Did you really mean 'pyopengl'?

Check warning on line 923 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L923

Did you really mean 'numpy'?
- LTXAV 텍스트 인코더의 최소 길이 및 메모리 추정 문제 수정
- torch 컴파일러 사용 시 동적 VRAM 비활성화로 충돌 예방

Expand All @@ -928,7 +928,7 @@
- 향상된 백업 처리 기능을 갖춘 comfy aimdo를 버전 0.2.2로 업데이트
- WanVideo 기반 세분화 모델 FlowRVS 지원 추가로 고급 영상 처리 가능
- 확장된 기능을 제공하는 NanoBanana2 API 노드 추가
- Aimdo 백업 처리를 수정하여 제로 복사 SFT 문제 방지

Check warning on line 931 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L931

Did you really mean 'Aimdo'?

**UI 및 프론트엔드**
- 프론트엔드를 버전 1.39.19로 패치하여 API 노드의 진행률 텍스트 직렬화 문제 해결
Expand All @@ -943,9 +943,9 @@
<Update label="v0.15.0" description="2026년 2월 24일">

**새로운 노드 및 기능**
- 이미지 워크플로우에서 정밀한 바운딩 박스 선택을 위한 BBox 위젯 추가

Check warning on line 946 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L946

Did you really mean 'BBox'?
- 오디오 처리 워크플로우를 위한 3밴드 이퀄라이저 노드 도입
- 기본 텍스트 생성 지원을 내장 모델과 함께 추가, 초기 지원 모델은 Gemma3와 Qwen 3임

Check warning on line 948 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L948

Did you really mean 'Qwen'?
- 고급 시각 효과를 위해 PyOpenGL을 사용하는 새로운 GLSL 셰이더 노드
- 텍스트 음성 변환 기능을 위한 ElevenLabs API 노드 추가
- FLOAT 입력에 대한 새로운 그라디언트 슬라이더 표시 모드 및 개선된 UI 제공
Expand All @@ -959,14 +959,14 @@
- Gemini가 더 나은 품질의 압축 해제 이미지를 반환하도록 강제함

**성능 및 버그 수정**
- requants 반환 제한으로 인해 FP8 동적 VRAM 워크플로우 성능 문제 해결

Check warning on line 962 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L962

Did you really mean 'requants'?
- 비디오 저장 작업 중 비연속 오디오 파형 충돌 문제 해결
- LTXAV 모델 로딩 문제 해결, fp8 체크포인트 지원 및 임베딩 커넥터 dtype 문제 포함

Check warning on line 964 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L964

Did you really mean 'dtype'?
- PyOpenGL 버전 < 3.1.4에 대한 호환성 개선

**UI 개선**
- 429개 위젯을 접을 수 있는 고급 UI로 표시
- 노드의 더 나은 구성 위해 essentials_category 추가

Check warning on line 969 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L969

Did you really mean 'essentials_category'?
- 개선된 분류를 통해 필수 하위 그래프 블루프린트 강화
- 프론트엔드를 버전 1.39.16으로 업데이트

Expand All @@ -990,7 +990,7 @@

**새로운 모델 지원**
- API 노드에 'viduq3-turbo' 모델 지원 추가
- 향상된 크리에이티브 워크플로우를 위한 Recraft V4 노드 추가

Check warning on line 993 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L993

Did you really mean 'Recraft'?

**업데이트**
- 워크플로우 템플릿을 v0.8.43으로 업데이트함
Expand All @@ -1001,7 +1001,7 @@
<Update label="v0.14.0" description="2026년 2월 17일">

**새로운 오픈소스 모델 지원**
* [**Qwen 2512 ControlNet / Fun ControlNet**](https://github.com/Comfy-Org/ComfyUI/pull/12359): 향상된 이미지 제어를 위한 Qwen 기반 ControlNet 모델 지원

Check warning on line 1004 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1004

Did you really mean 'Qwen'?

Check warning on line 1004 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1004

Did you really mean 'Qwen'?
* **NAG 구현**: 모든 Flux 기반 모델에 적용되는 네스테로프 가속 그래디언트

**새로운 노드 및 API**
Expand All @@ -1011,17 +1011,17 @@

**성능 및 기술 개선 사항**
* 왼쪽 패딩 텍스트 인코더 어텐션 마스크 지원
* Flux 모델을 위한 Torch RMSNorm + Hunyuan 비디오 리팩터링

Check warning on line 1014 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1014

Did you really mean 'RMSNorm'?

Check warning on line 1014 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1014

Did you really mean 'Hunyuan'?
* 동적 VRAM: 학습 수정 사항, vanilla-fp8 LoRA 품질 수정
* ModelPatcherDynamic: 비단말 가중치 강제 로드
* 오프로드 및 동적 VRAM 모드에서 LoRA 추출 수정
* Flux 모델에 torch RMSNorm 사용 및 Hunyuan 비디오 리팩터링

Check warning on line 1018 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1018

Did you really mean 'RMSNorm'?

Check warning on line 1018 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1018

Did you really mean 'Hunyuan'?

**파트너 노드 업데이트**
* **Magnific Upscalers** API 노드 활성화

Check warning on line 1021 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1021

Did you really mean 'Upscalers'?
* **Tencent 3D 노드**: ModelTo3DUV, 3DTextureEdit, 3DParts

Check warning on line 1022 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1022

Did you really mean 'Tencent'?
* **Bria RMBG** API 노드

Check warning on line 1023 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1023

Did you really mean 'Bria'?
* **Kling API 노드**: V3, O3 모델

Check warning on line 1024 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1024

Did you really mean 'Kling'?

**트레이너 개선 사항**
* 적절한 오프로딩을 통한 학습 (KohakuBlueleaf)
Expand All @@ -1044,8 +1044,8 @@
* 프런트엔드 1.38.13으로 업데이트됨

**파트너 노드 업데이트**
* **Kling API 노드**: 새로운 V3, O3 모델

Check warning on line 1047 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1047

Did you really mean 'Kling'?
* Moonvalley API 노드 수정

Check warning on line 1048 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1048

Did you really mean 'Moonvalley'?

</Update>

Expand Down Expand Up @@ -1097,8 +1097,8 @@

**API 노드**
* **HitPaw API 노드** ([#12117](https://github.com/Comfy-Org/ComfyUI/pull/12117))
* **Recraft** 스타일 노드

Check warning on line 1100 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1100

Did you really mean 'Recraft'?
* **Vidu**: Q3 모델, Extend 및 MultiFrame 노드

Check warning on line 1101 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1101

Did you really mean 'Vidu'?

**개발자**
* 에셋 파트 2 - 더 많은 엔드포인트
Expand All @@ -1125,15 +1125,15 @@
* [**Multi/InfiniteTalk**](https://github.com/Comfy-Org/ComfyUI/pull/10179): LTXV를 위한 대규모 다중 화자 지원
* [**Flux2 LyCORIS LoKr**](https://github.com/Comfy-Org/ComfyUI/pull/11997): Flux2용 LoKr(KronA 기반) LoRA 지원
* **LTX2 Tiny VAE**(taeltx_2) 지원
* **Z-Image Omni** 베이스 모델 지원

Check warning on line 1128 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1128

Did you really mean 'Omni'?
* **Flux2 Klein**용 ModelScope LoRA 형식

**새로운 노드 및 기능**
* **Magnific API 노드** ([#11986](https://github.com/Comfy-Org/ComfyUI/pull/11986)): Magnific을 통한 업스케일링
* **Tencent Hunyuan3D API 노드** ([#12026](https://github.com/Comfy-Org/ComfyUI/pull/12026))

Check warning on line 1133 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1133

Did you really mean 'Tencent'?
* **WaveSpeed API 노드** ([#11945](https://github.com/Comfy-Org/ComfyUI/pull/11945))
* 노드 검색 가능성 향상을 위한 노드 스키마의 **별칭 검색** 필드
* 프론트엔드에서 임의 입력을 위한 **kwargs 입력**

Check warning on line 1136 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1136

Did you really mean 'kwargs'?
* 작은 그리드 IC-LoRA를 위한 **RoPE 위치 조정**

**성능**
Expand All @@ -1143,7 +1143,7 @@

**모델 개선 사항**
* SaveCheckpoint로 저장된 Flux2 Klein 체크포인트가 이제 올바르게 로드됩니다.
* Qwen 3 0.6B 모델용 설정

Check warning on line 1146 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1146

Did you really mean 'Qwen'?
* 크로마 래디언스 패치 크기 동적 감지

</Update>
Expand All @@ -1156,9 +1156,9 @@
* **CLIP 비전 출력에 이미지 크기 추가** ([#11923](https://github.com/Comfy-Org/ComfyUI/pull/11923))

**파트너 노드 업데이트**
* **ByteDance Seedance 1.5 Pro** API 노드

Check warning on line 1159 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1159

Did you really mean 'Seedance'?
* **Bria Edit** 노드 ([#11978](https://github.com/Comfy-Org/ComfyUI/pull/11978))

Check warning on line 1160 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1160

Did you really mean 'Bria'?
* Autogrow 검증 개선

Check warning on line 1161 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1161

Did you really mean 'Autogrow'?

**버그 수정**
* LTXV VAE 인코딩을 위한 모노 오디오를 가짜 스테레오로 변환 ([#11965](https://github.com/Comfy-Org/ComfyUI/pull/11965))
Expand All @@ -1174,15 +1174,15 @@
* [**Z-Image ControlNet Lite**](https://github.com/Comfy-Org/ComfyUI/pull/11849): Alibaba-Pai Z-Image ControlNet의 경량화 버전

**새로운 노드**
* **Meshy 3D API 노드** ([#11843](https://github.com/Comfy-Org/ComfyUI/pull/11843))

Check warning on line 1177 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1177

Did you really mean 'Meshy'?
* **ResizeImageMaskNode**: crop-to-multiple 모드 추가
* **블루프린트 디렉토리** (내장 템플릿용)

**성능 및 기술**
* NVFP4 LoRA 최적화 (여러 차례)
* ProgressBar WebSocket 업데이트 제한
* Lanczos 그레이스케일 업스케일링 수정

Check warning on line 1184 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1184

Did you really mean 'Lanczos'?
* VAELoader 메타데이터 로딩

Check warning on line 1185 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1185

Did you really mean 'VAELoader'?
* 저장소 URL이 Comfy-Org/ComfyUI로 업데이트됨

</Update>
Expand All @@ -1199,7 +1199,7 @@
**새로운 기능**
* [**기본 에셋 지원**](https://github.com/Comfy-Org/ComfyUI/pull/11315): 모델 에셋 관리를 위한 기반 — 모델 CRUD 연산 엔드포인트
* **NVFP4 모델 LoRA** ([#11837](https://github.com/Comfy-Org/ComfyUI/pull/11837)): LoRA 지원이 이제 NVFP4 양자화 모델에서 작동합니다
* [**SigLIP 2 NAFlex**](https://github.com/Comfy-Org/ComfyUI/pull/11831): SigLIP 2 NAFlex CLIP 비전 모델 지원

Check warning on line 1202 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1202

Did you really mean 'NAFlex'?

Check warning on line 1202 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1202

Did you really mean 'NAFlex'?
* [**ImageCompare 노드**](https://github.com/Comfy-Org/ComfyUI/pull/11343): 두 이미지를 시각적으로 비교합니다
* [**JoinAudioChannels**](https://github.com/Comfy-Org/ComfyUI/pull/11728): 오디오 채널 결합 유틸리티

Expand All @@ -1207,7 +1207,7 @@
* Z-Image 모델을 위한 ModelScope LoRA ([#11805](https://github.com/Comfy-Org/ComfyUI/pull/11805))
* Chroma 및 텍스트 인코더에 대한 FP4/FP8 수정
* VAEEncodeForInpaint가 WAN VAE 튜플 downscale_ratio를 지원합니다
* 효율적인 timestep embed 처리를 통한 LTX2 VRAM 감소

Check warning on line 1210 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1210

Did you really mean 'timestep'?

**API 노드**
* **Vidu2** 노드 ([#11760](https://github.com/Comfy-Org/ComfyUI/pull/11760))
Expand Down Expand Up @@ -1249,21 +1249,21 @@
- CLI 인수를 통한 Sage Attention 3 지원 추가

**API 개선 사항**
- V3 API: DynamicCombo 및 Autogrow 기능 노출

Check warning on line 1252 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1252

Did you really mean 'Autogrow'?
- API 노드: Kling Omni 720p 해상도 지원, WAN2.6 ReferenceToVideo

Check warning on line 1253 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1253

Did you really mean 'Kling'?

Check warning on line 1253 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1253

Did you really mean 'Omni'?
- Tripo3D: face_limit 파라미터 처리 최적화

Check warning on line 1254 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1254

Did you really mean 'face_limit'?
- Vidu API 서명된 URL의 URL 인코딩 보존 문제 해결

Check warning on line 1255 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1255

Did you really mean 'Vidu'?

**버그 수정**
- 업스케일 모델 CPU 오프로딩 문제 해결
- LTXV2 텍스트 인코더 lowvram 호환성 문제 해결

Check warning on line 1259 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1259

Did you really mean 'lowvram'?
- fp8 처리 관련 문제 해결
- MPO 형식 이미지에서 첫 프레임만 사용하도록 수정
- PyTorch 버전에 따른 조건부 comfy-kitchen CUDA 지원

**기타**
- 디버깅을 위한 OOM 메모리 요약 추가
- Mahiro CFG 표시 이름 업데이트

Check warning on line 1266 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1266

Did you really mean 'Mahiro'?
- 워크플로우 템플릿 v0.7.67로 업데이트
- 코드 리팩토링 및 정리 (CLIP 전처리, comfy-kitchen 통합)
</Update>
Expand All @@ -1285,13 +1285,13 @@

**새로운 노드**
- 샘플링 제어를 위한 ManualSigmas 노드 추가
- Kling 모션 제어 노드 추가

Check warning on line 1288 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1288

Did you really mean 'Kling'?
- ResizeByLongerSide가 이제 동영상 처리를 지원함

**API 노드**
- Gemini API: 프롬프트 향상 강제 활성화
- 청구 방식을 달러에서 크레딧으로 변경
- "seededit"는 더 이상 사용되지 않으며, Seedream 노드의 표시 이름을 업데이트함

Check warning on line 1294 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1294

Did you really mean 'Seedream'?

**기술적 사항**
- 이미지 처리 노드를 V3 스키마로 변환
Expand All @@ -1301,7 +1301,7 @@
<Update label="v0.6.0" description="2025년 12월 24일">

**새로운 모델 지원**
- Qwen 이미지 레이어드

Check warning on line 1304 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1304

Did you really mean 'Qwen'?
- NewBie 이미지 Exp0.1

**새로운 기능**
Expand All @@ -1311,7 +1311,7 @@
- API 노드에서 기본적으로 워터마크 생성 비활성화
- 더 나은 훈련 효율성을 위해 해상도 버킷팅을 적용한 트레이너 재작업
- 향상된 AMD GPU 지원
- Kling O1 StartEndFrame 노드의 지속 시간 범위 확장

Check warning on line 1314 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1314

Did you really mean 'Kling'?
- Topaz 4k 영상 업스케일링 기능 업데이트

**버그 수정**
Expand All @@ -1331,11 +1331,11 @@

**새로운 모델 및 기능**
- 고급 모션 처리를 위한 WanMove 모델
- 자동 감지 기능이 포함된 Qwen Edit 2511

Check warning on line 1334 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1334

Did you really mean 'Qwen'?
- 인페인팅 기능이 추가된 Z-Image Fun Control Union 2.0

**API 노드**
- Kling Omni Image, TextToVideoWithAudio, ImageToVideoWithAudio

Check warning on line 1338 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1338

Did you really mean 'Kling'?

Check warning on line 1338 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1338

Did you really mean 'Omni'?
- Wan2.6 모델 통합
- 3D 모델 생성을 위한 Tripo3.0

Expand All @@ -1349,15 +1349,15 @@
<Update label="v0.4.0" description="2025년 12월 10일">

**모델 지원**
- Ovis 이미지

Check warning on line 1352 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1352

Did you really mean 'Ovis'?
- Kandinsky 5.0 (T2V/I2V/T2I 변형)
- Z-Image Alibaba PAI-Fun ControlNet 지원

Check warning on line 1354 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1354

Did you really mean 'Alibaba'?
- 개선된 Z-Image FP16 호환성
- CORS 헤더에 PATCH 메서드 및 크로마-라디언스-x0 모드 추가

**워크플로우 및 개선 사항**
- 컨텍스트 윈도우 수정 및 시간적 처리 강화
- Kling API @image 참조 형식

Check warning on line 1360 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1360

Did you really mean 'Kling'?
- ComfyUI-Manager pip 설치 지원

**프론트엔드 개선 사항**
Expand All @@ -1378,12 +1378,12 @@

**V3 스키마 마이그레이션**
- 3D, 오디오, 프리런치, 마스크 노드를 V3 스키마로 변환
- MatchType, DynamicCombo, Autogrow 지원 추가로 워크플로우 유연성 향상

Check warning on line 1381 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1381

Did you really mean 'Autogrow'?

**버그 수정 및 호환성**
- GPU에 직접 로드 시 텍스트 인코더 회귀 문제 해결
- Lumina 모델에서 transformer_options 클리어링 문제 해결

Check warning on line 1385 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1385

Did you really mean 'transformer_options'?
- Qwen 이미지 LoRA 학습, HunyuanVideo 1.5 meanflow distil 문제 해결

Check warning on line 1386 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1386

Did you really mean 'Qwen'?

Check warning on line 1386 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1386

Did you really mean 'meanflow'?
- EmptyAudio 노드 입력 유형을 정확히 조정해 오디오 워크플로우 통합 개선
- 원치 않는 전체 오프로딩을 방지하기 위해 VRAM 계산 정확도 개선
- Z-Image 워크플로우에서 "transformer." LoRA 접두사 지원
Expand Down Expand Up @@ -1420,8 +1420,8 @@

**파트너 노드**
- Veo3 첫 번째와 마지막 프레임 노드
- Kling FirstLastFrame 노드에 Kling v2.5 터보 추가

Check warning on line 1423 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1423

Did you really mean 'Kling'?

Check warning on line 1423 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1423

Did you really mean 'Kling'?
- Kling O1 모델 지원

Check warning on line 1424 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1424

Did you really mean 'Kling'?

</Update>

Expand Down Expand Up @@ -1479,8 +1479,8 @@

**API 노드 확장**
- **새로운 Topaz API 노드**: ComfyUI 내부에서 직접 Topaz 동영상 향상 워크플로우를 지원하도록 추가되었습니다.
- **Nano Banana Pro 통합**: 처리 능력을 강화한 Nano Banana Pro를 포함해 API 노드 컬렉션을 확장했습니다.

Check warning on line 1482 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1482

Did you really mean 'Nano'?

Check warning on line 1482 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1482

Did you really mean 'Nano'?
- **Kling Lip Sync 개선**: KlingLipSyncAudioToVideoNode에서 오디오 형식 변환 문제를 수정하여 MP3 형식 처리가 올바르게 이루어지도록 했습니다.

Check warning on line 1483 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1483

Did you really mean 'Kling'?

**이미지 처리 및 워크플로우 개선**
- **향상된 이미지 배치 처리**: ImageBatch 노드를 수정하여 채널 수가 다른 이미지를 처리하고, 필요 시 자동으로 알파 채널을 추가하도록 개선했습니다.
Expand Down Expand Up @@ -1509,7 +1509,7 @@
- 업데이트된 PR 템플릿과 Python 3.10 최소 버전 요구사항을 포함한 API 노드 개발 인프라를 개선했습니다.

**개발 및 인프라**
- 커스텀 노드 개발 시 코드 품질을 높이기 위해 pylint 구성 강화했습니다.

Check warning on line 1512 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1512

Did you really mean 'pylint'?
- 보다 안정적인 업데이트를 위해 릴리스 자동화 및 배포 프로세스를 개선했습니다.

</Update>
Expand All @@ -1518,7 +1518,7 @@

**메모리 및 성능 최적화**
- NVIDIA 및 AMD GPU용 **기본적으로 고정된 메모리 활성화**
- Flux, Qwen, LTX-Video 모델의 **VRAM 사용량 감소**

Check warning on line 1521 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1521

Did you really mean 'Qwen'?
- VRAM 사용량 증가 시 자동으로 메모리를 해제하는 스마트 모델 언로드 기능
- 오프로드 스트림에서 개선된 가중치 캐스팅 성능

Expand All @@ -1529,7 +1529,7 @@

**버그 수정**
- SingleStreamBlock/DoubleStreamBlock의 커스텀 노드 가져오기 오류 수정(임시 해결)
- Qwen ControlNet 회귀 문제 수정

Check warning on line 1532 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1532

Did you really mean 'Qwen'?
- 오프로드 지원 및 안정성 개선을 통한 양자화된 연산 강화
- 모든 모델 간 RoPE 함수 구현 통합

Expand All @@ -1550,7 +1550,7 @@
- 바이트 데이터와 None 출력을 적절히 처리하는 캐싱 시스템 개선

**API 노드 개선사항**
- API 노드를 V3 클라이언트 아키텍처로 이전: Luma, Minimax, Pixverse, Ideogram, StabilityAI, Pika, Recraft, Hypernetwork, OpenAI

Check warning on line 1553 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1553

Did you really mean 'Luma'?

Check warning on line 1553 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1553

Did you really mean 'Minimax'?

Check warning on line 1553 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1553

Did you really mean 'Pixverse'?

Check warning on line 1553 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1553

Did you really mean 'Recraft'?

Check warning on line 1553 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1553

Did you really mean 'Hypernetwork'?
- LTXV API 노드에 12초~20초 길이 옵션 추가
- DALL-E 2 노드의 img2img 작업 수정
- Rodin3D 노드를 개선하여 적절한 상대 경로 반환
Expand All @@ -1565,12 +1565,12 @@
<Update label="v0.3.67" description="2025년 10월 28일">

**API 노드**
- **LTXV API 통합**: Lightricks LTX 동영상 생성을 위한 새로운 LTXV API 노드 추가

Check warning on line 1568 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1568

Did you really mean 'Lightricks'?
- 비동기 작업 및 취소 지원 기능이 포함된 네트워크 클라이언트 V2 업그레이드
- Tripo 및 Gemini API 노드를 V3 스키마로 변환

Check warning on line 1570 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1570

Did you really mean 'Tripo'?

**성능 및 호환성**
- 최신 AMD GPU에서만 cudnn 비활성화하여 AMD GPU 지원 개선

Check warning on line 1573 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1573

Did you really mean 'cudnn'?
- API 노드의 Windows 특정 네트워크 문제를 수정해 더 나은 재시도 처리 가능하게 함

**핵심 개선 사항**
Expand All @@ -1596,7 +1596,7 @@
- 크로마 레디언스 처리 속도 향상 및 1 이상의 배치 크기 문제 해결

**API 노드**
- Veo 3.1 모델 지원 추가

Check warning on line 1599 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1599

Did you really mean 'Veo'?
- 비디오 워크플로우에서 고급 시간적 제어를 위한 TemporalScoreRescaling 노드 추가

**하드웨어 및 호환성**
Expand All @@ -1605,8 +1605,8 @@

**실행 및 스키마**
- ControlNet 노드를 V3 스키마로 변환
- EasyCache를 적절한 batch_slice 처리 방식으로 개선

Check warning on line 1608 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1608

Did you really mean 'batch_slice'?
- merge_nested_dicts 기능을 입력 순서에 맞게 개선

Check warning on line 1609 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1609

Did you really mean 'merge_nested_dicts'?
- 사용되지 않는 파일에 대한 폐기 경고 추가
</Update>

Expand All @@ -1616,18 +1616,18 @@
- 모델 다운스케일링, LoRA 추출, 합성, 랜덤 오퍼레이션, SD3/SLG, Flux, 업스케일 모델, HunyuanVideo 노드를 포함한 핵심 노드 카테고리를 V3 스키마로 마이그레이션했습니다.

**오디오 및 모델 개선사항**
- 고음질 오디오 워크플로우를 위한 MMaudio 16K VAE 지원 추가

Check warning on line 1619 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1619

Did you really mean 'MMaudio'?
- 모노 오디오가 잘못되어 스테레오로 저장되는 문제 수정
- 모델 샘플링 시그마 코드 재작성 및 FP8 스케일링된 LoRA 문제 해결
- 이전 버전의 Stable Diffusion 체크포인트를 최신 NumPy 버전에서 로드하는 문제 수정

**AMD GPU 최적화**
- SD/Flux VAE 연산에 대한 메모리 추정 개선
- ROCm 7.0+에서 RDNA4 PyTorch 어텐션 활성화

Check warning on line 1626 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1626

Did you really mean 'ROCm'?

**API 노드 업데이트**
- 가격 추출기 추가 및 Kling/Pika API 노드 개선
- aspect_ratio 지원을 추가한 Gemini 이미지 API 강화

Check warning on line 1630 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1630

Did you really mean 'aspect_ratio'?

**업데이트 사항**
- 템플릿 v0.1.95, 노드 문서 v0.3.0
Expand All @@ -1654,13 +1654,13 @@

**API 노드 업데이트**
- **클링 2.5 터보**: txt2video 및 img2video 노드 모두에서 kling-2-5-turbo 지원을 추가하고, 적절한 모드 설정을 수정했습니다.
- **API 노드 수정사항**: Gemini 노드의 base64 처리를 개선하고, Recraft API 노드 함수의 들여쓰기 문제를 수정했습니다.

Check warning on line 1657 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1657

Did you really mean 'Recraft'?

**노드 스키마 마이그레이션(V3)**
- **광범위한 V3 변환**: 오디오 인코더, GITS, 차분 확산, 최적 단계, PAG, LT, IP2P, 형태학, torch 컴파일, EPS, Pixverse, TomeSD, 편집 모델, Rodin, Stable3D 노드 등 수많은 노드 카테고리를 V3 스키마로 이전하여 워크플로우의 호환성을 높였습니다.

Check warning on line 1660 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1660

Did you really mean 'Pixverse'?

**개발자 경험 개선**
- **코드 품질**: comfy_api_nodes 폴더에 pylint 지원을 추가하고 example_node.py를 V3 스키마로 업데이트하여 커스텀 노드 개발을 더욱 일관되게 만들었습니다.

Check warning on line 1663 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1663

Did you really mean 'comfy_api_nodes'?

Check warning on line 1663 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1663

Did you really mean 'pylint'?
- **문서 업데이트**: Windows 사용자를 위한 야간 PyTorch 명령어를 포함한 AMD 설치 지침을 강화했습니다.

**프론트엔드 업데이트**
Expand All @@ -1672,13 +1672,13 @@

**API 노드**
- **Rodin3D-Gen2 매개변수 수정**
- **Seedance Pro 모델 지원**

Check warning on line 1675 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1675

Did you really mean 'Seedance'?
</Update>

<Update label="v0.3.61" description="2025년 9월 30일">

**API 노드**
- **Rodin3D Gen-2**: Rodin의 가장 강력한 이미지-to-3D 도구가 이제 ComfyUI에서 바로 사용 가능합니다!
- **Rodin3D Gen-2**: Rodin의 가장 강력한 이미지 기반 3D 생성 도구가 이제 ComfyUI에서 바로 사용 가능합니다!
- **WAN 이미지 투 이미지**: Wan2.5 이미지 투 이미지 API 노드로, 이미지 편집을 지원합니다.

**향상된 오디오 기능**
Expand All @@ -1690,7 +1690,7 @@

**메모리 및 성능 최적화**
- **메모리 누수 수정**: 모델 최종화자를 명시적으로 분리하여 메모리 누수를 해결함으로써 장시간 실행 워크플로우의 안정성을 향상시켰습니다.
- **샘플러 CFG 개선**: 샘플러 CFG 함수 인자에 'input_cond' 및 'input_uncond' 매개변수를 추가해 보다 유연한 조건부 제어를 가능하게 했습니다.

Check warning on line 1693 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1693

Did you really mean 'input_cond'?

Check warning on line 1693 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1693

Did you really mean 'input_uncond'?

</Update>

Expand All @@ -1698,17 +1698,17 @@

**새로운 모델 지원**
- **Wan2.2 애니메이트**: 캐릭터 교체 및 모션 전송 기능을 갖춘 Wan2.2 애니메이트 영상 생성 모델 지원
- **Qwen 이미지 편집 2509 지원**: 다중 이미지 편집, 더 높은 일관성, 기본 ControlNet 지원을 갖춘 업데이트된 Qwen 이미지 편집 2509 지원

Check warning on line 1701 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1701

Did you really mean 'Qwen'?

Check warning on line 1701 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1701

Did you really mean 'Qwen'?
- **HuMo 모델**: 오디오를 사용해 영상 생성을 제어하면서 입술 동기화를 유지하는 1.7B 및 17B HuMo 모델 모두 지원 추가
- **Chroma Radiance**: 픽셀 공간에서 이미지 생성을 수행하는 모델로, 이미지 생성 과정 중 손실을 줄여줍니다
- **Omnigen2 UMO LoRA**: Omnigen2 UMO LoRA 모델 지원 추가

**API 노드 추가**
- **Kling v2.1 지원**: KlingStartEndFrame 노드에 kling-v2-1 모델 추가

Check warning on line 1707 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1707

Did you really mean 'Kling'?
- **Seedream4 수정사항**: 부분 성공 시 오류를 무시하는 플래그를 수정하여 워크플로우를 더욱 안정적으로 만들었습니다

**노드 스키마 마이그레이션(V3)**
- **핵심 노드 업데이트**: Minimax API, Cosmos, 컨디셔닝, CFG, Canny 노드 등 여러 노드 카테고리를 V3 스키마로 마이그레이션했습니다

Check warning on line 1711 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1711

Did you really mean 'Minimax'?

**성능 및 기술적 개선**
- **FP8 연산**: gfx1200 하드웨어에서 기본적으로 FP8 연산을 활성화하여 처리 속도를 높였습니다
Expand All @@ -1721,8 +1721,8 @@

<Update label="v0.3.59" description="2025년 9월 10일">

**ByteDance Seedream 4.0 통합**

Check warning on line 1724 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1724

Did you really mean 'Seedream'?
- **새로운 Seedream 노드**: ByteDanceSeedream (4.0) 노드를 추가했습니다.

Check warning on line 1725 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1725

Did you really mean 'Seedream'?

</Update>

Expand All @@ -1746,7 +1746,7 @@
**워크플로우 유틸리티**
- **ImageScaleToMaxDimension 노드**: 지능형 이미지 스케일링을 위한 새로운 유틸리티
- **SEEDS 노이즈 시스템**: 개선된 알고리즘으로 업데이트된 노이즈 분해 기능
- **향상된 프롬프트 제어**: 인터럽트 핸들러가 이제 prompt_id 매개변수를 수락합니다.

Check warning on line 1749 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1749

Did you really mean 'prompt_id'?

**성능 및 아키텍처**
- **V3 스키마 마이그레이션**: 일부 핵심 노드를 V3 스키마로 변환했습니다.
Expand Down Expand Up @@ -1780,7 +1780,7 @@

**Wan2.2 5B 모델 통합**
- **Fun Control 모델 지원**: Wan2.2 5B fun control 모델 지원 추가.
- **Fun Inpaint 모델 지원**: Wan2.2 5B fun inpaint 모델 통합.

Check warning on line 1783 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1783

Did you really mean 'Inpaint'?

Check warning on line 1783 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1783

Did you really mean 'inpaint'?

</Update>

Expand All @@ -1805,7 +1805,7 @@
이번 릴리스에서는 ComfyUI 오디오 처리 기능을 추가하고, 성능 개선과 모델 호환성 업데이트를 포함했습니다:

**오디오 처리 업데이트**
- **Wav2vec2 오디오 인코더**: 오디오 인코더 모델로 네이티브 Wav2vec2 구현을 추가하여 멀티모달 애플리케이션의 오디오-to-임베딩 워크플로우를 가능하게 했습니다.
- **Wav2vec2 오디오 인코더**: 오디오 인코더 모델로 네이티브 Wav2vec2 구현을 추가하여 멀티모달 애플리케이션의 오디오 기반 임베딩 워크플로우를 가능하게 했습니다.
- **오디오 인코더 디렉토리**: models/audio_encoders 디렉토리를 추가했으며, 이는 Wan2.2 S2V용 오디오 인코더 디렉토리입니다.
- **AudioEncoderOutput V3 지원**: AudioEncoderOutput를 V3 노드 스키마와 호환되도록 만들어 현대적인 워크플로우 아키텍처와 원활한 통합을 보장했습니다.

Expand All @@ -1819,12 +1819,12 @@
- **음성 입력 처리 개선**: S2V 워크플로우에서 음성 입력의 음수 처리를 수정해 적절한 0값을 사용하도록 했습니다.

**샘플링 및 노드 개선**
- **DPM++ 2M SDE Heun (RES) 샘플러**: @Balladie가 개발한 새로운 고급 샘플러로, 세밀한 생성 제어를 위한 추가 샘플링 옵션을 제공합니다.

Check warning on line 1822 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1822

Did you really mean 'Heun'?
- **LatentConcat 노드**: 잠재 텐서를 연결하는 새 노드로, 고급 잠재 공간 조작 워크플로우를 가능하게 합니다.
- **EasyCache/LazyCache 안정성**: 샘플링 과정에서 텐서 속성(형태/데이터형/장치)이 변경될 때 발생하는 심각한 크래시를 수정해 워크플로우의 신뢰성을 보장했습니다.

**모델 호환성 개선**
- **ControlNet 유형 모델**: Qwen Edit 및 Kontext 워크플로우와 함께 작동하는 ControlNet 유형 모델의 호환성 문제를 개선했습니다.

Check warning on line 1827 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1827

Did you really mean 'Qwen'?

Check warning on line 1827 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1827

Did you really mean 'Kontext'?
- **Flux 메모리 최적화**: Flux 모델의 메모리 사용량 요소를 조정해 자원 활용을 더욱 효율적으로 만들었습니다.

**인프라 및 안정성**
Expand All @@ -1836,17 +1836,17 @@

<Update label="v0.3.52" description="2025년 8월 23일">

**향상된 모델 지원 및 Qwen 이미지 ControlNet 통합**

Check warning on line 1839 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1839

Did you really mean 'Qwen'?

이번 릴리스에서는 ControlNet 기능을 대폭 확장하고 모델 호환성을 개선하여 ComfyUI 워크플로우를 더욱 다재다능하고 안정적으로 만들었습니다:

**Qwen ControlNet 생태계**

Check warning on line 1843 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1843

Did you really mean 'Qwen'?
- **Diffsynth ControlNet 지원**: Canny 및 depth 조건부의 Qwen Diffsynth ControlNet 지원 추가로 정밀한 엣지 및 깊이 기반 이미지 제어 가능

Check warning on line 1844 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1844

Did you really mean 'Diffsynth'?

Check warning on line 1844 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1844

Did you really mean 'Qwen'?

Check warning on line 1844 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1844

Did you really mean 'Diffsynth'?
- **InstantX Qwen ControlNet**: 창의적 제어 옵션 확장을 위한 InstantX Qwen ControlNet 통합

Check warning on line 1845 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1845

Did you really mean 'Qwen'?

Check warning on line 1845 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1845

Did you really mean 'Qwen'?
- **인페인트 ControlNet/모델 패치**: 전용 Diffsynth 인페인트 ControlNet 지원으로 인페인트 기능 향상

Check warning on line 1846 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1846

Did you really mean 'Diffsynth'?

**노드 아키텍처 및 API 진화**
- **V3 아키텍처 마이그레이션**: 스트링 노드, Google Veo API, Ideogram API 노드를 V3 아키텍처로 업그레이드하여 더 나은 성능과 일관성 제공

Check warning on line 1849 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1849

Did you really mean 'Veo'?
- **향상된 API 노드**: OpenAI 챗 노드는 명확성을 위해 'OpenAI ChatGPT'로 이름 변경, Gemini 챗 노드에는 복사 버튼 기능 추가
- **향상된 사용성**: API 노드는 보다 명확한 라벨링과 강화된 상호작용 기능으로 더 나은 사용자 경험 제공

Expand All @@ -1870,7 +1870,7 @@
- **Qwen-Image-Edit 모델**: Qwen-Image-Edit 기본 지원
- **FluxKontextMultiReferenceLatentMethod 노드**: Flux 워크플로우를 위한 다중 참조 입력 노드
- **WAN 2.2 Fun Camera 모델 지원**: 카메라 제어를 통한 비디오 생성 지원
- **템플릿 업데이트**: 버전 0.1.62로 업그레이드, Wan2.2 Fun Camera 및 Qwen Image Edit 템플릿 추가

Check warning on line 1873 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1873

Did you really mean 'Qwen'?

**핵심 기능 개선**
- **컨텍스트 윈도우 지원**: 긴 시퀀스 생성 작업을 지원하도록 샘플링 코드 개선
Expand All @@ -1882,9 +1882,9 @@

**API 노드 지원 업데이트**
- **GPT-5 시리즈 모델**: 최신 GPT-5 모델 지원
- **Kling V2-1 및 V2-1-Master**: 업데이트된 비디오 생성 모델 기능

Check warning on line 1885 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1885

Did you really mean 'Kling'?
- **Minimax Hailuo 비디오 노드**: 새로운 비디오 생성 노드

Check warning on line 1886 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1886

Did you really mean 'Minimax'?

Check warning on line 1886 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1886

Did you really mean 'Hailuo'?
- **Vidu 비디오 노드**: Vidu API 노드 지원

Check warning on line 1887 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1887

Did you really mean 'Vidu'?

Check warning on line 1887 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1887

Did you really mean 'Vidu'?
- **Google 모델 업데이트**: 새로운 Google Gemini 모델 추가
- **OpenAI API 수정**: OpenAI API 노드 입력 이미지의 MIME 유형 오류 수정

Expand All @@ -1909,12 +1909,12 @@

**모델 통합 및 성능 개선**

이번 릴리스에서는 향상된 Qwen 지원, 비동기 API 기능, 복잡한 워크플로우를 위한 안정성 개선을 통해 ComfyUI의 모델 생태계를 확장했습니다:

Check warning on line 1912 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1912

Did you really mean 'Qwen'?

**Qwen 모델 생태계**

Check warning on line 1914 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1914

Did you really mean 'Qwen'?
- **Qwen 이미지 모델 지원**: 정교한 비전 워크플로우를 위한 적절한 LoRA 로딩 및 모델 병합 기능 포함된 통합 개선

Check warning on line 1915 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1915

Did you really mean 'Qwen'?
- **Qwen 모델 병합 노드**: Qwen 이미지 모델을 병합하는 전용 새 노드로, 크리에이터가 서로 다른 모델의 강점을 결합할 수 있도록 지원

Check warning on line 1916 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1916

Did you really mean 'Qwen'?

Check warning on line 1916 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1916

Did you really mean 'Qwen'?
- **SimpleTuner Lycoris LoRA 지원**: SimpleTuner에서 학습한 Lycoris LoRA의 Qwen-Image 모델과의 호환성 확장

Check warning on line 1917 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1917

Did you really mean 'Lycoris'?

Check warning on line 1917 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1917

Did you really mean 'Lycoris'?

**API 및 성능 인프라**
- **비동기 API 노드**: 비동기 API 노드 도입으로 차단되지 않는 워크플로우 실행이 가능해져 성능 향상
Expand All @@ -1922,18 +1922,18 @@
- **WAN 2.2 펀 컨트롤 지원**: WAN 2.2 펀 컨트롤 기능 지원 추가로 비디오 워크플로우의 창의적 제어 범위 확대

**하드웨어 최적화 및 호환성**
- **AMD GPU 개선**: FP16 정확도 처리 개선 및 성능 최적화로 AMD Radeon 지원 강화

Check warning on line 1925 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1925

Did you really mean 'Radeon'?
- **RDNA3 아키텍처 수정**: PyTorch attention을 사용하는 Flux 모델에서 gfx1201 GPU 관련 문제 해결
- **업데이트된 PyTorch 지원**: Python 3.13 및 CUDA 12.9에서 테스트 완료된 CUDA 및 ROCM PyTorch 버전 업그레이드

**개발자 경험 개선**
- **더 깔끔한 로깅**: 기능 플래그는 이제 세부 모드에서만 표시되어 콘솔의 혼잡도 감소
- **오디오 처리 안전성**: 토치오디오 가져오기 안전성 검사 강화로 오디오 종속성이 없을 때도 충돌 방지
- **Kling API 개선**: Kling Image API 노드의 이미지 유형 매개변수 처리 문제 해결

Check warning on line 1932 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1932

Did you really mean 'Kling'?

Check warning on line 1932 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1932

Did you really mean 'Kling'?

**워크플로우 이점**
- **비동기 워크플로우 실행**: 새로운 비동기 API 기능으로 외부 서비스 통합 시 더 반응성이 뛰어난 워크플로우 가능
- **모델 유연성**: 확장된 Qwen 지원으로 더욱 다양한 비전-언어 워크플로우와 향상된 LoRA 호환성 제공

Check warning on line 1936 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1936

Did you really mean 'Qwen'?
- **하드웨어 활용**: AMD GPU 최적화 및 업데이트된 PyTorch 지원으로 하드웨어 구성에 따른 성능 향상
- **배치 처리**: 고정된 RepeatLatentBatch로 복잡한 다차원 데이터 구조에서도 안정적인 작동 보장
- **비디오 제어**: WAN 2.2 펀 컨트롤 기능으로 비디오 생성 워크플로우의 고급 창의적 제어 가능
Expand All @@ -1951,8 +1951,8 @@
- **워크플로우 탐색**: 자주 접근하는 요소를 보다 효과적으로 정리하여 사용자 경험을 개선했습니다.

**모델 통합**
- **Qwen 비전 모델 지원**: 구성 옵션을 갖춘 Qwen 이미지 모델에 대한 초기 지원 제공

Check warning on line 1954 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1954

Did you really mean 'Qwen'?

Check warning on line 1954 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1954

Did you really mean 'Qwen'?
- **이미지 처리**: 향상된 Qwen 모델 통합으로 더욱 다채로운 이미지 분석 및 생성 워크플로우가 가능해졌습니다.

Check warning on line 1955 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1955

Did you really mean 'Qwen'?

**비디오 생성**
- **Veo3 비디오 생성**: 오디오 지원 기능이 통합된 Veo3 비디오 생성 노드 추가
Expand All @@ -1969,7 +1969,7 @@

**워크플로우 이점**
- **더 빠른 반복**: 최근 사용 항목 API를 통해 워크플로우 조립과 수정이 더욱 신속해졌습니다.
- **창의성 강화**: Qwen 비전 모델은 이미지 이해 및 조작 워크플로우에 새로운 가능성을 열어줍니다.

Check warning on line 1972 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L1972

Did you really mean 'Qwen'?
- **비디오 제작**: Veo3 통합으로 ComfyUI가 종합적인 멀티미디어 창작 플랫폼으로 변모했습니다.
- **신뢰성**: 메모리 최적화와 장치 관리 수정으로 복잡한 워크플로우에서도 안정적인 작동을 보장합니다.
- **성능**: VRAM 사용을 최적화해 리소스가 부족한 시스템에서도 보다 야심찬 프로젝트를 수행할 수 있습니다.
Expand Down Expand Up @@ -2038,9 +2038,9 @@
- **오디오 통합**: 멀티미디어 생성 워크플로우의 오디오 처리를 강화했습니다.

**하드웨어 지원**
- **Iluvatar CoreX 지원**: Iluvatar CoreX 가속기의 기본 지원을 추가했습니다.

Check warning on line 2041 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2041

Did you really mean 'Iluvatar'?

Check warning on line 2041 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2041

Did you really mean 'Iluvatar'?
- **Intel XPU 최적화**: 비동기 오프로드 기능을 포함한 XPU 지원 개선사항을 제공했습니다.
- **AMD ROCm 향상**: Torch 2.8에서 gfx1201에 대해 PyTorch 어텐션을 기본으로 활성화했습니다.

Check warning on line 2043 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2043

Did you really mean 'ROCm'?
- **CUDA 메모리 관리**: CUDA malloc을 CUDA 지원 PyTorch 설치에서만 활성화하도록 수정했습니다.

**샘플링 알고리즘 개선사항**
Expand All @@ -2052,12 +2052,12 @@
- **트레이닝 유연성**: 맞춤형 모델 트레이닝을 위한 더 나은 메모리 관리와 성능 최적화를 제공합니다.

**노드 및 워크플로우 개선사항**
- **Moonvalley V2V 노드**: 입력 검증 기능을 강화한 Moonvalley Marey V2V 노드를 추가했습니다.

Check warning on line 2055 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2055

Did you really mean 'Moonvalley'?

Check warning on line 2055 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2055

Did you really mean 'Moonvalley'?

Check warning on line 2055 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2055

Did you really mean 'Marey'?
- **네거티브 프롬프트 업데이트**: Moonvalley 노드의 네거티브 프롬프트 처리를 개선했습니다.

Check warning on line 2056 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2056

Did you really mean 'Moonvalley'?
- **히스토리 API 개선**: get_history API에 map_function 매개변수를 추가했습니다.

Check warning on line 2057 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2057

Did you really mean 'get_history'?

Check warning on line 2057 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2057

Did you really mean 'map_function'?

**API 및 시스템 개선사항**
- **프론트엔드 버전 추적**: /system_stats API 응답에 required_frontend_version 매개변수를 추가했습니다.

Check warning on line 2060 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2060

Did you really mean 'required_frontend_version'?
- **디바이스 정보**: 하드웨어 식별성을 높이기 위해 XPU 디바이스 이름 출력을 개선했습니다.
- **템플릿 업데이트**: 여러 템플릿 업데이트(0.1.40, 0.1.41)를 통해 호환성을 보장했습니다.

Expand All @@ -2076,7 +2076,7 @@

**샘플링 및 생성 기능**
- **SA-Solver 샘플러**: 새로운 재구성된 SA-Solver 샘플링 알고리즘이 추가되어 수치적 안정성이 향상되었습니다.
- **실험적 CFGNorm 노드**: 분류자 없는 가이던스 정규화를 통해 생성 품질을 더욱 세밀하게 제어할 수 있습니다.

Check warning on line 2079 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2079

Did you really mean 'CFGNorm'?
- **중첩된 듀얼 CFG 지원**: DualCFGGuider 노드에 중첩 스타일 구성 옵션이 추가되었습니다.
- **SamplingPercentToSigma 노드**: 샘플링 비율에서 정밀한 시그마 계산을 위한 새로운 유틸리티 노드입니다.

Expand All @@ -2087,9 +2087,9 @@

**플랫폼 및 성능 개선사항**
- **비동기 노드 지원**: 비동기 노드 함수를 완벽히 지원하며 이전보다 실행 속도가 최적화되었습니다.
- **크로마 유연성**: 크로마에서 patch_size 매개변수를 하드코딩하지 않도록 변경했습니다.

Check warning on line 2090 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2090

Did you really mean 'patch_size'?
- **LTXV VAE 디코더**: 이미지 품질 향상을 위해 기본 패딩 모드를 개선된 방식으로 전환했습니다.
- **Safetensors 메모리 관리**: mmap 문제를 해결하기 위한 대응책을 추가했습니다.

Check warning on line 2092 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2092

Did you really mean 'Safetensors'?

Check warning on line 2092 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2092

Did you really mean 'mmap'?

**API 및 통합 개선사항**
- **맞춤형 프롬프트 ID**: API에서 프롬프트 ID를 지정하여 워크플로우 추적성을 높일 수 있습니다.
Expand All @@ -2099,7 +2099,7 @@

**버그 수정 및 안정성**
- **MaskComposite 수정**: 대상 마스크가 2차원인 경우 발생하는 오류를 해결했습니다.
- **Fresca 입출력**: Fresca 모델 워크플로우의 입출력 처리를 정확히 수정했습니다.

Check warning on line 2102 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2102

Did you really mean 'Fresca'?

Check warning on line 2102 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2102

Did you really mean 'Fresca'?
- **참조 버그 수정**: Gemini 노드 구현에서 발생하는 잘못된 참조 버그를 해결했습니다.
- **라인 종료 표준화**: Windows 라인 종료를 자동으로 감지하고 제거했습니다.

Expand Down Expand Up @@ -2158,16 +2158,16 @@

이번 릴리스에서는 모델 호환성과 워크플로우 안정성 향상이 이루어졌습니다:

**확장된 모델 문서화**: Flux Kontext 및 Omnigen 2 모델에 대한 지원 문서 추가

Check warning on line 2161 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2161

Did you really mean 'Omnigen'?
**VAE 인코딩 개선**: VAE 인코딩 중 불필요한 임의 노이즈 주입 제거
**메모리 관리 수정**: Kontext 모델 사용에 영향을 미치던 메모리 추정 버그 해결

Check warning on line 2163 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2163

Did you really mean 'Kontext'?

</Update>

<Update label="v0.3.41" description="2025년 6월 17일">

**모델 지원 추가 사항**
- **Cosmos Predict2 지원**: 텍스트-to이미지(2B 및 14B 모델)와 이미지-to비디오 생성 워크플로우 구현
- **Cosmos Predict2 지원**: 텍스트 기반 이미지 생성(2B 및 14B 모델)와 이미지 기반 비디오 생성 워크플로우 구현
- **Flux 호환성**: Chroma 텍스트 인코더가 이제 일반 Flux 모델과도 작동합니다
- **LoRA 트레이닝 통합**: 가중치 어댑터 방식을 사용하는 새로운 기본 LoRA 트레이닝 노드

Expand All @@ -2178,17 +2178,17 @@

**샘플링 개선 사항**
- **Rectified Flow 샘플러**: SEEDS 및 다단계 DPM++ SDE 샘플러를 RF 지원과 함께 추가
- **ModelSamplingContinuousEDM**: 강화된 샘플링 제어를 위한 새로운 cosmos_rflow 옵션

Check warning on line 2181 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2181

Did you really mean 'cosmos_rflow'?
- **메모리 최적화**: Cosmos 모델에 대한 메모리 추정치 개선

**개발자 및 통합 기능**
- **SQLite 데이터베이스 지원**: 커스텀 노드를 위한 데이터 관리 기능 강화
- **PyProject.toml 통합**: pyproject 파일에서 자동으로 웹 폴더 등록

Check warning on line 2186 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2186

Did you really mean 'pyproject'?
- **프론트엔드 유연성**: semver 접미사 및 프리릴리즈 프론트엔드 버전 지원

Check warning on line 2187 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2187

Did you really mean 'semver'?
- **토큰라이저 개선**: tokenizer_data와 함께 구성 가능한 min_length 설정

Check warning on line 2188 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2188

Did you really mean 'min_length'?

**생활 편의성 개선 사항**
- **Kontext 종횡비 수정**: 위젯 전용 제한 해소

Check warning on line 2191 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2191

Did you really mean 'Kontext'?
- **SaveLora 일관성**: 모든 저장 노드에서 파일명 형식 표준화
- **파이썬 버전 경고**: 오래된 파이썬 설치에 대한 알림 추가
- **웹캠 캡처 수정사항**: IS_CHANGED 시그니처 수정
Expand Down Expand Up @@ -2223,7 +2223,7 @@

**버그 수정 및 안정성**
- **Pillow 호환성**: 사용되지 않는 API 호출을 업데이트했습니다
- **ROCm 지원**: AMD GPU 사용자를 위한 버전 감지 기능을 개선했습니다

Check warning on line 2226 in ko/changelog/index.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/changelog/index.mdx#L2226

Did you really mean 'ROCm'?
- **템플릿 업데이트**: 커스텀 노드 개발을 위한 프로젝트 템플릿을 강화했습니다

</Update>
22 changes: 11 additions & 11 deletions ko/get_started/first_generation.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -31,13 +31,13 @@
2. 모델 설치 안내
- 자동 모델 설치
- 수동 모델 설치
3. 첫 번째 텍스트-to-이미지 생성 완료
3. 첫 번째 텍스트 기반 이미지 생성 완료

## 텍스트-to-이미지란?
## 텍스트 기반 이미지 생성이란?

텍스트-to-이미지는 텍스트 설명을 바탕으로 이미지를 생성하는 기본적인 AI 드로잉 기능입니다. AI 아트 생성에서 가장 많이 사용되는 기능 중 하나입니다. 이 과정을 생각해보면, 당신의 요구사항(긍정적 및 부정적 프롬프트)을 아티스트(드로잉 모델)에게 전달하고, 아티스트는 당신이 원하는 것을 만들어냅니다. 텍스트-to-이미지에 대한 자세한 설명은 [텍스트 to 이미지](/ko/tutorials/basic/text-to-image) 챕터에서 다룹니다.
텍스트 기반 이미지 생성은 텍스트 설명을 바탕으로 이미지를 생성하는 기본적인 AI 드로잉 기능입니다. AI 아트 생성에서 가장 많이 사용되는 기능 중 하나입니다. 이 과정을 생각해보면, 당신의 요구사항(긍정적 및 부정적 프롬프트)을 아티스트(드로잉 모델)에게 전달하고, 아티스트는 당신이 원하는 것을 만들어냅니다. 텍스트 기반 이미지 생성에 대한 자세한 설명은 [텍스트 기반 이미지 생성](/ko/tutorials/basic/text-to-image) 챕터에서 다룹니다.

## ComfyUI 텍스트-to-이미지 워크플로드 튜토리얼
## ComfyUI 텍스트 기반 이미지 생성 워크플로드 튜토리얼

### 1. ComfyUI 실행하기

Expand All @@ -49,9 +49,9 @@

<InstallLink/>

### 2. 기본 텍스트-to-이미지 워크플로드 로딩하기
### 2. 기본 텍스트 기반 이미지 생성 워크플로드 로딩하기

ComfyUI는 보통 실행 시 기본 텍스트-to-이미지 워크플로드를 자동으로 로드합니다. 하지만 다양한 방법으로 워크플로드를 로드해 ComfyUI의 기본 작동 방식을 익힐 수 있습니다:
ComfyUI는 보통 실행 시 기본 텍스트 기반 이미지 생성 워크플로드를 자동으로 로드합니다. 하지만 다양한 방법으로 워크플로드를 로드해 ComfyUI의 기본 작동 방식을 익힐 수 있습니다:

<Tabs>
<Tab title="워크플로드 템플릿에서 로딩하기">
Expand All @@ -73,7 +73,7 @@
- 메뉴 **워크플로드** -> **열기**를 사용해 이미지를 열기

이 예제 이미지를 사용해 워크플로드를 로드해보세요:
![ComfyUI-텍스트 to 이미지 워크플로드](https://raw.githubusercontent.com/Comfy-Org/example_workflows/refs/heads/main/text-to-image-workflow.png)
![ComfyUI-텍스트 기반 이미지 생성 워크플로드](https://raw.githubusercontent.com/Comfy-Org/example_workflows/refs/heads/main/text-to-image-workflow.png)

</Tab>

Expand All @@ -82,7 +82,7 @@

이 예제 워크플로드를 다운로드해 로드해보세요:
<a className="prose" href="https://github.com/Comfy-Org/docs/blob/main/public/text-to-image.json" download style={{ display: 'inline-block', backgroundColor: '#0078D6', color: '#ffffff', padding: '10px 20px', borderRadius: '8px', borderColor: "transparent", textDecoration: 'none', fontWeight: 'bold'}}>
<p className="prose" style={{ margin: 0, fontSize: "0.8rem" }}>텍스트 to 이미지.json 다운로드</p>
<p className="prose" style={{ margin: 0, fontSize: "0.8rem" }}>text-to-image.json 다운로드</p>
</a>

다운로드 후 메뉴 **워크플로드** -> **열기**를 사용해 JSON 파일을 로드하세요.
Expand Down Expand Up @@ -180,16 +180,16 @@

![워크플로드 실행](/images/tutorial/gettingstarted/run_workflow.png)
1. **Load Checkpoint** 노드에서 **v1-5-pruned-emaonly-fp16.safetensors**가 선택되었는지 확인하세요
2. **실행**을 클릭하거나 **Ctrl + Enter**를 눌러 생성하세요

Check warning on line 183 in ko/get_started/first_generation.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/get_started/first_generation.mdx#L183

Did you really mean 'Ctrl'?

결과는 **Save Image** 노드에 나타납니다. 마우스 오른쪽 버튼을 클릭해 로컬에 저장하세요.

![ComfyUI 첫 번째 이미지 생성 결과](/images/tutorial/gettingstarted/result.png)

자세한 텍스트-to-이미지 지침은 종합 가이드를 참조하세요:
자세한 텍스트 기반 이미지 생성 지침은 종합 가이드를 참조하세요:

<Card title="ComfyUI 텍스트-to-이미지 워크플로드 가이드" icon="link" href="/ko/tutorials/basic/text-to-image">
자세한 텍스트-to-이미지 워크플로드 지침은 여기를 클릭하세요
<Card title="ComfyUI 텍스트 기반 이미지 생성 워크플로드 가이드" icon="link" href="/ko/tutorials/basic/text-to-image">
자세한 텍스트 기반 이미지 생성 워크플로드 지침은 여기를 클릭하세요
</Card>

## 문제 해결
Expand Down
34 changes: 17 additions & 17 deletions ko/tutorials/basic/text-to-image.mdx
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
title: "ComfyUI 텍스트-to-이미지 워크플로우"
description: "이 가이드는 AI 아트 생성에서 텍스트-to-이미지의 개념을 이해하고 ComfyUI에서 텍스트-to-이미지 워크플로우를 완료하는 데 도움을 드립니다."
sidebarTitle: "텍스트-to-이미지"
title: "ComfyUI 텍스트 기반 이미지 생성 워크플로우"
description: "이 가이드는 AI 아트 생성에서 텍스트 기반 이미지 생성의 개념을 이해하고 ComfyUI에서 텍스트 기반 이미지 생성 워크플로우를 완료하는 데 도움을 드립니다."
sidebarTitle: "텍스트 기반 이미지 생성"
translationSourceHash: b23cec9b
translationFrom: tutorials/basic/text-to-image.mdx
translationBlockHashes:
Expand All @@ -14,40 +14,40 @@
---


이 가이드는 ComfyUI의 텍스트-to-이미지 워크플로우를 소개하고 다양한 ComfyUI 노드의 기능과 사용법을 이해하도록 돕습니다.
이 가이드는 ComfyUI의 텍스트 기반 이미지 생성 워크플로우를 소개하고 다양한 ComfyUI 노드의 기능과 사용법을 이해하도록 돕습니다.

이 문서에서는 다음을 다룹니다:
- 텍스트-to-이미지 워크플로우 완료하기
- 텍스트 기반 이미지 생성 워크플로우 완료하기
- 확산 모델 원리에 대한 기본적인 이해하기
- 워크플로우 노드의 기능과 역할 알아보기
- SD1.5 모델에 대한 초기 이해하기

먼저 텍스트-to-이미지 워크플로우를 실행한 후 관련 개념에 대한 설명을 이어갑니다. 필요에 따라 관련 섹션을 선택해 주세요.
먼저 텍스트 기반 이미지 생성 워크플로우를 실행한 후 관련 개념에 대한 설명을 이어갑니다. 필요에 따라 관련 섹션을 선택해 주세요.

## 텍스트-to-이미지란?
## 텍스트 기반 이미지 생성이란?

**텍스트-to-이미지**는 텍스트 설명을 통해 이미지를 생성하는 AI 아트 생성의 핵심 과정으로, 그 중심에는 **확산 모델**이 있습니다.
**텍스트 기반 이미지 생성**은 텍스트 설명을 통해 이미지를 생성하는 AI 아트 생성의 핵심 과정으로, 그 중심에는 **확산 모델**이 있습니다.

텍스트-to-이미지 과정에는 다음과 같은 요소가 필요합니다:
텍스트 기반 이미지 생성 과정에는 다음과 같은 요소가 필요합니다:
- **아티스트**: 이미지 생성 모델
- **캔버스**: 잠재 공간
- **이미지 요구사항(프롬프트)**: 긍정적 프롬프트(이미지에 포함되기를 원하는 요소)와 부정적 프롬프트(포함되지 않기를 원하는 요소) 포함

이 텍스트-to-이미지 생성 과정은 간단히 말해, 당신의 요구사항(긍정적·부정적 프롬프트)을 **아티스트(이미지 모델)**에게 전달하면, 아티스트가 이러한 요구사항에 맞춰 원하는 이미지를 만들어내는 것입니다.
이 텍스트 기반 이미지 생성 과정은 간단히 말해, 당신의 요구사항(긍정적·부정적 프롬프트)을 **아티스트(이미지 모델)**에게 전달하면, 아티스트가 이러한 요구사항에 맞춰 원하는 이미지를 만들어내는 것입니다.

## ComfyUI 텍스트-to-이미지 워크플로우 예제 가이드
## ComfyUI 텍스트 기반 이미지 생성 워크플로우 예제 가이드

### 1. 준비 작업

`ComfyUI/models/checkpoints` 폴더에 최소 한 개의 SD1.5 모델 파일이 있는지 확인하세요. 예를 들어 [v1-5-pruned-emaonly-fp16.safetensors](https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/blob/main/v1-5-pruned-emaonly-fp16.safetensors)가 있습니다.

아직 설치하지 않았다면, [ComfyUI AI 아트 생성 시작하기](/ko/get_started/first_generation)의 모델 설치 섹션을 참고해 주세요.

### 2. 텍스트-to-이미지 워크플로우 로드하기
### 2. 텍스트 기반 이미지 생성 워크플로우 로드하기

아래 이미지를 다운로드한 후, **ComfyUI로 드래그하여 워크플로우를 로드**하세요:

<img src="https://raw.githubusercontent.com/Comfy-Org/example_workflows/main/text-to-image-workflow.png" alt="ComfyUI-텍스트-to-이미지 워크플로우" />
<img src="https://raw.githubusercontent.com/Comfy-Org/example_workflows/main/text-to-image-workflow.png" alt="ComfyUI-텍스트 기반 이미지 생성 워크플로우" />

<Tip>
워크플로우 JSON이 메타데이터에 포함된 이미지는 바로 ComfyUI로 드래그하거나, 메뉴 `Workflows` -> `Open (ctrl+o)`를 이용해 로드할 수 있습니다.
Expand Down Expand Up @@ -75,7 +75,7 @@

![CLIP Text Encoder](/images/comfy_core/conditioning/clip_text_encode.jpg)

KSampler 노드에 연결된 `Positive` 연결은 긍정적 프롬프트를, `Negative` 연결은 부정적 프롬프트를 나타냅니다.

Check warning on line 78 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L78

Did you really mean 'KSampler'?

SD1.5 모델을 위한 몇 가지 기본 프롬프팅 원칙은 다음과 같습니다:
- 가능하면 영어를 사용하세요.
Expand Down Expand Up @@ -129,9 +129,9 @@
blurry, low detail, cartoonish, unrealistic anatomy, out of focus, cluttered, flat lighting
```

## 텍스트-to-이미지 작동 원리
## 텍스트 기반 이미지 생성 작동 원리

텍스트-to-이미지 전체 과정은 **역확산 과정**으로 이해할 수 있습니다. 우리가 다운로드한 [v1-5-pruned-emaonly-fp16.safetensors](https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/blob/main/v1-5-pruned-emaonly-fp16.safetensors)는 사전 훈련된 모델로, **순수한 가우시안 노이즈로부터 목표 이미지를 생성**할 수 있습니다. 우리는 단지 우리의 프롬프트만 입력하면, 무작위 노이즈를 제거해 목표 이미지를 생성할 수 있습니다.
텍스트 기반 이미지 생성 전체 과정은 **역확산 과정**으로 이해할 수 있습니다. 우리가 다운로드한 [v1-5-pruned-emaonly-fp16.safetensors](https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/blob/main/v1-5-pruned-emaonly-fp16.safetensors)는 사전 훈련된 모델로, **순수한 가우시안 노이즈로부터 목표 이미지를 생성**할 수 있습니다. 우리는 단지 우리의 프롬프트만 입력하면, 무작위 노이즈를 제거해 목표 이미지를 생성할 수 있습니다.

```mermaid
graph LR
Expand All @@ -152,9 +152,9 @@
- [노이즈 제거 확산 암묵적 모델(DDIM)](https://arxiv.org/pdf/2010.02502)
- [잠재 확산 모델을 통한 고해상도 이미지 합성](https://arxiv.org/pdf/2112.10752)

## ComfyUI 텍스트-to-이미지 워크플로우 노드 설명
## ComfyUI 텍스트 기반 이미지 생성 워크플로우 노드 설명

![ComfyUI 텍스트-to-이미지 워크플로우 설명](/images/tutorial/basic/text-image-workflow.jpg)
![ComfyUI 텍스트 기반 이미지 생성 워크플로우 설명](/images/tutorial/basic/text-image-workflow.jpg)

### A. Load Checkpoint 노드

Expand All @@ -162,7 +162,7 @@

이 노드는 일반적으로 이미지 생성 모델을 로드하는 데 사용됩니다. `checkpoint`는 보통 세 가지 구성 요소를 포함합니다: `MODEL (UNet)`, `CLIP`, 그리고 `VAE`

- `MODEL (UNet)`: 확산 과정 중 노이즈 예측과 이미지 생성을 담당하는 UNet 모델

Check warning on line 165 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L165

Did you really mean 'UNet'?
- `CLIP`: 텍스트 프롬프트를 모델이 이해할 수 있는 벡터로 변환하는 텍스트 인코더로, 모델은 직접 텍스트 프롬프트를 이해할 수 없기 때문입니다.
- `VAE`: 이미지를 픽셀 공간과 잠재 공간 간에 변환하는 변분 AutoEncoder로, 확산 모델은 잠재 공간에서 작동하며 우리의 이미지는 픽셀 공간에 있습니다.

Expand All @@ -170,7 +170,7 @@

![Empty Latent Image](/images/comfy_core/latent/empty_latent_image.jpg)

KSampler 노드로 출력되는 잠재 공간을 정의합니다. Empty Latent Image 노드는 **순수한 노이즈 잠재 공간**을 구성합니다.

Check warning on line 173 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L173

Did you really mean 'KSampler'?

이 노드의 기능을 캔버스 크기 정의라고 생각할 수 있으며, 이는 최종 생성 이미지의 크기를 결정합니다.

Expand All @@ -179,16 +179,16 @@
![CLIP Text Encoder](/images/comfy_core/conditioning/clip_text_encode.jpg)

프롬프트를 인코딩하는 데 사용되며, 이는 이미지에 대한 귀하의 요구사항입니다.
- KSampler 노드에 연결된 `Positive` 조건 입력은 긍정적 프롬프트(이미지에 포함되기를 원하는 요소)를 나타냅니다.

Check warning on line 182 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L182

Did you really mean 'KSampler'?
- KSampler 노드에 연결된 `Negative` 조건 입력은 부정적 프롬프트(이미지에 포함되지 않기를 원하는 요소)를 나타냅니다.

Check warning on line 183 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L183

Did you really mean 'KSampler'?

프롬프트는 `Load Checkpoint` 노드의 `CLIP` 구성 요소에 의해 의미적 벡터로 인코딩되고, KSampler 노드에 조건으로 출력됩니다.

Check warning on line 185 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L185

Did you really mean 'KSampler'?

### D. KSampler 노드

Check warning on line 187 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L187

Did you really mean 'KSampler'?

![KSampler](/images/comfy_core/sampling/k_sampler.png)

Check warning on line 189 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L189

Did you really mean 'KSampler'?

**KSampler**는 전체 워크플로우의 핵심으로, 모든 노이즈 제거 과정이 이루어지며 결국 잠재 공간 이미지를 출력합니다.

Check warning on line 191 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L191

Did you really mean 'KSampler'?

```mermaid
graph LR
Expand All @@ -198,23 +198,23 @@
B --> E[노이즈 제거된 잠재]
```

KSampler 노드의 파라미터 설명은 다음과 같습니다:

Check warning on line 201 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L201

Did you really mean 'KSampler'?

| 파라미터 이름 | 설명 | 기능 |
|----------------------------|------------------------------------|-------------------------------------------------------------------------------------------------------------|
| **model** | 노이즈 제거에 사용되는 확산 모델 | 생성 이미지의 스타일과 품질을 결정합니다 |
| **positive** | 긍정적 프롬프트 조건 인코딩 | 지정된 요소를 포함하도록 생성을 유도합니다 |
| **negative** | 부정적 프롬프트 조건 인코딩 | 원치 않는 내용을 억제합니다 |
| **latent_image** | 노이즈 제거할 잠재 공간 이미지 | 노이즈 초기화를 위한 입력 캐리어 역할을 합니다 |

Check warning on line 208 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L208

Did you really mean 'latent_image'?
| **seed** | 노이즈 생성용 난수 시드 | 생성의 무작위성을 제어합니다 |
| **control_after_generate** | 생성 후 시드 제어 모드 | 배치 생성 시 시드 변동 패턴을 결정합니다 |

Check warning on line 210 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L210

Did you really mean 'control_after_generate'?
| **steps** | 노이즈 제거 반복 횟수 | 더 많은 단계는 더 세밀한 디테일을 제공하지만 처리 시간이 길어집니다 |
| **cfg** | 분류자 없는 안내 스케일 | 프롬프트 제약 강도를 제어합니다 (너무 높으면 과적합을 초래합니다) |

Check warning on line 212 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L212

Did you really mean 'cfg'?
| **sampler_name** | 샘플링 알고리즘 이름 | 노이즈 제거 경로의 수학적 방법을 결정합니다 |

Check warning on line 213 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L213

Did you really mean 'sampler_name'?
| **scheduler** | 스케줄러 유형 | 노이즈 감쇠율과 단계 크기 할당을 제어합니다 |
| **denoise** | 노이즈 제거 강도 계수 | 잠재 공간에 추가되는 노이즈 강도를 제어하며, 0.0은 원본 입력 특징을 유지하고, 1.0은 완전한 노이즈입니다 |

Check warning on line 215 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L215

Did you really mean 'denoise'?

KSampler 노드에서 잠재 공간은 `seed`를 초기화 파라미터로 사용해 랜덤 노이즈를 구성하고, 의미적 벡터 `Positive`와 `Negative`는 조건으로 확산 모델에 입력됩니다.

Check warning on line 217 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L217

Did you really mean 'KSampler'?

그런 다음, `steps` 파라미터로 지정된 노이즈 제거 단계 수에 따라 노이즈 제거가 수행됩니다. 각 노이즈 제거 단계는 `denoise` 파라미터로 지정된 노이즈 제거 강도 계수를 사용해 잠재 공간을 노이즈 제거하고 새로운 잠재 공간 이미지를 생성합니다.

Expand All @@ -222,7 +222,7 @@

![VAE 디코드](/images/comfy_core/latent/vae_decode.jpg)

**KSampler**에서 출력된 잠재 공간 이미지를 픽셀 공간 이미지로 변환합니다.

Check warning on line 225 in ko/tutorials/basic/text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/basic/text-to-image.mdx#L225

Did you really mean 'KSampler'?

### F. Save Image 노드

Expand Down
14 changes: 7 additions & 7 deletions ko/tutorials/flux/flux-1-text-to-image.mdx
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
title: "ComfyUI Flux.1 텍스트-to-이미지 워크플로우 예시"
description: "이 가이드는 Flux.1 모델에 대한 간략한 소개와 함께, 전체 버전 및 FP8 체크포인트 버전을 포함한 예시를 통해 Flux.1 모델을 이용한 텍스트-to-이미지 생성 방법을 안내합니다."
sidebarTitle: "Flux.1 텍스트-to-이미지"
title: "ComfyUI Flux.1 텍스트 기반 이미지 생성 워크플로우 예시"
description: "이 가이드는 Flux.1 모델에 대한 간략한 소개와 함께, 전체 버전 및 FP8 체크포인트 버전을 포함한 예시를 통해 Flux.1 모델을 이용한 텍스트 기반 이미지 생성 방법을 안내합니다."
sidebarTitle: "Flux.1 텍스트 기반 이미지 생성"
translationSourceHash: 45e7a7c3
translationFrom: tutorials/flux/flux-1-text-to-image.mdx
translationBlockHashes:
Expand All @@ -12,7 +12,7 @@


![Flux](/images/tutorial/flux/flux_example.png)
Flux는 120억 파라미터를 가진 오픈소스 텍스트-to-이미지 생성 모델 중 가장 큰 모델 중 하나로, 원본 파일 크기는 약 23GB입니다. 이 모델은 전 스테이블 디퓨전 팀원들이 설립한 [Black Forest Labs](https://blackforestlabs.ai/)에서 개발되었습니다.
Flux는 120억 파라미터를 가진 오픈소스 텍스트 기반 이미지 생성 모델 중 가장 큰 모델 중 하나로, 원본 파일 크기는 약 23GB입니다. 이 모델은 전 스테이블 디퓨전 팀원들이 설립한 [Black Forest Labs](https://blackforestlabs.ai/)에서 개발되었습니다.
Flux는 뛰어난 이미지 품질과 유연성으로 고화질의 다양한 이미지를 생성할 수 있습니다.

현재 Flux.1 모델에는 몇 가지 주요 버전이 있습니다:
Expand All @@ -25,7 +25,7 @@
- **파라미터 규모:** Flux는 120억 파라미터를 보유해 더욱 복잡한 패턴 관계를 포착하고 더 사실적이고 다양한 이미지를 생성합니다.
- **다양한 스타일 지원:** 다양한 스타일을 지원하며, 다양한 종류의 이미지에 대해 뛰어난 성능을 발휘합니다.

이 예시에서는 Flux.1 Dev와 Flux.1 Schnell 버전을 모두 사용해 텍스트-to-이미지 예시를 소개하며, 여기에는 전체 버전 모델과 간소화된 FP8 체크포인트 버전이 포함됩니다.
이 예시에서는 Flux.1 Dev와 Flux.1 Schnell 버전을 모두 사용해 텍스트 기반 이미지 생성 예시를 소개하며, 여기에는 전체 버전 모델과 간소화된 FP8 체크포인트 버전이 포함됩니다.

Check warning on line 28 in ko/tutorials/flux/flux-1-text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/flux/flux-1-text-to-image.mdx#L28

Did you really mean 'Schnell'?

- **Flux 전체 버전:** 최고의 성능을 자랑하지만 더 많은 VRAM 자원과 여러 모델 파일의 설치가 필요합니다.
- **Flux FP8 체크포인트:** 모델의 fp8 버전 한 개만 필요하지만, 전체 버전에 비해 품질이 약간 떨어집니다.
Expand All @@ -39,7 +39,7 @@
시작하기 전에 ComfyUI가 최신 버전으로 업데이트되었는지 확인하세요.
</Tip>

## Flux.1 전체 버전 텍스트-to-이미지 예시
## Flux.1 전체 버전 텍스트 기반 이미지 생성 예시

<Note>
[black-forest-labs/FLUX.1-dev](https://huggingface.co/black-forest-labs/FLUX.1-dev)에서 모델을 다운로드할 수 없는 경우, Huggingface에 로그인하고 해당 리포지토리의 라이선스 계약에 동의했는지 확인하세요.
Expand Down Expand Up @@ -105,7 +105,7 @@

아래 이미지를 다운로드해 ComfyUI에 드래그하여 워크플로우를 로드하세요.

![Flux Schnell 버전 워크플로우](https://raw.githubusercontent.com/Comfy-Org/example_workflows/main/flux/text-to-image/flux_schnell_t5fp8.png)

Check warning on line 108 in ko/tutorials/flux/flux-1-text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/flux/flux-1-text-to-image.mdx#L108

Did you really mean 'Schnell'?

<a className="prose" target='_blank' href="https://cloud.comfy.org/?template=flux_schnell&utm_source=docs" style={{ display: 'inline-block', backgroundColor: '#28a745', color: '#ffffff', padding: '10px 20px', borderRadius: '8px', borderColor: "transparent", textDecoration: 'none', fontWeight: 'bold'}}>
<p className="prose" style={{ margin: 0, fontSize: "0.8rem" }}>Comfy Cloud에서 실행</p>
Expand Down Expand Up @@ -140,7 +140,7 @@

#### 3. 워크플로우 실행 단계

![Flux Schnell 버전 워크플로우](/images/tutorial/flux/flow_diagram_flux_schnell_t5fp8.jpg)

Check warning on line 143 in ko/tutorials/flux/flux-1-text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/flux/flux-1-text-to-image.mdx#L143

Did you really mean 'Schnell'?

1. `DualCLIPLoader` 노드에 다음 모델이 로드되었는지 확인하세요:
- clip_name1: t5xxl_fp8_e4m3fn.safetensors
Expand All @@ -149,7 +149,7 @@
3. `Load VAE` 노드에 `ae.safetensors`가 로드되었는지 확인하세요.
4. `Queue` 버튼을 클릭하거나 단축키 `Ctrl(cmd) + Enter`를 사용해 워크플로우를 실행하세요.

## Flux.1 FP8 체크포인트 버전 텍스트-to-이미지 예시
## Flux.1 FP8 체크포인트 버전 텍스트 기반 이미지 생성 예시

fp8 버전은 원본 Flux.1 fp16 버전의 양자화된 버전입니다.
어느 정도는 이 버전의 품질이 fp16 버전보다 낮아지지만, VRAM 요구량도 줄어들고, 모델 파일을 하나만 설치하면 바로 실행해볼 수 있습니다.
Expand All @@ -168,11 +168,11 @@

해당 `Load Checkpoint` 노드가 `flux1-dev-fp8.safetensors`를 로드하도록 설정했는지 확인하고, 워크플로우를 실행해보세요.

### Flux.1 Schnell

Check warning on line 171 in ko/tutorials/flux/flux-1-text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/flux/flux-1-text-to-image.mdx#L171

Did you really mean 'Schnell'?

아래 이미지를 다운로드해 ComfyUI에 드래그하여 워크플로우를 로드하세요.

![Flux Schnell fp8 체크포인트 버전 워크플로우](https://raw.githubusercontent.com/Comfy-Org/example_workflows/main/flux/text-to-image/flux_schnell_fp8.png)

Check warning on line 175 in ko/tutorials/flux/flux-1-text-to-image.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/flux/flux-1-text-to-image.mdx#L175

Did you really mean 'Schnell'?

[flux1-schnell-fp8.safetensors](https://huggingface.co/Comfy-Org/flux1-schnell/resolve/main/flux1-schnell-fp8.safetensors?download=true)를 다운로드해 `ComfyUI/models/checkpoints/` 디렉토리에 저장하세요.

Expand Down
2 changes: 1 addition & 1 deletion ko/tutorials/image/hidream/hidream-e1.mdx
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
title: "ComfyUI 네이티브 HiDream-E1, E1.1 워크플로우 예시"
sidebarTitle: "HiDream-e1"
description: "이 가이드는 ComfyUI 네이티브 HiDream-I1 텍스트-to-이미지 워크플로우 예시를 이해하고 완료하는 데 도움을 줄 것입니다."
description: "이 가이드는 ComfyUI 네이티브 HiDream-I1 텍스트 기반 이미지 생성 워크플로우 예시를 이해하고 완료하는 데 도움을 줄 것입니다."
translationSourceHash: 76ea9fff
translationFrom: tutorials/image/hidream/hidream-e1.mdx
translationBlockHashes:
Expand Down Expand Up @@ -126,7 +126,7 @@
E1은 2025년 4월 28일에 출시된 모델입니다. 이 모델은 768*768 해상도만 지원합니다.

<Tip>
참고로, 이 워크플로우는 Google Colab L4 22.5GB VRAM에서 28개 샘플링 단계로 첫 번째 실행 시 약 500초, 두 번째 실행 시 약 370초가 소요됩니다.

Check warning on line 129 in ko/tutorials/image/hidream/hidream-e1.mdx

View check run for this annotation

Mintlify / Mintlify Validation (dripart) - vale-spellcheck

ko/tutorials/image/hidream/hidream-e1.mdx#L129

Did you really mean 'Colab'?
</Tip>

### 1. HiDream-e1 워크플로우
Expand Down
Loading
Loading