fe6936449d
This PR adds support for the Qwen-Image-Layered model, which decomposes an input image into semantically disentangled RGBA layers for layer-based editing workflows. ## Features - New CLI command: \`mflux-generate-qwen-layered\` - Decomposes images into N RGBA layers (default 4) - Supports 6-bit quantization for ~29GB memory usage (vs 55GB BF16) - Resolution buckets: 640 and 1024 ## Architecture - RGBA-VAE (4-channel) with 3D temporal convolutions for layer handling - Layer3D RoPE: 3D positional encoding [layer, height, width] - Uses base QwenTransformer with extended RoPE for multi-layer sequences - Condition image encoded as layer_index=-1 for proper decomposition ## New Files - \`src/mflux/models/qwen_layered/\` - Full model implementation - \`model/qwen_layered_vae/\` - RGBA-VAE encoder/decoder - \`model/qwen_layered_transformer/\` - Layer3D RoPE - \`weights/\` - Weight mapping and definitions - \`variants/i2l/\` - Image-to-Layers pipeline - \`cli/\` - Command-line interface ## Usage \`\`\`sh mflux-generate-qwen-layered \\ --image input.png \\ --layers 4 \\ --steps 50 \\ -q 6 \\ --output-dir ./layers \`\`\` ## Documentation Added comprehensive documentation to README.md including: - TOC entry - CLI argument reference - Usage examples and tips - Memory requirements Tested on M4 Max 48GB with 6-bit quantization.
160 lines
5.2 KiB
TOML
160 lines
5.2 KiB
TOML
[build-system]
|
|
requires = ["uv_build>=0.7.19,<0.8.0"]
|
|
build-backend = "uv_build"
|
|
|
|
[tool.uv.build-backend]
|
|
default-excludes = true
|
|
module-name = "mflux"
|
|
namespace = true
|
|
source-exclude = [
|
|
"**/assets/**",
|
|
"**/models/flux/variants/dreambooth/_example/images/**",
|
|
"**/optimization/_loss_derivation/**",
|
|
]
|
|
|
|
[project]
|
|
name = "mflux"
|
|
version = "0.13.3"
|
|
description = "A MLX port of FLUX based on the Huggingface Diffusers implementation."
|
|
readme = "README.md"
|
|
keywords = ["diffusers", "flux", "mlx"]
|
|
authors = [{ name = "Filip Strand", email = "strand.filip@gmail.com" }]
|
|
maintainers = [{ name = "Filip Strand", email = "strand.filip@gmail.com" }]
|
|
license = { file = "LICENSE" }
|
|
requires-python = ">=3.10"
|
|
dependencies = [
|
|
"huggingface-hub>=1.1.6,<2.0",
|
|
"matplotlib>=3.9.2,<4.0",
|
|
"mlx>=0.27.0,<0.31.0",
|
|
"numpy>=2.0.1,<3.0",
|
|
"opencv-python>=4.10.0,<5.0",
|
|
"piexif>=1.1.3,<2.0",
|
|
"pillow>=10.4.0",
|
|
"pillow>=10.4.0,<11.0; python_version<'3.13'",
|
|
"pillow>=11.0,<12.0; python_version>='3.13'",
|
|
"platformdirs>=4.0,<5.0",
|
|
"regex>=2024.11.6",
|
|
"requests>=2.32.4",
|
|
"safetensors>=0.4.4,<1.0",
|
|
"sentencepiece>=0.2.0,<1.0; python_version<'3.13'",
|
|
"sentencepiece>=0.2.1,<1.0; python_version>='3.13'",
|
|
"tokenizers>=0.20.3; python_version>='3.13'",
|
|
"toml>=0.10.2,<1.0",
|
|
"torch>=2.7.1",
|
|
"torch>=2.3.1,<3.0; python_version<'3.13'",
|
|
"torch>=2.8.0,<3.0; python_version>='3.13'",
|
|
"tqdm>=4.66.5,<5.0",
|
|
"transformers>=5.0.0rc0,<6.0",
|
|
"twine>=6.1.0,<7.0",
|
|
]
|
|
classifiers = [
|
|
"Intended Audience :: Developers",
|
|
"Operating System :: MacOS",
|
|
"Programming Language :: Python :: 3",
|
|
"Programming Language :: Python :: 3.10",
|
|
"Programming Language :: Python :: 3.11",
|
|
"Programming Language :: Python :: 3.12",
|
|
"Programming Language :: Python :: 3.13",
|
|
]
|
|
|
|
[project.optional-dependencies]
|
|
dev = [
|
|
"matplotlib>3.10,<4.0",
|
|
"pytest>=8.3.0,<9.0",
|
|
"pytest-timer>=1.0,<2.0",
|
|
"mlx==0.29.2",
|
|
]
|
|
|
|
[project.urls]
|
|
homepage = "https://github.com/filipstrand/mflux"
|
|
|
|
[project.scripts]
|
|
mflux-generate = "mflux.models.flux.cli.flux_generate:main"
|
|
mflux-generate-controlnet = "mflux.models.flux.cli.flux_generate_controlnet:main"
|
|
mflux-generate-in-context = "mflux.models.flux.cli.flux_generate_in_context_dev:main"
|
|
mflux-generate-in-context-edit = "mflux.models.flux.cli.flux_generate_in_context_edit:main"
|
|
mflux-generate-in-context-catvton = "mflux.models.flux.cli.flux_generate_in_context_catvton:main"
|
|
mflux-generate-fill = "mflux.models.flux.cli.flux_generate_fill:main"
|
|
mflux-generate-depth = "mflux.models.flux.cli.flux_generate_depth:main"
|
|
mflux-generate-redux = "mflux.models.flux.cli.flux_generate_redux:main"
|
|
mflux-generate-kontext = "mflux.models.flux.cli.flux_generate_kontext:main"
|
|
mflux-generate-qwen = "mflux.models.qwen.cli.qwen_image_generate:main"
|
|
mflux-generate-qwen-edit = "mflux.models.qwen.cli.qwen_image_edit_generate:main"
|
|
mflux-generate-qwen-layered = "mflux.models.qwen_layered.cli.qwen_image_layered_generate:main"
|
|
mflux-generate-fibo = "mflux.models.fibo.cli.fibo_generate:main"
|
|
mflux-generate-z-image-turbo = "mflux.models.z_image.cli.z_image_turbo_generate:main"
|
|
mflux-refine-fibo = "mflux.models.fibo_vlm.cli.fibo_refine:main"
|
|
mflux-inspire-fibo = "mflux.models.fibo_vlm.cli.fibo_inspire:main"
|
|
mflux-concept = "mflux.models.flux.cli.flux_concept:main"
|
|
mflux-concept-from-image = "mflux.models.flux.cli.flux_concept_from_image:main"
|
|
mflux-save = "mflux.models.common.cli.save:main"
|
|
mflux-save-depth = "mflux.models.depth_pro.cli.save_depth:main"
|
|
mflux-train = "mflux.models.common.cli.train:main"
|
|
mflux-upscale = "mflux.models.flux.cli.flux_upscale:main"
|
|
mflux-lora-library = "mflux.models.common.cli.lora_library:main"
|
|
mflux-info = "mflux.models.common.cli.info:main"
|
|
mflux-completions = "mflux.cli.completions.install:main"
|
|
|
|
|
|
[tool.ruff]
|
|
line-length = 120
|
|
indent-width = 4
|
|
target-version = "py310"
|
|
respect-gitignore = true
|
|
|
|
[tool.ruff.lint]
|
|
select = ["BLE", "E4", "E7", "E9", "F", "I", "ICN", "LOG", "PERF", "W"]
|
|
ignore = []
|
|
fixable = ["ALL"]
|
|
unfixable = []
|
|
dummy-variable-rgx = "^(_+|(_+[a-zA-Z0-9_]*[a-zA-Z0-9]+?))$"
|
|
|
|
[tool.ruff.format]
|
|
quote-style = "double"
|
|
indent-style = "space"
|
|
skip-magic-trailing-comma = false
|
|
line-ending = "auto"
|
|
docstring-code-format = false
|
|
docstring-code-line-length = "dynamic"
|
|
|
|
[tool.pytest.ini_options]
|
|
testpaths = ["tests"]
|
|
python_files = "test_*.py"
|
|
addopts = '-v --failed-first --showlocals --tb=long --full-trace -m "not high_memory_requirement"'
|
|
markers = [
|
|
"high_memory_requirement: marks tests that require high memory (deselect with '-m \"not high_memory_requirement\"')",
|
|
]
|
|
|
|
[tool.ruff.lint.isort]
|
|
case-sensitive = false
|
|
combine-as-imports = true
|
|
force-wrap-aliases = true
|
|
from-first = false
|
|
known-first-party = ["src"]
|
|
relative-imports-order = "furthest-to-closest"
|
|
section-order = [
|
|
"future",
|
|
"standard-library",
|
|
"third-party",
|
|
"first-party",
|
|
"local-folder",
|
|
]
|
|
|
|
[tool.mypy]
|
|
disable_error_code = [
|
|
"annotation-unchecked",
|
|
"arg-type",
|
|
"assignment",
|
|
"attr-defined",
|
|
"import-untyped",
|
|
"index",
|
|
"union-attr",
|
|
"return",
|
|
"return-value",
|
|
"var-annotated",
|
|
]
|
|
error_summary = true
|
|
ignore_missing_imports = true
|
|
implicit_optional = true
|
|
python_version = "3.12"
|