Equimo/pyproject.toml at main · RhizomeResearch/Equimo · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
[project]
name = "Equimo"
version = "2.7.0"
description = "Implementation of modern multimodal models in JAX/Equinox"
readme = "README.md"
requires-python = ">=3.12,<3.15"
license = "MIT AND Apache-2.0"
license-files = [
  "LICENSE.md",
  "NOTICE",
  "LICENSES/Apache-2.0.txt",
  "LICENSES/pretrained/README.md",
  "LICENSES/pretrained/ast-BSD-3-Clause.txt",
  "LICENSES/pretrained/convnext-Apache-2.0.txt",
  "LICENSES/pretrained/convnextv2-CC-BY-NC-4.0.txt",
  "LICENSES/pretrained/dinov2-Apache-2.0.txt",
  "LICENSES/pretrained/dinov3-License.md",
  "LICENSES/pretrained/eupe-FAIR-Noncommercial-Research-License.md",
  "LICENSES/pretrained/lingbot-vision-Apache-2.0.txt",
  "LICENSES/pretrained/siglip2-Apache-2.0.txt",
  "LICENSES/pretrained/t0-alpha-Apache-2.0.txt",
  "LICENSES/pretrained/tabpfn-3-License-v1.0.txt",
  "LICENSES/pretrained/tips-CC-BY-4.0.txt",
  "LICENSES/tfc-t0-APACHE-2.0.txt",
]
dependencies = [
  "banax>=0.1.2",
  "einops>=0.8.0",
  "equinox>=0.13.3",
  "jax>=0.11.1",
  "jaxlib>=0.11.1",
  "loguru>=0.7.3",
  "lz4>=4.4.3",
  "requests>=2.32.3",
]
classifiers = [
  "Development Status :: 5 - Production/Stable",

  "Intended Audience :: Developers",
  "Intended Audience :: Science/Research",

  "Topic :: Scientific/Engineering :: Artificial Intelligence",

  "Programming Language :: Python :: 3.12",
  "Programming Language :: Python :: 3.13",
  "Programming Language :: Python :: 3.14",
]

[project.optional-dependencies]
# Required for checkpoint-faithful AST waveform preprocessing
audio = [
  "torch>=2.11,<2.12",
  "torchaudio>=2.11,<2.12",
]
# Useful to convert models from PyTorch
torch = [
  "torch>=2.6.0",
  "timm>=1.0.15",
]
# Required for equimo.language tokenizers
language = [
  "tensorflow>=2.18.1",
  "tensorflow-text>=2.18.1",
]
# Required for some external utils
extras = [
  "matplotlib>=3.10.1",
  "pillow>=11.1.0",
  "scikit-learn>=1.6.1",
]

[build-system]
requires = ["uv_build>=0.10.12,<0.12.0"]
build-backend = "uv_build"

[tool.uv.build-backend]
source-include = [
  "CHANGELOG.md",
  "docs/licensing/t0-source-provenance.md",
  "docs/migration-v2.md",
  "docs/stability.md",
  "docs/timeseries.md",
  "models/huggingface/NOTICE",
  "models/huggingface/README.md",
  "models/huggingface/convnext-conversion.json",
]

[dependency-groups]
dev = [
    "jax2onnx>=0.17.0",
    "mdformat==1.0.0",
    "mdformat-footnote==0.1.3",
    "mdformat-gfm==1.0.0",
    "mdformat-myst==0.3.0",
    "onnxruntime>=1.28.0",
    "pre-commit>=4.5.1",
    "pytest>=9.0.2",
    "ruff>=0.15.7",
    "ty>=0.0.24",
]
reference = [
    "tabpfn>=8.1.0",
    "tfc-t0==0.2.1",
    "timm>=1.0.15",
    "torch>=2.6.0",
    "transformers>=5.13.1",
]

[tool.ruff.lint]
ignore = ["E731", "E741", "F722"]

[tool.ty.analysis]
# ty infers the return type of Equinox's unannotated metaclass `__call__` as the
# bare `Module` base, so every Equinox constructor call would lose its concrete
# class. Treating that base as `Any` restores the concrete module types.
replace-imports-with-any = ["equinox._module._module"]
# Optional extras are imported lazily and may be absent from the checked environment.
allowed-unresolved-imports = [
  "matplotlib.**",
  "PIL.**",
  "sklearn.**",
  "tensorflow.**",
  "tensorflow_text.**",
  "timm.**",
  "torch.**",
  "torchaudio.**",
]

[tool.pytest.ini_options]
addopts = "-m 'not reference_parity and not live_reference_parity'"
markers = [
  "reference_parity: parity against checked-in outputs from pinned upstream implementations",
  "live_reference_parity: parity requiring a separately supplied upstream checkpoint",
]