diff --git a/changelog/index.mdx b/changelog/index.mdx index ca3a706a..3d05093a 100644 --- a/changelog/index.mdx +++ b/changelog/index.mdx @@ -4,6 +4,51 @@ description: "Track ComfyUI's latest features, improvements, and bug fixes" icon: "clock-rotate-left" --- + + +**Audio & Video Processing Enhancements** + +This release significantly expands ComfyUI's multimedia capabilities with improved audio output formats and enhanced video processing workflows: + +## Audio Output Improvements +- **Expanded Format Support**: Audio output nodes now support Opus and MP3 formats, providing better compression options for audio generation workflows +- **Enhanced Audio Processing**: More flexibility in audio export formats for AI-generated music and sound design projects + +## Video Generation Advancements +- **3D Node Video Recording**: Added support for recording video as output for 3D nodes, enabling seamless 3D-to-video workflow creation +- **Video Duration API**: New `get_duration` method for ComfyUI VIDEO type improves video processing node development +- **Wan-FUN Camera Control**: New WanCameraImageToVideo node with camera control models for advanced video generation with precise camera movements +- **Video Input Validation**: Enhanced video input validation prevents workflow errors and improves stability + +## String Operations & Workflow Utilities +- **Basic String Operations**: New nodes for fundamental string manipulation tasks, essential for prompt engineering and text processing workflows +- **Concatenate with Delimiter**: Enhanced string concatenation node with customizable delimiter support for flexible text composition +- **Pixel Space Noise**: New node for adding pixel-level noise to images, useful for training data augmentation and artistic effects + +## Advanced Sampling & Guidance +- **APG Guidance Implementation**: Added Adaptive Projected Gradient (APG) guidance with momentum fixes for more stable and controlled generation +- **ImagePadForOutpaint**: Now returns proper 3-channel masks for better outpainting workflow compatibility + +## Model Support Extensions +- **ACE Step LoRA**: Official support for ACE Step LoRA models, expanding fine-tuning capabilities +- **Hunyuan Custom**: Initial implementation for Hunyuan model integration (experimental) +- **RMSNorm Compatibility**: Fixed issues with older PyTorch RMSNorm implementations for broader model compatibility + +## Developer Experience Improvements +- **Robust Client API**: Reworked client.py with enhanced error handling and API request logging for better debugging +- **Progress Display**: API nodes now show progress and result URLs directly in the interface +- **Task Failure Handling**: Improved error handling prevents unnecessary request retries on API task failures +- **T5TokenizerOption**: Added missing category for better node organization + +## Platform & Performance Updates +- **AMD GPU Support**: Updated ROCm instructions and compatibility information for AMD users +- **Windows Portable**: Removed outdated PyTorch 2.4 compatibility hack for cleaner Windows deployment +- **Memory Management**: Various optimizations for better memory usage across different hardware configurations + +These improvements make ComfyUI more versatile for multimedia AI workflows while maintaining stability and expanding creative possibilities for artists working with audio, video, and advanced text processing. + + + **Advanced Sampling & Model Control Enhancements** @@ -87,13 +132,7 @@ These changes enhance the reliability of advanced model workflows while maintain **Quality of Life Improvements** - **Kontext Aspect Ratio Fix**: Resolved widget-only limitation, now works properly in all connection modes -- **SaveLora Consistency**: Standardized filename format across all save nodes for better file organization -- **Python Version Warnings**: Added alerts for outdated Python installations to prevent compatibility issues -- **WebcamCapture Fixes**: Corrected IS_CHANGED signature for reliable live input workflows - -This release significantly expands ComfyUI's model ecosystem support while delivering crucial stability improvements and enhanced hardware compatibility across different platforms. - - +- **SaveLora Consistency**: Standardized filename format across all save nodes for better file This release brings powerful new workflow utilities and performance optimizations for ComfyUI creators: diff --git a/zh-CN/changelog/index.mdx b/zh-CN/changelog/index.mdx index 4715759d..55bacfa8 100644 --- a/zh-CN/changelog/index.mdx +++ b/zh-CN/changelog/index.mdx @@ -4,6 +4,50 @@ description: "跟踪ComfyUI的最新功能、改进和错误修复" icon: "clock-rotate-left" --- + + +**音频和视频处理增强** + +此版本通过改进的音频输出格式和增强的视频处理工作流程，显著扩展了ComfyUI的多媒体功能： + +## 音频输出改进 +- **扩展格式支持**：音频输出节点现在支持Opus和MP3格式，为音频生成工作流程提供更好的压缩选项 +- **增强音频处理**：为AI生成音乐和声音设计项目提供更灵活的音频导出格式 + +## 视频生成进步 +- **3D节点视频录制**：添加了对3D节点录制视频输出的支持，实现了无缝的3D转视频工作流程创建 +- **视频时长API**：ComfyUI VIDEO类型的新`get_duration`方法改善了视频处理节点开发 +- **Wan-FUN相机控制**：新的WanCameraImageToVideo节点具有相机控制模型，可进行精确相机移动的高级视频生成 +- **视频输入验证**：增强的视频输入验证防止工作流程错误并提高稳定性 + +## 字符串操作和工作流程实用工具 +- **基础字符串操作**：用于基本字符串操作任务的新节点，对提示词工程和文本处理工作流程至关重要 +- **带分隔符的连接**：增强的字符串连接节点，支持可自定义分隔符，用于灵活的文本组合 +- **像素空间噪声**：用于向图像添加像素级噪声的新节点，适用于训练数据增强和艺术效果 + +## 高级采样和引导 +- **APG引导实现**：添加了具有动量修复的自适应投影梯度（APG）引导，以实现更稳定和受控的生成 +- **ImagePadForOutpaint**：现在返回适当的3通道掩码，以获得更好的outpainting工作流程兼容性 + +## 模型支持扩展 +- **ACE Step LoRA**：官方支持ACE Step LoRA模型，扩展了微调功能 +- **Hunyuan Custom**：Hunyuan模型集成的初始实现（实验性） +- **RMSNorm兼容性**：修复了旧版PyTorch RMSNorm实现的问题，提供更广泛的模型兼容性 + +## 开发者体验改进 +- **强大的客户端API**：重写了client.py，增强了错误处理和API请求日志记录，以便更好地调试 +- **进度显示**：API节点现在直接在界面中显示进度和结果URL +- **任务失败处理**：改进的错误处理防止API任务失败时不必要的请求重试 +- **T5TokenizerOption**：添加了缺失的类别，以便更好地组织节点 + +## 平台和性能更新 +- **AMD GPU支持**：更新了ROCm说明和AMD用户的兼容性信息 +- **Windows便携版**：移除了过时的PyTorch 2.4兼容性hack，以实现更清洁的Windows部署 +- **内存管理**：针对不同硬件配置的内存使用进行了各种优化 + +这些改进使ComfyUI在多媒体AI工作流程中更加多功能，同时保持稳定性并为从事音频、视频和高级文本处理的艺术家扩展创作可能性。 + + **高级采样和模型控制增强**