Public/automatic
1
0
Fork 0
mirror of https://github.com/vladmandic/automatic synced 2026-09-19 01:04:32 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #4017 from vladmandic/dev

Browse Source 
merge dev
This commit is contained in:
Vladimir Mandic
2025-07-29 15:06:29 -04:00
committed by GitHub
parent 156148b820 3579b5b835
commit 891840f36a
789 changed files with 19033 additions and 55165 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.eslintrc.json
+13 -23
View File
@@ -1,20 +1,26 @@
{
"root": true,
"parserOptions": {
"ecmaVersion": 2020
"ecmaVersion": "latest",
"sourceType": "module"
},
"plugins": ["html", "json"],
"plugins": [
"html",
"json"
],
"extends": [
"plugin:json/recommended",
"plugin:node/recommended",
"eslint:recommended",
"airbnb-base"
"airbnb-base",
"plugin:css/recommended",
"plugin:json/recommended",
"plugin:node/recommended"
],
"env": {
"browser": true,
"commonjs": false,
"node": false,
"jquery": false,
"es2020": true
"es2021": true
},
"rules": {
"max-len": [1, 275, 3],
@@ -29,6 +35,7 @@
"no-loop-func":"off",
"no-mixed-operators":"off",
"no-param-reassign":"off",
"no-process-exit":"off",
"no-plusplus":"off",
"no-restricted-globals":"off",
"no-restricted-syntax":"off",
@@ -42,15 +49,12 @@
"node/shebang": "off"
},
"globals": {
// asssets
"panzoom": "readonly",
// logger.js
"log": "readonly",
"debug": "readonly",
"error": "readonly",
"xhrGet": "readonly",
"xhrPost": "readonly",
// script.js
"gradioApp": "readonly",
"executeCallbacks": "readonly",
"onAfterUiUpdate": "readonly",
@@ -66,11 +70,8 @@
"getUICurrentTabContent": "readonly",
"waitForFlag": "readonly",
"logFn": "readonly",
// contextmenus.js
"generateForever": "readonly",
// contributors.js
"showContributors": "readonly",
// ui.js
"opts": "writable",
"sortUIElements": "readonly",
"all_gallery_buttons": "readonly",
@@ -90,40 +91,29 @@
"toggleCompact": "readonly",
"setFontSize": "readonly",
"setTheme": "readonly",
// settings.js
"registerDragDrop": "readonly",
// extraNetworks.js
"getENActiveTab": "readonly",
"quickApplyStyle": "readonly",
"quickSaveStyle": "readonly",
"setupExtraNetworks": "readonly",
"showNetworks": "readonly",
// from python
"localization": "readonly",
// progressbar.js
"randomId": "readonly",
"requestProgress": "readonly",
"setRefreshInterval": "readonly",
// imageviewer.js
"modalPrevImage": "readonly",
"modalNextImage": "readonly",
"galleryClickEventHandler": "readonly",
"getExif": "readonly",
// logMonitor.js
"jobStatusEl": "readonly",
// loader.js
"removeSplash": "readonly",
// nvml.js
"initNVML": "readonly",
"disableNVML": "readonly",
// indexdb.js
"idbGet": "readonly",
"idbPut": "readonly",
"idbDel": "readonly",
"idbAdd": "readonly",
// changelog.js
"initChangelog": "readonly",
// notification.js
"sendNotification": "readonly"
},
"ignorePatterns": [
.gitignore
+1
View File
@@ -11,6 +11,7 @@ __pycache__
/webui-user.sh
/html/extensions.json
/html/themes.json
config_states
node_modules
pnpm-lock.yaml
package-lock.json
.gitmodules
-4
View File
@@ -2,10 +2,6 @@
path = wiki
url = https://github.com/vladmandic/sdnext.wiki
ignore = dirty
[submodule "modules/k-diffusion"]
path = modules/k-diffusion
url = https://github.com/crowsonkb/k-diffusion
ignore = dirty
[submodule "extensions-builtin/sd-extension-system-info"]
path = extensions-builtin/sd-extension-system-info
url = https://github.com/vladmandic/sd-extension-system-info
.pre-commit-config.yaml
+16 -31
View File
@@ -19,48 +19,33 @@ ci:
repos:
# Standard hooks
- repo: https://github.com/pre-commit/pre-commit-hooks
rev: v4.4.0
rev: v5.0.0
hooks:
- id: check-added-large-files
- id: check-case-conflict
- id: check-merge-conflict
- id: check-symlinks
- id: check-illegal-windows-names
- id: check-merge-conflict
- id: detect-private-key
- id: check-builtin-literals
- id: check-case-conflict
- id: check-symlinks
- id: check-yaml
args: ["--allow-multiple-documents"]
- id: debug-statements
- id: check-json
- id: check-toml
- id: check-xml
- id: end-of-file-fixer
- id: mixed-line-ending
- id: check-executables-have-shebangs
exclude: |
(?x)^(
.*.bat|
.*.ps1
)$
- id: trailing-whitespace
exclude: |
(?x)^(
.*\.md|
.github/ISSUE_TEMPLATE/.*\.yml
)$
- repo: https://github.com/charliermarsh/ruff-pre-commit
rev: 'v0.0.285'
hooks:
- id: ruff
args: [--fix, --exit-non-zero-on-fix]
- repo: local
hooks:
- id: pylint
name: pylint
entry: pylint
language: system
types: [python]
args: []
# Black, the code formatter, natively supports pre-commit
# - repo: https://github.com/psf/black
# rev: 23.7.0
# hooks:
# - id: black
# exclude: ^(docs)
# Changes tabs to spaces
# - repo: https://github.com/Lucas-C/pre-commit-hooks
# rev: v1.5.3
# hooks:
# - id: remove-tabs
# exclude: ^(docs)
.pylintrc
+40 -28
View File
@@ -7,54 +7,64 @@ fail-on=
fail-under=10
ignore=CVS
ignore-paths=/usr/lib/.*$,
venv,
.git,
.ruff_cache,
.vscode,
modules/apg,
modules/consistory,
modules/cfgzero,
modules/control/proc,
modules/control/units,
modules/ctrlx,
modules/dml,
modules/freescale,
modules/facelib,
modules/flash_attn_triton_amd,
modules/ggml,
modules/hidiffusion,
modules/hijack,
modules/instantir,
modules/hijack/ddpm_edit.py,
modules/intel,
modules/intel/ipex,
modules/intel/openvino,
modules/k-diffusion,
modules/flex2,
modules/framepack/pipeline,
modules/ldsr,
modules/hidream,
modules/meissonic,
modules/mod,
modules/omnigen,
modules/omnigen2,
modules/onnx_impl,
modules/pag,
modules/pixelsmith,
modules/postprocess/aurasr_arch.py,
modules/prompt_parser_xhinker.py,
modules/pulid/eva_clip,
modules/ras,
modules/rife,
modules/schedulers,
modules/taesd,
modules/teacache,
modules/todo,
modules/unipc,
modules/xadapter,
modules/cfgzero,
modules/infiniteyou,
modules/flash_attn_triton_amd,
scripts/softfill.py,
pipelines/bria,
pipelines/flex2,
pipelines/f_lite,
pipelines/hidream,
pipelines/meissonic,
pipelines/omnigen2,
pipelines/segmoe,
scripts/consistory,
scripts/ctrlx,
scripts/demofusion,
scripts/freescale,
scripts/infiniteyou,
scripts/instantir,
scripts/lbm,
scripts/layerdiffuse,
scripts/mod,
scripts/pixelsmith,
scripts/differential_diffusion.py,
scripts/pulid,
scripts/xadapter,
repositories,
extensions-builtin/Lora,
extensions-builtin/sd-webui-agent-scheduler,
extensions-builtin/sd-extension-chainner/nodes,
extensions-builtin/sd-webui-agent-scheduler,
extensions-builtin/sdnext-modernui/node_modules,
ignore-patterns=.*test*.py$,
.*_model.py$,
.*_arch.py$,
.*_model_arch.py*,
.*_model_arch_v2.py$,
.*_model_arch_v2.py$,
ignored-modules=
jobs=0
limit-inference-results=100
@@ -98,7 +108,7 @@ valid-metaclass-classmethod-first-arg=mcs
[DESIGN]
exclude-too-few-public-methods=
ignored-parents=
max-args=99
max-args=199
max-attributes=99
max-bool-expr=99
max-branches=199
@@ -158,8 +168,8 @@ disable=abstract-method,
consider-using-generator,
consider-using-get,
consider-using-in,
consider-using-min-builtin,
consider-using-max-builtin,
consider-using-min-builtin,
consider-using-sys-exit,
cyclic-import,
dangerous-default-value,
@@ -175,6 +185,7 @@ disable=abstract-method,
missing-class-docstring,
missing-function-docstring,
missing-module-docstring,
no-else-raise,
no-else-return,
not-callable,
pointless-string-statement,
@@ -185,16 +196,17 @@ disable=abstract-method,
too-many-instance-attributes,
too-many-locals,
too-many-nested-blocks,
too-many-statements,
too-many-positional-arguments,
too-many-statements,
unidiomatic-typecheck,
unknown-option-value,
unnecessary-dict-index-lookup,
unnecessary-dunder-call,
unnecessary-lambda,
unnecessary-lambda-assigment,
unnecessary-lambda,
unused-wildcard-import,
use-dict-literal,
use-symbolic-message-instead,
unknown-option-value,
useless-suppression,
wrong-import-position,
enable=c-extension-no-member
.ruff.toml
+24 -34
View File
@@ -3,44 +3,34 @@ exclude = [
".git",
".ruff_cache",
".vscode",
"modules/apg",
"modules/consistory",
"modules/cfgzero",
"modules/facelib",
"modules/flash_attn_triton_amd",
"modules/hidiffusion",
"modules/intel/ipex",
"modules/pag",
"modules/schedulers",
"modules/teacache",
"modules/control/proc",
"modules/control/units",
"modules/freescale",
"modules/flex2",
"modules/ggml",
"modules/hidiffusion",
"modules/hijack",
"modules/instantir",
"modules/intel/ipex",
"modules/intel/openvino",
"modules/k-diffusion",
"modules/ldsr",
"modules/meissonic",
"modules/mod",
"modules/omnigen",
"modules/omnigen2",
"modules/hidream",
"modules/pag",
"modules/pixelsmith",
"modules/control/units/xs_pipe.py",
"modules/postprocess/aurasr_arch.py",
"modules/prompt_parser_xhinker.py",
"modules/pulid/eva_clip",
"modules/ras",
"modules/rife",
"modules/schedulers",
"modules/segmoe",
"modules/taesd",
"modules/teacache",
"modules/todo",
"modules/unipc",
"modules/cfgzero",
"modules/xadapter",
"modules/infiniteyou",
"modules/flash_attn_triton_amd",
"scripts/softfill.py",
"pipelines/meissonic",
"pipelines/omnigen2",
"pipelines/segmoe",
"scripts/lbm",
"scripts/xadapter",
"scripts/pulid",
"scripts/instantir",
"scripts/freescale",
"scripts/consistory",
"repositories",
"extensions-builtin/Lora",
"extensions-builtin/sd-extension-chainner/nodes",
"extensions-builtin/sd-webui-agent-scheduler",
.vscode/launch.json
Vendored
+1 -2
View File
@@ -16,8 +16,7 @@
"--docs",
"--api-log",
"--log", "vscode.log",
"${command:pickArgs}",
]
"${command:pickArgs}"]
}
]
}
.vscode/settings.json
Vendored
+2 -9
View File
@@ -1,13 +1,6 @@
{
"python.analysis.extraPaths": [
".",
"./modules",
"./repositories/blip",
"./repositories/codeformer",
"./repositories/ldm",
"./repositories/taming"
],
"python.analysis.extraPaths": [".", "./modules", "./scripts", "./pipelines"],
"python.analysis.typeCheckingMode": "off",
"editor.formatOnSave": false,
"python.REPL.enableREPLSmartSend": false
}
}
CHANGELOG.md
+222 -3
View File
@@ -1,5 +1,226 @@
# Change Log for SD.Next
## Update for 2025-07-29
### Highlights for 2025-07-29
This is a big one: simply looking at number of changes, probably the biggest release since the project started!
Feature highlights include:
- [ModernUI](https://github.com/user-attachments/assets/6f156154-0b0a-4be2-94f0-979e9f679501) has quite some redesign which should make it more user friendly and easier to navigate plus several new UI themes
If you're still using **StandardUI**, give [ModernUI](https://vladmandic.github.io/sdnext-docs/Themes/) a try!
- New models such as [WanAI 2.2](https://wan.video/) in 5B and A14B variants for both *text-to-video* and *image-to-video* workflows as well as *text-to-image* workflow!
and also [FreePix F-Lite](https://huggingface.co/Freepik/F-Lite), [Bria 3.2](https://huggingface.co/briaai/BRIA-3.2) and [bigASP 2.5](https://civitai.com/models/1789765?modelVersionId=2025412)
- Redesigned [Video](https://vladmandic.github.io/sdnext-docs/Video) interface with support for general video models plus optimized [FramePack](https://vladmandic.github.io/sdnext-docs/FramePack) and [LTXVideo](https://vladmandic.github.io/sdnext-docs/LTX) support
- Fully integrated nudity detection and optional censorship with [NudeNet](https://vladmandic.github.io/sdnext-docs/NudeNet)
- New background replacement and relightning methods using **Latent Bridge Matching** and new **PixelArt** processing filter
- Enhanced auto-detection of default sampler types/settings results in avoiding common mistakes
- Additional **LLM/VLM** models available for captioning and prompt enhance
- Number of workflow and general quality-of-life improvements, especially around **Styles**, **Detailer**, **Preview**, **Batch**, **Control**
- Compute improvements
- [Wiki](https://github.com/vladmandic/automatic/wiki) & [Docs](https://vladmandic.github.io/sdnext-docs/) updates, especially new end-to-end [Parameters](https://vladmandic.github.io/sdnext-docs/Parameters/) page
In this release we finally break with legacy with the removal of the original [A1111](https://github.com/AUTOMATIC1111/stable-diffusion-webui/) codebase which has not been maintained for a while now
This plus major cleanup of codebase and external dependencies resulted in ~55k LoC (*lines-of-code*) reduction and spread over [~750 files](https://github.com/vladmandic/sdnext/pull/4017) in ~200 commits!
We also switched project license to [Apache-2.0](https://github.com/vladmandic/sdnext/blob/dev/LICENSE.txt) which means that SD.Next is now fully compatible with commercial and non-commercial use and redistribution regardless of modifications!
And (*as always*) many bugfixes and improvements to existing features!
For details, see [ChangeLog](https://github.com/vladmandic/automatic/blob/master/CHANGELOG.md)
> [!NOTE]
> We recommend clean install for this release due to sheer size of changes
> Although upgrades and existing installations are tested and should work fine!
![Screenshot](https://github.com/user-attachments/assets/6f156154-0b0a-4be2-94f0-979e9f679501)
[ReadMe](https://github.com/vladmandic/automatic/blob/master/README.md) | [ChangeLog](https://github.com/vladmandic/automatic/blob/master/CHANGELOG.md) | [Docs](https://vladmandic.github.io/sdnext-docs/) | [WiKi](https://github.com/vladmandic/automatic/wiki) | [Discord](https://discord.com/invite/sd-next-federal-batch-inspectors-1101998836328697867)
### Details for 2025-07-29
- **License**
- SD.Next [license](https://github.com/vladmandic/sdnext/blob/dev/LICENSE.txt) switched from **aGPL-v3.0** to **Apache-v2.0**
this means that SD.Next is now fully compatible with commercial and non-commercial use and redistribution regardless of modifications!
- **Models**
- [WanAI Wan 2.2](https://github.com/Wan-Video/Wan2.2) both 5B and A14B variants, for both T2V and I2V support
go to: *video -> generic -> wan -> pick variant*
optimized support with *VACE*, etc. will follow soon
*caution* Wan2.2 on its own is ~68GB, but also includes optional second-stage for later low-noise processing which is absolutely massive at additional ~54GB
you can enable second stage processing in *settings -> model options*, its disabled by default
*note*: quantization and offloading are highly recommended regardless of first-stage only or both stages!
- [WanAI Wan](https://wan.video/) T2V models for T2I workflows
Wan is originally designed for *video* workflows, but now also be used for *text-to-image* workflows!
supports *Wan-2.1 in 1.3B* and 14B variants and *Wan-2.2 in 5B and A14B* variants
supports all standard features such as quantization, offloading, TAESD preview generation, LoRA support etc.
can also load unet/transformer fine-tunes in safetensors format using UNET loader
simply select in *networks -> models -> reference*
*note* 1.3B model is a bit too small for good results and 14B is very large at 78GB even without second-stage so aggressive quantization and offloading are recommended
- [FreePix F-Lite](https://huggingface.co/Freepik/F-Lite) in *7B, 10B and Texture* variants
F-Lite is a 7B/10B model trained exclusively on copyright-safe and SFW content, trained on internal dataset comprising approximately 80 million copyright-safe images
available via *networks -> models -> reference*
- [Bria 3.2](https://huggingface.co/briaai/BRIA-3.2)
Bria is a smaller 4B parameter model built entirely on licensed data and safe for commercial use
*note*: this is a gated model, you need to [accept terms](https://huggingface.co/briaai/BRIA-3.2) and set your [huggingface token](https://vladmandic.github.io/sdnext-docs/Gated/)
available via *networks -> models -> reference*
- [bigASP 2.5](https://civitai.com/models/1789765)
bigASP is an experimental SDXL finetune using Flow matching method
load as usual, and leave sampler set to *Default*
or you can use following samplers: *UniPC, DPM, DEIS, SA*
required sampler settings: *prediction-method=flow-prediction*, *sigma-method=flowmatch*
recommended sampler settings: *flow-shift=1.0*
- [LBM: Latent Bridge Matching](https://github.com/gojasper/LBM)
very fast automatic image background replacement methods with relightning!
*simple*: automatic background replacement using [BiRefNet](https://github.com/ZhengPeng7/BiRefNet)
*relighting*: automatic background replacement with reglighting so source image fits desired background
with optional composite blending
available in *img2img or control -> scripts*
- add **FLUX.1-Kontext-Dev** inpaint workflow
- add **FLUX.1-Kontext-Dev** **Nunchaku** support
*note*: FLUX.1 Kontext is about 2-3x faster with Nunchaku vs standard execution!
- support **FLUX.1** all-in-one safetensors
- support for [Google Gemma 3n](https://huggingface.co/google/gemma-3n-E4B-it) E2B and E4B LLM/VLM models
available in **prompt enhance** and process **captioning**
- support for [HuggingFace SmolLM3](https://huggingface.co/HuggingFaceTB/SmolLM3-3B) 3B LLM model
available in **prompt enhance**
- add [fal AuraFlow 0.2](https://huggingface.co/fal/AuraFlow-v0.2) in addition to existing [fal AuraFlow 0.3](https://huggingface.co/fal/AuraFlow-v0.3) due to large differences in model behavior
available via *networks -> models -> reference*
- add integrated [NudeNet](https://vladmandic.github.io/sdnext-docs/NudeNet) as built-in functionality
*note*: used to be available as a separate [extension](https://github.com/vladmandic/sd-extension-nudenet)
- **Video**
- redesigned **Video** interface
- support for **Generic** video models
includes support for many video models without specific per-model optimizations
included: *Hunyuan, LTX, WAN, Mochi, Latte, Allegro, Cog*
supports quantization, offloading, frame interpolation, etc.
- support for optimized [FramePack](https://vladmandic.github.io/sdnext-docs/FramePack)
with *t2i, i2i, flf2v* workflows
LoRA support, prompt enhance, etc.
now fully integrated instead of being a separate extension
- support for optmized [LTXVideo](https://vladmandic.github.io/sdnext-docs/LTX)
with *t2i, i2i, v2v* workflows
optional native upsampling and video refine workflows
LoRA support with different conditioning types such as Canny/Depth/Pose, etc.
- support for post load quantization
- **UI**
- major update to modernui layout
- add new Windows-like *Blocks* UI theme
- redesign of the *Flat* UI theme
- enhanced look&feel for *Gallery* tab with better search and collapsible sections, thanks to @CalamitousFelicitousness
- **WIKI**
- new [Parameters](https://vladmandic.github.io/sdnext-docs/Parameters/) page that lists and explains all generation parameters
massive thanks to @CalamitousFelicitousness for bringing this to life!
- updated *Models, Video, LTX, FramePack, Styles*, etc.
- **Compute**
- support for [SageAttention2++](https://github.com/thu-ml/SageAttention)
provides 10-15% performance improvement over default SDPA for transformer-based models!
enable in *settings -> compute settings -> sdp options*
*note*: SD.Next will use either SageAttention v1/v2/v2++, depending which one is installed
until authors provide pre-build wheels for v2++, you need to install it manually or SD.Next will auto-install v1
- support for `torch.compile` for LLM: captioning/prompt-enhannce
- support for `torch.compile` with repeated-blocks
reduces time-to-compile 5x without loss of performance!
enable in *settings -> model compile -> repeated*
*note*: torch.compile is not compatible with balanced offload
- **Other**
- **Styles** can now include both generation params and server settings
see [Styles docs](https://vladmandic.github.io/sdnext-docs/Styles/) for details
- **TAESD** is now default preview type since its the only one that supports most new models
- support **TAESD** preview and remote VAE for **HunyuanDit**
- support **TAESD** preview and remote VAE for **AuraFlow**
- support **TAESD** preview for **WanAI**
- SD.Next now starts with *locked* state preventing model loading until startup is complete
- warn when modifying legacy settings that are no longer supported, but available for compatibilty
- warn on incompatible sampler and automatically restore default sampler
- **XYZ grid** can now work with control tab:
if controlnet or processor are selected in xyz grid, they will overwrite settings from first unit in control tab,
when using controlnet/processor selected in xyz grid, behavior is forced as control-only
also freely selectable are control strength, start and end values
- **Batch** warn on unprocessable images and skip operations on errors so that other images can still be processed
- **Metadata** improved parsing and detect foreign metadata
detect ComfyUI images
detect InvokeAI images
- **Detailer** add `expert` mode where list of detailer models can be converted to textbox for manual editing
see [docs](https://vladmandic.github.io/sdnext-docs/Detailer/) for more information
- **Detailer** add option to merge multiple results from each detailer model
for example, hands model can result in two hands each being processed separately or both hands can be merged into one composite job
- **Control** auto-update width/height on image upload
- **Control** auto-determine image save path depending on operations performed
- autodetect **V-prediction** models and override default sampler prediction type as needed
- **SDNQ**
- use inference context during quantization
- use static compile
- rename quantization type for text encoders `default` option to `Same as model`
- **API**
- add `/sdapi/v1/lock-checkpoint` endpoint that can be used to lock/unlock model changes
if model is locked, it cannot be changed using normal load or unload methods
- **Fixes**
- allow theme type `None` to be set in config
- installer dont cache installed state
- fix Cosmos-Predict2 retrying TAESD download
- better handle startup import errors
- fix traceback width preventing copy&paste
- fix ansi controle output from scripts/extensions
- fix diffusers models non-unique hash
- fix loading of manually downloaded diffuser models
- fix api `/sdapi/v1/embeddings` endpoint
- fix incorrect reporting of deleted and modified files
- fix SD3.x loader and TAESD preview
- fix xyz with control enabled
- fix control order of image save operations
- fix control batch-input processing
- fix modules merge save model
- fix torchvision bicubic upsample with ipex
- fix instantir pipeline
- fix prompt encoding if prompts within batch have different segment counts
- fix detailer min/max size
- fix loopback script
- fix networks tags display
- fix yolo refresh models
- cleanup control infotext
- allow upscaling with models that have implicit VAE processing
- framepack improve offloading
- improve prompt parser tokenizer loader
- improve scripts error handling
- improve infotext param parsing
- improve extensions ui search
- improve model type autodetection
- improve model auth check for hf repos
- improve Chroma prompt padding as per recommendations
- lock directml torch to `torch-directml==0.2.4.dev240913`
- lock directml transformers to `transformers==4.52.4`
- improve install of `sentencepiece` tokenizer
- add int8 matmul fallback for ipex with onednn qlinear
- **Refactoring**
*note*: none of the removals result in loss-of-functionality since all those features are already re-implemented
goal here is to remove legacy code, code duplication and reduce code complexity
- obsolete **original backend**
- remove majority of legacy **a1111** codebase
- remove legacy ldm codebase: `/repositories/ldm`
- remove legacy blip codebase: `/repositories/blip`
- remove legacy codeformer codebase: `/repositories/codeformer`
- remove legacy clip patch model: `/models/karlo`
- remove legacy model configs: `/configs/*.yaml`
- remove legacy submodule: `/modules/k-diffusion`
- remove legacy hypernetworks support: `/modules/hypernetworks`
- remove legacy lora support: `/extensions-builtin/Lora`
- remove legacy clip/blip interrogate module
- remove modern-ui remove `only-original` vs `only-diffusers` code paths
- refactor control processing and separate preprocessing and image save ops
- refactor modernui layouts to rely on accordions more than individual controls
- refactore pipeline apply/unapply optional components & features
- split monolithic `shared.py`
- cleanup `/modules`: move pipeline loaders to `/pipelines` root
- cleanup `/modules`: move code folders used by pipelines to `/pipelines/<pipeline>` folder
- cleanup `/modules`: move code folders used by scripts to `/scripts/<script>` folder
- cleanup `/modules`: global rename `modules.scripts` to avoid conflict with `/scripts`
- override `gradio` installer
- major refactoring of requirements and dependencies to unblock `numpy>=2.1.0`
- patch `insightface`
- patch `facelib`
- patch `numpy`
- stronger lint rules
add separate `npm run lint`, `npm run todo`, `npm run test`, `npm run format` macros
## Update for 2025-06-30
### Highlights for 2025-06-30
@@ -56,7 +277,7 @@ And (as always) many bugfixes and improvements to existing features!
- Control add setting to run hires with or without control
- Update OpenVINO to 2025.2.0
- Simplified and unified quantization enabled for options
- Add PixelArt filter to processing tab
- Add **PixelArt** filter to processing tab
- **SDNQ Quantization**
- Add `auto` quantization mode
- Add `modules_to_not_convert` support for post mode
@@ -3537,7 +3758,6 @@ Also new is support for **SDXL-Turbo** as well as new **Kandinsky 3** models and
- lightweight native implementation of T2I adapters which can guide generation towards specific image style
- supports most T2I models, not limited to SD 1.5
- models are auto-downloaded on first use
- for IP adapter support in *Original* backend, use standard *ControlNet* extension
- **AnimateDiff**
- lightweight native implementation of AnimateDiff models:
*AnimateDiff 1.4, 1.5 v1, 1.5 v2, AnimateFace*
@@ -3545,7 +3765,6 @@ Also new is support for **SDXL-Turbo** as well as new **Kandinsky 3** models and
- models are auto-downloaded on first use
- for video saving support, see video support section
- can be combined with IP-Adapter for even better results!
- for AnimateDiff support in *Original* backend, use standard *AnimateDiff* extension
- **HDR latent control**, based on [article](https://huggingface.co/blog/TimothyAlexisVass/explaining-the-sdxl-latent-space#long-prompts-at-high-guidance-scales-becoming-possible)
- in *Advanced* params
- allows control of *latent clamping*, *color centering* and *range maximization*
CITATION.cff
+1 -1
View File
@@ -21,7 +21,7 @@ repository-code: 'https://github.com/vladmandic/sdnext'
abstract: >-
SD.Next: Advanced Implementation of Stable Diffusion
and other diffusion models for text, image and video
generation
generation
keywords:
- stablediffusion diffusers sdnext
license: Apache-2.0
CODE_OF_CONDUCT
+10 -10
View File
@@ -1,17 +1,17 @@
# Code of Conduct
Use your best judgement
Use your best judgement
If it will possibly make others uncomfortable, do not post it
- Be respectful
Disagreement is not an opportunity to attack someone else's thoughts or opinions
Although views may differ, remember to approach every situation with patience and care
- Be considerate
Think about how your contribution will affect others in the community
- Be open minded
Embrace new people and new ideas. Our community is continually evolving and we welcome positive change
- Be respectful
Disagreement is not an opportunity to attack someone else's thoughts or opinions
Although views may differ, remember to approach every situation with patience and care
- Be considerate
Think about how your contribution will affect others in the community
- Be open minded
Embrace new people and new ideas. Our community is continually evolving and we welcome positive change
Be mindful of your language
Be mindful of your language
Any of the following behavior is unacceptable:
- Offensive comments of any kind
@@ -20,7 +20,7 @@ Any of the following behavior is unacceptable:
If you believe someone is violating the code of conduct, we ask that you report it
Participants asked to stop any harassing behavior are expected to comply immediately
Participants asked to stop any harassing behavior are expected to comply immediately
<br>
CONTRIBUTING
+20 -20
View File
@@ -4,24 +4,24 @@ Pull requests from everyone are welcome
Procedure for contributing:
- Select SD.Next `dev` branch:
<https://github.com/vladmandic/sdnext/tree/dev>
- Create a fork of the repository on github
In a top right corner of a GitHub, select "Fork"
Its recommended to fork latest version from main branch to avoid any possible conflicting code updates
- Clone your forked repository to your local system
`git clone https://github.com/<your-username>/<your-fork>`
- Make your changes
- Test your changes
- Lint your changes against code guidelines
- `ruff check`
- `pylint <folder>/<filename>.py`
- Push changes to your fork
- Submit a PR (pull request)
- Make sure that PR is against `dev` branch
- Update your fork before createing PR so that it is based on latest code
- Make sure that PR does NOT include any unrelated edits
- Make sure that PR does not include changes to submodules
- Select SD.Next `dev` branch:
<https://github.com/vladmandic/sdnext/tree/dev>
- Create a fork of the repository on github
In a top right corner of a GitHub, select "Fork"
Its recommended to fork latest version from main branch to avoid any possible conflicting code updates
- Clone your forked repository to your local system
`git clone https://github.com/<your-username>/<your-fork>`
- Make your changes
- Test your changes
- Lint your changes against code guidelines
- `ruff check`
- `pylint <folder>/<filename>.py`
- Push changes to your fork
- Submit a PR (pull request)
- Make sure that PR is against `dev` branch
- Update your fork before createing PR so that it is based on latest code
- Make sure that PR does NOT include any unrelated edits
- Make sure that PR does not include changes to submodules
Your pull request will be reviewed and pending review results, merged into `dev` branch
Dev merges to main are performed regularly and any PRs that are merged to `dev` will be included in the next main release
Your pull request will be reviewed and pending review results, merged into `dev` branch
Dev merges to main are performed regularly and any PRs that are merged to `dev` will be included in the next main release
LICENSE.txt
+201 -661
View File
@@ -1,661 +1,201 @@
GNU AFFERO GENERAL PUBLIC LICENSE
Version 3, 19 November 2007
Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
Preamble
The GNU Affero General Public License is a free, copyleft license for
software and other kinds of works, specifically designed to ensure
cooperation with the community in the case of network server software.
The licenses for most software and other practical works are designed
to take away your freedom to share and change the works. By contrast,
our General Public Licenses are intended to guarantee your freedom to
share and change all versions of a program--to make sure it remains free
software for all its users.
When we speak of free software, we are referring to freedom, not
price. Our General Public Licenses are designed to make sure that you
have the freedom to distribute copies of free software (and charge for
them if you wish), that you receive source code or can get it if you
want it, that you can change the software or use pieces of it in new
free programs, and that you know you can do these things.
Developers that use our General Public Licenses protect your rights
with two steps: (1) assert copyright on the software, and (2) offer
you this License which gives you legal permission to copy, distribute
and/or modify the software.
A secondary benefit of defending all users' freedom is that
improvements made in alternate versions of the program, if they
receive widespread use, become available for other developers to
incorporate. Many developers of free software are heartened and
encouraged by the resulting cooperation. However, in the case of
software used on network servers, this result may fail to come about.
The GNU General Public License permits making a modified version and
letting the public access it on a server without ever releasing its
source code to the public.
The GNU Affero General Public License is designed specifically to
ensure that, in such cases, the modified source code becomes available
to the community. It requires the operator of a network server to
provide the source code of the modified version running there to the
users of that server. Therefore, public use of a modified version, on
a publicly accessible server, gives the public access to the source
code of the modified version.
An older license, called the Affero General Public License and
published by Affero, was designed to accomplish similar goals. This is
a different license, not a version of the Affero GPL, but Affero has
released a new version of the Affero GPL which permits relicensing under
this license.
The precise terms and conditions for copying, distribution and
modification follow.
TERMS AND CONDITIONS
0. Definitions.
"This License" refers to version 3 of the GNU Affero General Public License.
"Copyright" also means copyright-like laws that apply to other kinds of
works, such as semiconductor masks.
"The Program" refers to any copyrightable work licensed under this
License. Each licensee is addressed as "you". "Licensees" and
"recipients" may be individuals or organizations.
To "modify" a work means to copy from or adapt all or part of the work
in a fashion requiring copyright permission, other than the making of an
exact copy. The resulting work is called a "modified version" of the
earlier work or a work "based on" the earlier work.
A "covered work" means either the unmodified Program or a work based
on the Program.
To "propagate" a work means to do anything with it that, without
permission, would make you directly or secondarily liable for
infringement under applicable copyright law, except executing it on a
computer or modifying a private copy. Propagation includes copying,
distribution (with or without modification), making available to the
public, and in some countries other activities as well.
To "convey" a work means any kind of propagation that enables other
parties to make or receive copies. Mere interaction with a user through
a computer network, with no transfer of a copy, is not conveying.
An interactive user interface displays "Appropriate Legal Notices"
to the extent that it includes a convenient and prominently visible
feature that (1) displays an appropriate copyright notice, and (2)
tells the user that there is no warranty for the work (except to the
extent that warranties are provided), that licensees may convey the
work under this License, and how to view a copy of this License. If
the interface presents a list of user commands or options, such as a
menu, a prominent item in the list meets this criterion.
1. Source Code.
The "source code" for a work means the preferred form of the work
for making modifications to it. "Object code" means any non-source
form of a work.
A "Standard Interface" means an interface that either is an official
standard defined by a recognized standards body, or, in the case of
interfaces specified for a particular programming language, one that
is widely used among developers working in that language.
The "System Libraries" of an executable work include anything, other
than the work as a whole, that (a) is included in the normal form of
packaging a Major Component, but which is not part of that Major
Component, and (b) serves only to enable use of the work with that
Major Component, or to implement a Standard Interface for which an
implementation is available to the public in source code form. A
"Major Component", in this context, means a major essential component
(kernel, window system, and so on) of the specific operating system
(if any) on which the executable work runs, or a compiler used to
produce the work, or an object code interpreter used to run it.
The "Corresponding Source" for a work in object code form means all
the source code needed to generate, install, and (for an executable
work) run the object code and to modify the work, including scripts to
control those activities. However, it does not include the work's
System Libraries, or general-purpose tools or generally available free
programs which are used unmodified in performing those activities but
which are not part of the work. For example, Corresponding Source
includes interface definition files associated with source files for
the work, and the source code for shared libraries and dynamically
linked subprograms that the work is specifically designed to require,
such as by intimate data communication or control flow between those
subprograms and other parts of the work.
The Corresponding Source need not include anything that users
can regenerate automatically from other parts of the Corresponding
Source.
The Corresponding Source for a work in source code form is that
same work.
2. Basic Permissions.
All rights granted under this License are granted for the term of
copyright on the Program, and are irrevocable provided the stated
conditions are met. This License explicitly affirms your unlimited
permission to run the unmodified Program. The output from running a
covered work is covered by this License only if the output, given its
content, constitutes a covered work. This License acknowledges your
rights of fair use or other equivalent, as provided by copyright law.
You may make, run and propagate covered works that you do not
convey, without conditions so long as your license otherwise remains
in force. You may convey covered works to others for the sole purpose
of having them make modifications exclusively for you, or provide you
with facilities for running those works, provided that you comply with
the terms of this License in conveying all material for which you do
not control copyright. Those thus making or running the covered works
for you must do so exclusively on your behalf, under your direction
and control, on terms that prohibit them from making any copies of
your copyrighted material outside their relationship with you.
Conveying under any other circumstances is permitted solely under
the conditions stated below. Sublicensing is not allowed; section 10
makes it unnecessary.
3. Protecting Users' Legal Rights From Anti-Circumvention Law.
No covered work shall be deemed part of an effective technological
measure under any applicable law fulfilling obligations under article
11 of the WIPO copyright treaty adopted on 20 December 1996, or
similar laws prohibiting or restricting circumvention of such
measures.
When you convey a covered work, you waive any legal power to forbid
circumvention of technological measures to the extent such circumvention
is effected by exercising rights under this License with respect to
the covered work, and you disclaim any intention to limit operation or
modification of the work as a means of enforcing, against the work's
users, your or third parties' legal rights to forbid circumvention of
technological measures.
4. Conveying Verbatim Copies.
You may convey verbatim copies of the Program's source code as you
receive it, in any medium, provided that you conspicuously and
appropriately publish on each copy an appropriate copyright notice;
keep intact all notices stating that this License and any
non-permissive terms added in accord with section 7 apply to the code;
keep intact all notices of the absence of any warranty; and give all
recipients a copy of this License along with the Program.
You may charge any price or no price for each copy that you convey,
and you may offer support or warranty protection for a fee.
5. Conveying Modified Source Versions.
You may convey a work based on the Program, or the modifications to
produce it from the Program, in the form of source code under the
terms of section 4, provided that you also meet all of these conditions:
a) The work must carry prominent notices stating that you modified
it, and giving a relevant date.
b) The work must carry prominent notices stating that it is
released under this License and any conditions added under section
7. This requirement modifies the requirement in section 4 to
"keep intact all notices".
c) You must license the entire work, as a whole, under this
License to anyone who comes into possession of a copy. This
License will therefore apply, along with any applicable section 7
additional terms, to the whole of the work, and all its parts,
regardless of how they are packaged. This License gives no
permission to license the work in any other way, but it does not
invalidate such permission if you have separately received it.
d) If the work has interactive user interfaces, each must display
Appropriate Legal Notices; however, if the Program has interactive
interfaces that do not display Appropriate Legal Notices, your
work need not make them do so.
A compilation of a covered work with other separate and independent
works, which are not by their nature extensions of the covered work,
and which are not combined with it such as to form a larger program,
in or on a volume of a storage or distribution medium, is called an
"aggregate" if the compilation and its resulting copyright are not
used to limit the access or legal rights of the compilation's users
beyond what the individual works permit. Inclusion of a covered work
in an aggregate does not cause this License to apply to the other
parts of the aggregate.
6. Conveying Non-Source Forms.
You may convey a covered work in object code form under the terms
of sections 4 and 5, provided that you also convey the
machine-readable Corresponding Source under the terms of this License,
in one of these ways:
a) Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by the
Corresponding Source fixed on a durable physical medium
customarily used for software interchange.
b) Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by a
written offer, valid for at least three years and valid for as
long as you offer spare parts or customer support for that product
model, to give anyone who possesses the object code either (1) a
copy of the Corresponding Source for all the software in the
product that is covered by this License, on a durable physical
medium customarily used for software interchange, for a price no
more than your reasonable cost of physically performing this
conveying of source, or (2) access to copy the
Corresponding Source from a network server at no charge.
c) Convey individual copies of the object code with a copy of the
written offer to provide the Corresponding Source. This
alternative is allowed only occasionally and noncommercially, and
only if you received the object code with such an offer, in accord
with subsection 6b.
d) Convey the object code by offering access from a designated
place (gratis or for a charge), and offer equivalent access to the
Corresponding Source in the same way through the same place at no
further charge. You need not require recipients to copy the
Corresponding Source along with the object code. If the place to
copy the object code is a network server, the Corresponding Source
may be on a different server (operated by you or a third party)
that supports equivalent copying facilities, provided you maintain
clear directions next to the object code saying where to find the
Corresponding Source. Regardless of what server hosts the
Corresponding Source, you remain obligated to ensure that it is
available for as long as needed to satisfy these requirements.
e) Convey the object code using peer-to-peer transmission, provided
you inform other peers where the object code and Corresponding
Source of the work are being offered to the general public at no
charge under subsection 6d.
A separable portion of the object code, whose source code is excluded
from the Corresponding Source as a System Library, need not be
included in conveying the object code work.
A "User Product" is either (1) a "consumer product", which means any
tangible personal property which is normally used for personal, family,
or household purposes, or (2) anything designed or sold for incorporation
into a dwelling. In determining whether a product is a consumer product,
doubtful cases shall be resolved in favor of coverage. For a particular
product received by a particular user, "normally used" refers to a
typical or common use of that class of product, regardless of the status
of the particular user or of the way in which the particular user
actually uses, or expects or is expected to use, the product. A product
is a consumer product regardless of whether the product has substantial
commercial, industrial or non-consumer uses, unless such uses represent
the only significant mode of use of the product.
"Installation Information" for a User Product means any methods,
procedures, authorization keys, or other information required to install
and execute modified versions of a covered work in that User Product from
a modified version of its Corresponding Source. The information must
suffice to ensure that the continued functioning of the modified object
code is in no case prevented or interfered with solely because
modification has been made.
If you convey an object code work under this section in, or with, or
specifically for use in, a User Product, and the conveying occurs as
part of a transaction in which the right of possession and use of the
User Product is transferred to the recipient in perpetuity or for a
fixed term (regardless of how the transaction is characterized), the
Corresponding Source conveyed under this section must be accompanied
by the Installation Information. But this requirement does not apply
if neither you nor any third party retains the ability to install
modified object code on the User Product (for example, the work has
been installed in ROM).
The requirement to provide Installation Information does not include a
requirement to continue to provide support service, warranty, or updates
for a work that has been modified or installed by the recipient, or for
the User Product in which it has been modified or installed. Access to a
network may be denied when the modification itself materially and
adversely affects the operation of the network or violates the rules and
protocols for communication across the network.
Corresponding Source conveyed, and Installation Information provided,
in accord with this section must be in a format that is publicly
documented (and with an implementation available to the public in
source code form), and must require no special password or key for
unpacking, reading or copying.
7. Additional Terms.
"Additional permissions" are terms that supplement the terms of this
License by making exceptions from one or more of its conditions.
Additional permissions that are applicable to the entire Program shall
be treated as though they were included in this License, to the extent
that they are valid under applicable law. If additional permissions
apply only to part of the Program, that part may be used separately
under those permissions, but the entire Program remains governed by
this License without regard to the additional permissions.
When you convey a copy of a covered work, you may at your option
remove any additional permissions from that copy, or from any part of
it. (Additional permissions may be written to require their own
removal in certain cases when you modify the work.) You may place
additional permissions on material, added by you to a covered work,
for which you have or can give appropriate copyright permission.
Notwithstanding any other provision of this License, for material you
add to a covered work, you may (if authorized by the copyright holders of
that material) supplement the terms of this License with terms:
a) Disclaiming warranty or limiting liability differently from the
terms of sections 15 and 16 of this License; or
b) Requiring preservation of specified reasonable legal notices or
author attributions in that material or in the Appropriate Legal
Notices displayed by works containing it; or
c) Prohibiting misrepresentation of the origin of that material, or
requiring that modified versions of such material be marked in
reasonable ways as different from the original version; or
d) Limiting the use for publicity purposes of names of licensors or
authors of the material; or
e) Declining to grant rights under trademark law for use of some
trade names, trademarks, or service marks; or
f) Requiring indemnification of licensors and authors of that
material by anyone who conveys the material (or modified versions of
it) with contractual assumptions of liability to the recipient, for
any liability that these contractual assumptions directly impose on
those licensors and authors.
All other non-permissive additional terms are considered "further
restrictions" within the meaning of section 10. If the Program as you
received it, or any part of it, contains a notice stating that it is
governed by this License along with a term that is a further
restriction, you may remove that term. If a license document contains
a further restriction but permits relicensing or conveying under this
License, you may add to a covered work material governed by the terms
of that license document, provided that the further restriction does
not survive such relicensing or conveying.
If you add terms to a covered work in accord with this section, you
must place, in the relevant source files, a statement of the
additional terms that apply to those files, or a notice indicating
where to find the applicable terms.
Additional terms, permissive or non-permissive, may be stated in the
form of a separately written license, or stated as exceptions;
the above requirements apply either way.
8. Termination.
You may not propagate or modify a covered work except as expressly
provided under this License. Any attempt otherwise to propagate or
modify it is void, and will automatically terminate your rights under
this License (including any patent licenses granted under the third
paragraph of section 11).
However, if you cease all violation of this License, then your
license from a particular copyright holder is reinstated (a)
provisionally, unless and until the copyright holder explicitly and
finally terminates your license, and (b) permanently, if the copyright
holder fails to notify you of the violation by some reasonable means
prior to 60 days after the cessation.
Moreover, your license from a particular copyright holder is
reinstated permanently if the copyright holder notifies you of the
violation by some reasonable means, this is the first time you have
received notice of violation of this License (for any work) from that
copyright holder, and you cure the violation prior to 30 days after
your receipt of the notice.
Termination of your rights under this section does not terminate the
licenses of parties who have received copies or rights from you under
this License. If your rights have been terminated and not permanently
reinstated, you do not qualify to receive new licenses for the same
material under section 10.
9. Acceptance Not Required for Having Copies.
You are not required to accept this License in order to receive or
run a copy of the Program. Ancillary propagation of a covered work
occurring solely as a consequence of using peer-to-peer transmission
to receive a copy likewise does not require acceptance. However,
nothing other than this License grants you permission to propagate or
modify any covered work. These actions infringe copyright if you do
not accept this License. Therefore, by modifying or propagating a
covered work, you indicate your acceptance of this License to do so.
10. Automatic Licensing of Downstream Recipients.
Each time you convey a covered work, the recipient automatically
receives a license from the original licensors, to run, modify and
propagate that work, subject to this License. You are not responsible
for enforcing compliance by third parties with this License.
An "entity transaction" is a transaction transferring control of an
organization, or substantially all assets of one, or subdividing an
organization, or merging organizations. If propagation of a covered
work results from an entity transaction, each party to that
transaction who receives a copy of the work also receives whatever
licenses to the work the party's predecessor in interest had or could
give under the previous paragraph, plus a right to possession of the
Corresponding Source of the work from the predecessor in interest, if
the predecessor has it or can get it with reasonable efforts.
You may not impose any further restrictions on the exercise of the
rights granted or affirmed under this License. For example, you may
not impose a license fee, royalty, or other charge for exercise of
rights granted under this License, and you may not initiate litigation
(including a cross-claim or counterclaim in a lawsuit) alleging that
any patent claim is infringed by making, using, selling, offering for
sale, or importing the Program or any portion of it.
11. Patents.
A "contributor" is a copyright holder who authorizes use under this
License of the Program or a work on which the Program is based. The
work thus licensed is called the contributor's "contributor version".
A contributor's "essential patent claims" are all patent claims
owned or controlled by the contributor, whether already acquired or
hereafter acquired, that would be infringed by some manner, permitted
by this License, of making, using, or selling its contributor version,
but do not include claims that would be infringed only as a
consequence of further modification of the contributor version. For
purposes of this definition, "control" includes the right to grant
patent sublicenses in a manner consistent with the requirements of
this License.
Each contributor grants you a non-exclusive, worldwide, royalty-free
patent license under the contributor's essential patent claims, to
make, use, sell, offer for sale, import and otherwise run, modify and
propagate the contents of its contributor version.
In the following three paragraphs, a "patent license" is any express
agreement or commitment, however denominated, not to enforce a patent
(such as an express permission to practice a patent or covenant not to
sue for patent infringement). To "grant" such a patent license to a
party means to make such an agreement or commitment not to enforce a
patent against the party.
If you convey a covered work, knowingly relying on a patent license,
and the Corresponding Source of the work is not available for anyone
to copy, free of charge and under the terms of this License, through a
publicly available network server or other readily accessible means,
then you must either (1) cause the Corresponding Source to be so
available, or (2) arrange to deprive yourself of the benefit of the
patent license for this particular work, or (3) arrange, in a manner
consistent with the requirements of this License, to extend the patent
license to downstream recipients. "Knowingly relying" means you have
actual knowledge that, but for the patent license, your conveying the
covered work in a country, or your recipient's use of the covered work
in a country, would infringe one or more identifiable patents in that
country that you have reason to believe are valid.
If, pursuant to or in connection with a single transaction or
arrangement, you convey, or propagate by procuring conveyance of, a
covered work, and grant a patent license to some of the parties
receiving the covered work authorizing them to use, propagate, modify
or convey a specific copy of the covered work, then the patent license
you grant is automatically extended to all recipients of the covered
work and works based on it.
A patent license is "discriminatory" if it does not include within
the scope of its coverage, prohibits the exercise of, or is
conditioned on the non-exercise of one or more of the rights that are
specifically granted under this License. You may not convey a covered
work if you are a party to an arrangement with a third party that is
in the business of distributing software, under which you make payment
to the third party based on the extent of your activity of conveying
the work, and under which the third party grants, to any of the
parties who would receive the covered work from you, a discriminatory
patent license (a) in connection with copies of the covered work
conveyed by you (or copies made from those copies), or (b) primarily
for and in connection with specific products or compilations that
contain the covered work, unless you entered into that arrangement,
or that patent license was granted, prior to 28 March 2007.
Nothing in this License shall be construed as excluding or limiting
any implied license or other defenses to infringement that may
otherwise be available to you under applicable patent law.
12. No Surrender of Others' Freedom.
If conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License. If you cannot convey a
covered work so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you may
not convey it at all. For example, if you agree to terms that obligate you
to collect a royalty for further conveying from those to whom you convey
the Program, the only way you could satisfy both those terms and this
License would be to refrain entirely from conveying the Program.
13. Remote Network Interaction; Use with the GNU General Public License.
Notwithstanding any other provision of this License, if you modify the
Program, your modified version must prominently offer all users
interacting with it remotely through a computer network (if your version
supports such interaction) an opportunity to receive the Corresponding
Source of your version by providing access to the Corresponding Source
from a network server at no charge, through some standard or customary
means of facilitating copying of software. This Corresponding Source
shall include the Corresponding Source for any work covered by version 3
of the GNU General Public License that is incorporated pursuant to the
following paragraph.
Notwithstanding any other provision of this License, you have
permission to link or combine any covered work with a work licensed
under version 3 of the GNU General Public License into a single
combined work, and to convey the resulting work. The terms of this
License will continue to apply to the part which is the covered work,
but the work with which it is combined will remain governed by version
3 of the GNU General Public License.
14. Revised Versions of this License.
The Free Software Foundation may publish revised and/or new versions of
the GNU Affero General Public License from time to time. Such new versions
will be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.
Each version is given a distinguishing version number. If the
Program specifies that a certain numbered version of the GNU Affero General
Public License "or any later version" applies to it, you have the
option of following the terms and conditions either of that numbered
version or of any later version published by the Free Software
Foundation. If the Program does not specify a version number of the
GNU Affero General Public License, you may choose any version ever published
by the Free Software Foundation.
If the Program specifies that a proxy can decide which future
versions of the GNU Affero General Public License can be used, that proxy's
public statement of acceptance of a version permanently authorizes you
to choose that version for the Program.
Later license versions may give you additional or different
permissions. However, no additional obligations are imposed on any
author or copyright holder as a result of your choosing to follow a
later version.
15. Disclaimer of Warranty.
THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
16. Limitation of Liability.
IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
SUCH DAMAGES.
17. Interpretation of Sections 15 and 16.
If the disclaimer of warranty and limitation of liability provided
above cannot be given local legal effect according to their terms,
reviewing courts shall apply local law that most closely approximates
an absolute waiver of all civil liability in connection with the
Program, unless a warranty or assumption of liability accompanies a
copy of the Program in return for a fee.
END OF TERMS AND CONDITIONS
How to Apply These Terms to Your New Programs
If you develop a new program, and you want it to be of the greatest
possible use to the public, the best way to achieve this is to make it
free software which everyone can redistribute and change under these terms.
To do so, attach the following notices to the program. It is safest
to attach them to the start of each source file to most effectively
state the exclusion of warranty; and each file should have at least
the "copyright" line and a pointer to where the full notice is found.
<one line to give the program's name and a brief idea of what it does.>
Copyright (C) <year> <name of author>
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as published
by the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <https://www.gnu.org/licenses/>.
Also add information on how to contact you by electronic and paper mail.
If your software can interact with users remotely through a computer
network, you should also make sure that it provides a way for users to
get its source. For example, if your program is a web application, its
interface could display a "Source" link that leads users to an archive
of the code. There are many ways you could offer source, and different
solutions will be better for different programs; see section 13 for the
specific requirements.
You should also get your employer (if you work as a programmer) or school,
if any, to sign a "copyright disclaimer" for the program, if necessary.
For more information on this, and how to apply and follow the GNU AGPL, see
<https://www.gnu.org/licenses/>.
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
APPENDIX: How to apply the Apache License to your work.
To apply the Apache License to your work, attach the following
boilerplate notice, with the fields enclosed by brackets "[]"
replaced with your own identifying information. (Don't include
the brackets!) The text should be enclosed in the appropriate
comment syntax for the file format. We also recommend that a
file or class name and description of purpose be included on the
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright [yyyy] [name of copyright owner]
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
README.md
+2 -2
View File
@@ -77,8 +77,8 @@ Plus Docker container receipes for: [CUDA, ROCm, Intel IPEX and OpenVINO](https:
- Get started with **SD.Next** by following the [installation instructions](https://vladmandic.github.io/sdnext-docs/Installation/)
- For more details, check out [advanced installation](https://vladmandic.github.io/sdnext-docs/Advanced-Install/) guide
- List and explanation of [command line arguments](https://vladmandic.github.io/sdnext-docs/CLI-Arguments/)
- Install walkthrough [video](https://www.youtube.com/watch?v=nWTnTyFTuAs)
- List and explanation of [command line arguments](https://vladmandic.github.io/sdnext-docs/CLI-Arguments/)
- Install walkthrough [video](https://www.youtube.com/watch?v=nWTnTyFTuAs)
> [!TIP]
> And for platform specific information, check out
SECURITY.md
+2 -2
View File
@@ -29,8 +29,8 @@ Any code commit is validated before merge
`SD.Next` library can establish external connections *only* for following purposes and *only* when explicitly configured by user:
- Download extensions and themes indexes from automatically updated indexes
- Download extensions and themes indexes from automatically updated indexes
- Download required packages and repositories from GitHub during installation/upgrade
- Download installed/enabled extensions
- Download models from CivitAI and/or Huggingface when instructed by user
- Submit benchmark info upon user interaction
- Submit benchmark info upon user interaction
TODO.md
+30 -28
View File
@@ -6,51 +6,52 @@ Main ToDo list can be found at [GitHub projects](https://github.com/users/vladma
## Future Candidates
- Refactor: Move `model_*` stuff into subfolder
- Refactor: sampler options
- Common repo for `T5` and `CLiP`
- Upgrade: unblock `numpy`: see `gradio`
- Upgrade: unblock `pydantic`: see <https://github.com/Cschlaefli/automatic>
- [Modular pipelines and guiders](https://github.com/huggingface/diffusers/issues/11915)
- Refactor: Sampler options
- Feature: Diffusers [group offloading](https://github.com/vladmandic/sdnext/issues/4049)
- Feature: Common repo for `T5` and `CLiP`
- Feature: LoRA add OMI format support for SD35/FLUX.1
- Video: Generic API support
- Video: LTX TeaCache and others
- Video: LTX API
- Video: LTX PromptEnhance
- Video: LTX Conditioning preprocess
- [WanAI-2.1 VACE](https://huggingface.co/Wan-AI/Wan2.1-VACE-14B)(https://github.com/huggingface/diffusers/pull/11582)
- [SkyReels-v2](https://github.com/SkyworkAI/SkyReels-V2)(https://github.com/huggingface/diffusers/pull/11518)
- [Cosmos-Predict2-Video](https://huggingface.co/nvidia/Cosmos-Predict2-2B-Video2World)(https://github.com/huggingface/diffusers/pull/11695)
### Complete Features
### Blocked items
- Python==3.13 improved support
- Video: API support
- LoRA: add OMI format support for SD35/FLUX.1
- Upgrade: unblock `pydantic` and `albumentations`
- see <https://github.com/Cschlaefli/automatic>
- blocked by `insightface`
### Under Consideration
- [IPAdapter negative guidance](https://github.com/huggingface/diffusers/discussions/7167)
- [IPAdapter composition](https://huggingface.co/ostris/ip-composition-adapter)
- [Refactor attention](https://github.com/huggingface/diffusers/pull/11311)
- [STG](https://github.com/huggingface/diffusers/blob/main/examples/community/README.md#spatiotemporal-skip-guidance)
- [LBM](https://github.com/gojasper/LBM)
- [SmoothCache](https://github.com/huggingface/diffusers/issues/11135)
- [MagCache](https://github.com/lllyasviel/FramePack/pull/673/files)
- [HiDream GGUF](https://github.com/huggingface/diffusers/pull/11550)
- [Diffusers guiders](https://github.com/huggingface/diffusers/pull/11311)
- [Nunchaku PulID](https://github.com/mit-han-lab/nunchaku/pull/274)
- [Dream0 guidance](https://huggingface.co/ByteDance/DreamO)
- [S3Diff diffusion upscaler](https://github.com/ArcticHare105/S3Diff)
- [SUPIR upscaler](https://github.com/Fanghua-Yu/SUPIR)
- Remove: Agent Scheduler
- Remove: CodeFormer
- Remove: GFPGAN
- ModernUI: Lite vs Expert mode
- [Canvas](https://konvajs.org/)
### Monitoring
- [TensorRT](https://github.com/huggingface/diffusers/pull/11173)
### Future Considerations
- [TensorRT](https://github.com/huggingface/diffusers/pull/11173)
### New models
#### Stable
- [Diffusers-0.34.0](https://github.com/huggingface/diffusers/releases/tag/v0.34.0)
- [WanAI-2.1 VACE](https://huggingface.co/Wan-AI/Wan2.1-VACE-14B)(https://github.com/huggingface/diffusers/pull/11582)
- [LTXVideo-0.9.7](https://github.com/Lightricks/LTX-Video?tab=readme-ov-file#diffusers-integration)(https://github.com/huggingface/diffusers/pull/11516)
- [Cosmos-Predict2-Video](https://huggingface.co/nvidia/Cosmos-Predict2-2B-Video2World)(https://github.com/huggingface/diffusers/pull/11695)
#### Pending
- [Magi](https://github.com/SandAI-org/MAGI-1)(https://github.com/huggingface/diffusers/pull/11713)
- [SEVA](https://github.com/huggingface/diffusers/pull/11440)
- [SkyReels-v2](https://github.com/SkyworkAI/SkyReels-V2)(https://github.com/huggingface/diffusers/pull/11518)
#### External:Unified/MultiModal
- [Bagel](https://huggingface.co/ByteDance-Seed/BAGEL-7B-MoT)(https://github.com/bytedance-seed/bagel)
- [Ming](https://github.com/inclusionAI/Ming)
- [Liquid](https://github.com/FoundationVision/Liquid)
#### External:Image2Image/Editing
@@ -69,11 +70,12 @@ Main ToDo list can be found at [GitHub projects](https://github.com/users/vladma
## Code TODO
> pnpm lint | grep W0511 | awk -F'TODO ' '{print "- "$NF}' | sed 's/ (fixme)//g'
> pnpm lint | grep W0511 | awk -F'TODO ' '{print "- "$NF}' | sed 's/ (fixme)//g' | sort
- control: support scripts via api
- fc: autodetect distilled based on model
- fc: autodetect tensor format based on model
- flux: loader for civitai nf4 models
- hypertile: vae breaks when using non-standard sizes
- install: enable ROCm for windows when available
- loader: load receipe
@@ -81,11 +83,11 @@ Main ToDo list can be found at [GitHub projects](https://github.com/users/vladma
- lora: add other quantization types
- lora: add t5 key support for sd35/f1
- lora: maybe force imediate quantization
- lora: support pre-quantized flux
- model load: cogview4: balanced offload does not work for GlmModel
- model load: add ChromaFillPipeline, ChromaControlPipeline, ChromaImg2ImgPipeline etc when available
- model load: force-reloading entire model as loading transformers only leads to massive memory usage
- model loader: implement model in-memory caching
- model load: implement model in-memory caching
- modernui: monkey-patch for missing tabs.select event
- modules/lora/lora_extract.py:188:9: W0511: TODO: lora: support pre-quantized flux
- nunchaku: batch support
- nunchaku: cache-dir for transformer and t5 loader
- processing: remove duplicate mask params
- resize image: enable full VAE mode for resize-latent
cli/api-txt2img.py
+1 -1
View File
@@ -63,7 +63,7 @@ def generate(args): # pylint: disable=redefined-outer-name
info = data['info']
log.info(f'image received: size={image.size} time={t1-t0:.2f} info="{info}"')
if args.output:
image.save(args.output)
image.save(args.output, exif=image._getexif())
log.info(f'image saved: size={image.size} filename={args.output}')
else:
log.warning(f'no images received: {data}')
cli/create-previews.py
-45
View File
@@ -61,10 +61,6 @@ options = Map({
'lora': {
'strength': 1.0,
},
'hypernetwork': {
'keyword': '',
'strength': 1.0,
},
})
@@ -243,46 +239,6 @@ async def lyco(params):
log.info({ 'lyco preview created': model, 'image': fn, 'images': len(images), 'grid': [image.width, image.height], 'time': round(t, 2), 'its': round(its, 2) })
async def hypernetwork(params):
opt = await get('/sdapi/v1/options')
folder = opt['hypernetwork_dir']
if not os.path.exists(folder):
log.error({ 'hypernetwork directory not found': folder })
return
models = [os.path.splitext(f)[0] for f in Path(folder).glob('**/*.pt')]
log.info({ 'hypernetworks': len(models) })
for model in models:
if preview_exists(folder, model) and len(params.input) == 0: # if model preview exists and not manually included
log.info({ 'hypernetwork preview exists': model })
continue
fn = os.path.join(folder, model + options.format)
images = []
labels = []
t0 = time.time()
keyword = options.hypernetwork.keyword
options.generate.prompt = options.prompt.replace('<keyword>', options.hypernetwork.keyword)
options.generate.prompt = options.generate.prompt.replace('<embedding>', '')
options.generate.prompt = f' <hypernet:{model}:{options.hypernetwork.strength}> ' + options.generate.prompt
log.info({ 'hypernetwork generating': model, 'keyword': keyword, 'prompt': options.generate.prompt })
data = await generate(options = options, quiet=True)
if 'image' in data:
for img in data['image']:
images.append(img)
labels.append(keyword)
else:
log.error({ 'hypernetwork': model, 'keyword': keyword, 'error': data })
t1 = time.time()
if len(images) == 0:
log.error({ 'model': model, 'error': 'no images generated' })
continue
image = grid(images = images, labels = labels, border = 8)
log.info({ 'saving preview': fn, 'images': len(images), 'size': [image.width, image.height] })
image.save(fn)
t = t1 - t0
its = 1.0 * options.generate.steps * len(images) / t
log.info({ 'hypernetwork preview created': model, 'image': fn, 'images': len(images), 'grid': [image.width, image.height], 'time': round(t, 2), 'its': round(its, 2) })
async def embedding(params):
opt = await get('/sdapi/v1/options')
folder = opt['embeddings_dir']
@@ -327,7 +283,6 @@ async def create_previews(params):
await preview_models(params)
await lora(params)
await lyco(params)
await hypernetwork(params)
await embedding(params)
await close()
cli/generate.json
-1
View File
@@ -24,7 +24,6 @@
{
"upscaler_1": "SwinIR_4x",
"upscaler_2": "None",
"upscale_first": false,
"upscaling_resize": 0,
"gfpgan_visibility": 0,
"codeformer_visibility": 0,
configs/Dockerfile.cuda
+1 -1
View File
@@ -39,7 +39,7 @@ ENV SD_MODELSDIR="/mnt/models"
ENV SD_DOCKER=true
# tcmalloc is not required but it is highly recommended
ENV LD_PRELOAD=libtcmalloc.so.4
ENV LD_PRELOAD=libtcmalloc.so.4
# sdnext will run all necessary pip install ops and then exit
RUN ["python", "/app/launch.py", "--debug", "--uv", "--use-cuda", "--log", "sdnext.log", "--test", "--optional"]
# preinstall additional packages to avoid installation during runtime
configs/alt-diffusion-inference.yaml
-72
View File
@@ -1,72 +0,0 @@
model:
base_learning_rate: 1.0e-04
target: ldm.models.diffusion.ddpm.LatentDiffusion
params:
linear_start: 0.00085
linear_end: 0.0120
num_timesteps_cond: 1
log_every_t: 200
timesteps: 1000
first_stage_key: "jpg"
cond_stage_key: "txt"
image_size: 64
channels: 4
cond_stage_trainable: false # Note: different from the one we trained before
conditioning_key: crossattn
monitor: val/loss_simple_ema
scale_factor: 0.18215
use_ema: False
scheduler_config: # 10000 warmup steps
target: ldm.lr_scheduler.LambdaLinearScheduler
params:
warm_up_steps: [ 10000 ]
cycle_lengths: [ 10000000000000 ] # incredibly large number to prevent corner cases
f_start: [ 1.e-6 ]
f_max: [ 1. ]
f_min: [ 1. ]
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
image_size: 32 # unused
in_channels: 4
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_heads: 8
use_spatial_transformer: True
transformer_depth: 1
context_dim: 768
use_checkpoint: True
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: []
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: modules.xlmr.BertSeriesModelWithTransformation
params:
name: "XLMR-Large"
configs/flux/tokenizer_2/tokenizer.json
+1 -1
View File
@@ -129425,4 +129425,4 @@
],
"byte_fallback": false
}
}
}
configs/instruct-pix2pix.yaml
-98
View File
@@ -1,98 +0,0 @@
# File modified by authors of InstructPix2Pix from original (https://github.com/CompVis/stable-diffusion).
# See more details in LICENSE.
model:
base_learning_rate: 1.0e-04
target: modules.hijack.ddpm_edit.LatentDiffusion
params:
linear_start: 0.00085
linear_end: 0.0120
num_timesteps_cond: 1
log_every_t: 200
timesteps: 1000
first_stage_key: edited
cond_stage_key: edit
# image_size: 64
# image_size: 32
image_size: 16
channels: 4
cond_stage_trainable: false # Note: different from the one we trained before
conditioning_key: hybrid
monitor: val/loss_simple_ema
scale_factor: 0.18215
use_ema: false
scheduler_config: # 10000 warmup steps
target: ldm.lr_scheduler.LambdaLinearScheduler
params:
warm_up_steps: [ 0 ]
cycle_lengths: [ 10000000000000 ] # incredibly large number to prevent corner cases
f_start: [ 1.e-6 ]
f_max: [ 1. ]
f_min: [ 1. ]
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
image_size: 32 # unused
in_channels: 8
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_heads: 8
use_spatial_transformer: True
transformer_depth: 1
context_dim: 768
use_checkpoint: True
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: []
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: ldm.modules.encoders.modules.FrozenCLIPEmbedder
data:
target: main.DataModuleFromConfig
params:
batch_size: 128
num_workers: 1
wrap: false
validation:
target: edit_dataset.EditDataset
params:
path: data/clip-filtered-dataset
cache_dir: data/
cache_name: data_10k
split: val
min_text_sim: 0.2
min_image_sim: 0.75
min_direction_sim: 0.2
max_samples_per_prompt: 1
min_resize_res: 512
max_resize_res: 512
crop_res: 512
output_as_edit: False
real_input: True
configs/sd3/tokenizer_3/tokenizer.json
+1 -1
View File
@@ -129425,4 +129425,4 @@
],
"byte_fallback": false
}
}
}
configs/sd_xl_base.yaml
-98
View File
@@ -1,98 +0,0 @@
model:
target: sgm.models.diffusion.DiffusionEngine
params:
scale_factor: 0.13025
disable_first_stage_autocast: True
denoiser_config:
target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
params:
num_idx: 1000
weighting_config:
target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
scaling_config:
target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
discretization_config:
target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
network_config:
target: sgm.modules.diffusionmodules.openaimodel.UNetModel
params:
adm_in_channels: 2816
num_classes: sequential
use_checkpoint: True
in_channels: 4
out_channels: 4
model_channels: 320
attention_resolutions: [4, 2]
num_res_blocks: 2
channel_mult: [1, 2, 4]
num_head_channels: 64
use_spatial_transformer: True
use_linear_in_transformer: True
transformer_depth: [1, 2, 10] # note: the first is unused (due to attn_res starting at 2) 32, 16, 8 --> 64, 32, 16
context_dim: 2048
spatial_transformer_attn_type: softmax-xformers
legacy: False
conditioner_config:
target: sgm.modules.GeneralConditioner
params:
emb_models:
# crossattn cond
- is_trainable: False
input_key: txt
target: sgm.modules.encoders.modules.FrozenCLIPEmbedder
params:
layer: hidden
layer_idx: 11
# crossattn and vector cond
- is_trainable: False
input_key: txt
target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
params:
arch: ViT-bigG-14
version: laion2b_s39b_b160k
freeze: True
layer: penultimate
always_return_pooled: True
legacy: False
# vector cond
- is_trainable: False
input_key: original_size_as_tuple
target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
params:
outdim: 256 # multiplied by two
# vector cond
- is_trainable: False
input_key: crop_coords_top_left
target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
params:
outdim: 256 # multiplied by two
# vector cond
- is_trainable: False
input_key: target_size_as_tuple
target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
params:
outdim: 256 # multiplied by two
first_stage_config:
target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
attn_type: vanilla-xformers
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult: [1, 2, 4, 4]
num_res_blocks: 2
attn_resolutions: []
dropout: 0.0
lossconfig:
target: torch.nn.Identity
configs/sd_xl_refiner.yaml
-91
View File
@@ -1,91 +0,0 @@
model:
target: sgm.models.diffusion.DiffusionEngine
params:
scale_factor: 0.13025
disable_first_stage_autocast: True
denoiser_config:
target: sgm.modules.diffusionmodules.denoiser.DiscreteDenoiser
params:
num_idx: 1000
weighting_config:
target: sgm.modules.diffusionmodules.denoiser_weighting.EpsWeighting
scaling_config:
target: sgm.modules.diffusionmodules.denoiser_scaling.EpsScaling
discretization_config:
target: sgm.modules.diffusionmodules.discretizer.LegacyDDPMDiscretization
network_config:
target: sgm.modules.diffusionmodules.openaimodel.UNetModel
params:
adm_in_channels: 2560
num_classes: sequential
use_checkpoint: True
in_channels: 4
out_channels: 4
model_channels: 384
attention_resolutions: [4, 2]
num_res_blocks: 2
channel_mult: [1, 2, 4, 4]
num_head_channels: 64
use_spatial_transformer: True
use_linear_in_transformer: True
transformer_depth: 4
context_dim: [1280, 1280, 1280, 1280] # 1280
spatial_transformer_attn_type: softmax-xformers
legacy: False
conditioner_config:
target: sgm.modules.GeneralConditioner
params:
emb_models:
# crossattn and vector cond
- is_trainable: False
input_key: txt
target: sgm.modules.encoders.modules.FrozenOpenCLIPEmbedder2
params:
arch: ViT-bigG-14
version: laion2b_s39b_b160k
legacy: False
freeze: True
layer: penultimate
always_return_pooled: True
# vector cond
- is_trainable: False
input_key: original_size_as_tuple
target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
params:
outdim: 256 # multiplied by two
# vector cond
- is_trainable: False
input_key: crop_coords_top_left
target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
params:
outdim: 256 # multiplied by two
# vector cond
- is_trainable: False
input_key: aesthetic_score
target: sgm.modules.encoders.modules.ConcatTimestepEmbedderND
params:
outdim: 256 # multiplied by one
first_stage_config:
target: sgm.models.autoencoder.AutoencoderKLInferenceWrapper
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
attn_type: vanilla-xformers
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult: [1, 2, 4, 4]
num_res_blocks: 2
attn_resolutions: []
dropout: 0.0
lossconfig:
target: torch.nn.Identity
configs/v1-inference.yaml
-70
View File
@@ -1,70 +0,0 @@
model:
base_learning_rate: 1.0e-04
target: ldm.models.diffusion.ddpm.LatentDiffusion
params:
linear_start: 0.00085
linear_end: 0.0120
num_timesteps_cond: 1
log_every_t: 200
timesteps: 1000
first_stage_key: "jpg"
cond_stage_key: "txt"
image_size: 64
channels: 4
cond_stage_trainable: false # Note: different from the one we trained before
conditioning_key: crossattn
monitor: val/loss_simple_ema
scale_factor: 0.18215
use_ema: False
scheduler_config: # 10000 warmup steps
target: ldm.lr_scheduler.LambdaLinearScheduler
params:
warm_up_steps: [ 10000 ]
cycle_lengths: [ 10000000000000 ] # incredibly large number to prevent corner cases
f_start: [ 1.e-6 ]
f_max: [ 1. ]
f_min: [ 1. ]
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
image_size: 32 # unused
in_channels: 4
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_heads: 8
use_spatial_transformer: True
transformer_depth: 1
context_dim: 768
use_checkpoint: True
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: []
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: ldm.modules.encoders.modules.FrozenCLIPEmbedder
configs/v1-inpainting-inference.yaml
-70
View File
@@ -1,70 +0,0 @@
model:
base_learning_rate: 7.5e-05
target: ldm.models.diffusion.ddpm.LatentInpaintDiffusion
params:
linear_start: 0.00085
linear_end: 0.0120
num_timesteps_cond: 1
log_every_t: 200
timesteps: 1000
first_stage_key: "jpg"
cond_stage_key: "txt"
image_size: 64
channels: 4
cond_stage_trainable: false # Note: different from the one we trained before
conditioning_key: hybrid # important
monitor: val/loss_simple_ema
scale_factor: 0.18215
finetune_keys: null
scheduler_config: # 10000 warmup steps
target: ldm.lr_scheduler.LambdaLinearScheduler
params:
warm_up_steps: [ 2500 ] # NOTE for resuming. use 10000 if starting from scratch
cycle_lengths: [ 10000000000000 ] # incredibly large number to prevent corner cases
f_start: [ 1.e-6 ]
f_max: [ 1. ]
f_min: [ 1. ]
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
image_size: 32 # unused
in_channels: 9 # 4 data + 4 downscaled image + 1 mask
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_heads: 8
use_spatial_transformer: True
transformer_depth: 1
context_dim: 768
use_checkpoint: True
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: []
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: ldm.modules.encoders.modules.FrozenCLIPEmbedder
configs/v2-1-stable-unclip-h-inference.yaml
-80
View File
@@ -1,80 +0,0 @@
model:
base_learning_rate: 1.0e-04
target: ldm.models.diffusion.ddpm.ImageEmbeddingConditionedLatentDiffusion
params:
embedding_dropout: 0.25
parameterization: "v"
linear_start: 0.00085
linear_end: 0.0120
log_every_t: 200
timesteps: 1000
first_stage_key: "jpg"
cond_stage_key: "txt"
image_size: 96
channels: 4
cond_stage_trainable: false
conditioning_key: crossattn-adm
scale_factor: 0.18215
monitor: val/loss_simple_ema
use_ema: False
embedder_config:
target: ldm.modules.encoders.modules.FrozenOpenCLIPImageEmbedder
noise_aug_config:
target: ldm.modules.encoders.modules.CLIPEmbeddingNoiseAugmentation
params:
timestep_dim: 1024
noise_schedule_config:
timesteps: 1000
beta_schedule: squaredcos_cap_v2
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
num_classes: "sequential"
adm_in_channels: 2048
use_checkpoint: True
image_size: 32 # unused
in_channels: 4
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_head_channels: 64 # need to fix for flash-attn
use_spatial_transformer: True
use_linear_in_transformer: True
transformer_depth: 1
context_dim: 1024
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
attn_type: "vanilla-xformers"
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: [ ]
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
params:
freeze: True
layer: "penultimate"
configs/v2-1-stable-unclip-l-inference.yaml
-83
View File
@@ -1,83 +0,0 @@
model:
base_learning_rate: 1.0e-04
target: ldm.models.diffusion.ddpm.ImageEmbeddingConditionedLatentDiffusion
params:
embedding_dropout: 0.25
parameterization: "v"
linear_start: 0.00085
linear_end: 0.0120
log_every_t: 200
timesteps: 1000
first_stage_key: "jpg"
cond_stage_key: "txt"
image_size: 96
channels: 4
cond_stage_trainable: false
conditioning_key: crossattn-adm
scale_factor: 0.18215
monitor: val/loss_simple_ema
use_ema: False
embedder_config:
target: ldm.modules.encoders.modules.ClipImageEmbedder
params:
model: "ViT-L/14"
noise_aug_config:
target: ldm.modules.encoders.modules.CLIPEmbeddingNoiseAugmentation
params:
clip_stats_path: "checkpoints/karlo_models/ViT-L-14_stats.th"
timestep_dim: 768
noise_schedule_config:
timesteps: 1000
beta_schedule: squaredcos_cap_v2
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
num_classes: "sequential"
adm_in_channels: 1536
use_checkpoint: True
image_size: 32 # unused
in_channels: 4
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_head_channels: 64 # need to fix for flash-attn
use_spatial_transformer: True
use_linear_in_transformer: True
transformer_depth: 1
context_dim: 1024
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
attn_type: "vanilla-xformers"
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: [ ]
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
params:
freeze: True
layer: "penultimate"
configs/v2-inference-512-base.yaml
-67
View File
@@ -1,67 +0,0 @@
model:
base_learning_rate: 1.0e-4
target: ldm.models.diffusion.ddpm.LatentDiffusion
params:
linear_start: 0.00085
linear_end: 0.0120
num_timesteps_cond: 1
log_every_t: 200
timesteps: 1000
first_stage_key: "jpg"
cond_stage_key: "txt"
image_size: 64
channels: 4
cond_stage_trainable: false
conditioning_key: crossattn
monitor: val/loss_simple_ema
scale_factor: 0.18215
use_ema: False # we set this to false because this is an inference only config
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
use_checkpoint: True
use_fp16: True
image_size: 32 # unused
in_channels: 4
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_head_channels: 64 # need to fix for flash-attn
use_spatial_transformer: True
use_linear_in_transformer: True
transformer_depth: 1
context_dim: 1024
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
#attn_type: "vanilla-xformers"
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: []
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
params:
freeze: True
layer: "penultimate"
configs/v2-inference-768-v.yaml
-68
View File
@@ -1,68 +0,0 @@
model:
base_learning_rate: 1.0e-4
target: ldm.models.diffusion.ddpm.LatentDiffusion
params:
parameterization: "v"
linear_start: 0.00085
linear_end: 0.0120
num_timesteps_cond: 1
log_every_t: 200
timesteps: 1000
first_stage_key: "jpg"
cond_stage_key: "txt"
image_size: 64
channels: 4
cond_stage_trainable: false
conditioning_key: crossattn
monitor: val/loss_simple_ema
scale_factor: 0.18215
use_ema: False # we set this to false because this is an inference only config
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
use_checkpoint: True
use_fp16: True
image_size: 32 # unused
in_channels: 4
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_head_channels: 64 # need to fix for flash-attn
use_spatial_transformer: True
use_linear_in_transformer: True
transformer_depth: 1
context_dim: 1024
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
#attn_type: "vanilla-xformers"
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: []
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
params:
freeze: True
layer: "penultimate"
configs/v2-inpaint.yaml
-158
View File
@@ -1,158 +0,0 @@
model:
base_learning_rate: 5.0e-05
target: ldm.models.diffusion.ddpm.LatentInpaintDiffusion
params:
linear_start: 0.00085
linear_end: 0.0120
num_timesteps_cond: 1
log_every_t: 200
timesteps: 1000
first_stage_key: "jpg"
cond_stage_key: "txt"
image_size: 64
channels: 4
cond_stage_trainable: false
conditioning_key: hybrid
scale_factor: 0.18215
monitor: val/loss_simple_ema
finetune_keys: null
use_ema: False
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
use_checkpoint: True
image_size: 32 # unused
in_channels: 9
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_head_channels: 64 # need to fix for flash-attn
use_spatial_transformer: True
use_linear_in_transformer: True
transformer_depth: 1
context_dim: 1024
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
#attn_type: "vanilla-xformers"
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: [ ]
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
params:
freeze: True
layer: "penultimate"
data:
target: ldm.data.laion.WebDataModuleFromConfig
params:
tar_base: null # for concat as in LAION-A
p_unsafe_threshold: 0.1
filter_word_list: "data/filters.yaml"
max_pwatermark: 0.45
batch_size: 8
num_workers: 6
multinode: True
min_size: 512
train:
shards:
- "pipe:aws s3 cp s3://stability-aws/laion-a-native/part-0/{00000..18699}.tar -"
- "pipe:aws s3 cp s3://stability-aws/laion-a-native/part-1/{00000..18699}.tar -"
- "pipe:aws s3 cp s3://stability-aws/laion-a-native/part-2/{00000..18699}.tar -"
- "pipe:aws s3 cp s3://stability-aws/laion-a-native/part-3/{00000..18699}.tar -"
- "pipe:aws s3 cp s3://stability-aws/laion-a-native/part-4/{00000..18699}.tar -" #{00000-94333}.tar"
shuffle: 10000
image_key: jpg
image_transforms:
- target: torchvision.transforms.Resize
params:
size: 512
interpolation: 3
- target: torchvision.transforms.RandomCrop
params:
size: 512
postprocess:
target: ldm.data.laion.AddMask
params:
mode: "512train-large"
p_drop: 0.25
# NOTE use enough shards to avoid empty validation loops in workers
validation:
shards:
- "pipe:aws s3 cp s3://deep-floyd-s3/datasets/laion_cleaned-part5/{93001..94333}.tar - "
shuffle: 0
image_key: jpg
image_transforms:
- target: torchvision.transforms.Resize
params:
size: 512
interpolation: 3
- target: torchvision.transforms.CenterCrop
params:
size: 512
postprocess:
target: ldm.data.laion.AddMask
params:
mode: "512train-large"
p_drop: 0.25
lightning:
find_unused_parameters: True
modelcheckpoint:
params:
every_n_train_steps: 5000
callbacks:
metrics_over_trainsteps_checkpoint:
params:
every_n_train_steps: 10000
image_logger:
target: main.ImageLogger
params:
enable_autocast: False
disabled: False
batch_frequency: 1000
max_images: 4
increase_log_steps: False
log_first_step: False
log_images_kwargs:
use_ema_scope: False
inpaint: False
plot_progressive_rows: False
plot_diffusion_rows: False
N: 4
unconditional_guidance_scale: 5.0
unconditional_guidance_label: [""]
ddim_steps: 50 # todo check these out for depth2img,
ddim_eta: 0.0 # todo check these out for depth2img,
trainer:
benchmark: True
val_check_interval: 5000000
num_sanity_val_steps: 0
accumulate_grad_batches: 1
configs/v2-midas-inference.yaml
-74
View File
@@ -1,74 +0,0 @@
model:
base_learning_rate: 5.0e-07
target: ldm.models.diffusion.ddpm.LatentDepth2ImageDiffusion
params:
linear_start: 0.00085
linear_end: 0.0120
num_timesteps_cond: 1
log_every_t: 200
timesteps: 1000
first_stage_key: "jpg"
cond_stage_key: "txt"
image_size: 64
channels: 4
cond_stage_trainable: false
conditioning_key: hybrid
scale_factor: 0.18215
monitor: val/loss_simple_ema
finetune_keys: null
use_ema: False
depth_stage_config:
target: ldm.modules.midas.api.MiDaSInference
params:
model_type: "dpt_hybrid"
unet_config:
target: ldm.modules.diffusionmodules.openaimodel.UNetModel
params:
use_checkpoint: True
image_size: 32 # unused
in_channels: 5
out_channels: 4
model_channels: 320
attention_resolutions: [ 4, 2, 1 ]
num_res_blocks: 2
channel_mult: [ 1, 2, 4, 4 ]
num_head_channels: 64 # need to fix for flash-attn
use_spatial_transformer: True
use_linear_in_transformer: True
transformer_depth: 1
context_dim: 1024
legacy: False
first_stage_config:
target: ldm.models.autoencoder.AutoencoderKL
params:
embed_dim: 4
monitor: val/rec_loss
ddconfig:
#attn_type: "vanilla-xformers"
double_z: true
z_channels: 4
resolution: 256
in_channels: 3
out_ch: 3
ch: 128
ch_mult:
- 1
- 2
- 4
- 4
num_res_blocks: 2
attn_resolutions: [ ]
dropout: 0.0
lossconfig:
target: torch.nn.Identity
cond_stage_config:
target: ldm.modules.encoders.modules.FrozenOpenCLIPEmbedder
params:
freeze: True
layer: "penultimate"
extensions-builtin/Lora/extra_networks_lora.py
-152
View File
@@ -1,152 +0,0 @@
import re
import time
import numpy as np
import networks
import lora_patches
from modules import extra_networks, shared
# from https://github.com/cheald/sd-webui-loractl/blob/master/loractl/lib/utils.py
def get_stepwise(param, step, steps):
def sorted_positions(raw_steps):
steps = [[float(s.strip()) for s in re.split("[@~]", x)]
for x in re.split("[,;]", str(raw_steps))]
if len(steps[0]) == 1: # If we just got a single number, just return it
return steps[0][0]
steps = [[s[0], s[1] if len(s) == 2 else 1] for s in steps] # Add implicit 1s to any steps which don't have a weight
steps.sort(key=lambda k: k[1]) # Sort by index
steps = [list(v) for v in zip(*steps)]
return steps
def calculate_weight(m, step, max_steps, step_offset=2):
if isinstance(m, list):
if m[1][-1] <= 1.0:
step = (step) / (max_steps - step_offset) if max_steps > 0 else 1.0
v = np.interp(step, m[1], m[0])
return v
else:
return m
stepwise = calculate_weight(sorted_positions(param), step, steps)
return stepwise
class ExtraNetworkLora(extra_networks.ExtraNetwork):
def __init__(self):
super().__init__('lora')
self.active = False
self.model = None
self.errors = {}
networks.originals = lora_patches.LoraPatches()
def prompt(self, p):
if shared.opts.lora_apply_tags == 0:
return
all_tags = []
for loaded in networks.loaded_networks:
page = [en for en in shared.extra_networks if en.name == 'lora'][0]
item = page.create_item(loaded.name)
tags = (item or {}).get("tags", {})
loaded.tags = list(tags)
if len(loaded.tags) == 0:
loaded.tags.append(loaded.name)
if shared.opts.lora_apply_tags > 0:
loaded.tags = loaded.tags[:shared.opts.lora_apply_tags]
all_tags.extend(loaded.tags)
if len(all_tags) > 0:
shared.log.debug(f"Network load: type=LoRA tags={all_tags} max={shared.opts.lora_apply_tags} apply")
all_tags = ', '.join(all_tags)
p.extra_generation_params["LoRA tags"] = all_tags
if '_tags_' in p.prompt:
p.prompt = p.prompt.replace('_tags_', all_tags)
else:
p.prompt = f"{p.prompt}, {all_tags}"
if p.all_prompts is not None:
for i in range(len(p.all_prompts)):
if '_tags_' in p.all_prompts[i]:
p.all_prompts[i] = p.all_prompts[i].replace('_tags_', all_tags)
else:
p.all_prompts[i] = f"{p.all_prompts[i]}, {all_tags}"
def infotext(self, p):
names = [i.name for i in networks.loaded_networks]
if len(names) > 0:
p.extra_generation_params["LoRA networks"] = ", ".join(names)
if shared.opts.lora_add_hashes_to_infotext:
network_hashes = []
for item in networks.loaded_networks:
if not item.network_on_disk.shorthash:
continue
network_hashes.append(item.network_on_disk.shorthash)
if len(network_hashes) > 0:
p.extra_generation_params["LoRA hashes"] = ", ".join(network_hashes)
def parse(self, p, params_list, step=0):
names = []
te_multipliers = []
unet_multipliers = []
dyn_dims = []
for params in params_list:
assert params.items
names.append(params.positional[0])
te_multiplier = params.named.get("te", params.positional[1] if len(params.positional) > 1 else shared.opts.extra_networks_default_multiplier)
if isinstance(te_multiplier, str) and "@" in te_multiplier:
te_multiplier = get_stepwise(te_multiplier, step, p.steps)
else:
te_multiplier = float(te_multiplier)
unet_multiplier = [params.positional[2] if len(params.positional) > 2 else te_multiplier] * 3
unet_multiplier = [params.named.get("unet", unet_multiplier[0])] * 3
unet_multiplier[0] = params.named.get("in", unet_multiplier[0])
unet_multiplier[1] = params.named.get("mid", unet_multiplier[1])
unet_multiplier[2] = params.named.get("out", unet_multiplier[2])
for i in range(len(unet_multiplier)):
if isinstance(unet_multiplier[i], str) and "@" in unet_multiplier[i]:
unet_multiplier[i] = get_stepwise(unet_multiplier[i], step, p.steps)
else:
unet_multiplier[i] = float(unet_multiplier[i])
dyn_dim = int(params.positional[3]) if len(params.positional) > 3 else None
dyn_dim = int(params.named["dyn"]) if "dyn" in params.named else dyn_dim
te_multipliers.append(te_multiplier)
unet_multipliers.append(unet_multiplier)
dyn_dims.append(dyn_dim)
return names, te_multipliers, unet_multipliers, dyn_dims
def activate(self, p, params_list, step=0):
t0 = time.time()
self.errors.clear()
if self.active:
if self.model != shared.opts.sd_model_checkpoint: # reset if model changed
self.active = False
if len(params_list) > 0 and not self.active: # activate patches once
shared.log.debug(f'Activate network: type=LoRA model="{shared.opts.sd_model_checkpoint}"')
networks.originals.apply() # apply patches
self.active = True
self.model = shared.opts.sd_model_checkpoint
names, te_multipliers, unet_multipliers, dyn_dims = self.parse(p, params_list, step)
networks.load_networks(names, te_multipliers, unet_multipliers, dyn_dims)
t1 = time.time()
if len(networks.loaded_networks) > 0 and step == 0:
self.infotext(p)
self.prompt(p)
shared.log.info(f'Network load: type=LoRA apply={[n.name for n in networks.loaded_networks]} method=legacy te={te_multipliers} unet={unet_multipliers} dims={dyn_dims} load={t1-t0:.2f}')
def deactivate(self, p):
t0 = time.time()
if shared.native and len(networks.diffuser_loaded) > 0:
if hasattr(shared.sd_model, "unload_lora_weights") and hasattr(shared.sd_model, "text_encoder"):
if not (shared.compiled_model_state is not None and shared.compiled_model_state.is_compiled is True):
try:
if shared.opts.lora_fuse_diffusers:
shared.sd_model.unfuse_lora()
shared.sd_model.unload_lora_weights() # fails for non-CLIP models
except Exception:
pass
t1 = time.time()
networks.timer['restore'] += t1 - t0
if self.active and networks.debug:
shared.log.debug(f"Network end: type=LoRA load={networks.timer['load']:.2f} apply={networks.timer['apply']:.2f} restore={networks.timer['restore']:.2f}")
if self.errors:
for k, v in self.errors.items():
shared.log.error(f'LoRA: name="{k}" errors={v}')
self.errors.clear()
extensions-builtin/Lora/lora.py
-8
View File
@@ -1,8 +0,0 @@
import networks
list_available_loras = networks.list_available_networks
available_loras = networks.available_networks
available_lora_aliases = networks.available_network_aliases
available_lora_hash_lookup = networks.available_network_hash_lookup
forbidden_lora_aliases = networks.forbidden_network_aliases
loaded_loras = networks.loaded_networks
extensions-builtin/Lora/lora_convert.py
-457
View File
@@ -1,457 +0,0 @@
import os
import re
import bisect
from typing import Dict
import torch
from modules import shared
debug = os.environ.get('SD_LORA_DEBUG', None) is not None
suffix_conversion = {
"attentions": {},
"resnets": {
"conv1": "in_layers_2",
"conv2": "out_layers_3",
"norm1": "in_layers_0",
"norm2": "out_layers_0",
"time_emb_proj": "emb_layers_1",
"conv_shortcut": "skip_connection",
}
}
re_digits = re.compile(r"\d+")
re_x_proj = re.compile(r"(.*)_([qkv]_proj)$")
re_compiled = {}
def make_unet_conversion_map() -> Dict[str, str]:
unet_conversion_map_layer = []
for i in range(3): # num_blocks is 3 in sdxl
# loop over downblocks/upblocks
for j in range(2):
# loop over resnets/attentions for downblocks
hf_down_res_prefix = f"down_blocks.{i}.resnets.{j}."
sd_down_res_prefix = f"input_blocks.{3 * i + j + 1}.0."
unet_conversion_map_layer.append((sd_down_res_prefix, hf_down_res_prefix))
if i < 3:
# no attention layers in down_blocks.3
hf_down_atn_prefix = f"down_blocks.{i}.attentions.{j}."
sd_down_atn_prefix = f"input_blocks.{3 * i + j + 1}.1."
unet_conversion_map_layer.append((sd_down_atn_prefix, hf_down_atn_prefix))
for j in range(3):
# loop over resnets/attentions for upblocks
hf_up_res_prefix = f"up_blocks.{i}.resnets.{j}."
sd_up_res_prefix = f"output_blocks.{3 * i + j}.0."
unet_conversion_map_layer.append((sd_up_res_prefix, hf_up_res_prefix))
# if i > 0: commentout for sdxl
# no attention layers in up_blocks.0
hf_up_atn_prefix = f"up_blocks.{i}.attentions.{j}."
sd_up_atn_prefix = f"output_blocks.{3 * i + j}.1."
unet_conversion_map_layer.append((sd_up_atn_prefix, hf_up_atn_prefix))
if i < 3:
# no downsample in down_blocks.3
hf_downsample_prefix = f"down_blocks.{i}.downsamplers.0.conv."
sd_downsample_prefix = f"input_blocks.{3 * (i + 1)}.0.op."
unet_conversion_map_layer.append((sd_downsample_prefix, hf_downsample_prefix))
# no upsample in up_blocks.3
hf_upsample_prefix = f"up_blocks.{i}.upsamplers.0."
sd_upsample_prefix = f"output_blocks.{3 * i + 2}.{2}." # change for sdxl
unet_conversion_map_layer.append((sd_upsample_prefix, hf_upsample_prefix))
hf_mid_atn_prefix = "mid_block.attentions.0."
sd_mid_atn_prefix = "middle_block.1."
unet_conversion_map_layer.append((sd_mid_atn_prefix, hf_mid_atn_prefix))
for j in range(2):
hf_mid_res_prefix = f"mid_block.resnets.{j}."
sd_mid_res_prefix = f"middle_block.{2 * j}."
unet_conversion_map_layer.append((sd_mid_res_prefix, hf_mid_res_prefix))
unet_conversion_map_resnet = [
# (stable-diffusion, HF Diffusers)
("in_layers.0.", "norm1."),
("in_layers.2.", "conv1."),
("out_layers.0.", "norm2."),
("out_layers.3.", "conv2."),
("emb_layers.1.", "time_emb_proj."),
("skip_connection.", "conv_shortcut."),
]
unet_conversion_map = []
for sd, hf in unet_conversion_map_layer:
if "resnets" in hf:
for sd_res, hf_res in unet_conversion_map_resnet:
unet_conversion_map.append((sd + sd_res, hf + hf_res))
else:
unet_conversion_map.append((sd, hf))
for j in range(2):
hf_time_embed_prefix = f"time_embedding.linear_{j + 1}."
sd_time_embed_prefix = f"time_embed.{j * 2}."
unet_conversion_map.append((sd_time_embed_prefix, hf_time_embed_prefix))
for j in range(2):
hf_label_embed_prefix = f"add_embedding.linear_{j + 1}."
sd_label_embed_prefix = f"label_emb.0.{j * 2}."
unet_conversion_map.append((sd_label_embed_prefix, hf_label_embed_prefix))
unet_conversion_map.append(("input_blocks.0.0.", "conv_in."))
unet_conversion_map.append(("out.0.", "conv_norm_out."))
unet_conversion_map.append(("out.2.", "conv_out."))
sd_hf_conversion_map = {sd.replace(".", "_")[:-1]: hf.replace(".", "_")[:-1] for sd, hf in unet_conversion_map}
return sd_hf_conversion_map
class KeyConvert:
def __init__(self):
if not shared.native:
self.converter = self.original
self.is_sd2 = 'model_transformer_resblocks' in shared.sd_model.network_layer_mapping
else:
self.converter = self.diffusers
self.is_sdxl = True if shared.sd_model_type == "sdxl" else False
self.UNET_CONVERSION_MAP = make_unet_conversion_map() if self.is_sdxl else None
self.LORA_PREFIX_UNET = "lora_unet_"
self.LORA_PREFIX_TEXT_ENCODER = "lora_te_"
self.OFT_PREFIX_UNET = "oft_unet_"
# SDXL: must starts with LORA_PREFIX_TEXT_ENCODER
self.LORA_PREFIX_TEXT_ENCODER1 = "lora_te1_"
self.LORA_PREFIX_TEXT_ENCODER2 = "lora_te2_"
def original(self, key):
key = convert_diffusers_name_to_compvis(key, self.is_sd2)
sd_module = shared.sd_model.network_layer_mapping.get(key, None)
if sd_module is None:
m = re_x_proj.match(key)
if m:
sd_module = shared.sd_model.network_layer_mapping.get(m.group(1), None)
# SDXL loras seem to already have correct compvis keys, so only need to replace "lora_unet" with "diffusion_model"
if sd_module is None and "lora_unet" in key:
key = key.replace("lora_unet", "diffusion_model")
sd_module = shared.sd_model.network_layer_mapping.get(key, None)
elif sd_module is None and "lora_te1_text_model" in key:
key = key.replace("lora_te1_text_model", "0_transformer_text_model")
sd_module = shared.sd_model.network_layer_mapping.get(key, None)
# some SD1 Loras also have correct compvis keys
if sd_module is None:
key = key.replace("lora_te1_text_model", "transformer_text_model")
sd_module = shared.sd_model.network_layer_mapping.get(key, None)
# SegMoE begin
expert_key = key + "_experts_0"
expert_module = shared.sd_model.network_layer_mapping.get(expert_key, None)
if expert_module is not None:
sd_module = expert_module
key = expert_key
if sd_module is None:
key = key.replace("_net_", "_experts_0_net_")
sd_module = shared.sd_model.network_layer_mapping.get(key, None)
key = key if isinstance(key, list) else [key]
sd_module = sd_module if isinstance(sd_module, list) else [sd_module]
if "_experts_0" in key[0]:
i = expert_module = 1
while expert_module is not None:
expert_key = key[0].replace("_experts_0", f"_experts_{i}")
expert_module = shared.sd_model.network_layer_mapping.get(expert_key, None)
if expert_module is not None:
key.append(expert_key)
sd_module.append(expert_module)
i += 1
# SegMoE end
return key, sd_module
def diffusers(self, key):
if self.is_sdxl:
if "diffusion_model" in key: # Fix NTC Slider naming error
key = key.replace("diffusion_model", "lora_unet")
map_keys = list(self.UNET_CONVERSION_MAP.keys()) # prefix of U-Net modules
map_keys.sort()
search_key = key.replace(self.LORA_PREFIX_UNET, "").replace(self.OFT_PREFIX_UNET, "").replace(self.LORA_PREFIX_TEXT_ENCODER1, "").replace(self.LORA_PREFIX_TEXT_ENCODER2, "")
position = bisect.bisect_right(map_keys, search_key)
map_key = map_keys[position - 1]
if search_key.startswith(map_key):
key = key.replace(map_key, self.UNET_CONVERSION_MAP[map_key]).replace("oft", "lora") # pylint: disable=unsubscriptable-object
if "lycoris" in key and "transformer" in key:
key = key.replace("lycoris", "lora_transformer")
sd_module = shared.sd_model.network_layer_mapping.get(key, None)
if sd_module is None:
sd_module = shared.sd_model.network_layer_mapping.get(key.replace("guidance", "timestep"), None) # FLUX1 fix
# SegMoE begin
expert_key = key + "_experts_0"
expert_module = shared.sd_model.network_layer_mapping.get(expert_key, None)
if expert_module is not None:
sd_module = expert_module
key = expert_key
if sd_module is None:
key = key.replace("_net_", "_experts_0_net_")
sd_module = shared.sd_model.network_layer_mapping.get(key, None)
key = key if isinstance(key, list) else [key]
sd_module = sd_module if isinstance(sd_module, list) else [sd_module]
if "_experts_0" in key[0]:
i = expert_module = 1
while expert_module is not None:
expert_key = key[0].replace("_experts_0", f"_experts_{i}")
expert_module = shared.sd_model.network_layer_mapping.get(expert_key, None)
if expert_module is not None:
key.append(expert_key)
sd_module.append(expert_module)
i += 1
# SegMoE end
if debug and sd_module is None:
raise RuntimeError(f"LoRA key not found in network_layer_mapping: key={key} mapping={shared.sd_model.network_layer_mapping.keys()}")
return key, sd_module
def __call__(self, key):
return self.converter(key)
def convert_diffusers_name_to_compvis(key, is_sd2):
def match(match_list, regex_text):
regex = re_compiled.get(regex_text)
if regex is None:
regex = re.compile(regex_text)
re_compiled[regex_text] = regex
r = re.match(regex, key)
if not r:
return False
match_list.clear()
match_list.extend([int(x) if re.match(re_digits, x) else x for x in r.groups()])
return True
m = []
if match(m, r"lora_unet_conv_in(.*)"):
return f'diffusion_model_input_blocks_0_0{m[0]}'
if match(m, r"lora_unet_conv_out(.*)"):
return f'diffusion_model_out_2{m[0]}'
if match(m, r"lora_unet_time_embedding_linear_(\d+)(.*)"):
return f"diffusion_model_time_embed_{m[0] * 2 - 2}{m[1]}"
if match(m, r"lora_unet_down_blocks_(\d+)_(attentions|resnets)_(\d+)_(.+)"):
suffix = suffix_conversion.get(m[1], {}).get(m[3], m[3])
return f"diffusion_model_input_blocks_{1 + m[0] * 3 + m[2]}_{1 if m[1] == 'attentions' else 0}_{suffix}"
if match(m, r"lora_unet_mid_block_(attentions|resnets)_(\d+)_(.+)"):
suffix = suffix_conversion.get(m[0], {}).get(m[2], m[2])
return f"diffusion_model_middle_block_{1 if m[0] == 'attentions' else m[1] * 2}_{suffix}"
if match(m, r"lora_unet_up_blocks_(\d+)_(attentions|resnets)_(\d+)_(.+)"):
suffix = suffix_conversion.get(m[1], {}).get(m[3], m[3])
return f"diffusion_model_output_blocks_{m[0] * 3 + m[2]}_{1 if m[1] == 'attentions' else 0}_{suffix}"
if match(m, r"lora_unet_down_blocks_(\d+)_downsamplers_0_conv"):
return f"diffusion_model_input_blocks_{3 + m[0] * 3}_0_op"
if match(m, r"lora_unet_up_blocks_(\d+)_upsamplers_0_conv"):
return f"diffusion_model_output_blocks_{2 + m[0] * 3}_{2 if m[0]>0 else 1}_conv"
if match(m, r"lora_te_text_model_encoder_layers_(\d+)_(.+)"):
if is_sd2:
if 'mlp_fc1' in m[1]:
return f"model_transformer_resblocks_{m[0]}_{m[1].replace('mlp_fc1', 'mlp_c_fc')}"
elif 'mlp_fc2' in m[1]:
return f"model_transformer_resblocks_{m[0]}_{m[1].replace('mlp_fc2', 'mlp_c_proj')}"
else:
return f"model_transformer_resblocks_{m[0]}_{m[1].replace('self_attn', 'attn')}"
return f"transformer_text_model_encoder_layers_{m[0]}_{m[1]}"
if match(m, r"lora_te2_text_model_encoder_layers_(\d+)_(.+)"):
if 'mlp_fc1' in m[1]:
return f"1_model_transformer_resblocks_{m[0]}_{m[1].replace('mlp_fc1', 'mlp_c_fc')}"
elif 'mlp_fc2' in m[1]:
return f"1_model_transformer_resblocks_{m[0]}_{m[1].replace('mlp_fc2', 'mlp_c_proj')}"
else:
return f"1_model_transformer_resblocks_{m[0]}_{m[1].replace('self_attn', 'attn')}"
return key
# Taken from https://github.com/huggingface/diffusers/blob/main/src/diffusers/loaders/lora_conversion_utils.py
# Modified from 'lora_A' and 'lora_B' to 'lora_down' and 'lora_up'
# Added early exit
# The utilities under `_convert_kohya_flux_lora_to_diffusers()`
# are taken from https://github.com/kohya-ss/sd-scripts/blob/a61cf73a5cb5209c3f4d1a3688dd276a4dfd1ecb/networks/convert_flux_lora.py
# All credits go to `kohya-ss`.
def _convert_kohya_flux_lora_to_diffusers(state_dict):
def _convert_to_ai_toolkit(sds_sd, ait_sd, sds_key, ait_key):
if sds_key + ".lora_down.weight" not in sds_sd:
return
down_weight = sds_sd.pop(sds_key + ".lora_down.weight")
# scale weight by alpha and dim
rank = down_weight.shape[0]
alpha = sds_sd.pop(sds_key + ".alpha").item() # alpha is scalar
scale = alpha / rank # LoRA is scaled by 'alpha / rank' in forward pass, so we need to scale it back here
# calculate scale_down and scale_up to keep the same value. if scale is 4, scale_down is 2 and scale_up is 2
scale_down = scale
scale_up = 1.0
while scale_down * 2 < scale_up:
scale_down *= 2
scale_up /= 2
ait_sd[ait_key + ".lora_down.weight"] = down_weight * scale_down
ait_sd[ait_key + ".lora_up.weight"] = sds_sd.pop(sds_key + ".lora_up.weight") * scale_up
def _convert_to_ai_toolkit_cat(sds_sd, ait_sd, sds_key, ait_keys, dims=None):
if sds_key + ".lora_down.weight" not in sds_sd:
return
down_weight = sds_sd.pop(sds_key + ".lora_down.weight")
up_weight = sds_sd.pop(sds_key + ".lora_up.weight")
sd_lora_rank = down_weight.shape[0]
# scale weight by alpha and dim
alpha = sds_sd.pop(sds_key + ".alpha")
scale = alpha / sd_lora_rank
# calculate scale_down and scale_up
scale_down = scale
scale_up = 1.0
while scale_down * 2 < scale_up:
scale_down *= 2
scale_up /= 2
down_weight = down_weight * scale_down
up_weight = up_weight * scale_up
# calculate dims if not provided
num_splits = len(ait_keys)
if dims is None:
dims = [up_weight.shape[0] // num_splits] * num_splits
else:
assert sum(dims) == up_weight.shape[0]
# check upweight is sparse or not
is_sparse = False
if sd_lora_rank % num_splits == 0:
ait_rank = sd_lora_rank // num_splits
is_sparse = True
i = 0
for j in range(len(dims)):
for k in range(len(dims)):
if j == k:
continue
is_sparse = is_sparse and torch.all(
up_weight[i : i + dims[j], k * ait_rank : (k + 1) * ait_rank] == 0
)
i += dims[j]
# if is_sparse:
# print(f"weight is sparse: {sds_key}")
# make ai-toolkit weight
ait_down_keys = [k + ".lora_down.weight" for k in ait_keys]
ait_up_keys = [k + ".lora_up.weight" for k in ait_keys]
if not is_sparse:
# down_weight is copied to each split
ait_sd.update({k: down_weight for k in ait_down_keys})
# up_weight is split to each split
ait_sd.update({k: v for k, v in zip(ait_up_keys, torch.split(up_weight, dims, dim=0))}) # noqa: C416 # pylint: disable=unnecessary-comprehension
else:
# down_weight is chunked to each split
ait_sd.update({k: v for k, v in zip(ait_down_keys, torch.chunk(down_weight, num_splits, dim=0))}) # noqa: C416 # pylint: disable=unnecessary-comprehension
# up_weight is sparse: only non-zero values are copied to each split
i = 0
for j in range(len(dims)):
ait_sd[ait_up_keys[j]] = up_weight[i : i + dims[j], j * ait_rank : (j + 1) * ait_rank].contiguous()
i += dims[j]
def _convert_sd_scripts_to_ai_toolkit(sds_sd):
ait_sd = {}
for i in range(19):
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_img_attn_proj",
f"transformer.transformer_blocks.{i}.attn.to_out.0",
)
_convert_to_ai_toolkit_cat(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_img_attn_qkv",
[
f"transformer.transformer_blocks.{i}.attn.to_q",
f"transformer.transformer_blocks.{i}.attn.to_k",
f"transformer.transformer_blocks.{i}.attn.to_v",
],
)
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_img_mlp_0",
f"transformer.transformer_blocks.{i}.ff.net.0.proj",
)
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_img_mlp_2",
f"transformer.transformer_blocks.{i}.ff.net.2",
)
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_img_mod_lin",
f"transformer.transformer_blocks.{i}.norm1.linear",
)
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_txt_attn_proj",
f"transformer.transformer_blocks.{i}.attn.to_add_out",
)
_convert_to_ai_toolkit_cat(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_txt_attn_qkv",
[
f"transformer.transformer_blocks.{i}.attn.add_q_proj",
f"transformer.transformer_blocks.{i}.attn.add_k_proj",
f"transformer.transformer_blocks.{i}.attn.add_v_proj",
],
)
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_txt_mlp_0",
f"transformer.transformer_blocks.{i}.ff_context.net.0.proj",
)
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_txt_mlp_2",
f"transformer.transformer_blocks.{i}.ff_context.net.2",
)
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_double_blocks_{i}_txt_mod_lin",
f"transformer.transformer_blocks.{i}.norm1_context.linear",
)
for i in range(38):
_convert_to_ai_toolkit_cat(
sds_sd,
ait_sd,
f"lora_unet_single_blocks_{i}_linear1",
[
f"transformer.single_transformer_blocks.{i}.attn.to_q",
f"transformer.single_transformer_blocks.{i}.attn.to_k",
f"transformer.single_transformer_blocks.{i}.attn.to_v",
f"transformer.single_transformer_blocks.{i}.proj_mlp",
],
dims=[3072, 3072, 3072, 12288],
)
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_single_blocks_{i}_linear2",
f"transformer.single_transformer_blocks.{i}.proj_out",
)
_convert_to_ai_toolkit(
sds_sd,
ait_sd,
f"lora_unet_single_blocks_{i}_modulation_lin",
f"transformer.single_transformer_blocks.{i}.norm.linear",
)
if len(sds_sd) > 0:
return None
return ait_sd
return _convert_sd_scripts_to_ai_toolkit(state_dict)
extensions-builtin/Lora/lora_extract.py
-261
View File
@@ -1,261 +0,0 @@
import os
import time
import json
import datetime
import torch
from safetensors.torch import save_file
import gradio as gr
from rich import progress as p
from modules import shared, devices
from modules.ui_common import create_refresh_button
from modules.call_queue import wrap_gradio_gpu_call
class SVDHandler:
def __init__(self, maxrank=0, rank_ratio=1):
self.network_name: str = None
self.U: torch.Tensor = None
self.S: torch.Tensor = None
self.Vh: torch.Tensor = None
self.maxrank: int = maxrank
self.rank_ratio: float = rank_ratio
self.rank: int = 0
self.out_size: int = None
self.in_size: int = None
self.kernel_size: tuple[int, int] = None
self.conv2d: bool = False
def decompose(self, weight, backupweight):
self.conv2d = len(weight.size()) == 4
self.kernel_size = None if not self.conv2d else weight.size()[2:4]
self.out_size, self.in_size = weight.size()[0:2]
diffweight = weight.clone().to(devices.device)
diffweight -= backupweight.to(devices.device)
if self.conv2d:
if self.conv2d and self.kernel_size != (1, 1):
diffweight = diffweight.flatten(start_dim=1)
else:
diffweight = diffweight.squeeze()
self.U, self.S, self.Vh = torch.svd_lowrank(diffweight.to(device=devices.device, dtype=torch.float), self.maxrank, 2)
# del diffweight
self.U = self.U.to(device=devices.cpu, dtype=torch.bfloat16)
self.S = self.S.to(device=devices.cpu, dtype=torch.bfloat16)
self.Vh = self.Vh.t().to(device=devices.cpu, dtype=torch.bfloat16) # svd_lowrank outputs a transposed matrix
def findrank(self):
if self.rank_ratio < 1:
S_squared = self.S.pow(2)
S_fro_sq = float(torch.sum(S_squared))
sum_S_squared = torch.cumsum(S_squared, dim=0) / S_fro_sq
index = int(torch.searchsorted(sum_S_squared, self.rank_ratio ** 2)) + 1
index = max(1, min(index, len(self.S) - 1))
self.rank = index
if self.maxrank > 0:
self.rank = min(self.rank, self.maxrank)
else:
self.rank = min(self.in_size, self.out_size, self.maxrank)
def makeweights(self):
self.findrank()
up = self.U[:, :self.rank] @ torch.diag(self.S[:self.rank])
down = self.Vh[:self.rank, :]
if self.conv2d and self.kernel_size is not None:
up = up.reshape(self.out_size, self.rank, 1, 1)
down = down.reshape(self.rank, self.in_size, self.kernel_size[0], self.kernel_size[1]) # pylint: disable=unsubscriptable-object
return_dict = {f'{self.network_name}.lora_up.weight': up.contiguous(),
f'{self.network_name}.lora_down.weight': down.contiguous(),
f'{self.network_name}.alpha': torch.tensor(down.shape[0]),
}
return return_dict
def loaded_lora():
if not shared.sd_loaded:
return ""
loaded = set()
if hasattr(shared.sd_model, 'unet'):
for _name, module in shared.sd_model.unet.named_modules():
current = getattr(module, "network_current_names", None)
if current is not None:
current = [item[0] for item in current]
loaded.update(current)
return list(loaded)
def loaded_lora_str():
return ", ".join(loaded_lora())
def make_meta(fn, maxrank, rank_ratio):
meta = {
"model_spec.sai_model_spec": "1.0.0",
"model_spec.title": os.path.splitext(os.path.basename(fn))[0],
"model_spec.author": "SD.Next",
"model_spec.implementation": "https://github.com/vladmandic/sdnext",
"model_spec.date": datetime.datetime.now().astimezone().replace(microsecond=0).isoformat(),
"model_spec.base_model": shared.opts.sd_model_checkpoint,
"model_spec.dtype": str(devices.dtype),
"model_spec.base_lora": json.dumps(loaded_lora()),
"model_spec.config": f"maxrank={maxrank} rank_ratio={rank_ratio}",
}
if shared.sd_model_type == "sdxl":
meta["model_spec.architecture"] = "stable-diffusion-xl-v1-base/lora" # sai standard
meta["ss_base_model_version"] = "sdxl_base_v1-0" # kohya standard
elif shared.sd_model_type == "sd":
meta["model_spec.architecture"] = "stable-diffusion-v1/lora"
meta["ss_base_model_version"] = "sd_v1"
elif shared.sd_model_type == "f1":
meta["model_spec.architecture"] = "flux-1-dev/lora"
meta["ss_base_model_version"] = "flux1"
elif shared.sd_model_type == "chroma":
meta["model_spec.architecture"] = "chroma/lora"
meta["ss_base_model_version"] = "chroma"
elif shared.sd_model_type == "sc":
meta["model_spec.architecture"] = "stable-cascade-v1-prior/lora"
return meta
def make_lora(fn, maxrank, auto_rank, rank_ratio, modules, overwrite):
if not shared.sd_loaded or not shared.native:
msg = "LoRA extract: model not loaded"
shared.log.warning(msg)
yield msg
return
if loaded_lora() == "":
msg = "LoRA extract: no LoRA detected"
shared.log.warning(msg)
yield msg
return
if not fn:
msg = "LoRA extract: target filename required"
shared.log.warning(msg)
yield msg
return
t0 = time.time()
maxrank = int(maxrank)
rank_ratio = 1 if not auto_rank else rank_ratio
shared.log.debug(f'LoRA extract: modules={modules} maxrank={maxrank} auto={auto_rank} ratio={rank_ratio} fn="{fn}"')
shared.state.begin('LoRA extract')
with p.Progress(p.TextColumn('[cyan]LoRA extract'), p.BarColumn(), p.TaskProgressColumn(), p.TimeRemainingColumn(), p.TimeElapsedColumn(), p.TextColumn('[cyan]{task.description}'), console=shared.console) as progress:
if 'te' in modules and getattr(shared.sd_model, 'text_encoder', None) is not None:
modules = shared.sd_model.text_encoder.named_modules()
task = progress.add_task(description="te1 decompose", total=len(list(modules)))
for name, module in shared.sd_model.text_encoder.named_modules():
progress.update(task, advance=1)
weights_backup = getattr(module, "network_weights_backup", None)
if weights_backup is None or getattr(module, "network_current_names", None) is None:
continue
prefix = "lora_te1_" if hasattr(shared.sd_model, 'text_encoder_2') else "lora_te_"
module.svdhandler = SVDHandler(maxrank, rank_ratio)
module.svdhandler.network_name = prefix + name.replace(".", "_")
with devices.inference_context():
module.svdhandler.decompose(module.weight, weights_backup)
progress.remove_task(task)
t1 = time.time()
if 'te' in modules and getattr(shared.sd_model, 'text_encoder_2', None) is not None:
modules = shared.sd_model.text_encoder_2.named_modules()
task = progress.add_task(description="te2 decompose", total=len(list(modules)))
for name, module in shared.sd_model.text_encoder_2.named_modules():
progress.update(task, advance=1)
weights_backup = getattr(module, "network_weights_backup", None)
if weights_backup is None or getattr(module, "network_current_names", None) is None:
continue
module.svdhandler = SVDHandler(maxrank, rank_ratio)
module.svdhandler.network_name = "lora_te2_" + name.replace(".", "_")
with devices.inference_context():
module.svdhandler.decompose(module.weight, weights_backup)
progress.remove_task(task)
t2 = time.time()
if 'unet' in modules and getattr(shared.sd_model, 'unet', None) is not None:
modules = shared.sd_model.unet.named_modules()
task = progress.add_task(description="unet decompose", total=len(list(modules)))
for name, module in shared.sd_model.unet.named_modules():
progress.update(task, advance=1)
weights_backup = getattr(module, "network_weights_backup", None)
if weights_backup is None or getattr(module, "network_current_names", None) is None:
continue
module.svdhandler = SVDHandler(maxrank, rank_ratio)
module.svdhandler.network_name = "lora_unet_" + name.replace(".", "_")
with devices.inference_context():
module.svdhandler.decompose(module.weight, weights_backup)
progress.remove_task(task)
t3 = time.time()
lora_state_dict = {}
for sub in ['text_encoder', 'text_encoder_2', 'unet', 'transformer']:
submodel = getattr(shared.sd_model, sub, None)
if submodel is not None:
modules = submodel.named_modules()
task = progress.add_task(description=f"{sub} exctract", total=len(list(modules)))
for _name, module in submodel.named_modules():
progress.update(task, advance=1)
if not hasattr(module, "svdhandler"):
continue
lora_state_dict.update(module.svdhandler.makeweights())
del module.svdhandler
progress.remove_task(task)
t4 = time.time()
if not os.path.isabs(fn):
fn = os.path.join(shared.cmd_opts.lora_dir, fn)
if not fn.endswith('.safetensors'):
fn += '.safetensors'
if os.path.exists(fn):
if overwrite:
os.remove(fn)
else:
msg = f'LoRA extract: fn="{fn}" file exists'
shared.log.warning(msg)
yield msg
return
shared.state.end()
meta = make_meta(fn, maxrank, rank_ratio)
shared.log.debug(f'LoRA metadata: {meta}')
try:
save_file(tensors=lora_state_dict, metadata=meta, filename=fn)
except Exception as e:
msg = f'LoRA extract error: fn="{fn}" {e}'
shared.log.error(msg)
yield msg
return
t5 = time.time()
shared.log.debug(f'LoRA extract: time={t5-t0:.2f} te1={t1-t0:.2f} te2={t2-t1:.2f} unet={t3-t2:.2f} save={t5-t4:.2f}')
keys = list(lora_state_dict.keys())
msg = f'LoRA extract: fn="{fn}" keys={len(keys)}'
shared.log.info(msg)
yield msg
def create_ui():
def gr_show(visible=True):
return {"visible": visible, "__type__": "update"}
with gr.Tab(label="Extract LoRA"):
with gr.Row():
loaded = gr.Textbox(placeholder="Press refresh to query loaded LoRA", label="Loaded LoRA", interactive=False)
create_refresh_button(loaded, lambda: None, lambda: {'value': loaded_lora_str()}, "testid")
with gr.Group():
with gr.Row():
modules = gr.CheckboxGroup(label="Modules to extract", value=['unet'], choices=['te', 'unet'])
with gr.Row():
auto_rank = gr.Checkbox(value=False, label="Automatically determine rank")
rank_ratio = gr.Slider(label="Autorank ratio", value=1, minimum=0, maximum=1, step=0.05, visible=False)
rank = gr.Slider(label="Maximum rank", value=32, minimum=1, maximum=256)
with gr.Row():
filename = gr.Textbox(label="LoRA target filename")
overwrite = gr.Checkbox(value=False, label="Overwrite existing file")
with gr.Row():
extract = gr.Button(value="Extract LoRA", variant='primary')
status = gr.HTML(value="", show_label=False)
auto_rank.change(fn=lambda x: gr_show(x), inputs=[auto_rank], outputs=[rank_ratio])
extract.click(
fn=wrap_gradio_gpu_call(make_lora, extra_outputs=[]),
inputs=[filename, rank, auto_rank, rank_ratio, modules, overwrite],
outputs=[status]
)
extensions-builtin/Lora/lora_patches.py
-77
View File
@@ -1,77 +0,0 @@
import sys
import torch
import networks
from modules import patches, shared, model_quant
class LoraPatches:
def __init__(self):
self.active = False
self.Linear_forward = None
self.Linear_load_state_dict = None
self.Conv2d_forward = None
self.Conv2d_load_state_dict = None
self.GroupNorm_forward = None
self.GroupNorm_load_state_dict = None
self.LayerNorm_forward = None
self.LayerNorm_load_state_dict = None
self.MultiheadAttention_forward = None
self.MultiheadAttention_load_state_dict = None
# optional quant forwards
self.Linear4bit_forward = None # bitsandbytes
self.QLinear_forward = None # optimum.quanto
self.QConv2d_forward = None # optimum.quanto
def handle_quant(self, apply: bool):
if 'bitsandbytes' in sys.modules: # lora should not be first to initialize quantization
bnb = model_quant.load_bnb(silent=True)
if bnb is not None:
if apply:
self.Linear4bit_forward = patches.patch(__name__, bnb.nn.Linear4bit, 'forward', networks.network_Linear4bit_forward)
else:
self.Linear4bit_forward = patches.undo(__name__, bnb.nn.Linear4bit, 'forward') # pylint: disable=E1128
if 'optimum.quanto' in sys.modules:
quanto = model_quant.load_quanto(silent=True)
if quanto is not None:
if apply:
self.QLinear_forward = patches.patch(__name__, quanto.nn.QLinear, 'forward', networks.network_QLinear_forward)
self.QConv2d_forward = patches.patch(__name__, quanto.nn.QConv2d, 'forward', networks.network_QConv2d_forward)
else:
self.QLinear_forward = patches.undo(__name__, quanto.nn.QLinear, 'forward') # pylint: disable=E1128
self.QConv2d_forward = patches.undo(__name__, quanto.nn.QConv2d, 'forward') # pylint: disable=E1128
def apply(self):
if self.active or shared.opts.lora_force_diffusers:
return
self.Linear_forward = patches.patch(__name__, torch.nn.Linear, 'forward', networks.network_Linear_forward)
self.Linear_load_state_dict = patches.patch(__name__, torch.nn.Linear, '_load_from_state_dict', networks.network_Linear_load_state_dict)
self.Conv2d_forward = patches.patch(__name__, torch.nn.Conv2d, 'forward', networks.network_Conv2d_forward)
self.Conv2d_load_state_dict = patches.patch(__name__, torch.nn.Conv2d, '_load_from_state_dict', networks.network_Conv2d_load_state_dict)
self.GroupNorm_forward = patches.patch(__name__, torch.nn.GroupNorm, 'forward', networks.network_GroupNorm_forward)
self.GroupNorm_load_state_dict = patches.patch(__name__, torch.nn.GroupNorm, '_load_from_state_dict', networks.network_GroupNorm_load_state_dict)
self.LayerNorm_forward = patches.patch(__name__, torch.nn.LayerNorm, 'forward', networks.network_LayerNorm_forward)
self.LayerNorm_load_state_dict = patches.patch(__name__, torch.nn.LayerNorm, '_load_from_state_dict', networks.network_LayerNorm_load_state_dict)
self.MultiheadAttention_forward = patches.patch(__name__, torch.nn.MultiheadAttention, 'forward', networks.network_MultiheadAttention_forward)
self.MultiheadAttention_load_state_dict = patches.patch(__name__, torch.nn.MultiheadAttention, '_load_from_state_dict', networks.network_MultiheadAttention_load_state_dict)
self.handle_quant(apply=True)
networks.timer['load'] = 0
networks.timer['apply'] = 0
networks.timer['restore'] = 0
self.active = True
def undo(self):
if not self.active or shared.opts.lora_force_diffusers:
return
self.Linear_forward = patches.undo(__name__, torch.nn.Linear, 'forward') # pylint: disable=E1128
self.Linear_load_state_dict = patches.undo(__name__, torch.nn.Linear, '_load_from_state_dict') # pylint: disable=E1128
self.Conv2d_forward = patches.undo(__name__, torch.nn.Conv2d, 'forward') # pylint: disable=E1128
self.Conv2d_load_state_dict = patches.undo(__name__, torch.nn.Conv2d, '_load_from_state_dict') # pylint: disable=E1128
self.GroupNorm_forward = patches.undo(__name__, torch.nn.GroupNorm, 'forward') # pylint: disable=E1128
self.GroupNorm_load_state_dict = patches.undo(__name__, torch.nn.GroupNorm, '_load_from_state_dict') # pylint: disable=E1128
self.LayerNorm_forward = patches.undo(__name__, torch.nn.LayerNorm, 'forward') # pylint: disable=E1128
self.LayerNorm_load_state_dict = patches.undo(__name__, torch.nn.LayerNorm, '_load_from_state_dict') # pylint: disable=E1128
self.MultiheadAttention_forward = patches.undo(__name__, torch.nn.MultiheadAttention, 'forward') # pylint: disable=E1128
self.MultiheadAttention_load_state_dict = patches.undo(__name__, torch.nn.MultiheadAttention, '_load_from_state_dict') # pylint: disable=E1128
self.handle_quant(apply=False)
patches.originals.pop(__name__, None)
self.active = False
extensions-builtin/Lora/lyco_helpers.py
-66
View File
@@ -1,66 +0,0 @@
import torch
def make_weight_cp(t, wa, wb):
temp = torch.einsum('i j k l, j r -> i r k l', t, wb)
return torch.einsum('i j k l, i r -> r j k l', temp, wa)
def rebuild_conventional(up, down, shape, dyn_dim=None):
up = up.reshape(up.size(0), -1)
down = down.reshape(down.size(0), -1)
if dyn_dim is not None:
up = up[:, :dyn_dim]
down = down[:dyn_dim, :]
return (up @ down).reshape(shape)
def rebuild_cp_decomposition(up, down, mid):
up = up.reshape(up.size(0), -1)
down = down.reshape(down.size(0), -1)
return torch.einsum('n m k l, i n, m j -> i j k l', mid, up, down)
# copied from https://github.com/KohakuBlueleaf/LyCORIS/blob/dev/lycoris/modules/lokr.py
def factorization(dimension: int, factor:int=-1) -> tuple[int, int]:
'''
return a tuple of two value of input dimension decomposed by the number closest to factor
second value is higher or equal than first value.
In LoRA with Kroneckor Product, first value is a value for weight scale.
secon value is a value for weight.
Becuase of non-commutative property, A⊗B ≠ B⊗A. Meaning of two matrices is slightly different.
examples)
factor
-1 2 4 8 16 ...
127 -> 1, 127 127 -> 1, 127 127 -> 1, 127 127 -> 1, 127 127 -> 1, 127
128 -> 8, 16 128 -> 2, 64 128 -> 4, 32 128 -> 8, 16 128 -> 8, 16
250 -> 10, 25 250 -> 2, 125 250 -> 2, 125 250 -> 5, 50 250 -> 10, 25
360 -> 8, 45 360 -> 2, 180 360 -> 4, 90 360 -> 8, 45 360 -> 12, 30
512 -> 16, 32 512 -> 2, 256 512 -> 4, 128 512 -> 8, 64 512 -> 16, 32
1024 -> 32, 32 1024 -> 2, 512 1024 -> 4, 256 1024 -> 8, 128 1024 -> 16, 64
'''
if factor > 0 and (dimension % factor) == 0:
m = factor
n = dimension // factor
if m > n:
n, m = m, n
return m, n
if factor < 0:
factor = dimension
m, n = 1, dimension
length = m + n
while m<n:
new_m = m + 1
while dimension%new_m != 0:
new_m += 1
new_n = dimension // new_m
if new_m + new_n > length or new_m>factor:
break
m, n = new_m, new_n
if m > n:
n, m = m, n
return m, n
extensions-builtin/Lora/network.py
-187
View File
@@ -1,187 +0,0 @@
from __future__ import annotations
import os
from collections import namedtuple
import enum
from modules import sd_models, hashes, shared
NetworkWeights = namedtuple('NetworkWeights', ['network_key', 'sd_key', 'w', 'sd_module'])
metadata_tags_order = {"ss_sd_model_name": 1, "ss_resolution": 2, "ss_clip_skip": 3, "ss_num_train_images": 10, "ss_tag_frequency": 20}
class SdVersion(enum.Enum):
Unknown = 1
SD1 = 2
SD2 = 3
SD3 = 3
SDXL = 4
SC = 5
F1 = 6
class NetworkOnDisk:
def __init__(self, name, filename):
self.name = name
self.filename = filename
if filename.startswith(shared.cmd_opts.lora_dir):
self.fullname = os.path.splitext(filename[len(shared.cmd_opts.lora_dir):].strip("/"))[0]
else:
self.fullname = name
self.metadata = {}
self.is_safetensors = os.path.splitext(filename)[1].lower() == ".safetensors"
if self.is_safetensors:
self.metadata = sd_models.read_metadata_from_safetensors(filename)
if self.metadata:
m = {}
for k, v in sorted(self.metadata.items(), key=lambda x: metadata_tags_order.get(x[0], 999)):
m[k] = v
self.metadata = m
self.alias = self.metadata.get('ss_output_name', self.name)
# self.set_hash(self.metadata.get('sshs_model_hash') or hashes.sha256_from_cache(self.filename, "lora/" + self.name, use_addnet_hash=self.is_safetensors) or '')
sha256 = hashes.sha256_from_cache(self.filename, "lora/" + self.name) or hashes.sha256_from_cache(self.filename, "lora/" + self.name, use_addnet_hash=True) or self.metadata.get('sshs_model_hash')
self.set_hash(sha256)
self.sd_version = self.detect_version()
def detect_version(self):
base = str(self.metadata.get('ss_base_model_version', "")).lower()
arch = str(self.metadata.get('modelspec.architecture', "")).lower()
if base.startswith("sd_v1"):
return 'sd1'
if base.startswith("sdxl"):
return 'xl'
if base.startswith("stable_cascade"):
return 'sc'
if base.startswith("sd3"):
return 'sd3'
if base.startswith("flux"):
return 'f1'
if arch.startswith("stable-diffusion-v1"):
return 'sd1'
if arch.startswith("stable-diffusion-xl"):
return 'xl'
if arch.startswith("stable-cascade"):
return 'sc'
if arch.startswith("flux"):
return 'f1'
if "v1-5" in str(self.metadata.get('ss_sd_model_name', "")):
return 'sd1'
if str(self.metadata.get('ss_v2', "")) == "True":
return 'sd2'
if 'flux' in self.name.lower():
return 'f1'
if 'xl' in self.name.lower():
return 'xl'
return ''
def set_hash(self, v):
self.hash = v or ''
self.shorthash = self.hash[0:8]
def read_hash(self):
if not self.hash:
self.set_hash(hashes.sha256(self.filename, "lora/" + self.name, use_addnet_hash=self.is_safetensors) or '')
def get_alias(self):
import networks
return self.name if shared.opts.lora_preferred_name == "filename" or self.alias.lower() in networks.forbidden_network_aliases else self.alias
class Network: # LoraModule
def __init__(self, name, network_on_disk: NetworkOnDisk):
self.name = name
self.network_on_disk = network_on_disk
self.te_multiplier = 1.0
self.unet_multiplier = [1.0] * 3
self.dyn_dim = None
self.modules = {}
self.bundle_embeddings = {}
self.mtime = None
self.mentioned_name = None
self.tags = None
"""the text that was used to add the network to prompt - can be either name or an alias"""
class ModuleType:
def create_module(self, net: Network, weights: NetworkWeights) -> Network | None: # pylint: disable=W0613
return None
class NetworkModule:
def __init__(self, net: Network, weights: NetworkWeights):
self.network = net
self.network_key = weights.network_key
self.sd_key = weights.sd_key
self.sd_module = weights.sd_module
if hasattr(self.sd_module, 'weight'):
self.shape = self.sd_module.weight.shape
self.dim = None
self.bias = weights.w.get("bias")
self.alpha = weights.w["alpha"].item() if "alpha" in weights.w else None
self.scale = weights.w["scale"].item() if "scale" in weights.w else None
self.dora_scale = weights.w.get("dora_scale", None)
self.dora_norm_dims = len(self.shape) - 1
def multiplier(self):
unet_multiplier = 3 * [self.network.unet_multiplier] if not isinstance(self.network.unet_multiplier, list) else self.network.unet_multiplier
if 'transformer' in self.sd_key[:20]:
return self.network.te_multiplier
if "down_blocks" in self.sd_key:
return unet_multiplier[0]
if "mid_block" in self.sd_key:
return unet_multiplier[1]
if "up_blocks" in self.sd_key:
return unet_multiplier[2]
else:
return unet_multiplier[0]
def calc_scale(self):
if self.scale is not None:
return self.scale
if self.dim is not None and self.alpha is not None:
return self.alpha / self.dim
return 1.0
def apply_weight_decompose(self, updown, orig_weight):
# Match the device/dtype
orig_weight = orig_weight.to(updown.dtype)
dora_scale = self.dora_scale.to(device=orig_weight.device, dtype=updown.dtype)
updown = updown.to(orig_weight.device)
merged_scale1 = updown + orig_weight
merged_scale1_norm = (
merged_scale1.transpose(0, 1)
.reshape(merged_scale1.shape[1], -1)
.norm(dim=1, keepdim=True)
.reshape(merged_scale1.shape[1], *[1] * self.dora_norm_dims)
.transpose(0, 1)
)
dora_merged = (
merged_scale1 * (dora_scale / merged_scale1_norm)
)
final_updown = dora_merged - orig_weight
return final_updown
def finalize_updown(self, updown, orig_weight, output_shape, ex_bias=None):
if self.bias is not None:
updown = updown.reshape(self.bias.shape)
updown += self.bias.to(orig_weight.device, dtype=orig_weight.dtype)
updown = updown.reshape(output_shape)
if len(output_shape) == 4:
updown = updown.reshape(output_shape)
if orig_weight.size().numel() == updown.size().numel():
updown = updown.reshape(orig_weight.shape)
if ex_bias is not None:
ex_bias = ex_bias * self.multiplier()
if self.dora_scale is not None:
updown = self.apply_weight_decompose(updown, orig_weight)
return updown * self.calc_scale() * self.multiplier(), ex_bias
def calc_updown(self, target):
raise NotImplementedError
def forward(self, x, y):
raise NotImplementedError
extensions-builtin/Lora/network_full.py
-26
View File
@@ -1,26 +0,0 @@
import network
class ModuleTypeFull(network.ModuleType):
def create_module(self, net: network.Network, weights: network.NetworkWeights):
if all(x in weights.w for x in ["diff"]):
return NetworkModuleFull(net, weights)
return None
class NetworkModuleFull(network.NetworkModule): # pylint: disable=abstract-method
def __init__(self, net: network.Network, weights: network.NetworkWeights):
super().__init__(net, weights)
self.weight = weights.w.get("diff")
self.ex_bias = weights.w.get("diff_b")
def calc_updown(self, target):
output_shape = self.weight.shape
updown = self.weight.to(target.device, dtype=target.dtype)
if self.ex_bias is not None:
ex_bias = self.ex_bias.to(target.device, dtype=target.dtype)
else:
ex_bias = None
return self.finalize_updown(updown, target, output_shape, ex_bias)
extensions-builtin/Lora/network_glora.py
-30
View File
@@ -1,30 +0,0 @@
import network
class ModuleTypeGLora(network.ModuleType):
def create_module(self, net: network.Network, weights: network.NetworkWeights):
if all(x in weights.w for x in ["a1.weight", "a2.weight", "alpha", "b1.weight", "b2.weight"]):
return NetworkModuleGLora(net, weights)
return None
# adapted from https://github.com/KohakuBlueleaf/LyCORIS
class NetworkModuleGLora(network.NetworkModule): # pylint: disable=abstract-method
def __init__(self, net: network.Network, weights: network.NetworkWeights):
super().__init__(net, weights)
if hasattr(self.sd_module, 'weight'):
self.shape = self.sd_module.weight.shape
self.w1a = weights.w["a1.weight"]
self.w1b = weights.w["b1.weight"]
self.w2a = weights.w["a2.weight"]
self.w2b = weights.w["b2.weight"]
def calc_updown(self, target): # pylint: disable=arguments-differ
w1a = self.w1a.to(target.device, dtype=target.dtype)
w1b = self.w1b.to(target.device, dtype=target.dtype)
w2a = self.w2a.to(target.device, dtype=target.dtype)
w2b = self.w2b.to(target.device, dtype=target.dtype)
output_shape = [w1a.size(0), w1b.size(1)]
updown = (w2b @ w1b) + ((target @ w2a) @ w1a)
return self.finalize_updown(updown, target, output_shape)
extensions-builtin/Lora/network_hada.py
-46
View File
@@ -1,46 +0,0 @@
import lyco_helpers
import network
class ModuleTypeHada(network.ModuleType):
def create_module(self, net: network.Network, weights: network.NetworkWeights):
if all(x in weights.w for x in ["hada_w1_a", "hada_w1_b", "hada_w2_a", "hada_w2_b"]):
return NetworkModuleHada(net, weights)
return None
class NetworkModuleHada(network.NetworkModule): # pylint: disable=abstract-method
def __init__(self, net: network.Network, weights: network.NetworkWeights):
super().__init__(net, weights)
if hasattr(self.sd_module, 'weight'):
self.shape = self.sd_module.weight.shape
self.w1a = weights.w["hada_w1_a"]
self.w1b = weights.w["hada_w1_b"]
self.dim = self.w1b.shape[0]
self.w2a = weights.w["hada_w2_a"]
self.w2b = weights.w["hada_w2_b"]
self.t1 = weights.w.get("hada_t1")
self.t2 = weights.w.get("hada_t2")
def calc_updown(self, target):
w1a = self.w1a.to(target.device, dtype=target.dtype)
w1b = self.w1b.to(target.device, dtype=target.dtype)
w2a = self.w2a.to(target.device, dtype=target.dtype)
w2b = self.w2b.to(target.device, dtype=target.dtype)
output_shape = [w1a.size(0), w1b.size(1)]
if self.t1 is not None:
output_shape = [w1a.size(1), w1b.size(1)]
t1 = self.t1.to(target.device, dtype=target.dtype)
updown1 = lyco_helpers.make_weight_cp(t1, w1a, w1b)
output_shape += t1.shape[2:]
else:
if len(w1b.shape) == 4:
output_shape += w1b.shape[2:]
updown1 = lyco_helpers.rebuild_conventional(w1a, w1b, output_shape)
if self.t2 is not None:
t2 = self.t2.to(target.device, dtype=target.dtype)
updown2 = lyco_helpers.make_weight_cp(t2, w2a, w2b)
else:
updown2 = lyco_helpers.rebuild_conventional(w2a, w2b, output_shape)
updown = updown1 * updown2
return self.finalize_updown(updown, target, output_shape)
extensions-builtin/Lora/network_ia3.py
-25
View File
@@ -1,25 +0,0 @@
import network
class ModuleTypeIa3(network.ModuleType):
def create_module(self, net: network.Network, weights: network.NetworkWeights):
if all(x in weights.w for x in ["weight"]):
return NetworkModuleIa3(net, weights)
return None
class NetworkModuleIa3(network.NetworkModule): # pylint: disable=abstract-method
def __init__(self, net: network.Network, weights: network.NetworkWeights):
super().__init__(net, weights)
self.w = weights.w["weight"]
self.on_input = weights.w["on_input"].item()
def calc_updown(self, target):
w = self.w.to(target.device, dtype=target.dtype)
output_shape = [w.size(0), target.size(1)]
if self.on_input:
output_shape.reverse()
else:
w = w.reshape(-1, 1)
updown = target * w
return self.finalize_updown(updown, target, output_shape)
extensions-builtin/Lora/network_lokr.py
-57
View File
@@ -1,57 +0,0 @@
import torch
import lyco_helpers
import network
class ModuleTypeLokr(network.ModuleType):
def create_module(self, net: network.Network, weights: network.NetworkWeights):
has_1 = "lokr_w1" in weights.w or ("lokr_w1_a" in weights.w and "lokr_w1_b" in weights.w)
has_2 = "lokr_w2" in weights.w or ("lokr_w2_a" in weights.w and "lokr_w2_b" in weights.w)
if has_1 and has_2:
return NetworkModuleLokr(net, weights)
return None
def make_kron(orig_shape, w1, w2):
if len(w2.shape) == 4:
w1 = w1.unsqueeze(2).unsqueeze(2)
w2 = w2.contiguous()
return torch.kron(w1, w2).reshape(orig_shape)
class NetworkModuleLokr(network.NetworkModule): # pylint: disable=abstract-method
def __init__(self, net: network.Network, weights: network.NetworkWeights):
super().__init__(net, weights)
self.w1 = weights.w.get("lokr_w1")
self.w1a = weights.w.get("lokr_w1_a")
self.w1b = weights.w.get("lokr_w1_b")
self.dim = self.w1b.shape[0] if self.w1b is not None else self.dim
self.w2 = weights.w.get("lokr_w2")
self.w2a = weights.w.get("lokr_w2_a")
self.w2b = weights.w.get("lokr_w2_b")
self.dim = self.w2b.shape[0] if self.w2b is not None else self.dim
self.t2 = weights.w.get("lokr_t2")
def calc_updown(self, target):
if self.w1 is not None:
w1 = self.w1.to(target.device, dtype=target.dtype)
else:
w1a = self.w1a.to(target.device, dtype=target.dtype)
w1b = self.w1b.to(target.device, dtype=target.dtype)
w1 = w1a @ w1b
if self.w2 is not None:
w2 = self.w2.to(target.device, dtype=target.dtype)
elif self.t2 is None:
w2a = self.w2a.to(target.device, dtype=target.dtype)
w2b = self.w2b.to(target.device, dtype=target.dtype)
w2 = w2a @ w2b
else:
t2 = self.t2.to(target.device, dtype=target.dtype)
w2a = self.w2a.to(target.device, dtype=target.dtype)
w2b = self.w2b.to(target.device, dtype=target.dtype)
w2 = lyco_helpers.make_weight_cp(t2, w2a, w2b)
output_shape = [w1.size(0) * w2.size(0), w1.size(1) * w2.size(1)]
if len(target.shape) == 4:
output_shape = target.shape
updown = make_kron(output_shape, w1, w2)
return self.finalize_updown(updown, target, output_shape)
extensions-builtin/Lora/network_lora.py
-75
View File
@@ -1,75 +0,0 @@
import torch
import diffusers.models.lora as diffusers_lora
import lyco_helpers
import network
from modules import devices
class ModuleTypeLora(network.ModuleType):
def create_module(self, net: network.Network, weights: network.NetworkWeights):
if all(x in weights.w for x in ["lora_up.weight", "lora_down.weight"]):
return NetworkModuleLora(net, weights)
return None
class NetworkModuleLora(network.NetworkModule):
def __init__(self, net: network.Network, weights: network.NetworkWeights):
super().__init__(net, weights)
self.up_model = self.create_module(weights.w, "lora_up.weight")
self.down_model = self.create_module(weights.w, "lora_down.weight")
self.mid_model = self.create_module(weights.w, "lora_mid.weight", none_ok=True)
self.dim = weights.w["lora_down.weight"].shape[0]
def create_module(self, weights, key, none_ok=False):
weight = weights.get(key)
if weight is None and none_ok:
return None
linear_modules = [torch.nn.Linear, torch.nn.modules.linear.NonDynamicallyQuantizableLinear, torch.nn.MultiheadAttention, diffusers_lora.LoRACompatibleLinear]
is_linear = type(self.sd_module) in linear_modules or self.sd_module.__class__.__name__ in {"NNCFLinear", "QLinear", "Linear4bit"}
is_conv = type(self.sd_module) in [torch.nn.Conv2d, diffusers_lora.LoRACompatibleConv] or self.sd_module.__class__.__name__ in {"NNCFConv2d", "QConv2d"}
if is_linear:
weight = weight.reshape(weight.shape[0], -1)
module = torch.nn.Linear(weight.shape[1], weight.shape[0], bias=False)
elif is_conv and (key == "lora_down.weight" or key == "dyn_up"):
if len(weight.shape) == 2:
weight = weight.reshape(weight.shape[0], -1, 1, 1)
if weight.shape[2] != 1 or weight.shape[3] != 1:
module = torch.nn.Conv2d(weight.shape[1], weight.shape[0], self.sd_module.kernel_size, self.sd_module.stride, self.sd_module.padding, bias=False)
else:
module = torch.nn.Conv2d(weight.shape[1], weight.shape[0], (1, 1), bias=False)
elif is_conv and key == "lora_mid.weight":
module = torch.nn.Conv2d(weight.shape[1], weight.shape[0], self.sd_module.kernel_size, self.sd_module.stride, self.sd_module.padding, bias=False)
elif is_conv and (key == "lora_up.weight" or key == "dyn_down"):
module = torch.nn.Conv2d(weight.shape[1], weight.shape[0], (1, 1), bias=False)
else:
raise AssertionError(f'Lora unsupported: layer={self.network_key} type={type(self.sd_module).__name__}')
with torch.no_grad():
if weight.shape != module.weight.shape:
weight = weight.reshape(module.weight.shape)
module.weight.copy_(weight)
module.weight.requires_grad_(False)
return module
def calc_updown(self, target): # pylint: disable=W0237
target_dtype = target.dtype if target.dtype != torch.uint8 else self.up_model.weight.dtype
up = self.up_model.weight.to(target.device, dtype=target_dtype)
down = self.down_model.weight.to(target.device, dtype=target_dtype)
output_shape = [up.size(0), down.size(1)]
if self.mid_model is not None:
# cp-decomposition
mid = self.mid_model.weight.to(target.device, dtype=target_dtype)
updown = lyco_helpers.rebuild_cp_decomposition(up, down, mid)
output_shape += mid.shape[2:]
else:
if len(down.shape) == 4:
output_shape += down.shape[2:]
updown = lyco_helpers.rebuild_conventional(up, down, output_shape, self.network.dyn_dim)
return self.finalize_updown(updown, target, output_shape)
def forward(self, x, y):
self.up_model.to(device=devices.device)
self.down_model.to(device=devices.device)
if hasattr(y, "scale"):
return y(scale=1) + self.up_model(self.down_model(x)) * self.multiplier() * self.calc_scale()
return y + self.up_model(self.down_model(x)) * self.multiplier() * self.calc_scale()
extensions-builtin/Lora/network_norm.py
-24
View File
@@ -1,24 +0,0 @@
import network
class ModuleTypeNorm(network.ModuleType):
def create_module(self, net: network.Network, weights: network.NetworkWeights):
if all(x in weights.w for x in ["w_norm", "b_norm"]):
return NetworkModuleNorm(net, weights)
return None
class NetworkModuleNorm(network.NetworkModule): # pylint: disable=abstract-method
def __init__(self, net: network.Network, weights: network.NetworkWeights):
super().__init__(net, weights)
self.w_norm = weights.w.get("w_norm")
self.b_norm = weights.w.get("b_norm")
def calc_updown(self, target):
output_shape = self.w_norm.shape
updown = self.w_norm.to(target.device, dtype=target.dtype)
if self.b_norm is not None:
ex_bias = self.b_norm.to(target.device, dtype=target.dtype)
else:
ex_bias = None
return self.finalize_updown(updown, target, output_shape, ex_bias)
extensions-builtin/Lora/network_oft.py
-81
View File
@@ -1,81 +0,0 @@
import torch
import network
from lyco_helpers import factorization
from einops import rearrange
class ModuleTypeOFT(network.ModuleType):
def create_module(self, net: network.Network, weights: network.NetworkWeights):
if all(x in weights.w for x in ["oft_blocks"]) or all(x in weights.w for x in ["oft_diag"]):
return NetworkModuleOFT(net, weights)
return None
# Supports both kohya-ss' implementation of COFT https://github.com/kohya-ss/sd-scripts/blob/main/networks/oft.py
# and KohakuBlueleaf's implementation of OFT/COFT https://github.com/KohakuBlueleaf/LyCORIS/blob/dev/lycoris/modules/diag_oft.py
class NetworkModuleOFT(network.NetworkModule): # pylint: disable=abstract-method
def __init__(self, net: network.Network, weights: network.NetworkWeights):
super().__init__(net, weights)
self.lin_module = None
self.org_module: list[torch.Module] = [self.sd_module]
self.scale = 1.0
# kohya-ss
if "oft_blocks" in weights.w.keys():
self.is_kohya = True
self.oft_blocks = weights.w["oft_blocks"] # (num_blocks, block_size, block_size)
self.alpha = weights.w["alpha"] # alpha is constraint
self.dim = self.oft_blocks.shape[0] # lora dim
# LyCORIS
elif "oft_diag" in weights.w.keys():
self.is_kohya = False
self.oft_blocks = weights.w["oft_diag"]
# self.alpha is unused
self.dim = self.oft_blocks.shape[1] # (num_blocks, block_size, block_size)
is_linear = type(self.sd_module) in [torch.nn.Linear, torch.nn.modules.linear.NonDynamicallyQuantizableLinear]
is_conv = type(self.sd_module) in [torch.nn.Conv2d]
is_other_linear = type(self.sd_module) in [torch.nn.MultiheadAttention] # unsupported
if is_linear:
self.out_dim = self.sd_module.out_features
elif is_conv:
self.out_dim = self.sd_module.out_channels
elif is_other_linear:
self.out_dim = self.sd_module.embed_dim
if self.is_kohya:
self.constraint = self.alpha * self.out_dim
self.num_blocks = self.dim
self.block_size = self.out_dim // self.dim
else:
self.constraint = None
self.block_size, self.num_blocks = factorization(self.out_dim, self.dim)
def calc_updown(self, target):
oft_blocks = self.oft_blocks.to(target.device, dtype=target.dtype)
eye = torch.eye(self.block_size, device=target.device)
constraint = self.constraint.to(target.device)
if self.is_kohya:
block_Q = oft_blocks - oft_blocks.transpose(1, 2) # ensure skew-symmetric orthogonal matrix
norm_Q = torch.norm(block_Q.flatten()).to(target.device)
new_norm_Q = torch.clamp(norm_Q, max=constraint)
block_Q = block_Q * ((new_norm_Q + 1e-8) / (norm_Q + 1e-8))
mat1 = eye + block_Q
mat2 = (eye - block_Q).float().inverse()
oft_blocks = torch.matmul(mat1, mat2)
R = oft_blocks.to(target.device, dtype=target.dtype)
# This errors out for MultiheadAttention, might need to be handled up-stream
merged_weight = rearrange(target, '(k n) ... -> k n ...', k=self.num_blocks, n=self.block_size)
merged_weight = torch.einsum(
'k n m, k n ... -> k m ...',
R,
merged_weight
)
merged_weight = rearrange(merged_weight, 'k m ... -> (k m) ...')
updown = merged_weight.to(target.device, dtype=target.dtype) - target
output_shape = target.shape
return self.finalize_updown(updown, target, output_shape)
extensions-builtin/Lora/network_overrides.py
-48
View File
@@ -1,48 +0,0 @@
from modules import shared
maybe_diffusers = [ # forced if lora_maybe_diffusers is enabled
'aaebf6360f7d', # sd15-lcm
'3d18b05e4f56', # sdxl-lcm
'b71dcb732467', # sdxl-tcd
'813ea5fb1c67', # sdxl-turbo
# not really needed, but just in case
'5a48ac366664', # hyper-sd15-1step
'ee0ff23dcc42', # hyper-sd15-2step
'e476eb1da5df', # hyper-sd15-4step
'ecb844c3f3b0', # hyper-sd15-8step
'1ab289133ebb', # hyper-sd15-8step-cfg
'4f494295edb1', # hyper-sdxl-8step
'ca14a8c621f8', # hyper-sdxl-8step-cfg
'1c88f7295856', # hyper-sdxl-4step
'fdd5dcd1d88a', # hyper-sdxl-2step
'8cca3706050b', # hyper-sdxl-1step
]
force_diffusers = [ # forced always
'816d0eed49fd', # flash-sdxl
'c2ec22757b46', # flash-sd15
]
force_models = [ # forced always
'sc',
'kandinsky',
'hunyuandit',
'auraflow',
]
force_classes = [ # forced always
]
def check_override(shorthash=''):
force = False
force = force or (shared.sd_model_type in force_models)
force = force or (shared.sd_model.__class__.__name__ in force_classes)
if len(shorthash) < 4:
return force
force = force or (any(x.startswith(shorthash) for x in maybe_diffusers) if shared.opts.lora_maybe_diffusers else False)
force = force or any(x.startswith(shorthash) for x in force_diffusers)
if force and shared.opts.lora_maybe_diffusers:
shared.log.debug('LoRA override: force diffusers')
return force
extensions-builtin/Lora/networks.py
-604
View File
@@ -1,604 +0,0 @@
from typing import Union, List
import os
import re
import time
import concurrent
import lora_patches
import network
import network_lora
import network_hada
import network_ia3
import network_oft
import network_lokr
import network_full
import network_norm
import network_glora
import network_overrides
import lora_convert
import torch
import diffusers.models.lora
from modules import shared, devices, sd_models, sd_models_compile, errors, scripts, files_cache, model_quant
debug = os.environ.get('SD_LORA_DEBUG', None) is not None
originals: lora_patches.LoraPatches = None
extra_network_lora = None
available_networks = {}
available_network_aliases = {}
loaded_networks: List[network.Network] = []
timer = { 'load': 0, 'apply': 0, 'restore': 0, 'deactivate': 0 }
# networks_in_memory = {}
lora_cache = {}
diffuser_loaded = []
diffuser_scales = []
available_network_hash_lookup = {}
forbidden_network_aliases = {}
re_network_name = re.compile(r"(.*)\s*\([0-9a-fA-F]+\)")
module_types = [
network_lora.ModuleTypeLora(),
network_hada.ModuleTypeHada(),
network_ia3.ModuleTypeIa3(),
network_oft.ModuleTypeOFT(),
network_lokr.ModuleTypeLokr(),
network_full.ModuleTypeFull(),
network_norm.ModuleTypeNorm(),
network_glora.ModuleTypeGLora(),
]
convert_diffusers_name_to_compvis = lora_convert.convert_diffusers_name_to_compvis # supermerger compatibility item
def assign_network_names_to_compvis_modules(sd_model):
if sd_model is None:
return
sd_model = getattr(shared.sd_model, "pipe", shared.sd_model) # wrapped model compatiblility
network_layer_mapping = {}
if shared.native:
if hasattr(sd_model, 'text_encoder') and sd_model.text_encoder is not None:
for name, module in sd_model.text_encoder.named_modules():
prefix = "lora_te1_" if hasattr(sd_model, 'text_encoder_2') else "lora_te_"
network_name = prefix + name.replace(".", "_")
network_layer_mapping[network_name] = module
module.network_layer_name = network_name
if hasattr(sd_model, 'text_encoder_2'):
for name, module in sd_model.text_encoder_2.named_modules():
network_name = "lora_te2_" + name.replace(".", "_")
network_layer_mapping[network_name] = module
module.network_layer_name = network_name
if hasattr(sd_model, 'unet'):
for name, module in sd_model.unet.named_modules():
network_name = "lora_unet_" + name.replace(".", "_")
network_layer_mapping[network_name] = module
module.network_layer_name = network_name
if hasattr(sd_model, 'transformer'):
for name, module in sd_model.transformer.named_modules():
network_name = "lora_transformer_" + name.replace(".", "_")
network_layer_mapping[network_name] = module
if "norm" in network_name and "linear" not in network_name:
continue
module.network_layer_name = network_name
else:
if not hasattr(sd_model, 'cond_stage_model'):
sd_model.network_layer_mapping = {}
return
for name, module in sd_model.cond_stage_model.wrapped.named_modules():
network_name = name.replace(".", "_")
network_layer_mapping[network_name] = module
module.network_layer_name = network_name
for name, module in sd_model.model.named_modules():
network_name = name.replace(".", "_")
network_layer_mapping[network_name] = module
module.network_layer_name = network_name
sd_model.network_layer_mapping = network_layer_mapping
def load_diffusers(name, network_on_disk, lora_scale=shared.opts.extra_networks_default_multiplier) -> network.Network:
t0 = time.time()
name = name.replace(".", "_")
#cached = lora_cache.get(name, None)
shared.log.debug(f'Network load: type=LoRA name="{name}" file="{network_on_disk.filename}" detected={network_on_disk.sd_version} method=diffusers scale={lora_scale} fuse={shared.opts.lora_fuse_diffusers}')
# if cached is not None:
# return cached
if not shared.native:
return None
if not hasattr(shared.sd_model, 'load_lora_weights'):
shared.log.error(f'Network load: type=LoRA class={shared.sd_model.__class__} does not implement load lora')
return None
try:
shared.sd_model.load_lora_weights(network_on_disk.filename, adapter_name=name)
except Exception as e:
if 'already in use' in str(e):
pass
else:
if 'The following keys have not been correctly renamed' in str(e):
shared.log.error(f'Network load: type=LoRA name="{name}" diffusers unsupported format')
else:
shared.log.error(f'Network load: type=LoRA name="{name}" {e}')
if debug:
errors.display(e, "LoRA")
return None
if name not in diffuser_loaded:
diffuser_loaded.append(name)
diffuser_scales.append(lora_scale)
net = network.Network(name, network_on_disk)
net.mtime = os.path.getmtime(network_on_disk.filename)
# lora_cache[name] = net
t1 = time.time()
timer['load'] += t1 - t0
return net
def load_network(name, network_on_disk) -> network.Network:
if not shared.sd_loaded:
return None
t0 = time.time()
cached = lora_cache.get(name, None)
if debug:
shared.log.debug(f'Network load: type=LoRA name="{name}" file="{network_on_disk.filename}" type=lora {"cached" if cached else ""}')
if cached is not None:
return cached
net = network.Network(name, network_on_disk)
net.mtime = os.path.getmtime(network_on_disk.filename)
sd = sd_models.read_state_dict(network_on_disk.filename, what='network')
if shared.sd_model_type in ['f1', 'chroma']: # if kohya flux lora, convert state_dict
sd = lora_convert._convert_kohya_flux_lora_to_diffusers(sd) or sd # pylint: disable=protected-access
assign_network_names_to_compvis_modules(shared.sd_model)
keys_failed_to_match = {}
matched_networks = {}
bundle_embeddings = {}
convert = lora_convert.KeyConvert()
for key_network, weight in sd.items():
parts = key_network.split('.')
if parts[0] == "bundle_emb":
emb_name, vec_name = parts[1], key_network.split(".", 2)[-1]
emb_dict = bundle_embeddings.get(emb_name, {})
emb_dict[vec_name] = weight
bundle_embeddings[emb_name] = emb_dict
if len(parts) > 5: # messy handler for diffusers peft lora
key_network_without_network_parts = '_'.join(parts[:-2])
if not key_network_without_network_parts.startswith('lora_'):
key_network_without_network_parts = 'lora_' + key_network_without_network_parts
network_part = '.'.join(parts[-2:]).replace('lora_A', 'lora_down').replace('lora_B', 'lora_up')
else:
key_network_without_network_parts, network_part = key_network.split(".", 1)
key, sd_module = convert(key_network_without_network_parts) # Now returns lists
if sd_module[0] is None:
if "bundle_emb" not in key_network:
keys_failed_to_match[key_network] = key
continue
for k, module in zip(key, sd_module):
if k not in matched_networks:
matched_networks[k] = network.NetworkWeights(network_key=key_network, sd_key=k, w={}, sd_module=module)
matched_networks[k].w[network_part] = weight
network_types = []
for key, weights in matched_networks.items():
net_module = None
for nettype in module_types:
net_module = nettype.create_module(net, weights)
if net_module is not None:
network_types.append(nettype.__class__.__name__)
break
if net_module is None:
shared.log.error(f'LoRA unhandled: name={name} key={key} weights={weights.w.keys()}')
else:
net.modules[key] = net_module
if len(keys_failed_to_match) > 0:
shared.log.warning(f'Network load: type=LoRA name="{name}" type={set(network_types)} unmatched={len(keys_failed_to_match)} matched={len(matched_networks)}')
if debug:
shared.log.debug(f'Network load: type=LoRA name="{name}" unmatched={keys_failed_to_match}')
else:
shared.log.debug(f'Network load: type=LoRA name="{name}" type={set(network_types)} keys={len(matched_networks)}')
if len(matched_networks) == 0:
return None
lora_cache[name] = net
t1 = time.time()
net.bundle_embeddings = bundle_embeddings
timer['load'] += t1 - t0
return net
def load_networks(names, te_multipliers=None, unet_multipliers=None, dyn_dims=None):
networks_on_disk: list[network.NetworkOnDisk] = [available_network_aliases.get(name, None) for name in names]
if any(x is None for x in networks_on_disk):
list_available_networks()
networks_on_disk: list[network.NetworkOnDisk] = [available_network_aliases.get(name, None) for name in names]
failed_to_load_networks = []
recompile_model = False
if shared.compiled_model_state is not None and shared.compiled_model_state.is_compiled:
if len(names) == len(shared.compiled_model_state.lora_model):
for i, name in enumerate(names):
if shared.compiled_model_state.lora_model[i] != f"{name}:{te_multipliers[i] if te_multipliers else shared.opts.extra_networks_default_multiplier}":
recompile_model = True
shared.compiled_model_state.lora_model = []
break
if not recompile_model:
if len(loaded_networks) > 0 and debug:
shared.log.debug('Model Compile: Skipping LoRa loading')
return
else:
recompile_model = True
shared.compiled_model_state.lora_model = []
if recompile_model:
backup_cuda_compile = shared.opts.cuda_compile
sd_models.unload_model_weights(op='model')
shared.opts.cuda_compile = []
sd_models.reload_model_weights(op='model')
shared.opts.cuda_compile = backup_cuda_compile
loaded_networks.clear()
diffuser_loaded.clear()
diffuser_scales.clear()
for i, (network_on_disk, name) in enumerate(zip(networks_on_disk, names)):
net = None
if network_on_disk is not None:
shorthash = getattr(network_on_disk, 'shorthash', '').lower()
if debug:
shared.log.debug(f'Network load: type=LoRA name="{name}" file="{network_on_disk.filename}" hash="{shorthash}"')
try:
if recompile_model:
shared.compiled_model_state.lora_model.append(f"{name}:{te_multipliers[i] if te_multipliers else shared.opts.extra_networks_default_multiplier}")
if shared.native and (shared.opts.lora_force_diffusers or network_overrides.check_override(shorthash)): # OpenVINO only works with Diffusers LoRa loading
net = load_diffusers(name, network_on_disk, lora_scale=te_multipliers[i] if te_multipliers else shared.opts.extra_networks_default_multiplier)
else:
net = load_network(name, network_on_disk)
if net is not None:
net.mentioned_name = name
network_on_disk.read_hash()
except Exception as e:
shared.log.error(f'Network load: type=LoRA file="{network_on_disk.filename}" {e}')
if debug:
errors.display(e, 'LoRA')
continue
if net is None:
failed_to_load_networks.append(name)
shared.log.error(f'Network load: type=LoRA name="{name}" detected={network_on_disk.sd_version if network_on_disk is not None else None} failed')
continue
if shared.native:
shared.sd_model.embedding_db.load_diffusers_embedding(None, net.bundle_embeddings)
net.te_multiplier = te_multipliers[i] if te_multipliers else shared.opts.extra_networks_default_multiplier
net.unet_multiplier = unet_multipliers[i] if unet_multipliers else shared.opts.extra_networks_default_multiplier
net.dyn_dim = dyn_dims[i] if dyn_dims else shared.opts.extra_networks_default_multiplier
loaded_networks.append(net)
while len(lora_cache) > shared.opts.lora_in_memory_limit:
name = next(iter(lora_cache))
lora_cache.pop(name, None)
if len(diffuser_loaded) > 0:
shared.log.debug(f'Network load: type=LoRA loaded={diffuser_loaded} available={shared.sd_model.get_list_adapters()} active={shared.sd_model.get_active_adapters()} scales={diffuser_scales}')
try:
shared.sd_model.set_adapters(adapter_names=diffuser_loaded, adapter_weights=diffuser_scales)
if shared.opts.lora_fuse_diffusers:
shared.sd_model.fuse_lora(adapter_names=diffuser_loaded, lora_scale=1.0, fuse_unet=True, fuse_text_encoder=True) # fuse uses fixed scale since later apply does the scaling
shared.sd_model.unload_lora_weights()
except Exception as e:
shared.log.error(f'Network load: type=LoRA {e}')
if debug:
errors.display(e, 'LoRA')
if len(loaded_networks) > 0 and debug:
shared.log.debug(f'Network load: type=LoRA loaded={len(loaded_networks)} cache={list(lora_cache)}')
devices.torch_gc()
if recompile_model:
shared.log.info("Network load: type=LoRA recompiling model")
backup_lora_model = shared.compiled_model_state.lora_model
if 'Model' in shared.opts.cuda_compile:
shared.sd_model = sd_models_compile.compile_diffusers(shared.sd_model)
shared.compiled_model_state.lora_model = backup_lora_model
if shared.opts.diffusers_offload_mode == "balanced":
sd_models.apply_balanced_offload(shared.sd_model)
def network_restore_weights_from_backup(self: Union[torch.nn.Conv2d, torch.nn.Linear, torch.nn.GroupNorm, torch.nn.LayerNorm, torch.nn.MultiheadAttention, diffusers.models.lora.LoRACompatibleLinear, diffusers.models.lora.LoRACompatibleConv]):
t0 = time.time()
weights_backup = getattr(self, "network_weights_backup", None)
bias_backup = getattr(self, "network_bias_backup", None)
if weights_backup is None and bias_backup is None:
t1 = time.time()
timer['restore'] += t1 - t0
return
# if debug:
# shared.log.debug('LoRA restore weights')
if weights_backup is not None:
if isinstance(self, torch.nn.MultiheadAttention):
self.in_proj_weight.copy_(weights_backup[0])
self.out_proj.weight.copy_(weights_backup[1])
elif hasattr(self, "qweight") and hasattr(self, "freeze"):
self.weight = torch.nn.Parameter(weights_backup.to(self.weight.device, copy=True))
self.freeze()
elif getattr(self, "quant_type", None) in ['nf4', 'fp4']:
bnb = model_quant.load_bnb('Network load: type=LoRA', silent=True)
if bnb is not None:
device = self.weight.device
self.weight = bnb.nn.Params4bit(weights_backup, quant_state=self.quant_state, quant_type=self.quant_type, blocksize=self.blocksize)
self.weight.to(device)
else:
self.weight.copy_(weights_backup)
else:
self.weight.copy_(weights_backup)
if bias_backup is not None:
if isinstance(self, torch.nn.MultiheadAttention):
self.out_proj.bias.copy_(bias_backup)
else:
self.bias.copy_(bias_backup)
else:
if isinstance(self, torch.nn.MultiheadAttention):
self.out_proj.bias = None
else:
self.bias = None
t1 = time.time()
timer['restore'] += t1 - t0
def maybe_backup_weights(self: Union[torch.nn.Conv2d, torch.nn.Linear, torch.nn.GroupNorm, torch.nn.LayerNorm, torch.nn.MultiheadAttention, diffusers.models.lora.LoRACompatibleLinear, diffusers.models.lora.LoRACompatibleConv], wanted_names, current_names): # pylint: disable=W0613
weights_backup = getattr(self, "network_weights_backup", None)
if weights_backup is None and wanted_names != (): # pylint: disable=C1803
if isinstance(self, torch.nn.MultiheadAttention):
weights_backup = (self.in_proj_weight.clone().to(devices.cpu), self.out_proj.weight.clone().to(devices.cpu))
elif getattr(self.weight, "quant_type", None) in ['nf4', 'fp4']:
bnb = model_quant.load_bnb('Network load: type=LoRA', silent=True)
if bnb is not None:
with devices.inference_context():
weights_backup = bnb.functional.dequantize_4bit(self.weight, quant_state=self.weight.quant_state, quant_type=self.weight.quant_type, blocksize=self.weight.blocksize,).to(devices.cpu)
self.quant_state = self.weight.quant_state
self.quant_type = self.weight.quant_type
self.blocksize = self.weight.blocksize
else:
weights_backup = self.weight.clone().to(devices.cpu)
else:
weights_backup = self.weight.clone().to(devices.cpu)
self.network_weights_backup = weights_backup
bias_backup = getattr(self, "network_bias_backup", None)
if bias_backup is None:
if isinstance(self, torch.nn.MultiheadAttention) and self.out_proj.bias is not None:
bias_backup = self.out_proj.bias.clone().to(devices.cpu)
elif getattr(self, 'bias', None) is not None:
bias_backup = self.bias.clone().to(devices.cpu)
else:
bias_backup = None
self.network_bias_backup = bias_backup
def network_apply_weights(self: Union[torch.nn.Conv2d, torch.nn.Linear, torch.nn.GroupNorm, torch.nn.LayerNorm, torch.nn.MultiheadAttention, diffusers.models.lora.LoRACompatibleLinear, diffusers.models.lora.LoRACompatibleConv]):
"""
Applies the currently selected set of networks to the weights of torch layer self.
If weights already have this particular set of networks applied, does nothing.
If not, restores orginal weights from backup and alters weights according to networks.
"""
network_layer_name = getattr(self, 'network_layer_name', None)
if network_layer_name is None:
return
t0 = time.time()
current_names = getattr(self, "network_current_names", ())
wanted_names = tuple((x.name, x.te_multiplier, x.unet_multiplier, x.dyn_dim) for x in loaded_networks)
if any([net.modules.get(network_layer_name, None) for net in loaded_networks]): # noqa: C419 # pylint: disable=R1729
maybe_backup_weights(self, wanted_names, current_names)
if current_names != wanted_names:
network_restore_weights_from_backup(self)
for net in loaded_networks:
# default workflow where module is known and has weights
module = net.modules.get(network_layer_name, None)
if module is not None and hasattr(self, 'weight'):
try:
with devices.inference_context():
weight = self.weight # calculate quant weights once
updown, ex_bias = module.calc_updown(weight)
if len(weight.shape) == 4 and weight.shape[1] == 9:
# inpainting model. zero pad updown to make channel[1] 4 to 9
updown = torch.nn.functional.pad(updown, (0, 0, 0, 0, 0, 5)) # pylint: disable=not-callable
if getattr(self.weight, "quant_type", None) in ['nf4', 'fp4']: # or self.weight.numel() != updown.numel():
bnb = model_quant.load_bnb('Network load: type=LoRA', silent=True)
if bnb is not None:
device = self.weight.device
weight = bnb.functional.dequantize_4bit(self.weight, quant_state=self.weight.quant_state, quant_type=self.weight.quant_type, blocksize=self.weight.blocksize)
self.weight = bnb.nn.Params4bit(weight + updown, quant_state=self.quant_state, quant_type=shared.opts.lora_quant.lower(), blocksize=self.blocksize)
self.weight.to(device)
else:
self.weight = torch.nn.Parameter(weight + updown)
else:
self.weight = torch.nn.Parameter(weight + updown)
if hasattr(self, "qweight") and hasattr(self, "freeze"):
self.freeze()
if ex_bias is not None and hasattr(self, 'bias'):
if self.bias is None:
self.bias = torch.nn.Parameter(ex_bias)
else:
self.bias += ex_bias
except RuntimeError as e:
extra_network_lora.errors[net.name] = extra_network_lora.errors.get(net.name, 0) + 1
if debug:
module_name = net.modules.get(network_layer_name, None)
shared.log.error(f'LoRA apply weight name="{net.name}" module="{module_name}" layer="{network_layer_name}" {e}')
errors.display(e, 'LoRA')
raise RuntimeError('LoRA apply weight') from e
continue
# alternative workflow looking at _*_proj layers
module_q = net.modules.get(network_layer_name + "_q_proj", None)
module_k = net.modules.get(network_layer_name + "_k_proj", None)
module_v = net.modules.get(network_layer_name + "_v_proj", None)
module_out = net.modules.get(network_layer_name + "_out_proj", None)
if isinstance(self, torch.nn.MultiheadAttention) and module_q and module_k and module_v and module_out:
try:
with devices.inference_context():
updown_q, _ = module_q.calc_updown(self.in_proj_weight)
updown_k, _ = module_k.calc_updown(self.in_proj_weight)
updown_v, _ = module_v.calc_updown(self.in_proj_weight)
updown_qkv = torch.vstack([updown_q, updown_k, updown_v])
updown_out, ex_bias = module_out.calc_updown(self.out_proj.weight)
self.in_proj_weight += updown_qkv
self.out_proj.weight += updown_out
if ex_bias is not None:
if self.out_proj.bias is None:
self.out_proj.bias = torch.nn.Parameter(ex_bias)
else:
self.out_proj.bias += ex_bias
except RuntimeError as e:
if debug:
shared.log.debug(f'LoRA network="{net.name}" layer="{network_layer_name}" {e}')
extra_network_lora.errors[net.name] = extra_network_lora.errors.get(net.name, 0) + 1
continue
if module is None:
continue
shared.log.warning(f'LoRA network="{net.name}" layer="{network_layer_name}" unsupported operation')
extra_network_lora.errors[net.name] = extra_network_lora.errors.get(net.name, 0) + 1
self.network_current_names = wanted_names
t1 = time.time()
timer['apply'] += t1 - t0
def network_forward(module, input, original_forward): # pylint: disable=W0622
"""
Old way of applying Lora by executing operations during layer's forward.
Stacking many loras this way results in big performance degradation.
"""
if len(loaded_networks) == 0:
return original_forward(module, input)
input = devices.cond_cast_unet(input)
network_restore_weights_from_backup(module)
network_reset_cached_weight(module)
y = original_forward(module, input)
network_layer_name = getattr(module, 'network_layer_name', None)
for lora in loaded_networks:
module = lora.modules.get(network_layer_name, None)
if module is None:
continue
y = module.forward(input, y)
return y
def network_reset_cached_weight(self: Union[torch.nn.Conv2d, torch.nn.Linear]):
self.network_current_names = ()
self.network_weights_backup = None
def network_Linear_forward(self, input): # pylint: disable=W0622
network_apply_weights(self)
return originals.Linear_forward(self, input)
def network_QLinear_forward(self, input): # pylint: disable=W0622
network_apply_weights(self)
return torch.nn.functional.linear(input, self.qweight, bias=self.bias)
def network_Linear_load_state_dict(self, *args, **kwargs):
network_reset_cached_weight(self)
return originals.Linear_load_state_dict(self, *args, **kwargs)
def network_Linear4bit_forward(self, input): # pylint: disable=W0622
network_apply_weights(self)
return originals.Linear4bit_forward(self, input)
#
# def network_Linear4bit_load_state_dict(self, *args, **kwargs):
# network_reset_cached_weight(self)
# return originals.Linear4bit_load_state_dict(self, *args, **kwargs)
def network_Conv2d_forward(self, input): # pylint: disable=W0622
network_apply_weights(self)
return originals.Conv2d_forward(self, input)
def network_QConv2d_forward(self, input): # pylint: disable=W0622
network_apply_weights(self)
return self._conv_forward(input, self.qweight, self.bias) # pylint: disable=protected-access
def network_Conv2d_load_state_dict(self, *args, **kwargs):
network_reset_cached_weight(self)
return originals.Conv2d_load_state_dict(self, *args, **kwargs)
def network_GroupNorm_forward(self, input): # pylint: disable=W0622
network_apply_weights(self)
return originals.GroupNorm_forward(self, input)
def network_GroupNorm_load_state_dict(self, *args, **kwargs):
network_reset_cached_weight(self)
return originals.GroupNorm_load_state_dict(self, *args, **kwargs)
def network_LayerNorm_forward(self, input): # pylint: disable=W0622
network_apply_weights(self)
return originals.LayerNorm_forward(self, input)
def network_LayerNorm_load_state_dict(self, *args, **kwargs):
network_reset_cached_weight(self)
return originals.LayerNorm_load_state_dict(self, *args, **kwargs)
def network_MultiheadAttention_forward(self, *args, **kwargs):
network_apply_weights(self)
return originals.MultiheadAttention_forward(self, *args, **kwargs)
def network_MultiheadAttention_load_state_dict(self, *args, **kwargs):
network_reset_cached_weight(self)
return originals.MultiheadAttention_load_state_dict(self, *args, **kwargs)
def list_available_networks():
t0 = time.time()
available_networks.clear()
available_network_aliases.clear()
forbidden_network_aliases.clear()
available_network_hash_lookup.clear()
forbidden_network_aliases.update({"none": 1, "Addams": 1})
if not os.path.exists(shared.cmd_opts.lora_dir):
shared.log.warning(f'LoRA directory not found: path="{shared.cmd_opts.lora_dir}"')
def add_network(filename):
if not os.path.isfile(filename):
return
name = os.path.splitext(os.path.basename(filename))[0]
name = name.replace('.', '_')
try:
entry = network.NetworkOnDisk(name, filename)
available_networks[entry.name] = entry
if entry.alias in available_network_aliases:
forbidden_network_aliases[entry.alias.lower()] = 1
if shared.opts.lora_preferred_name == 'filename':
available_network_aliases[entry.name] = entry
else:
available_network_aliases[entry.alias] = entry
if entry.shorthash:
available_network_hash_lookup[entry.shorthash] = entry
except OSError as e: # should catch FileNotFoundError and PermissionError etc.
shared.log.error(f'LoRA: filename="{filename}" {e}')
candidates = list(files_cache.list_files(shared.cmd_opts.lora_dir, ext_filter=[".pt", ".ckpt", ".safetensors"]))
with concurrent.futures.ThreadPoolExecutor(max_workers=shared.max_workers) as executor:
for fn in candidates:
executor.submit(add_network, fn)
t1 = time.time()
shared.log.info(f'Available LoRAs: path="{shared.cmd_opts.lora_dir}" items={len(available_networks)} folders={len(forbidden_network_aliases)} time={t1 - t0:.2f}')
def infotext_pasted(infotext, params): # pylint: disable=W0613
if "AddNet Module 1" in [x[1] for x in scripts.scripts_txt2img.infotext_fields]:
return # if the other extension is active, it will handle those fields, no need to do anything
added = []
for k in params:
if not k.startswith("AddNet Model "):
continue
num = k[13:]
if params.get("AddNet Module " + num) != "LoRA":
continue
name = params.get("AddNet Model " + num)
if name is None:
continue
m = re_network_name.match(name)
if m:
name = m.group(1)
multiplier = params.get("AddNet Weight A " + num, "1.0")
added.append(f"<lora:{name}:{multiplier}>")
if added:
params["Prompt"] += "\n" + "".join(added)
list_available_networks()
extensions-builtin/Lora/scripts/lora_script.py
-66
View File
@@ -1,66 +0,0 @@
import re
import networks
import lora # pylint: disable=unused-import
from lora_extract import create_ui
from network import NetworkOnDisk
from ui_extra_networks_lora import ExtraNetworksPageLora
from extra_networks_lora import ExtraNetworkLora
from modules import script_callbacks, extra_networks, ui_extra_networks, ui_models, shared # pylint: disable=unused-import
re_lora = re.compile("<lora:([^:]+):")
def before_ui():
ui_extra_networks.register_page(ExtraNetworksPageLora())
networks.extra_network_lora = ExtraNetworkLora()
extra_networks.register_extra_network(networks.extra_network_lora)
ui_models.extra_ui.append(create_ui)
def create_lora_json(obj: NetworkOnDisk):
return {
"name": obj.name,
"alias": obj.alias,
"path": obj.filename,
"metadata": obj.metadata,
}
def api_networks(_, app):
@app.get("/sdapi/v1/loras")
async def get_loras():
return [create_lora_json(obj) for obj in networks.available_networks.values()]
@app.post("/sdapi/v1/refresh-loras")
async def refresh_loras():
return networks.list_available_networks()
def infotext_pasted(infotext, d): # pylint: disable=unused-argument
hashes = d.get("Lora hashes", None)
if hashes is None:
return
def network_replacement(m):
alias = m.group(1)
shorthash = hashes.get(alias)
if shorthash is None:
return m.group(0)
network_on_disk = networks.available_network_hash_lookup.get(shorthash)
if network_on_disk is None:
return m.group(0)
return f'<lora:{network_on_disk.get_alias()}:'
hashes = [x.strip().split(':', 1) for x in hashes.split(",")]
hashes = {x[0].strip().replace(",", ""): x[1].strip() for x in hashes}
d["Prompt"] = re.sub(re_lora, network_replacement, d["Prompt"])
if shared.opts.lora_legacy:
shared.log.debug('Register network: type=LoRA method=legacy')
script_callbacks.on_app_started(api_networks)
script_callbacks.on_before_ui(before_ui)
script_callbacks.on_model_loaded(networks.assign_network_names_to_compvis_modules)
script_callbacks.on_infotext_pasted(networks.infotext_pasted)
script_callbacks.on_infotext_pasted(infotext_pasted)
extensions-builtin/Lora/ui_extra_networks_lora.py
-123
View File
@@ -1,123 +0,0 @@
import os
import json
import concurrent
import networks
from modules import shared, ui_extra_networks
debug = os.environ.get('SD_LORA_DEBUG', None) is not None
class ExtraNetworksPageLora(ui_extra_networks.ExtraNetworksPage):
def __init__(self):
super().__init__('Lora')
self.list_time = 0
shared.log.warning('Networks: type=lora method=legacy')
def refresh(self):
networks.list_available_networks()
def get_tags(self, l, info):
tags = {}
try:
if l.metadata is not None:
modelspec_tags = l.metadata.get('modelspec.tags', {})
possible_tags = l.metadata.get('ss_tag_frequency', {}) # tags from model metedata
if isinstance(possible_tags, str):
possible_tags = {}
if isinstance(modelspec_tags, str):
modelspec_tags = {}
if len(list(modelspec_tags)) > 0:
possible_tags.update(modelspec_tags)
for k, v in possible_tags.items():
words = k.split('_', 1) if '_' in k else [v, k]
words = [str(w).replace('.json', '') for w in words]
if words[0] == '{}':
words[0] = 0
tag = ' '.join(words[1:]).lower()
tags[tag] = words[0]
def find_version():
found_versions = []
current_hash = l.hash[:8].upper()
all_versions = info.get('modelVersions', [])
for v in info.get('modelVersions', []):
for f in v.get('files', []):
if any(h.startswith(current_hash) for h in f.get('hashes', {}).values()):
found_versions.append(v)
if len(found_versions) == 0:
found_versions = all_versions
return found_versions
for v in find_version(): # trigger words from info json
possible_tags = v.get('trainedWords', [])
if isinstance(possible_tags, list):
for tag_str in possible_tags:
for tag in tag_str.split(','):
tag = tag.strip().lower()
if tag not in tags:
tags[tag] = 0
possible_tags = info.get('tags', []) # tags from info json
if not isinstance(possible_tags, list):
possible_tags = list(possible_tags.values())
for tag in possible_tags:
tag = tag.strip().lower()
if tag not in tags:
tags[tag] = 0
except Exception:
pass
bad_chars = [';', ':', '<', ">", "*", '?', '\'', '\"', '(', ')', '[', ']', '{', '}', '\\', '/']
clean_tags = {}
for k, v in tags.items():
tag = ''.join(i for i in k if i not in bad_chars).strip()
clean_tags[tag] = v
clean_tags.pop('img', None)
clean_tags.pop('dataset', None)
return clean_tags
def create_item(self, name):
l = networks.available_networks.get(name)
if l is None:
shared.log.warning(f'Networks: type=lora registered={len(list(networks.available_networks))} file="{name}" not registered')
return None
try:
# path, _ext = os.path.splitext(l.filename)
name = os.path.splitext(os.path.relpath(l.filename, shared.cmd_opts.lora_dir))[0]
item = {
"type": 'Lora',
"name": name,
"filename": l.filename,
"hash": l.shorthash,
"prompt": json.dumps(f" <lora:{l.get_alias()}:{shared.opts.extra_networks_default_multiplier}>"),
"metadata": json.dumps(l.metadata, indent=4) if l.metadata else None,
"mtime": os.path.getmtime(l.filename),
"size": os.path.getsize(l.filename),
"version": l.sd_version,
}
info = self.find_info(l.filename)
item["info"] = info
item["description"] = self.find_description(l.filename, info) # use existing info instead of double-read
item["tags"] = self.get_tags(l, info)
return item
except Exception as e:
shared.log.error(f'Networks: type=lora file="{name}" {e}')
if debug:
from modules import errors
errors.display('e', 'Lora')
return None
def list_items(self):
items = []
with concurrent.futures.ThreadPoolExecutor(max_workers=shared.max_workers) as executor:
future_items = {executor.submit(self.create_item, net): net for net in networks.available_networks}
for future in concurrent.futures.as_completed(future_items):
item = future.result()
if item is not None:
items.append(item)
self.update_all_previews(items)
return items
def allowed_directories_for_previews(self):
return [shared.cmd_opts.lora_dir, shared.cmd_opts.lyco_dir]
extensions-builtin/sd-extension-chainner
+1 -1
Submodule extensions-builtin/sd-extension-chainner updated: c12e8cda4e...716b1ee7dc
extensions-builtin/sd-extension-system-info
+1 -1
Submodule extensions-builtin/sd-extension-system-info updated: 539625f289...0760f3bcee
extensions-builtin/sdnext-modernui
+1 -1
Submodule extensions-builtin/sdnext-modernui updated: 6ad9a291b5...63d6d509b8
extensions-builtin/stable-diffusion-webui-rembg
+1 -1
Submodule extensions-builtin/stable-diffusion-webui-rembg updated: fbf4215c3a...3ec8f9eb79
html/art-styles.json
+1 -1
View File
File diff suppressed because one or more lines are too long
html/locale_de.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "Zoe-Tiefe"
}
]
}
}
html/locale_en.json
+10 -2
View File
@@ -10,7 +10,7 @@
{"id":"","label":"⟲","localized":"","hint":"Refresh"},
{"id":"","label":"✕","localized":"","hint":"Close"},
{"id":"","label":"⊜","localized":"","hint":"Fill"},
{"id":"","label":"⌾","localized":"","hint":"Load model as refiner model when selected, otherwise load as base model"},
{"id":"","label":"※","localized":"","hint":"Load model as refiner model when selected, otherwise load as base model"},
{"id":"","label":"🔎︎","localized":"","hint":"Scan CivitAI for missing metadata and previews"},
{"id":"","label":"☲","localized":"","hint":"Change view type"},
{"id":"","label":"📐","localized":"","hint":"Measure"},
@@ -19,7 +19,15 @@
{"id":"","label":"🖼️","localized":"","hint":"Show preview"},
{"id":"","label":"♻","localized":"","hint":"Interrogate image"},
{"id":"","label":"↶","localized":"","hint":"Apply selected style to prompt"},
{"id":"","label":"↷","localized":"","hint":"Save current prompt to style"}
{"id":"","label":"↷","localized":"","hint":"Save current prompt to style"},
{"id":"","label":"","localized":"","hint":"Sort by name, ascending"},
{"id":"","label":"","localized":"","hint":"Sort by name, descending"},
{"id":"","label":"","localized":"","hint":"Sort by size, ascending"},
{"id":"","label":"","localized":"","hint":"Sort by size, descending"},
{"id":"","label":"","localized":"","hint":"Sort by resolution, ascending"},
{"id":"","label":"","localized":"","hint":"Sort by resolution, descending"},
{"id":"","label":"","localized":"","hint":"Sort by time, ascending"},
{"id":"","label":"","localized":"","hint":"Sort by time, descending"}
],
"main": [
{"id":"","label":"Prompt","localized":"","hint":"Describe image you want to generate"},
html/locale_es.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "profundidad zoe"
}
]
}
}
html/locale_fr.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "Profondeur ZOE"
}
]
}
}
html/locale_hr.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "zoe dubina"
}
]
}
}
html/locale_it.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "Profondità ZOE"
}
]
}
}
html/locale_ja.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "Zoe深度"
}
]
}
}
html/locale_ko.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "Zoe 깊이"
}
]
}
}
html/locale_pt.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "Profundidade ZOE"
}
]
}
}
html/locale_ru.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "Zoe глубина"
}
]
}
}
html/locale_zh.json
+1 -1
View File
@@ -7553,4 +7553,4 @@
"hint": "ZOE深度"
}
]
}
}
html/override_en.json
+2 -2
View File
@@ -1,2 +1,2 @@
[
]
[
]
html/override_hr.json
+1 -1
View File
@@ -1,3 +1,3 @@
[
{ "id": "", "label": "Reprocess", "localized": "Ponovi", "hint": "Ponovno obradite prethodne generacije koristeći različite parametre" }
]
]
html/override_ko.json
+1 -1
View File
File diff suppressed because one or more lines are too long
html/reference.json
+77 -13
View File
@@ -50,7 +50,7 @@
"preview": "dreamshaperXL_v21TurboDPMSDE.jpg",
"desc": "Showcase finetuned model based on Stable diffusion XL",
"extras": "sampler: DPM SDE, steps: 8, cfg_scale: 2.0"
},
},
"SDXS DreamShaper 512": {
"path": "IDKiro/sdxs-512-dreamshaper",
@@ -189,42 +189,93 @@
"extras": "sampler: Default, cfg_scale: 3.5"
},
"Wan-AI Wan2.1 1.3B": {
"path": "Wan-AI/Wan2.1-T2V-1.3B-Diffusers",
"preview": "Wan-AI--Wan2.1.jpg",
"desc": "Wan is an advanced and powerful visual generation model developed by Tongyi Lab of Alibaba Group. It can generate videos based on text, images, and other control signals. The Wan2.1 series models are now fully open-source.",
"skip": true,
"extras": "sampler: Default"
},
"Wan-AI Wan2.1 14B": {
"path": "Wan-AI/Wan2.1-T2V-14B-Diffusers",
"preview": "Wan-AI--Wan2.1.jpg",
"desc": "Wan is an advanced and powerful visual generation model developed by Tongyi Lab of Alibaba Group. It can generate videos based on text, images, and other control signals. The Wan2.1 series models are now fully open-source.",
"skip": true,
"extras": "sampler: Default"
},
"Wan-AI Wan2.2 5B": {
"path": "Wan-AI/Wan2.2-TI2V-5B-Diffusers",
"preview": "Wan-AI--Wan2.2.jpg",
"desc": "Wan2.2, offering more powerful capabilities, better performance, and superior visual quality. With Wan2.2, we have focused on incorporating the following technical innovations: MoE Architecture, Data Scalling, Cinematic Aesthetics, Efficient High-Definition Hybrid",
"skip": true,
"extras": "sampler: Default"
},
"Wan-AI Wan2.2 A14B": {
"path": "Wan-AI/Wan2.2-T2V-A14B-Diffusers",
"preview": "Wan-AI--Wan2.2.jpg",
"desc": "Wan2.2, offering more powerful capabilities, better performance, and superior visual quality. With Wan2.2, we have focused on incorporating the following technical innovations: MoE Architecture, Data Scalling, Cinematic Aesthetics, Efficient High-Definition Hybrid",
"skip": true,
"extras": "sampler: Default"
},
"Freepik F-Lite": {
"path": "Freepik/F-Lite",
"preview": "Freepik--F-Lite.jpg",
"desc": "F Lite is a 10B parameter diffusion model created by Freepik and Fal, trained exclusively on copyright-safe and SFW content. The model was trained on Freepik's internal dataset comprising approximately 80 million copyright-safe images, making it the first publicly available model of this scale trained exclusively on legally compliant and SFW content.",
"skip": true,
"extras": "sampler: Default, cfg_scale: 3.5"
},
"Freepik F-Lite Texture": {
"path": "Freepik/F-Lite-Texture",
"preview": "Freepik--F-Lite.jpg",
"desc": "F Lite is a 10B parameter diffusion model created by Freepik and Fal, trained exclusively on copyright-safe and SFW content. The model was trained on Freepik's internal dataset comprising approximately 80 million copyright-safe images, making it the first publicly available model of this scale trained exclusively on legally compliant and SFW content.",
"skip": true,
"extras": "sampler: Default, cfg_scale: 3.5"
},
"Freepik F-Lite 7B": {
"path": "Freepik/F-Lite-7B",
"preview": "Freepik--F-Lite.jpg",
"desc": "F Lite is a 10B parameter diffusion model created by Freepik and Fal, trained exclusively on copyright-safe and SFW content. The model was trained on Freepik's internal dataset comprising approximately 80 million copyright-safe images, making it the first publicly available model of this scale trained exclusively on legally compliant and SFW content.",
"skip": true,
"extras": "sampler: Default, cfg_scale: 3.5"
},
"NVLabs Sana 1.5 1.6B 1k": {
"path": "Efficient-Large-Model/SANA1.5_1.6B_1024px_diffusers",
"desc": "Sana is an efficient model with scaling of training-time and inference time techniques. SANA-1.5 delivers: efficient model growth from 1.6B Sana-1.0 model to 4.8B, achieving similar or better performance than training from scratch and saving 60% training cost; efficient model depth pruning, slimming any model size as you want; powerful VLM selection based inference scaling, smaller model+inference scaling > larger model.",
"preview": "Efficient-Large-Model--Sana_1600M_1024px_diffusers.jpg",
"skip": true
},
},
"NVLabs Sana 1.5 4.8B 1k": {
"path": "Efficient-Large-Model/SANA1.5_4.8B_1024px_diffusers",
"desc": "Sana is an efficient model with scaling of training-time and inference time techniques. SANA-1.5 delivers: efficient model growth from 1.6B Sana-1.0 model to 4.8B, achieving similar or better performance than training from scratch and saving 60% training cost; efficient model depth pruning, slimming any model size as you want; powerful VLM selection based inference scaling, smaller model+inference scaling > larger model.",
"preview": "Efficient-Large-Model--Sana_1600M_1024px_diffusers.jpg",
"skip": true
},
},
"NVLabs Sana 1.5 1.6B 1k Sprint": {
"path": "Efficient-Large-Model/Sana_Sprint_1.6B_1024px_diffusers",
"desc": "SANA-Sprint is an ultra-efficient diffusion model for text-to-image (T2I) generation, reducing inference steps from 20 to 1-4 while achieving state-of-the-art performance.",
"preview": "Efficient-Large-Model--Sana_1600M_1024px_diffusers.jpg",
"skip": true
},
},
"NVLabs Sana 1.0 1.6B 4k": {
"path": "Efficient-Large-Model/Sana_1600M_4Kpx_BF16_diffusers",
"desc": "Sana is a text-to-image framework that can efficiently generate images up to 4096 × 4096 resolution. Sana can synthesize high-resolution, high-quality images with strong text-image alignment at a remarkably fast speed, deployable on laptop GPU.",
"preview": "Efficient-Large-Model--Sana_1600M_1024px_diffusers.jpg",
"skip": true
},
},
"NVLabs Sana 1.0 1.6B 2k": {
"path": "Efficient-Large-Model/Sana_1600M_2Kpx_BF16_diffusers",
"desc": "Sana is a text-to-image framework that can efficiently generate images up to 4096 × 4096 resolution. Sana can synthesize high-resolution, high-quality images with strong text-image alignment at a remarkably fast speed, deployable on laptop GPU.",
"preview": "Efficient-Large-Model--Sana_1600M_1024px_diffusers.jpg",
"skip": true
},
},
"NVLabs Sana 1.0 1.6B 1k": {
"path": "Efficient-Large-Model/Sana_1600M_1024px_diffusers",
"desc": "Sana is a text-to-image framework that can efficiently generate images up to 4096 × 4096 resolution. Sana can synthesize high-resolution, high-quality images with strong text-image alignment at a remarkably fast speed, deployable on laptop GPU.",
"preview": "Efficient-Large-Model--Sana_1600M_1024px_diffusers.jpg",
"skip": true
},
},
"NVLabs Sana 1.0 0.6B 0.5k": {
"path": "Efficient-Large-Model/Sana_600M_512px_diffusers",
"desc": "Sana is a text-to-image framework that can efficiently generate images up to 4096 × 4096 resolution. Sana can synthesize high-resolution, high-quality images with strong text-image alignment at a remarkably fast speed, deployable on laptop GPU.",
@@ -250,21 +301,27 @@
"desc": "OmniGen is a unified image generation model that can generate a wide range of images from multi-modal prompts. It is designed to be simple, flexible and easy to use.",
"preview": "Shitao--OmniGen-v1.jpg",
"skip": true
},
},
"VectorSpaceLab OmniGen v2": {
"path": "OmniGen2/OmniGen2",
"desc": "OmniGen2 is a powerful and efficient unified multimodal model. Unlike OmniGen v1, OmniGen2 features two distinct decoding pathways for text and image modalities, utilizing unshared parameters and a decoupled image tokenizer.",
"preview": "OmniGen2--OmniGen2.jpg",
"skip": true
},
},
"AuraFlow 0.3": {
"path": "fal/AuraFlow-v0.3",
"desc": "AuraFlow v0.3 is the fully open-sourced flow-based text-to-image generation model. The model was trained with more compute compared to the previous version, AuraFlow-v0.2. Compared to AuraFlow-v0.2, the model is fine-tuned on more aesthetic datasets and now supports various aspect ratio, (now width and height up to 1536 pixels).",
"preview": "fal--AuraFlow-v0.3.jpg",
"skip": true
},
},
"AuraFlow 0.2": {
"path": "fal/AuraFlow-v0.2",
"desc": "AuraFlow v0.2 is the fully open-sourced largest flow-based text-to-image generation model. The model was trained with more compute compared to the previous version, AuraFlow-v0.1",
"preview": "fal--AuraFlow-v0.3.jpg",
"skip": true
},
"Segmind Vega": {
"path": "huggingface/segmind/Segmind-Vega",
@@ -334,7 +391,7 @@
"skip": true,
"extras": "sampler: Default, cfg_scale: 2.0"
},
"Tencent HunyuanDiT 1.2": {
"path": "Tencent-Hunyuan/HunyuanDiT-v1.2-Diffusers",
"desc": "Hunyuan-DiT : A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understanding.",
@@ -348,7 +405,7 @@
"preview": "Alpha-VLLM--Lumina-Next-SFT-diffusers.jpg",
"skip": true,
"extras": "sampler: Default"
},
},
"AlphaVLLM Lumina 2": {
"path": "Alpha-VLLM/Lumina-Image-2.0",
"desc": "A Unified and Efficient Image Generative Model. Lumina-Image-2.0 is a 2 billion parameter flow-based diffusion transformer capable of generating images from text descriptions.",
@@ -392,7 +449,7 @@
"preview": "Kwai-Kolors--Kolors-diffusers.jpg",
"skip": true,
"extras": "width: 1024, height: 1024"
},
},
"Kandinsky 2.1": {
"path": "kandinsky-community/kandinsky-2-1",
@@ -477,6 +534,13 @@
"skip": true
},
"Bria 3.2": {
"path": "briaai/BRIA-3.2",
"desc": "Bria 3.2 is the next-generation commercial-ready text-to-image model. With just 4 billion parameters, it provides exceptional aesthetics and text rendering, evaluated to provide on par results to leading open-source models, and outperforming other licensed models.",
"preview": "briaai--BRIA-3.2.jpg",
"skip": true
},
"Meissonic": {
"path": "MeissonFlow/Meissonic",
"desc": "Meissonic is a non-autoregressive mask image modeling text-to-image synthesis model that can generate high-resolution images. It is designed to run on consumer graphics cards.",
html/swagger.css
+1 -1
View File
@@ -865,4 +865,4 @@
linear-gradient(270deg, #696969 30%, rgba(0, 0, 0, 0) 31%);
background-color: #b6b6b6;
}
}
}
installer.py
+137 -65
View File
@@ -1,4 +1,4 @@
from functools import lru_cache
from typing import List, Optional
import os
import sys
import json
@@ -105,6 +105,16 @@ def install_traceback(suppress: list = []):
# setup console and file logging
def setup_logging():
from functools import partial, partialmethod
from logging.handlers import RotatingFileHandler
from rich.theme import Theme
from rich.logging import RichHandler
from rich.console import Console
from rich.padding import Padding
from rich.segment import Segment
from rich import box
from rich import print as rprint
from rich.pretty import install as pretty_install
class RingBuffer(logging.StreamHandler):
def __init__(self, capacity):
@@ -129,6 +139,7 @@ def setup_logging():
def get(self):
return self.buffer
class LogFilter(logging.Filter):
def __init__(self):
super().__init__()
@@ -136,14 +147,35 @@ def setup_logging():
def filter(self, record):
return len(record.getMessage()) > 2
def override_padding(self, console, options): # pylint: disable=redefined-outer-name
style = console.get_style(self.style)
width = options.max_width
self.left = 0
render_options = options.update_width(width - self.left - self.right)
if render_options.height is not None:
render_options = render_options.update_height(height=render_options.height - self.top - self.bottom)
lines = console.render_lines(self.renderable, render_options, style=style, pad=False)
_Segment = Segment
left = _Segment(" " * self.left, style) if self.left else None
right = [_Segment.line()]
blank_line: Optional[List[Segment]] = None
if self.top:
blank_line = [_Segment(f'{" " * width}\n', style)]
yield from blank_line * self.top
if left:
for line in lines:
yield left
yield from line
yield from right
else:
for line in lines:
yield from line
yield from right
if self.bottom:
blank_line = blank_line or [_Segment(f'{" " * width}\n', style)]
yield from blank_line * self.bottom
t_start = time.time()
from functools import partial, partialmethod
from logging.handlers import RotatingFileHandler
from rich.theme import Theme
from rich.logging import RichHandler
from rich.console import Console
from rich import print as rprint
from rich.pretty import install as pretty_install
if args.log:
global log_file # pylint: disable=global-statement
@@ -160,12 +192,15 @@ def setup_logging():
global console # pylint: disable=global-statement
theme = Theme({
"traceback.border": "black",
"traceback.border.syntax_error": "black",
"inspect.value.border": "black",
"traceback.border.syntax_error": "dark_red",
"logging.level.info": "blue_violet",
"logging.level.debug": "purple4",
"logging.level.trace": "dark_blue",
})
Padding.__rich_console__ = override_padding
box.ROUNDED = box.SIMPLE
console = Console(
log_time=True,
log_time_format='%H:%M:%S-%f',
@@ -174,6 +209,7 @@ def setup_logging():
safe_box=True,
theme=theme,
)
logging.basicConfig(level=logging.ERROR, format='%(asctime)s | %(name)s | %(levelname)s | %(module)s | %(message)s', handlers=[logging.NullHandler()]) # redirect default logger to null
pretty_install(console=console)
install_traceback()
@@ -254,7 +290,6 @@ def print_profile(profiler: cProfile.Profile, msg: str):
profile(profiler, msg)
@lru_cache()
def package_version(package):
try:
return pkg_resources.get_distribution(package).version
@@ -262,7 +297,6 @@ def package_version(package):
return None
@lru_cache()
def package_spec(package):
spec = pkg_resources.working_set.by_key.get(package, None) # more reliable than importlib
if spec is None:
@@ -273,7 +307,6 @@ def package_spec(package):
# check if package is installed
@lru_cache()
def installed(package, friendly: str = None, reload = False, quiet = False): # pylint: disable=redefined-outer-name
t_start = time.time()
ok = True
@@ -341,12 +374,11 @@ def run(cmd: str, arg: str):
return txt
@lru_cache()
def pip(arg: str, ignore: bool = False, quiet: bool = True, uv = True):
t_start = time.time()
originalArg = arg
arg = arg.replace('>=', '==')
package = arg.replace("install", "").replace("--upgrade", "").replace("--no-deps", "").replace("--force", "").replace(" ", " ").strip()
package = arg.replace("install", "").replace("--upgrade", "").replace("--no-deps", "").replace("--force-reinstall", "").replace(" ", " ").strip()
uv = uv and args.uv and not package.startswith('git+')
pipCmd = "uv pip" if uv else "pip"
if not quiet and '-r ' not in arg:
@@ -360,7 +392,7 @@ def pip(arg: str, ignore: bool = False, quiet: bool = True, uv = True):
if len(result.stderr) > 0:
if uv and result.returncode != 0:
err = result.stderr.decode(encoding="utf8", errors="ignore")
log.warning('Install: cannot use uv, fallback to pip')
log.warning(f'Install: cmd="{pipCmd}" args="{all_args}" cannot use uv, fallback to pip')
debug(f'Install: uv pip error: {err}')
return pip(originalArg, ignore, quiet, uv=False)
else:
@@ -376,7 +408,6 @@ def pip(arg: str, ignore: bool = False, quiet: bool = True, uv = True):
# install package using pip if not already installed
@lru_cache()
def install(package, friendly: str = None, ignore: bool = False, reinstall: bool = False, no_deps: bool = False, quiet: bool = False, force: bool = False):
t_start = time.time()
res = ''
@@ -385,7 +416,7 @@ def install(package, friendly: str = None, ignore: bool = False, reinstall: bool
quick_allowed = False
if (args.reinstall) or (reinstall) or (not installed(package, friendly, quiet=quiet)):
deps = '' if not no_deps else '--no-deps '
cmd = f"install{' --upgrade' if not args.uv else ''}{' --force' if force else ''} {deps}{package}"
cmd = f"install{' --upgrade' if not args.uv else ''}{' --force-reinstall' if force else ''} {deps}{package}"
res = pip(cmd, ignore=ignore, uv=package != "uv" and not package.startswith('git+'))
try:
importlib.reload(pkg_resources)
@@ -396,7 +427,6 @@ def install(package, friendly: str = None, ignore: bool = False, reinstall: bool
# execute git command
@lru_cache()
def git(arg: str, folder: str = None, ignore: bool = False, optional: bool = False): # pylint: disable=unused-argument
t_start = time.time()
if args.skip_git:
@@ -407,20 +437,22 @@ def git(arg: str, folder: str = None, ignore: bool = False, optional: bool = Fal
if git_cmd != "git":
git_cmd = os.path.abspath(git_cmd)
result = subprocess.run(f'"{git_cmd}" {arg}', check=False, shell=True, env=os.environ, stdout=subprocess.PIPE, stderr=subprocess.PIPE, cwd=folder or '.')
txt = result.stdout.decode(encoding="utf8", errors="ignore")
stdout = result.stdout.decode(encoding="utf8", errors="ignore")
if len(result.stderr) > 0:
txt += ('\n' if len(txt) > 0 else '') + result.stderr.decode(encoding="utf8", errors="ignore")
txt = txt.strip()
stdout += ('\n' if len(stdout) > 0 else '') + result.stderr.decode(encoding="utf8", errors="ignore")
stdout = stdout.strip()
if result.returncode != 0 and not ignore:
if "couldn't find remote ref" in txt: # not a git repo
return txt
if "couldn't find remote ref" in stdout: # not a git repo
log.error(f'Git: folder="{folder}" could not identify repository')
elif "no submodule mapping found" in stdout:
log.warning(f'Git: folder="{folder}" submodules changed')
elif 'or stash them' in stdout:
log.error(f'Git: folder="{folder}" local changes detected')
else:
log.error(f'Git: folder="{folder}" arg="{arg}" output={stdout}')
errors.append(f'git: {folder}')
log.error(f'Git: {folder} / {arg}')
if 'or stash them' in txt:
log.error(f'Git local changes detected: check details log="{log_file}"')
log.debug(f'Git output: {txt}')
ts('git', t_start)
return txt
return stdout
# reattach as needed as head can get detached
@@ -559,17 +591,17 @@ def check_python(supported_minors=[], experimental_minors=[], reason=None):
# check diffusers version
def check_diffusers():
t_start = time.time()
if args.skip_all or args.skip_git or args.experimental:
if args.skip_all or args.skip_git:
return
sha = '00f95b9755718aabb65456e791b8408526ae6e76' # diffusers commit hash
sha = '56d438727036b0918b30bbe3110c5fe1634ed19d' # diffusers commit hash
pkg = pkg_resources.working_set.by_key.get('diffusers', None)
minor = int(pkg.version.split('.')[1] if pkg is not None else 0)
cur = opts.get('diffusers_version', '') if minor > 0 else ''
if (minor == 0) or (cur != sha):
if minor == 0:
minor = int(pkg.version.split('.')[1] if pkg is not None else -1)
cur = opts.get('diffusers_version', '') if minor > -1 else ''
if (minor == -1) or ((cur != sha) and (not args.experimental)):
if minor == -1:
log.info(f'Diffusers install: commit={sha}')
else:
log.info(f'Diffusers update: package={pkg} current={cur} target={sha}')
log.info(f'Diffusers update: current={pkg.version} hash={cur} target={sha}')
pip('uninstall --yes diffusers', ignore=True, quiet=True, uv=False)
pip(f'install --upgrade git+https://github.com/huggingface/diffusers@{sha}', ignore=False, quiet=True, uv=False)
global diffusers_commit # pylint: disable=global-statement
@@ -577,6 +609,26 @@ def check_diffusers():
ts('diffusers', t_start)
# check transformers version
def check_transformers():
t_start = time.time()
if args.skip_all or args.skip_git or args.experimental:
return
pkg = pkg_resources.working_set.by_key.get('transformers', None)
if args.use_directml:
target = '4.52.4'
else:
target = '4.53.2'
if (pkg is None) or ((pkg.version != target) and (not args.experimental)):
if pkg is None:
log.info(f'Transformers install: version={target}')
else:
log.info(f'Transformers update: current={pkg.version} target={target}')
pip('uninstall --yes transformers', ignore=True, quiet=True, uv=False)
pip(f'install --upgrade transformers=={target}', ignore=False, quiet=True, uv=False)
ts('transformers', t_start)
# check onnx version
def check_onnx():
t_start = time.time()
@@ -799,12 +851,12 @@ def install_torch_addons():
install('DeepCache')
if opts.get('cuda_compile_backend', '') == 'olive-ai':
install('olive-ai')
if opts.get('optimum_quanto_weights', False):
if len(opts.get('optimum_quanto_weights', [])):
install('optimum-quanto==0.2.7', 'optimum-quanto')
if opts.get('torchao_quantization', False):
if len(opts.get('torchao_quantization', [])):
install('torchao==0.10.0', 'torchao')
if opts.get('samples_format', 'jpg') == 'jxl' or opts.get('grid_format', 'jpg') == 'jxl':
install('pillow-jxl-plugin==1.3.3', 'pillow-jxl-plugin')
install('pillow-jxl-plugin==1.3.4', 'pillow-jxl-plugin')
if not args.experimental:
uninstall('wandb', quiet=True)
ts('addons', t_start)
@@ -827,6 +879,7 @@ def check_cudnn():
# check torch version
def check_torch():
log.info('Verifying torch installation')
t_start = time.time()
if args.skip_torch:
log.info('Torch: skip tests')
@@ -871,7 +924,7 @@ def check_torch():
torch_command = os.environ.get('TORCH_COMMAND', 'torch torchvision')
elif allow_directml and args.use_directml and ('arm' not in machine and 'aarch' not in machine):
log.info('DirectML: selected')
torch_command = os.environ.get('TORCH_COMMAND', 'torch==2.4.1 torchvision torch-directml')
torch_command = os.environ.get('TORCH_COMMAND', 'torch==2.4.1 torchvision torch-directml==0.2.4.dev240913')
if 'torch' in torch_command and not args.version:
install(torch_command, 'torch torchvision')
install('onnxruntime-directml', 'onnxruntime-directml', ignore=True)
@@ -955,10 +1008,10 @@ def check_modified_files():
files = [x for x in files if len(x) > 0 and (not x.startswith('extensions')) and (not x.startswith('wiki')) and (not x.endswith('.json')) and ('.log' not in x)]
deleted = [x for x in files if not os.path.exists(x)]
if len(deleted) > 0:
log.warning(f'Deleted files: {files}')
files = [x for x in files if os.path.exists(x) and not os.path.isdir(x)]
if len(files) > 0:
log.warning(f'Modified files: {files}')
log.warning(f'Deleted files: {deleted}')
modified = [x for x in files if os.path.exists(x) and not os.path.isdir(x)]
if len(modified) > 0:
log.warning(f'Modified files: {modified}')
except Exception:
pass
ts('files', t_start)
@@ -1163,31 +1216,43 @@ def ensure_base_requirements():
update_setuptools()
# used by installler itself so must be installed before requirements
install('rich', 'rich', quiet=True)
install('rich==14.0.0', 'rich', quiet=True)
install('psutil', 'psutil', quiet=True)
install('requests', 'requests', quiet=True)
install('requests==2.32.3', 'requests', quiet=True)
ts('base', t_start)
def install_gradio():
# pip install gradio==3.43.2 installs:
# aiofiles-23.2.1 altair-5.5.0 annotated-types-0.7.0 anyio-4.9.0 attrs-25.3.0 certifi-2025.6.15 charset_normalizer-3.4.2 click-8.2.1 contourpy-1.3.2 cycler-0.12.1 fastapi-0.115.14 ffmpy-0.6.0 filelock-3.18.0 fonttools-4.58.4 fsspec-2025.5.1 gradio-3.43.2 gradio-client-0.5.0 h11-0.16.0 hf-xet-1.1.5 httpcore-1.0.9 httpx-0.28.1 huggingface-hub-0.33.1 idna-3.10 importlib-resources-6.5.2 jinja2-3.1.6 jsonschema-4.24.0 jsonschema-specifications-2025.4.1 kiwisolver-1.4.8 markupsafe-2.1.5 matplotlib-3.10.3 narwhals-1.45.0 numpy-1.26.4 orjson-3.10.18 packaging-25.0 pandas-2.3.0 pillow-10.4.0 pydantic-2.11.7 pydantic-core-2.33.2 pydub-0.25.1 pyparsing-3.2.3 python-dateutil-2.9.0.post0 python-multipart-0.0.20 pytz-2025.2 pyyaml-6.0.2 referencing-0.36.2 requests-2.32.4 rpds-py-0.25.1 semantic-version-2.10.0 six-1.17.0 sniffio-1.3.1 starlette-0.46.2 tqdm-4.67.1 typing-extensions-4.14.0 typing-inspection-0.4.1 tzdata-2025.2 urllib3-2.5.0 uvicorn-0.35.0 websockets-11.0.3
install('gradio==3.43.2', no_deps=True)
install('gradio-client==0.5.0', no_deps=True, quiet=True)
install('dctorch==0.1.2', no_deps=True, quiet=True)
pkgs = ['fastapi', 'websockets', 'aiofiles', 'ffmpy', 'pydub', 'uvicorn', 'semantic-version', 'altair', 'python-multipart', 'matplotlib']
for pkg in pkgs:
if not installed(pkg, quiet=True):
install(pkg, quiet=True)
def install_optional():
t_start = time.time()
log.info('Installing optional requirements...')
install('git+https://github.com/Disty0/BasicSR@2b6a12c28e0c81bfb13b7e984144f0b0f5461484', 'basicsr')
install('git+https://github.com/Disty0/GFPGAN@09b1190eabbc77e5f15c61fa7c38a2064b403e20', 'gfpgan')
install('clean-fid')
install('pillow-jxl-plugin==1.3.3', ignore=True)
install('optimum-quanto==0.2.7', ignore=True)
install('torchao==0.10.0', ignore=True)
install('bitsandbytes==0.45.5', ignore=True)
install('pynvml', ignore=True)
install('ultralytics==8.3.40', ignore=True)
install('Cython', ignore=True)
install('insightface==0.7.3', ignore=True) # problematic build
install('albumentations==1.4.3', ignore=True)
install('pydantic==1.10.21', ignore=True)
install('clean-fid', quiet=True)
install('pillow-jxl-plugin==1.3.4', ignore=True, quiet=True)
install('optimum-quanto==0.2.7', ignore=True, quiet=True)
install('torchao==0.10.0', ignore=True, quiet=True)
install('bitsandbytes==0.46.1', ignore=True, quiet=True)
install('pynvml', ignore=True, quiet=True)
install('ultralytics==8.3.40', ignore=True, quiet=True)
install('Cython', ignore=True, quiet=True)
install('git+https://github.com/deepinsight/insightface@554a05561cb71cfebb4e012dfea48807f845a0c2#subdirectory=python-package', 'insightface') # insightface==0.7.3 with patches
install('albumentations==1.4.3', ignore=True, quiet=True)
install('pydantic==1.10.21', ignore=True, quiet=True)
reload('pydantic', '1.10.21')
install('gguf', ignore=True)
install('av', ignore=True)
install('av', ignore=True, quiet=True)
try:
import gguf
scripts_dir = os.path.join(os.path.dirname(gguf.__file__), '..', 'scripts')
@@ -1198,6 +1263,21 @@ def install_optional():
ts('optional', t_start)
def install_sentencepiece():
if installed('sentencepiece', quiet=True):
pass
elif int(sys.version_info.minor) >= 13:
backup_cmake_policy = os.environ.get('CMAKE_POLICY_VERSION_MINIMUM', None)
backup_cxxflags = os.environ.get('CXXFLAGS', None)
os.environ.setdefault('CMAKE_POLICY_VERSION_MINIMUM', '3.5')
os.environ.setdefault('CXXFLAGS', '-include cstdint')
install('git+https://github.com/google/sentencepiece#subdirectory=python', 'sentencepiece')
os.environ.setdefault('CMAKE_POLICY_VERSION_MINIMUM', backup_cmake_policy)
os.environ.setdefault('CXXFLAGS', backup_cxxflags)
else:
install('sentencepiece', 'sentencepiece')
def install_requirements():
t_start = time.time()
if args.profile:
@@ -1207,14 +1287,6 @@ def install_requirements():
return
if int(sys.version_info.minor) >= 13:
install('audioop-lts')
# gcc 15 patch
backup_cmake_policy = os.environ.get('CMAKE_POLICY_VERSION_MINIMUM', None)
backup_cxxflags = os.environ.get('CXXFLAGS', None)
os.environ.setdefault('CMAKE_POLICY_VERSION_MINIMUM', '3.5')
os.environ.setdefault('CXXFLAGS', '-include cstdint')
install('git+https://github.com/google/sentencepiece#subdirectory=python', 'sentencepiece')
os.environ.setdefault('CMAKE_POLICY_VERSION_MINIMUM', backup_cmake_policy)
os.environ.setdefault('CXXFLAGS', backup_cxxflags)
if not installed('diffusers', quiet=True): # diffusers are not installed, so run initial installation
global quick_allowed # pylint: disable=global-statement
quick_allowed = False
javascript/base.css
+7 -7
View File
@@ -10,7 +10,7 @@
.token-counter div { display: inline; }
.token-counter span { padding: 0.1em 0.75em; }
/* tooltips and statuses */
/* tooltips and statuses */
.infotext { overflow-wrap: break-word; }
.tooltip { display: block; position: fixed; top: 1em; right: 1em; padding: 0.5em; background: var(--input-background-fill); color: var(--body-text-color); border: 1pt solid var(--button-primary-border-color);
width: 22em; min-height: 1.3em; font-size: 0.8em; transition: opacity 0.2s ease-in; pointer-events: none; opacity: 0; z-index: 999; }
@@ -98,14 +98,14 @@ table.settings-value-table td { padding: 0.4em; border: 1px solid #ccc; max-widt
.extra-network-cards .card:hover .overlay { background: rgba(0, 0, 0, 0.40); }
.extra-network-cards .card:hover .preview { box-shadow: none; filter: grayscale(100%); }
.extra-network-cards .card:hover .overlay { background: rgba(0, 0, 0, 0.40); }
.extra-network-cards .card .overlay .tags { margin: 4px; display: none; overflow-wrap: break-word; }
.extra-network-cards .card .overlay .tag { padding: 2px; margin: 2px; background: var(--neutral-700); cursor: pointer; display: inline-block; }
.extra-network-cards .card .tags { margin: 4px; display: none; overflow-wrap: break-word; }
.extra-network-cards .card .tag { padding: 2px; margin: 2px; background: var(--neutral-700); cursor: pointer; display: inline-block; }
.extra-network-cards .card .actions > span { padding: 4px; }
.extra-network-cards .card:hover .actions { display: block; }
.extra-network-cards .card:hover .overlay .tags { display: block; }
.extra-network-cards .card:hover .tags { display: block; }
.extra-network-cards .card .actions { font-size: 3em; display: none; text-align-last: right; cursor: pointer; font-variant: unicase; position: absolute; z-index: 100; right: 0; height: 0.7em; width: 100%; background: rgba(0, 0, 0, 0.40); }
#txt2img_description, #img2img_description { max-height: 63px; overflow-y: auto !important; }
#txt2img_description > label > textarea, #img2img_description > label > textarea { font-size: 0.9em }
#txt2img_description, #img2img_description, #control_description, #video_description { max-height: 63px; overflow-y: auto !important; }
#txt2img_description>label>textarea, #img2img_description>label>textarea, #control_description>label>textarea, #video_description>label>textarea { font-size: 0.9em }
.extra-details { position: fixed; bottom: 50%; left: 50%; transform: translate(-50%, 50%); padding: 0.8em; border: var(--block-border-width) solid var(--highlight-color) !important; z-index: 100; box-shadow: var(--button-shadow); }
.extra-details > div { overflow-y: auto; min-height: 40vh; max-height: 80vh; align-self: flex-start; }
@@ -124,7 +124,7 @@ table.settings-value-table td { padding: 0.4em; border: 1px solid #ccc; max-widt
#tab-browser-search { padding: 0; }
#tab-browser-status { align-content: center; text-align: right; background: var(--input-background-fill); padding-right: 1em; margin-left: -1em; }
div:has(>#tab-browser-folders) { flex-grow: 0 !important; background-color: var(--input-background-fill); min-width: max-content !important; }
.browser-separator { background-color: var(--input-background-fill); font-size: larger; padding: 0.5em; display: block !important; }
.browser-separator { background-color: var(--input-background-fill); font-size: larger; padding: 0.5em; display: block !important; }
/* loader */
.splash { position: fixed; top: 0; left: 0; width: 100vw; height: 100vh; z-index: 1000; display: block; text-align: center; }
javascript/black-teal-reimagined.css
+4 -4
View File
@@ -900,7 +900,7 @@ svg.feather.feather-image,
}
/* Tags Styles */
.extra-network-cards .card .overlay .tags {
.extra-network-cards .card .tags {
display: none;
overflow-wrap: anywhere;
position: absolute;
@@ -913,7 +913,7 @@ svg.feather.feather-image,
}
/* Individual Tag Styles */
.extra-network-cards .card .overlay .tag {
.extra-network-cards .card .tag {
padding: 2px;
margin: 2px;
background: rgba(70, 70, 70, 0.60);
@@ -961,7 +961,7 @@ svg.feather.feather-image,
display: block;
}
.extra-network-cards .card:hover .overlay .tags {
.extra-network-cards .card:hover .tags {
display: block;
}
@@ -1192,4 +1192,4 @@ svg.feather.feather-image,
--button-transition: none;
--size-9: 64px;
--size-14: 64px;
}
}
javascript/black-teal.css
+3 -2
View File
@@ -76,12 +76,14 @@ body, button, input, select, textarea { font-family: var(--font); }
button { max-width: 400px; white-space: nowrap; }
img { background-color: var(--background-color); }
input[type='range'] { display: block; margin: 0; padding: 0; height: 1em; background-color: transparent; overflow: hidden; cursor: pointer; box-shadow: 0 0 0 0 transparent; -webkit-appearance: none; appearance: none; }
input[type='range'] { display: block; margin: 0; padding: 0; height: 0.8em; background-color: transparent; overflow: hidden; cursor: pointer; box-shadow: 0 0 0 0 transparent; -webkit-appearance: none; appearance: none; }
input[type='range']::-webkit-slider-thumb { height: .9em; width: .9em; background-color: hsl(180, 54%, 61%); box-shadow: var(--range-shadow); border-radius: var(--radius-xs); }
input[type='range']::-webkit-slider-runnable-track, input[type='range']::-webkit-slider-thumb { -webkit-appearance: none; }
input[type='range']::-moz-range-thumb { height: .9em; width: .9em; background-color: hsl(180, 54%, 61%); box-shadow: var(--range-shadow); }
input[type='range']::-moz-range-track, input[type='range']::-webkit-slider-runnable-track { border: none; background: none; width: 100%; height: 100%; }
.gradio-dropdown .wrap-inner { padding: 2px !important; }
:root { scrollbar-color: var(--highlight-color) #303030; }
::-webkit-scrollbar { width: 12px; height: 12px; }
::-webkit-scrollbar-track { background: #303030; }
@@ -157,7 +159,6 @@ textarea[rows="1"] { height: 33px !important; width: 99% !important; padding: 8p
#txt2img_progress_row > div { min-width: var(--left-column); max-width: var(--left-column); }
#txt2img_settings { min-width: var(--left-column); max-width: var(--left-column); background-color: var(--neutral-950); padding-top: 16px; }
#pnginfo_html2_info { margin-top: -18px; background-color: var(--input-background-fill); padding: var(--input-padding) }
#txt2img_styles_row, #img2img_styles_row, #control_styles_row { margin-top: -6px; }
.block > span { margin-bottom: 0 !important; margin-top: var(--spacing-lg); }
/* based on gradio built-in dark theme */
javascript/changelog.js
+4
View File
@@ -60,6 +60,10 @@ function changelogNavigate(found) {
async function initChangelog() {
const search = gradioApp().querySelector('#changelog_search > label> textarea');
const md = gradioApp().getElementById('changelog_markdown');
if (!search || !md) {
error('initChangelog', 'Missing search or markdown elements');
return;
}
const searchChangelog = async (e) => {
if (changelogElements.length < 100) changelogElements = getAllChildren(md);
const found = [];
javascript/emerald-paradise.css
+1 -1
View File
@@ -14,7 +14,7 @@
--inactive-color: var(--primary--800);
--body-text-color: var(--neutral-100);
--body-text-color-subdued: var(--neutral-300);
--background-color: var(--primary-100);
--background-color: var(--primary-100);
--background-fill-primary: var(--input-background-fill);
--input-padding: 8px;
--input-background-fill: var(--primary-200);
javascript/extraNetworks.js
+4 -2
View File
@@ -101,7 +101,6 @@ function readCardDescription(page, item) {
const description = gradioApp().querySelector(`#${tabname}_description > label > textarea`);
if (description) {
description.value = data?.description?.trim() || '';
// description.focus();
updateInput(description);
}
setENState({ op: 'readCardDescription', page, item });
@@ -488,7 +487,6 @@ function setupExtraNetworksForTab(tabname) {
en.style.top = '13em';
en.style.transition = 'width 0.3s ease';
en.style.zIndex = 100;
// gradioApp().getElementById(`${tabname}_settings`).parentNode.style.width = `${100 - 2 - window.opts.extra_networks_sidebar_width}vw`;
gradioApp().getElementById(`${tabname}_settings`).parentNode.style.width = `calc(100vw - 2em - min(${window.opts.extra_networks_sidebar_width}vw, 50vw))`;
} else {
en.style.position = 'relative';
@@ -506,6 +504,10 @@ function setupExtraNetworksForTab(tabname) {
if (window.opts.extra_networks_card_cover === 'sidebar') en.style.width = 0;
gradioApp().getElementById(`${tabname}_settings`).parentNode.style.width = 'unset';
}
if (tabname === 'video') {
gradioApp().getElementById('framepack_settings').parentNode.style.width = gradioApp().getElementById(`${tabname}_settings`).parentNode.style.width;
gradioApp().getElementById('ltx_settings').parentNode.style.width = gradioApp().getElementById(`${tabname}_settings`).parentNode.style.width;
}
}
});
intersectionObserver.observe(en); // monitor visibility
javascript/gallery.js
+254 -51
View File
@@ -1,4 +1,11 @@
/* eslint-disable max-classes-per-file */
// Known issues
// Images flash on the screen before they get processed and separator is properly closed, especially when root/subfolder has large amount of files
// Search is a bit wonky, I tried to get the separators to hide if 0 hits in seperator are found, but no luck so
// Sorting huge amount of images is slow, might look at optimising, I don't think it's a regression.
// TODO
// Setting to enable or disable separator state persistence
let ws;
let url;
@@ -6,7 +13,9 @@ let currentImage;
let pruneImagesTimer;
let outstanding = 0;
let lastSort = 0;
let lastSortName = 'none';
let lastSortName = 'None';
// Store separator states for the session
const separatorStates = new Map();
const el = {
folders: undefined,
files: undefined,
@@ -15,6 +24,8 @@ const el = {
btnSend: undefined,
};
const SUPPORTED_EXTENSIONS = ['jpg', 'jpeg', 'png', 'webp', 'tiff', 'jp2', 'jxl', 'gif', 'mp4', 'mkv', 'avi', 'mjpeg', 'mpg', 'avr'];
// HTML Elements
class GalleryFolder extends HTMLElement {
@@ -28,9 +39,10 @@ class GalleryFolder extends HTMLElement {
const style = document.createElement('style'); // silly but necessasry since we're inside shadowdom
if (window.opts.theme_type === 'Modern') {
style.textContent = `
.gallery-folder { cursor: pointer; padding: 8px 6px 8px 6px; background-color: var(--sd-secondary-color); }
.gallery-folder { cursor: pointer; padding: 8px 6px 8px 6px; background-color: var(--sd-button-normal-color); border-radius: var(--sd-border-radius); text-align: left; min-width: 12em;}
.gallery-folder:hover { background-color: var(--button-primary-background-fill-hover); }
.gallery-folder-selected { background-color: var(--sd-button-selected-color); color: var(--sd-button-selected-text-color); }
.gallery-folder-icon { font-size: 1.2em; color: var(--sd-button-icon-color); margin-right: 1em; filter: drop-shadow(1px 1px 2px black); float: left; }
`;
} else {
style.textContent = `
@@ -42,7 +54,7 @@ class GalleryFolder extends HTMLElement {
this.shadow.appendChild(style);
const div = document.createElement('div');
div.className = 'gallery-folder';
div.textContent = `\uf44a ${this.name}`;
div.innerHTML = `<span class="gallery-folder-icon">\uf03e</span> ${this.name}`;
div.addEventListener('click', () => {
for (const folder of el.folders.children) {
if (folder.name === this.name) folder.shadow.children[1].classList.add('gallery-folder-selected');
@@ -71,25 +83,110 @@ async function createThumb(img) {
return dataURL;
}
async function handleSeparator(separator) {
separator.classList.toggle('gallery-separator-hidden');
const nowHidden = separator.classList.contains('gallery-separator-hidden');
// Store the state (true = open, false = closed)
separatorStates.set(separator.title, !nowHidden);
// Update arrow and count
const arrow = separator.querySelector('.gallery-separator-arrow');
arrow.style.transform = nowHidden ? 'rotate(0deg)' : 'rotate(90deg)';
const all = Array.from(el.files.children);
for (const f of all) {
if (!f.name) continue; // Skip separators
// Check if file belongs to this exact directory
const fileDir = f.name.match(/(.*)[\/\\]/);
const fileDirPath = fileDir ? fileDir[1] : '';
if (separator.title.length > 0 && fileDirPath === separator.title) {
f.style.display = nowHidden ? 'none' : 'unset';
}
}
// Note: Count is not updated here on manual toggle, as it reflects the total.
// If I end up implementing it, the search function will handle dynamic count updates.
}
async function addSeparators() {
document.querySelectorAll('.gallery-separator').forEach((node) => el.files.removeChild(node));
const all = Array.from(el.files.children);
let lastDir;
let isFirstSeparator = true; // Flag to open the first separator by default
// First pass: create separators
for (const f of all) {
let dir = f.name.match(/(.*)[\/\\]/);
let dir = f.name?.match(/(.*)[\/\\]/);
if (!dir) dir = '';
else dir = dir[1];
if (dir !== lastDir) {
lastDir = dir;
if (dir.length > 0) {
// Count files in this directory
let fileCount = 0;
for (const file of all) {
if (!file.name) continue;
const fileDir = file.name.match(/(.*)[\/\\]/);
const fileDirPath = fileDir ? fileDir[1] : '';
if (fileDirPath === dir) fileCount++;
}
const sep = document.createElement('div');
sep.className = 'gallery-separator';
sep.innerText = dir;
sep.title = dir;
// Default to open for the first separator if no state is saved, otherwise closed.
const isOpen = separatorStates.has(dir) ? separatorStates.get(dir) : isFirstSeparator;
separatorStates.set(dir, isOpen); // Ensure it's in the map
if (isFirstSeparator) isFirstSeparator = false; // Subsequent separators will default to closed
if (!isOpen) {
sep.classList.add('gallery-separator-hidden');
}
// Create arrow span
const arrow = document.createElement('span');
arrow.className = 'gallery-separator-arrow';
arrow.textContent = '▶';
arrow.style.transform = isOpen ? 'rotate(90deg)' : 'rotate(0deg)';
// Create directory name span
const dirName = document.createElement('span');
dirName.className = 'gallery-separator-name';
dirName.textContent = dir;
dirName.title = dir; // Show full path on hover
// Create count span
const count = document.createElement('span');
count.className = 'gallery-separator-count';
count.textContent = `${fileCount} files`;
sep.dataset.totalFiles = fileCount; // Store total count for search filtering
sep.appendChild(arrow);
sep.appendChild(dirName);
sep.appendChild(count);
sep.onclick = () => handleSeparator(sep);
el.files.insertBefore(sep, f);
}
}
}
// Second pass: hide files in closed directories
for (const f of all) {
if (!f.name) continue; // Skip separators
const dir = f.name.match(/(.*)[\/\\]/);
if (dir && dir[1]) {
const dirPath = dir[1];
const isOpen = separatorStates.get(dirPath);
if (isOpen === false) {
f.style.display = 'none';
}
}
}
}
async function delayFetchThumb(fn) {
@@ -129,11 +226,17 @@ class GalleryFile extends HTMLElement {
if (this.shadow.children.length > 0) {
return;
}
const ext = this.name.split('.').pop().toLowerCase();
if (!['jpg', 'jpeg', 'png', 'gif', 'webp', 'jxl', 'svg', 'mp4'].includes(ext)) {
// console.error(`gallery: type=${ext} file=${this.name} unsupported`);
return;
// Check separator state early to hide the element immediately
const dir = this.name.match(/(.*)[\/\\]/);
if (dir && dir[1]) {
const dirPath = dir[1];
const isOpen = separatorStates.get(dirPath);
if (isOpen === false) {
this.style.display = 'none';
}
}
this.hash = await getHash(`${this.folder}/${this.name}/${this.size}/${this.mtime}`); // eslint-disable-line no-use-before-define
const style = document.createElement('style');
const width = opts.browser_fixed_width ? `${opts.extra_networks_card_size}px` : 'unset';
@@ -214,7 +317,13 @@ class GalleryFile extends HTMLElement {
return; // avoid double-adding
}
this.title = img.title;
this.style.display = this.title.toLowerCase().includes(el.search.value.toLowerCase()) ? 'unset' : 'none';
// Final visibility check based on search term.
const shouldDisplayBasedOnSearch = this.title.toLowerCase().includes(el.search.value.toLowerCase());
if (this.style.display !== 'none') { // Only proceed if not already hidden by a closed separator
this.style.display = shouldDisplayBasedOnSearch ? 'unset' : 'none';
}
this.shadow.appendChild(style);
this.shadow.appendChild(img);
}
@@ -237,6 +346,12 @@ async function getHash(str, algo = 'SHA-256') {
}
}
// Helper function to update status with sort mode
function updateStatusWithSort(message) {
const sortIndicator = `Sort: ${lastSortName} | `;
el.status.innerText = sortIndicator + message;
}
async function wsConnect(socket, timeout = 5000) {
const intrasleep = 100;
const ttl = timeout / intrasleep;
@@ -251,15 +366,49 @@ async function wsConnect(socket, timeout = 5000) {
return isOpened();
}
async function gallerySearch(evt) {
async function gallerySearch() {
if (el.search.busy) clearTimeout(el.search.busy);
el.search.busy = setTimeout(async () => {
let numFound = 0;
const all = Array.from(el.files.children);
const str = el.search.value.toLowerCase();
const r = /^(.+)([=<>])(.*)/;
const t0 = performance.now();
for (const f of all) {
const str = el.search.value.toLowerCase();
const allFiles = Array.from(el.files.children).filter((node) => node.name);
const allSeparators = Array.from(el.files.children).filter((node) => node.classList.contains('gallery-separator'));
// If search is cleared, restore original view
if (str === '') {
allSeparators.forEach((sep) => {
sep.style.display = 'flex';
const isOpen = separatorStates.has(sep.title) ? separatorStates.get(sep.title) : false;
const countSpan = sep.querySelector('.gallery-separator-count');
if (countSpan && sep.dataset.totalFiles) {
countSpan.textContent = `${sep.dataset.totalFiles} files`;
}
const arrow = sep.querySelector('.gallery-separator-arrow');
sep.classList.toggle('gallery-separator-hidden', !isOpen);
if (arrow) arrow.style.transform = isOpen ? 'rotate(90deg)' : 'rotate(0deg)';
});
allFiles.forEach((f) => {
const dir = f.name.match(/(.*)[\/\\]/);
const dirPath = (dir && dir[1]) ? dir[1] : '';
const isOpen = separatorStates.get(dirPath);
f.style.display = (!dirPath || isOpen) ? 'unset' : 'none';
});
updateStatusWithSort(`Filter | Cleared | ${allFiles.length.toLocaleString()} images`);
return;
}
// --- Search logic ---
let totalFound = 0;
const directoryMatches = new Map();
const fileMatches = new WeakSet();
const r = /^(.+)([=<>])(.*)/;
for (const f of allFiles) {
let isMatch = false;
if (r.test(str)) {
const match = str.match(r);
const key = match[1].trim();
@@ -267,20 +416,46 @@ async function gallerySearch(evt) {
let val = match[3].trim();
if (key === 'mtime') val = new Date(val);
if (((op === '=') && (f[key] === val)) || ((op === '>') && (f[key] > val)) || ((op === '<') && (f[key] < val))) {
f.style.display = 'unset';
numFound++;
} else {
f.style.display = 'none';
isMatch = true;
}
} else if (f.title?.toLowerCase().includes(str) || f.exif?.toLowerCase().includes(str)) {
f.style.display = 'unset';
numFound++;
} else {
f.style.display = 'none';
isMatch = true;
}
if (isMatch) {
fileMatches.add(f);
totalFound++;
const dir = f.name.match(/(.*)[\/\\]/);
const dirPath = (dir && dir[1]) ? dir[1] : '';
directoryMatches.set(dirPath, (directoryMatches.get(dirPath) || 0) + 1);
}
const t1 = performance.now();
el.status.innerText = `Filter | ${f.folder} | ${numFound.toLocaleString()}/${all.length.toLocaleString()} images | ${Math.floor(t1 - t0).toLocaleString()}ms`;
}
// Update separators based on search results
for (const sep of allSeparators) {
const dirPath = sep.title;
const foundCount = directoryMatches.get(dirPath) || 0;
if (foundCount > 0) {
sep.style.display = 'flex'; // Show separator
sep.classList.remove('gallery-separator-hidden'); // Force open
const arrow = sep.querySelector('.gallery-separator-arrow');
if (arrow) arrow.style.transform = 'rotate(90deg)';
// Removed file count update during search as it was buggy.
} else {
sep.style.display = 'none'; // Hide separator
}
}
// Update file visibility
for (const f of allFiles) {
f.style.display = fileMatches.has(f) ? 'unset' : 'none';
}
const t1 = performance.now();
updateStatusWithSort(`Filter | ${totalFound.toLocaleString()} / ${allFiles.length.toLocaleString()} images | ${Math.floor(t1 - t0).toLocaleString()}ms`);
}, 250);
}
@@ -299,53 +474,53 @@ async function gallerySort(btn) {
const arr = Array.from(el.files.children).filter((node) => node.name); // filter out separators
if (arr.length === 0) return; // no files to sort
if (btn) lastSort = btn.charCodeAt(0);
lastSortName = 'none';
lastSortName = 'None';
const fragment = document.createDocumentFragment();
switch (lastSort) {
case 61789: // name asc
lastSortName = 'name asc';
lastSortName = 'Name Ascending';
arr
.sort((a, b) => a.name.localeCompare(b.name))
.forEach((node) => fragment.appendChild(node));
break;
case 61790: // name dsc
lastSortName = 'name dsc';
lastSortName = 'Name Descending';
arr
.sort((b, a) => a.name.localeCompare(b.name))
.forEach((node) => fragment.appendChild(node));
break;
case 61792: // size asc
lastSortName = 'size asc';
lastSortName = 'Size Ascending';
arr
.sort((a, b) => a.size - b.size)
.forEach((node) => fragment.appendChild(node));
break;
case 61793: // size dsc
lastSortName = 'size dsc';
lastSortName = 'Size Descending';
arr
.sort((b, a) => a.size - b.size)
.forEach((node) => fragment.appendChild(node));
break;
case 61794: // resolution asc
lastSortName = 'resolution asc';
lastSortName = 'Resolution Ascending';
arr
.sort((a, b) => a.width * a.height - b.width * b.height)
.forEach((node) => fragment.appendChild(node));
break;
case 61795: // resolution dsc
lastSortName = 'resolution dsc';
lastSortName = 'Resolution Descending';
arr
.sort((b, a) => a.width * a.height - b.width * b.height)
.forEach((node) => fragment.appendChild(node));
break;
case 61662:
lastSortName = 'modified asc';
lastSortName = 'Modified Ascending';
arr
.sort((a, b) => a.mtime - b.mtime)
.forEach((node) => fragment.appendChild(node));
break;
case 61661:
lastSortName = 'modified dsc';
lastSortName = 'Modified Descending';
arr
.sort((b, a) => a.mtime - b.mtime)
.forEach((node) => fragment.appendChild(node));
@@ -357,35 +532,55 @@ async function gallerySort(btn) {
el.files.innerHTML = '';
el.files.appendChild(fragment);
addSeparators();
// After sorting and adding separators, ensure files respect separator states
const all = Array.from(el.files.children);
for (const f of all) {
if (!f.name) continue; // Skip separators
const dir = f.name.match(/(.*)[\/\\]/);
if (dir && dir[1]) {
const dirPath = dir[1];
const isOpen = separatorStates.get(dirPath);
if (isOpen === false) {
f.style.display = 'none';
}
}
}
const t1 = performance.now();
log(`gallerySort: char=${lastSort} len=${arr.length} time=${Math.floor(t1 - t0)} sort=${lastSortName}`);
el.status.innerText = `Sort | ${lastSortName} | ${arr.length.toLocaleString()} images | ${Math.floor(t1 - t0).toLocaleString()}ms`;
updateStatusWithSort(`${arr.length.toLocaleString()} images | ${Math.floor(t1 - t0).toLocaleString()}ms`);
}
async function fetchFilesHT(evt) {
const t0 = performance.now();
const fragment = document.createDocumentFragment();
el.status.innerText = `Folder | ${evt.target.name} | in-progress`;
updateStatusWithSort(`Folder: ${evt.target.name} | in-progress`);
let numFiles = 0;
const res = await fetch(`${window.api}/browser/files?folder=${encodeURI(evt.target.name)}`);
if (!res || res.status !== 200) {
el.status.innerText = `Folder | ${evt.target.name} | failed: ${res?.statusText}`;
updateStatusWithSort(`Folder: ${evt.target.name} | failed: ${res?.statusText}`);
return;
}
const jsonData = await res.json();
for (const line of jsonData) {
const data = decodeURI(line).split('##F##');
numFiles++;
const f = new GalleryFile(data[0], data[1]);
fragment.appendChild(f);
const fileName = data[1];
const ext = fileName.split('.').pop().toLowerCase();
if (SUPPORTED_EXTENSIONS.includes(ext)) {
numFiles++;
const f = new GalleryFile(data[0], fileName);
fragment.appendChild(f);
}
}
el.files.appendChild(fragment);
const t1 = performance.now();
log(`gallery: folder=${evt.target.name} num=${numFiles} time=${Math.floor(t1 - t0)}ms`);
el.status.innerText = `Folder | ${evt.target.name} | ${numFiles.toLocaleString()} images | ${Math.floor(t1 - t0).toLocaleString()}ms`;
updateStatusWithSort(`Folder: ${evt.target.name} | ${numFiles.toLocaleString()} images | ${Math.floor(t1 - t0).toLocaleString()}ms`);
addSeparators();
}
@@ -406,25 +601,29 @@ async function fetchFilesWS(evt) { // fetch file-by-file list over websockets
await fetchFilesHT(evt); // fallback to http
return;
}
el.status.innerText = `Folder | ${evt.target.name}`;
updateStatusWithSort(`Folder: ${evt.target.name}`);
const t0 = performance.now();
let numFiles = 0;
let t1 = performance.now();
let fragment = document.createDocumentFragment();
ws.onmessage = (event) => {
numFiles++;
t1 = performance.now();
const data = decodeURI(event.data).split('##F##');
if (data[0] === '#END#') {
ws.close();
} else {
const file = new GalleryFile(data[0], data[1]);
fragment.appendChild(file);
if (numFiles % 100 === 0) {
el.status.innerText = `Folder | ${evt.target.name} | ${numFiles.toLocaleString()} images | in-progress | ${Math.floor(t1 - t0).toLocaleString()}ms`;
el.files.appendChild(fragment);
fragment = document.createDocumentFragment();
const fileName = data[1];
const ext = fileName.split('.').pop().toLowerCase();
if (SUPPORTED_EXTENSIONS.includes(ext)) {
const file = new GalleryFile(data[0], fileName);
numFiles++;
fragment.appendChild(file);
if (numFiles % 100 === 0) {
updateStatusWithSort(`Folder: ${evt.target.name} | ${numFiles.toLocaleString()} images | in-progress | ${Math.floor(t1 - t0).toLocaleString()}ms`);
el.files.appendChild(fragment);
fragment = document.createDocumentFragment();
}
}
}
};
@@ -432,7 +631,7 @@ async function fetchFilesWS(evt) { // fetch file-by-file list over websockets
el.files.appendChild(fragment);
// gallerySort();
log(`gallery: folder=${evt.target.name} num=${numFiles} time=${Math.floor(t1 - t0)}ms`);
el.status.innerText = `Folder | ${evt.target.name} | ${numFiles.toLocaleString()} images | ${Math.floor(t1 - t0).toLocaleString()}ms`;
updateStatusWithSort(`Folder: ${evt.target.name} | ${numFiles.toLocaleString()} images | ${Math.floor(t1 - t0).toLocaleString()}ms`);
addSeparators();
};
ws.onerror = (event) => {
@@ -469,6 +668,10 @@ async function initGallery() { // triggered on gradio change to monitor when ui
el.files = gradioApp().getElementById('tab-gallery-files');
el.status = gradioApp().getElementById('tab-gallery-status');
el.search = gradioApp().querySelector('#tab-gallery-search textarea');
if (!el.folders || !el.files || !el.status || !el.search) {
error('initGallery', 'Missing gallery elements');
return;
}
el.search.addEventListener('input', gallerySearch);
el.btnSend = gradioApp().getElementById('tab-gallery-send-image');
document.getElementById('tab-gallery-files').style.height = opts.logmonitor_show ? '75vh' : '85vh';
javascript/ibmplexmono-nerdfont-medium.ttf
BIN
View File
Binary file not shown.
javascript/invoked.css
+5 -5
View File
@@ -6,15 +6,15 @@
--primary-100: #2b303b;
--primary-200: #15181e;
--primary-300: #0a0c0e;
--primary-400: #566176;
--primary-500: #483a90;
--primary-700: #6b7994;
--primary-800: #5b49b3;
--primary-400: #566176;
--primary-500: #483a90;
--primary-700: #6b7994;
--primary-800: #5b49b3;
--highlight-color: var(--primary-500);
--inactive-color: var(--primary--800);
--body-text-color: var(--neutral-100);
--body-text-color-subdued: var(--neutral-300);
--background-color: var(--primary-100);
--background-color: var(--primary-100);
--background-fill-primary: var(--input-background-fill);
--input-padding: 8px;
--input-background-fill: var(--primary-200);
javascript/login.js
+1 -1
View File
@@ -12,7 +12,7 @@ const loginCSS = `
const loginHTML = `
<div id="loginDiv" style="margin: 15% auto; max-width: 200px; padding: 2em; background: var(--background-fill-secondary);">
<h2>Login</h2>
<h2>Login</h2>
<label for="username" style="margin-top: 0.5em">Username</label>
<input type="text" id="loginUsername" name="username" style="width: 92%; padding: 0.5em; margin-top: 0.5em">
<label for="password" style="margin-top: 0.5em">Password</label>
javascript/orchid-dreams.css
+1 -1
View File
@@ -14,7 +14,7 @@
--inactive-color: var(--primary--800);
--body-text-color: var(--neutral-100);
--body-text-color-subdued: var(--neutral-300);
--background-color: var(--primary-100);
--background-color: var(--primary-100);
--background-fill-primary: var(--input-background-fill);
--input-padding: 8px;
--input-background-fill: var(--primary-200);
javascript/panZoom.js
+3 -3
View File
@@ -1093,7 +1093,7 @@ function makeDomController(domElement, options) {
domElement.scrollTop = 0;
if (!options.disableKeyboardInteraction) {
owner.setAttribute('tabindex', 0);
if (owner) owner.setAttribute('tabindex', 0);
}
var api = {
@@ -1287,7 +1287,7 @@ function makeSvgController(svgElement, options) {
}
if (!options.disableKeyboardInteraction) {
owner.setAttribute('tabindex', 0);
if (owner) owner.setAttribute('tabindex', 0);
}
var api = {
@@ -1339,7 +1339,7 @@ function makeSvgController(svgElement, options) {
}
function applyTransform(transform) {
svgElement.setAttribute('transform', 'matrix(' +
if (svgElement) svgElement.setAttribute('transform', 'matrix(' +
transform.scale + ' 0 0 ' +
transform.scale + ' ' +
transform.x + ' ' + transform.y + ')')
javascript/sdnext.css
+1967 -396
View File
File diff suppressed because it is too large Load Diff
javascript/settings.js
+4
View File
@@ -181,6 +181,10 @@ async function initSettings() {
if (settingsInitialized) return;
settingsInitialized = true;
const tabNavElements = gradioApp().querySelector('#settings > .tab-nav');
if (!tabNavElements) {
error('initSettings', 'No tab nav elements found');
return;
}
const tabNavButtons = gradioApp().querySelectorAll('#settings > .tab-nav > button');
const tabElements = gradioApp().querySelectorAll('#settings > div:not(.tab-nav)');
const observer = new MutationObserver((mutations) => {
javascript/timeless-beige.css
+1 -1
View File
@@ -14,7 +14,7 @@
--inactive-color: var(--primary--800);
--body-text-color: var(--neutral-100);
--body-text-color-subdued: var(--neutral-300);
--background-color: var(--primary-100);
--background-color: var(--primary-100);
--background-fill-primary: var(--input-background-fill);
--input-padding: 8px;
--input-background-fill: var(--primary-200);
javascript/ui.js
+28
View File
@@ -264,6 +264,34 @@ function submit_video(...args) {
return res;
}
function submit_framepack(...args) {
const id = randomId();
log('submitFramepack', id);
requestProgress(id, null, null);
window.submit_state = '';
args[0] = id;
return args;
}
function submit_ltx(...args) {
const id = randomId();
log('submitFramepack', id);
requestProgress(id, null, null);
window.submit_state = '';
args[0] = id;
return args;
}
function submit_video_wrapper(...args) {
const modernEl = gradioApp().querySelector('.video_output.fade-in');
let id = modernEl ? modernEl.id : args[0];
id = id.replace('video-selector-', '');
log('submitVideoWrapper', id);
const btn = gradioApp().getElementById(`${id}_generate_btn`);
if (btn) btn.click();
else console.log('submit_video_wrapper: no button found', id);
}
function submit_postprocessing(...args) {
log('SubmitExtras');
clearGallery('extras');
launch.py
+6 -3
View File
@@ -112,8 +112,8 @@ def check_run(command): # compatbility function
@lru_cache()
def is_installed(package): # compatbility function
return installer.installed(package)
def is_installed(pkg): # compatbility function
return installer.installed(pkg)
@lru_cache()
@@ -169,7 +169,7 @@ def clean_server():
modules_to_remove = ['webui', 'modules', 'scripts', 'gradio',
'onnx', 'torch', 'pytorch', 'lightning', 'tensor', 'diffusers', 'transformers', 'tokenize', 'safetensors', 'gguf', 'accelerate', 'peft', 'triton', 'huggingface',
'PIL', 'cv2', 'timm', 'numpy', 'scipy', 'sympy', 'sklearn', 'skimage', 'sqlalchemy', 'flash_attn', 'bitsandbytes', 'xformers', 'matplotlib', 'optimum', 'pandas', 'pi', 'git', 're', 'altair',
'framepack', 'nudenet', 'agent_scheduler', 'basicsr', 'k_diffusion', 'gfpgan', 'war',
'framepack', 'nudenet', 'agent_scheduler', 'basicsr', 'gfpgan', 'war',
'fastapi', 'urllib', 'uvicorn', 'web', 'http', 'google', 'starlette', 'socket']
removed_removed = []
for module_loaded in modules_loaded:
@@ -260,9 +260,12 @@ def main():
installer.set_environment()
if args.uv:
installer.install("uv", "uv")
installer.install_gradio()
installer.check_torch()
installer.check_onnx()
installer.check_transformers()
installer.check_diffusers()
installer.install_sentencepiece()
installer.check_modified_files()
if args.test:
installer.log.info('Startup: test mode')
models/Reference/Freepik--F-Lite.jpg
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 51 KiB

models/Reference/Wan-AI--Wan2.1.jpg
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 25 KiB

models/Reference/Wan-AI--Wan2.2.jpg
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 34 KiB

models/Reference/briaai--BRIA-3.2.jpg
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 31 KiB

models/karlo/ViT-L-14_stats.th
BIN
View File
Binary file not shown.
modules/api/api.py
+9 -6
View File
@@ -5,7 +5,7 @@ from fastapi import FastAPI, APIRouter, Depends, Request
from fastapi.security import HTTPBasic, HTTPBasicCredentials
from fastapi.exceptions import HTTPException
from modules import errors, shared, postprocessing
from modules.api import models, endpoints, script, helpers, server, nvml, generate, process, control, gallery, loras, docs
from modules.api import models, endpoints, script, helpers, server, nvml, generate, process, control, docs
errors.install()
@@ -79,7 +79,6 @@ class Api:
self.add_api_route("/sdapi/v1/upscalers", endpoints.get_upscalers, methods=["GET"], response_model=List[models.ItemUpscaler])
self.add_api_route("/sdapi/v1/sd-models", endpoints.get_sd_models, methods=["GET"], response_model=List[models.ItemModel])
self.add_api_route("/sdapi/v1/controlnets", endpoints.get_controlnets, methods=["GET"], response_model=List[str])
self.add_api_route("/sdapi/v1/hypernetworks", endpoints.get_hypernetworks, methods=["GET"], response_model=List[models.ItemHypernetwork])
self.add_api_route("/sdapi/v1/face-restorers", endpoints.get_detailers, methods=["GET"], response_model=List[models.ItemDetailer])
self.add_api_route("/sdapi/v1/prompt-styles", endpoints.get_prompt_styles, methods=["GET"], response_model=List[models.ItemStyle])
self.add_api_route("/sdapi/v1/embeddings", endpoints.get_embeddings, methods=["GET"], response_model=models.ResEmbeddings)
@@ -95,19 +94,23 @@ class Api:
self.add_api_route("/sdapi/v1/refresh-checkpoints", endpoints.post_refresh_checkpoints, methods=["POST"])
self.add_api_route("/sdapi/v1/unload-checkpoint", endpoints.post_unload_checkpoint, methods=["POST"])
self.add_api_route("/sdapi/v1/reload-checkpoint", endpoints.post_reload_checkpoint, methods=["POST"])
self.add_api_route("/sdapi/v1/lock-checkpoint", endpoints.post_lock_checkpoint, methods=["POST"])
self.add_api_route("/sdapi/v1/refresh-vae", endpoints.post_refresh_vae, methods=["POST"])
self.add_api_route("/sdapi/v1/latents", endpoints.get_latent_history, methods=["GET"], response_model=List[str])
self.add_api_route("/sdapi/v1/latents", endpoints.post_latent_history, methods=["POST"], response_model=int)
# lora api
if shared.native:
self.add_api_route("/sdapi/v1/lora", loras.get_lora, methods=["GET"], response_model=dict)
self.add_api_route("/sdapi/v1/loras", loras.get_loras, methods=["GET"], response_model=List[dict])
self.add_api_route("/sdapi/v1/refresh-loras", loras.post_refresh_loras, methods=["POST"])
from modules.api import loras
loras.register_api()
# gallery api
from modules.api import gallery
gallery.register_api(self.app)
# nudenet api
from modules.api import nudenet
nudenet.register_api()
def add_api_route(self, path: str, endpoint, **kwargs):
if (shared.cmd_opts.auth or shared.cmd_opts.auth_file) and shared.cmd_opts.api_only:
modules/api/endpoints.py
+16 -24
View File
@@ -17,19 +17,16 @@ def get_upscalers():
return [{"name": upscaler.name, "model_name": upscaler.scaler.model_name, "model_path": upscaler.data_path, "model_url": None, "scale": upscaler.scale} for upscaler in shared.sd_upscalers]
def get_sd_models():
from modules import sd_checkpoint, sd_models_config
from modules import sd_checkpoint
checkpoints = []
for v in sd_checkpoint.checkpoints_list.values():
checkpoints.append({"title": v.title, "model_name": v.name, "filename": v.filename, "type": v.type, "hash": v.shorthash, "sha256": v.sha256, "config": sd_models_config.find_checkpoint_config_near_filename(v)})
checkpoints.append({"title": v.title, "model_name": v.name, "filename": v.filename, "type": v.type, "hash": v.shorthash, "sha256": v.sha256})
return checkpoints
def get_controlnets(model_type: Optional[str] = None):
from modules.control.units.controlnet import api_list_models
return api_list_models(model_type)
def get_hypernetworks():
return [{"name": name, "path": shared.hypernetworks[name]} for name in shared.hypernetworks]
def get_detailers():
return [{"name":x.name(), "cmd_dir": getattr(x, "cmd_dir", None)} for x in shared.detailers]
@@ -37,15 +34,10 @@ def get_prompt_styles():
return [{ 'name': v.name, 'prompt': v.prompt, 'negative_prompt': v.negative_prompt, 'extra': v.extra, 'filename': v.filename, 'preview': v.preview} for v in shared.prompt_styles.styles.values()]
def get_embeddings():
from modules import sd_hijack
db = sd_hijack.model_hijack.embedding_db
def convert_embedding(embedding):
return {"step": embedding.step, "sd_checkpoint": embedding.sd_checkpoint, "sd_checkpoint_name": embedding.sd_checkpoint_name, "shape": embedding.shape, "vectors": embedding.vectors}
def convert_embeddings(embeddings):
return {embedding.name: convert_embedding(embedding) for embedding in embeddings.values()}
return {"loaded": convert_embeddings(db.word_embeddings), "skipped": convert_embeddings(db.skipped_embeddings)}
db = getattr(shared.sd_model, 'embedding_db', None) if shared.sd_loaded else None
if db is None:
return models.ResEmbeddings(loaded=[], skipped=[])
return models.ResEmbeddings(loaded=list(db.word_embeddings.keys()), skipped=list(db.skipped_embeddings.keys()))
def get_extra_networks(page: Optional[str] = None, name: Optional[str] = None, filename: Optional[str] = None, title: Optional[str] = None, fullname: Optional[str] = None, hash: Optional[str] = None): # pylint: disable=redefined-builtin
res = []
@@ -76,21 +68,14 @@ def get_extra_networks(page: Optional[str] = None, name: Optional[str] = None, f
def get_interrogate():
from modules.interrogate.openclip import refresh_clip_models
return ['clip', 'deepdanbooru'] + refresh_clip_models()
return ['deepdanbooru'] + refresh_clip_models()
def post_interrogate(req: models.ReqInterrogate):
if req.image is None or len(req.image) < 64:
raise HTTPException(status_code=404, detail="Image not found")
image = helpers.decode_base64_to_image(req.image)
image = image.convert('RGB')
if req.model == "clip":
try:
from modules.interrogate import openclip
caption = openclip.interrogator.interrogate(image)
except Exception as e:
caption = str(e)
return models.ResInterrogate(caption=caption)
elif req.model == "deepdanbooru" or req.model == 'deepbooru':
if req.model == "deepdanbooru" or req.model == 'deepbooru':
from modules.interrogate import deepbooru
caption = deepbooru.model.tag(image)
return models.ResInterrogate(caption=caption)
@@ -123,11 +108,18 @@ def post_unload_checkpoint():
sd_models.unload_model_weights(op='refiner')
return {}
def post_reload_checkpoint():
def post_reload_checkpoint(force:bool=False):
from modules import sd_models
if force:
sd_models.unload_model_weights(op='model')
sd_models.reload_model_weights()
return {}
def post_lock_checkpoint(lock:bool=False):
from modules import modeldata
modeldata.model_data.locked = lock
return {}
def get_checkpoint():
if not shared.sd_loaded or shared.sd_model is None:
checkpoint = {
modules/api/generate.py
+7 -7
View File
@@ -1,6 +1,6 @@
from threading import Lock
from fastapi.responses import JSONResponse
from modules import errors, shared, scripts, ui
from modules import errors, shared, scripts_manager, ui
from modules.api import models, script, helpers
from modules.processing import StableDiffusionProcessingTxt2Img, StableDiffusionProcessingImg2Img, process_images
@@ -85,7 +85,7 @@ class APIGenerate():
def post_text2img(self, txt2imgreq: models.ReqTxt2Img):
self.prepare_face_module(txt2imgreq)
script_runner = scripts.scripts_txt2img
script_runner = scripts_manager.scripts_txt2img
if not script_runner.scripts:
script_runner.initialize_scripts(False)
ui.create_ui(None)
@@ -113,10 +113,10 @@ class APIGenerate():
script_args = script.init_script_args(p, txt2imgreq, self.default_script_arg_txt2img, selectable_scripts, selectable_script_idx, script_runner)
p.script_args = tuple(script_args) # Need to pass args as tuple here
if selectable_scripts is not None:
processed = scripts.scripts_txt2img.run(p, *script_args) # Need to pass args as list here
processed = scripts_manager.scripts_txt2img.run(p, *script_args) # Need to pass args as list here
else:
processed = process_images(p)
processed = scripts.scripts_txt2img.after(p, processed, *script_args)
processed = scripts_manager.scripts_txt2img.after(p, processed, *script_args)
p.close()
shared.state.end(api=False)
if processed is None or processed.images is None or len(processed.images) == 0:
@@ -135,7 +135,7 @@ class APIGenerate():
mask = img2imgreq.mask
if mask:
mask = helpers.decode_base64_to_image(mask)
script_runner = scripts.scripts_img2img
script_runner = scripts_manager.scripts_img2img
if not script_runner.scripts:
script_runner.initialize_scripts(True)
ui.create_ui(None)
@@ -165,10 +165,10 @@ class APIGenerate():
script_args = script.init_script_args(p, img2imgreq, self.default_script_arg_img2img, selectable_scripts, selectable_script_idx, script_runner)
p.script_args = tuple(script_args) # Need to pass args as tuple here
if selectable_scripts is not None:
processed = scripts.scripts_img2img.run(p, *script_args) # Need to pass args as list here
processed = scripts_manager.scripts_img2img.run(p, *script_args) # Need to pass args as list here
else:
processed = process_images(p)
processed = scripts.scripts_img2img.after(p, processed, *script_args)
processed = scripts_manager.scripts_img2img.after(p, processed, *script_args)
p.close()
shared.state.end(api=False)
if processed is None or processed.images is None or len(processed.images) == 0:

Some files were not shown because too many files have changed in this diff Show More