mirror of
https://github.com/vladmandic/automatic
synced 2026-09-18 00:34:33 +02:00
CalamitousFelicitousness
84683a1d84
Replaces anima_lora.py's bespoke try_load_lora / group_keys / resolve_network_key with thin wrappers binding native_loader's generics to anima's prefix tuples and resolve_targets, mirroring flux2 / zimage / chroma / ernie. The hand-rolled apply_lora_alphas bake-with-balance pass goes away; alpha / scale / dora_scale flow through NetworkWeights.w to NetworkModule.calc_scale at apply time. native_loader gains an optional network_prefix kwarg (str or Callable[[prefix_used], str], default "lora_transformer_") used when constructing network_key. Anima passes a callable picking lora_transformer_ / lora_llm_adapter_ / lora_te_ per matched prefix. Single-component siblings keep the default and are unchanged. network.NetworkModule.apply_weight_decompose grows a dual-path DoRA convention detector. The pre-fix implementation only handled per-input dora_scale (DoRA paper / kohya, shape (1, in)), silently broadcasting per-output LyCORIS / PEFT dora_scale (shape (out, 1)) into an incoherent element-wise rescaling. Detection is structural: (out, 1, ...) routes to per-output; everything else (including the square-weight 1D ambiguity) defaults to per-input for legacy compat. Pre-existing bug surfaced by the LoKR+DoRA LyCORIS files Anima now loads. Behavior changes: - LoHA via the generic try_load_loha (NetworkModuleHada); covers scenery-anima-base and any other LyCORIS .hada_w* export. - Kohya lora_te_ prefix recognized. The legacy resolver only matched BFL text_encoders.qwen3_06b.transformer.model. and silently dropped lora_te_layers_N_* keys (41% of BlueArcStyle's bases were unloaded). - LoKR+DoRA LyCORIS files now apply correctly; the per-output dora_scale is honored instead of silently scrambled. Adds test/test-anima-native-adapters.py: 37 offline tests across all five prefixes (LoRA + LoHA), every COSMOS_2_FLAT_RENAME entry, DoRA threading, marker disambiguation, try_load_chain dispatch, calc_updown sanity, and both DoRA conventions (per-input / per-output / 1D ambiguous). Adapter mock mirrors AnimaLLMAdapter's real module tree.
SD.Next: All-in-one WebUI
SD.Next is a powerful, open-source WebUI app for AI image and video generation, built on Stable Diffusion and supporting dozens of advanced models. Create, caption, and process images and videos with a modern, cross-platform interface—perfect for artists, researchers, and AI enthusiasts.
Table of contents
- Documentation
- Features & Capabilities
- Supported AI Models
- Supported Platforms & Hardware
- Getting started
Screenshot: Desktop interface
Screenshot: Mobile interface
Features and Capabilities
SD.Next is feature-rich with a focus on performance, flexibility, and user experience. Key features include:
- Multi-platform!
- Many diffusion models!
- Fully localized to ~15 languages and with support for many UI themes!
- Desktop and Mobile support!
- Platform specific auto-detection and tuning performed on install
- Built in installer with automatic updates and dependency management
Unique features
SD.Next includes many features not found in other WebUIs, such as:
- SDNQ: State-of-the-Art quantization engine Use pre-quantized or run with quantization on-the-fly for up to 4x VRAM reduction with no or minimal quality and performance impact
- Balanced Offload: Dynamically balance CPU and GPU memory to run larger models on limited hardware
- Captioning with 150+ OpenCLiP models, Tagger with WaifuDiffusion and DeepDanbooru models, and 25+ built-in VLMs
- Image Processing with full image correction color-grading suite of tools
Supported AI Models
SD.Next supports broad range of models: supported models and model specs
Supported Platforms and Hardware
- nVidia GPUs using CUDA libraries on both Windows and Linux
- AMD GPUs using ROCm libraries on both Linux and Windows
- AMD GPUs on Windows using ZLUDA libraries
- Intel Arc GPUs using OneAPI with IPEX XPU libraries on both Windows and Linux
- Any CPU/GPU or device compatible with OpenVINO libraries on both Windows and Linux
- Any GPU compatible with DirectX on Windows using DirectML libraries
- Apple M1/M2 on OSX using built-in support in Torch with MPS optimizations
- ONNX/Olive
Plus Docker container recipes for: CUDA, ROCm, Intel IPEX and OpenVINO
Getting started
- Get started with SD.Next by following the installation instructions
- For more details, check out advanced installation guide
- List and explanation of command line arguments
- Install walkthrough video
Tip
And for platform specific information, check out
WSL | Intel Arc | DirectML | OpenVINO | ONNX & Olive | ZLUDA | AMD ROCm | MacOS | nVidia | Docker
Quick Start
git clone https://github.com/vladmandic/sdnext
cd sdnext
./webui.sh # Linux/Mac
webui.bat # Windows
webui.ps1 # PowerShell
Warning
If you run into issues, check out troubleshooting and debugging guides
Community and Support
If you're unsure how to use a feature, best place to start is Docs and if its not there,
check ChangeLog for when feature was first introduced as it will always have a short note on how to use it
And for any question, reach out on Discord or open an issue or discussion
Contributing
Please see Contributing for details on how to contribute to this project
License & Credits
- SD.Next is licensed under the Apache License 2.0
- Main credit goes to Automatic1111 WebUI for the original codebase