diff --git a/CHANGELOG.md b/CHANGELOG.md index 7c4cab49a..7572a142d 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -5,7 +5,7 @@ ### Highlights for 2026-03-23 This release brings massive code refactoring to modernize codebase and removal of some obsolete features. Leaner & Faster! -And since its a bit quieter period when it comes to new models, notable additions would be : *FireRed-Image-Edit* *SkyWorks-UniPic-3* and *Anima-Preview-2* +And since its a bit quieter period when it comes to new models, notable additions would be : *FireRed-Image-Edit* *SkyWorks-UniPic-3* and new *Anima-Preview* If you're on Windows platform, we have a brand new [All-in-one Installer & Launcher](https://github.com/vladmandic/sdnext-launcher): simply download [exe or zip](https://github.com/vladmandic/sdnext-launcher/releases) and done! @@ -62,8 +62,11 @@ But also many smaller quality-of-life improvements - for full details, see [Chan instead of being used implicitly via quantization - improve image **metadata** parser for foreign metadata (e.g. XMP) - **Compute** + - **ROCm** advanced configuration and tuning, thanks @resonantsky + see *main interface -> scripts -> rocm advanced config* - **ROCm** support for additional AMD GPUs: `gfx103X`, thanks @crashingalexsan - - **Cuda** `torch==2.10` removed support for `rtx1000` series, use following before first startup: + - **Cuda** `torch==2.10` removed support for `rtx1000` series and older GPUs + use following before first startup to force installation of `torch==2.9.1` with `cuda==12.6`: > `set TORCH_COMMAND='torch==2.9.1 torchvision==0.24.1 torchaudio==2.9.1 --index-url https://download.pytorch.org/whl/cu126'` - **UI** - new panel: **server info** with detailed runtime informaton diff --git a/TODO.md b/TODO.md index 1f7667751..f20835b9c 100644 --- a/TODO.md +++ b/TODO.md @@ -4,7 +4,6 @@ - Implement: `unload_auxiliary_models` - Add notes: **Enso** -- Update: **ROCm** - Tips: **Color Grading** - Regen: **Localization** @@ -63,6 +62,7 @@ TODO: Investigate which models are diffusers-compatible and prioritize! ### Image-Edit +- [Bria FIBO-Edit](https://huggingface.co/briaai/Fibo-Edit-RMBG): Fully JSON-based instruction-following image editing framework - [Meituan LongCat-Image-Edit-Turbo](https://huggingface.co/meituan-longcat/LongCat-Image-Edit-Turbo):6B instruction-following image editing with high visual consistency - [VIBE Image-Edit](https://huggingface.co/iitolstykh/VIBE-Image-Edit): (Sana+Qwen-VL)Fast visual instruction-based image editing framework - [LucyEdit](https://github.com/huggingface/diffusers/pull/12340):Instruction-guided video editing while preserving motion and identity diff --git a/installer.py b/installer.py index 090bc0748..ac5f9c7e0 100644 --- a/installer.py +++ b/installer.py @@ -894,7 +894,7 @@ def check_torch(): elif torch.version.hip and allow_rocm: torch_info.set(type='rocm', hip=torch.version.hip) else: - log.warning('Unknown Torch backend') + log.warning('Torch backend: cannot detect type') log.info(f"Torch backend: {torch_info}") for device in [torch.cuda.device(i) for i in range(torch.cuda.device_count())]: gpu = {