From e5b6b743dc67d742d4e9cbcdec2fc1b9ed61533b Mon Sep 17 00:00:00 2001 From: coderfromthenorth93 <213232275+coderfromthenorth93@users.noreply.github.com> Date: Mon, 6 Oct 2025 16:21:31 -0400 Subject: [PATCH 1/3] Add/update changelog/index.mdx --- changelog/index.mdx | 30 +++++++++++++++++++++++++++++- 1 file changed, 29 insertions(+), 1 deletion(-) diff --git a/changelog/index.mdx b/changelog/index.mdx index 914f2a5f..579ba3f7 100644 --- a/changelog/index.mdx +++ b/changelog/index.mdx @@ -4,6 +4,35 @@ description: "Track ComfyUI's latest features, improvements, and bug fixes" icon: "clock-rotate-left" --- + + +**Model Compatibility Enhancements** +- **HunyuanVAE Support**: Added support for the new HunyuanVAE, expanding model compatibility for advanced image generation workflows +- **Epsilon Scaling Node**: Introduced new Epsilon Scaling node for exposure bias correction, providing better control over image generation fine-tuning + +**Memory and Performance Optimizations** +- **VAE Memory Leak Fixes**: Resolved critical VRAM leaks in both VAE tiled fallback operations and WAN VAE cache error handling, improving stability for long-running workflows +- **AMD Performance Boost**: Enabled TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL by default for enhanced AMD GPU performance, plus added smart memory disable option for portable installations + +**API Node Updates** +- **Kling 2.5 Turbo**: Added support for kling-2-5-turbo in both txt2video and img2video nodes, with proper mode configuration fixes +- **API Node Fixes**: Improved Gemini node base64 handling and fixed indentation issues in Recraft API node functions + +**Node Schema Migration (V3)** +- **Extensive V3 Conversion**: Migrated numerous node categories to V3 schema including audio encoder, GITS, differential diffusion, optimal steps, PAG, LT, IP2P, morphology, torch compile, EPS, Pixverse, TomeSD, edit model, Rodin, and Stable3D nodes for improved workflow compatibility + +**Audio System Updates** +- **Dependency Cleanup**: Removed soundfile dependency and eliminated torchaudio load/save operations, streamlining the audio processing pipeline + +**Developer Experience Improvements** +- **Code Quality**: Added pylint support for comfy_api_nodes folder and updated example_node.py to use V3 schema, making custom node development more consistent +- **Documentation Updates**: Enhanced AMD installation instructions with nightly PyTorch commands for Windows users + +**Frontend Updates** +- **Frontend Version**: Updated to version 1.27.7 with latest UI improvements and bug fixes + + + **API Node** @@ -53,7 +82,6 @@ icon: "clock-rotate-left" - **Frontend Version Update**: Updated to version 1.26.13 - **ByteDance Seedream 4.0 Integration** From 0d75b0204974e13a0e9a2132afe3ed0d6bea2392 Mon Sep 17 00:00:00 2001 From: coderfromthenorth93 <213232275+coderfromthenorth93@users.noreply.github.com> Date: Mon, 6 Oct 2025 16:21:32 -0400 Subject: [PATCH 2/3] Add/update zh-CN/changelog/index.mdx --- zh-CN/changelog/index.mdx | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+) diff --git a/zh-CN/changelog/index.mdx b/zh-CN/changelog/index.mdx index fc0c239d..2fc8ded5 100644 --- a/zh-CN/changelog/index.mdx +++ b/zh-CN/changelog/index.mdx @@ -4,6 +4,34 @@ description: "跟踪 ComfyUI 的最新功能、改进和错误修复" icon: "clock-rotate-left" --- + + +**模型兼容性增强** +- **HunyuanVAE 支持**：新增对 HunyuanVAE 的支持，扩展了高级图像生成工作流的模型兼容性 +- **Epsilon Scaling 节点**：引入新的 Epsilon Scaling 节点用于曝光偏差校正，为图像生成微调提供更好的控制 + +**内存和性能优化** +- **VAE 内存泄漏修复**：解决了 VAE 瓦片回退操作和 WAN VAE 缓存错误处理中的关键 VRAM 泄漏问题，提高了长时间运行工作流的稳定性 +- **AMD 性能提升**：默认启用 TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL 以增强 AMD GPU 性能，并为便携式安装添加了智能内存禁用选项 + +**API 节点更新** +- **Kling 2.5 Turbo**：在 txt2video 和 img2video 节点中新增对 kling-2-5-turbo 的支持，并修复了模式配置问题 +- **API 节点修复**：改进了 Gemini 节点的 base64 处理，并修复了 Recraft API 节点函数中的缩进问题 + +**节点架构迁移（V3）** +- **大规模 V3 转换**：将众多节点类别迁移到 V3 架构，包括音频编码器、GITS、差分扩散、最优步数、PAG、LT、IP2P、形态学、torch compile、EPS、Pixverse、TomeSD、编辑模型、Rodin 和 Stable3D 节点，提高了工作流兼容性 + +**音频系统更新** +- **依赖清理**：移除了 soundfile 依赖并取消了 torchaudio 加载/保存操作，简化了音频处理管道 + +**开发者体验改进** +- **代码质量**：为 comfy_api_nodes 文件夹添加了 pylint 支持，并更新 example_node.py 使用 V3 架构，使自定义节点开发更加一致 +- **文档更新**：为 Windows 用户增强了 AMD 安装说明，添加了 nightly PyTorch 命令 + +**前端更新** +- **前端版本**：更新到版本 1.27.7，包含最新的 UI 改进和错误修复 + + - **Rodin3D-Gen2 参数修复** - **Seedance Pro 模型支持** From f077adc99e14983f00ff57631ee85a7ef98e6083 Mon Sep 17 00:00:00 2001 From: ComfyUI Wiki Date: Tue, 7 Oct 2025 20:44:04 +0800 Subject: [PATCH 3/3] Update changelog --- changelog/index.mdx | 13 +++++-------- zh-CN/changelog/index.mdx | 11 ++++++----- 2 files changed, 11 insertions(+), 13 deletions(-) diff --git a/changelog/index.mdx b/changelog/index.mdx index 579ba3f7..7db4d65e 100644 --- a/changelog/index.mdx +++ b/changelog/index.mdx @@ -8,11 +8,11 @@ icon: "clock-rotate-left" **Model Compatibility Enhancements** - **HunyuanVAE Support**: Added support for the new HunyuanVAE, expanding model compatibility for advanced image generation workflows -- **Epsilon Scaling Node**: Introduced new Epsilon Scaling node for exposure bias correction, providing better control over image generation fine-tuning +- **Epsilon Scaling Node**: Introduced new Epsilon Scaling node that reduces exposure bias in diffusion models by scaling predicted noise, improving generation quality (based on the paper [Elucidating the Exposure Bias in Diffusion Models](https://arxiv.org/abs/2308.15321)) **Memory and Performance Optimizations** -- **VAE Memory Leak Fixes**: Resolved critical VRAM leaks in both VAE tiled fallback operations and WAN VAE cache error handling, improving stability for long-running workflows -- **AMD Performance Boost**: Enabled TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL by default for enhanced AMD GPU performance, plus added smart memory disable option for portable installations +- **VAE Memory Leak Fix**: Fixed VRAM leak caused by Python call stack holding tensor references during VAE OOM exception handling, significantly improving tiled fallback reliability on low VRAM devices +- **AMD Support**: Enabled TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL by default **API Node Updates** - **Kling 2.5 Turbo**: Added support for kling-2-5-turbo in both txt2video and img2video nodes, with proper mode configuration fixes @@ -21,16 +21,13 @@ icon: "clock-rotate-left" **Node Schema Migration (V3)** - **Extensive V3 Conversion**: Migrated numerous node categories to V3 schema including audio encoder, GITS, differential diffusion, optimal steps, PAG, LT, IP2P, morphology, torch compile, EPS, Pixverse, TomeSD, edit model, Rodin, and Stable3D nodes for improved workflow compatibility -**Audio System Updates** -- **Dependency Cleanup**: Removed soundfile dependency and eliminated torchaudio load/save operations, streamlining the audio processing pipeline - **Developer Experience Improvements** - **Code Quality**: Added pylint support for comfy_api_nodes folder and updated example_node.py to use V3 schema, making custom node development more consistent - **Documentation Updates**: Enhanced AMD installation instructions with nightly PyTorch commands for Windows users **Frontend Updates** -- **Frontend Version**: Updated to version 1.27.7 with latest UI improvements and bug fixes - +- **Subgraph Publish**: Allows publishing subgraphs to the node library +- **Node Selection Toolbox Redesign**: Redesigned the node selection toolbox diff --git a/zh-CN/changelog/index.mdx b/zh-CN/changelog/index.mdx index 2fc8ded5..ca2de455 100644 --- a/zh-CN/changelog/index.mdx +++ b/zh-CN/changelog/index.mdx @@ -8,11 +8,11 @@ icon: "clock-rotate-left" **模型兼容性增强** - **HunyuanVAE 支持**：新增对 HunyuanVAE 的支持，扩展了高级图像生成工作流的模型兼容性 -- **Epsilon Scaling 节点**：引入新的 Epsilon Scaling 节点用于曝光偏差校正，为图像生成微调提供更好的控制 +- **Epsilon Scaling 节点**：新增 Epsilon Scaling 节点，通过缩放预测噪声来减少扩散模型的曝光偏差，提升生成质量（基于论文 [Elucidating the Exposure Bias in Diffusion Models](https://arxiv.org/abs/2308.15321)） **内存和性能优化** -- **VAE 内存泄漏修复**：解决了 VAE 瓦片回退操作和 WAN VAE 缓存错误处理中的关键 VRAM 泄漏问题，提高了长时间运行工作流的稳定性 -- **AMD 性能提升**：默认启用 TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL 以增强 AMD GPU 性能，并为便携式安装添加了智能内存禁用选项 +- **VAE 内存泄漏修复**：修复了 VAE OOM 异常处理中 Python 调用栈持有张量引用导致的 VRAM 泄漏，显著提升低显存设备上瓦片回退的可靠性 +- **AMD 相关支持**：默认启用 TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL **API 节点更新** - **Kling 2.5 Turbo**：在 txt2video 和 img2video 节点中新增对 kling-2-5-turbo 的支持，并修复了模式配置问题 @@ -26,10 +26,11 @@ icon: "clock-rotate-left" **开发者体验改进** - **代码质量**：为 comfy_api_nodes 文件夹添加了 pylint 支持，并更新 example_node.py 使用 V3 架构，使自定义节点开发更加一致 -- **文档更新**：为 Windows 用户增强了 AMD 安装说明，添加了 nightly PyTorch 命令 +- **文档更新**：为 Windows 用户增加了 AMD 安装说明，添加了 nightly PyTorch 命令 **前端更新** -- **前端版本**：更新到版本 1.27.7，包含最新的 UI 改进和错误修复 +- **子图发布工功能**： 允许将子图发布到节点库 +- **选择工具箱重设计**：重新设计了节点选择工具箱