pyproject.toml

[project]
name = "open-instruct"
version = "0.1.0"
description = "Train open, instruction-following language models"
readme = "README.md"
requires-python = "==3.12.*"
dependencies = [
    "accelerate>=1.10.1",
    "antlr4-python3-runtime==4.11",
    "backoff>=2.2.1",
    "bitsandbytes>=0.44.1; platform_system != 'Darwin'",
    "datasets>=4.0.0",
    "debugpy>=1.8.13",
    "deepspeed>=0.18.3",
    "hf-transfer>=0.1.8",
    "litellm>=1.72.0,<1.75.2", # avoid needing backoff https://github.com/BerriAI/litellm/issues/13827
    "matplotlib>=3.9.3",
    "multiprocess<0.70.18", # avoid AttributeError in ResourceTracker.__del__
    "nltk>=3.9.1",
    "numpy>=2",
    "ai2-olmo-core",
    "nvitop>=1.4.2",
    "packaging>=24.2",
    "peft>=0.13.2",
    "pyarrow>=20.0.0",
    "pyyaml>=6.0.2",
    "ray[default]>=2.49.2",
    "setuptools>=75.6.0,<80.0.0",
    "tensorboard>=2.18.0",
    "torch>=2.10.0",
    "transformers>=5.4.0",
    "vllm>=0.19.0; platform_system != 'Darwin'",
    "wandb==0.23.1",
    "langdetect==1.0.9",
    "immutabledict==1.2.0",
    "flash-attn>=2.8.3; platform_system != 'Darwin' and platform_machine != 'aarch64'",
    # The flat FA3 wheel index can publish a newer aarch64-only build than the latest x86_64 build.
    # Pin the Linux x86_64 wheel directly so uv doesn't resolve to an aarch64-only version.
    "flash-attn-3 @ https://github.com/windreamer/flash-attention3-wheels/releases/download/2026.03.19-850211f/flash_attn_3-3.0.0%2B20260318.cu128torch2100cxx11abitrue.8afc61-cp39-abi3-linux_x86_64.whl ; platform_system == 'Linux' and platform_machine == 'x86_64'",
    "flash-attn-4 @ https://github.com/Dao-AILab/flash-attention/releases/download/fa4-v4.0.0.beta6/flash_attn_4-4.0.0b5-py3-none-any.whl ; platform_system != 'Darwin'",
    "liger-kernel>=0.7.0; platform_system != 'Darwin'",
    "fastapi>=0.100.0",
    "uvicorn>=0.20.0",
    "mcp>=1.9.0",
    "openenv-core>=0.2.1",
    "docker>=7.0.0",
    "flash-linear-attention>=0.4.2",
]

[build-system]
requires = ["setuptools"]
build-backend = "setuptools.build_meta"

[tool.setuptools]
py-modules = ["open_instruct"]

[tool.uv.extra-build-dependencies]
flash-attn = [{ requirement = "torch", match-runtime = true }]
flash-attn-3 = [{ requirement = "torch", match-runtime = true }]

[tool.uv.extra-build-variables]
flash-attn = { FLASH_ATTENTION_SKIP_CUDA_BUILD = "TRUE" }
flash-attn-3 = { FLASH_ATTENTION_SKIP_CUDA_BUILD = "TRUE" }

# pytorch related setups
[tool.uv.sources]
ai2-olmo-core = { git = "https://github.com/allenai/OLMo-core.git", rev = "002958a8f15a" }
flash-attn = [
  { url = "https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/download/v0.9.0/flash_attn-2.8.3+cu128torch2.10-cp312-cp312-linux_x86_64.whl", marker = "platform_system == 'Linux' and platform_machine == 'x86_64'" },
]

torch = [
  { index = "pytorch-cu128", marker = "platform_system == 'Linux' and platform_machine != 'aarch64'"},
  { index = "pytorch-cu130", marker = "platform_system == 'Linux' and platform_machine == 'aarch64'"},
]
[[tool.uv.index]]
name = "pytorch-cu128"
url = "https://download.pytorch.org/whl/cu128"
explicit = true

[[tool.uv.index]]
name = "pytorch-cu130"
url = "https://download.pytorch.org/whl/cu130"
explicit = true

[[tool.uv.index]]
name = "pytorch-cpu"
url = "https://download.pytorch.org/whl/cpu"
explicit = true

[project.optional-dependencies]
code = [
    "pydantic>=2.0.0",
    "requests>=2.28.0",
]
dr-tulu = [
    "dr_agent",
    "authlib",
    "scipy",
]

[tool.uv]
preview = true
python-preference = "only-managed"
link-mode = "hardlink"
override-dependencies = [
    "transformers>=5.4.0",
]
environments = [
    "sys_platform == 'linux' and platform_machine == 'x86_64'",
    "sys_platform == 'linux' and platform_machine == 'aarch64'",
    "sys_platform == 'darwin'",
]

[dependency-groups]
dev = [
    "beaker-py>=2.5.7",
    "mkdocs-material>=9.6.8",
    "pytest>=8.3.4",
    "ruff>=0.11.13",
    "ty>=0.0.1a13",
    "parameterized>=0.9.0",
    "rich>=13.7.0",
    "responses>=0.25.8",
    "pre-commit>=4.5.0",
]

[tool.pytest.ini_options]
addopts = "--ignore=oe-eval-internal/"
filterwarnings = [
    "ignore:.*sentry_sdk.Hub.*:DeprecationWarning",
    "ignore:builtin type SwigPy.*:DeprecationWarning",
]


[tool.black]
line-length = 119
target-version = ['py310']

[tool.ruff]
target-version = "py310"
line-length = 119
src = ["open_instruct"]
exclude = ["wandb"]

[tool.ruff.format]
# Use black-compatible formatting
quote-style = "double"
indent-style = "space"
line-ending = "auto"
# Preserve original parentheses style to minimize changes
skip-magic-trailing-comma = true
# Enable black-compatible string normalization
docstring-code-format = false

[tool.ruff.lint]
# Enable rules equivalent to previous autoflake and flake8 configs:
# F = Pyflakes (covers autoflake functionality)
# E = pycodestyle errors (flake8)
# W = pycodestyle warnings (flake8)
# I = isort
# SIM = flake8-simplify
# UP = pyupgrade
# B = flake8-bugbear
select = ["F", "E", "W", "I", "SIM", "UP", "B", "PLC0415"]
ignore = [
    "B905", # zip() without explicit strict= parameter
    "C408", # dict() calls (stylistic)
    "C901", # function complexity
    "E501", # Line too long (handled by line-length setting)
]

[tool.ruff.lint.isort]
known-first-party = ["open_instruct", "mason"]
# case insensitive to match isort --profile black
case-sensitive = false
# Disable split-on-trailing-comma to work with skip-magic-trailing-comma
split-on-trailing-comma = false

# Ignore unresolved imports because some packages (e.g. vllm) are not available on macOS
# Ignore possibly-missing-attribute for torch.distributed which has incomplete type stubs
[tool.ty.rules]
unresolved-import = "ignore"
possibly-missing-attribute = "ignore"

[tool.ty.src]
include = [
    "mason.py",
    "open_instruct/actor_manager.py",
    "open_instruct/benchmark_generators.py",
    "open_instruct/code_utils/api.py",
    "open_instruct/context_window_checker.py",
    "open_instruct/dataset_processor.py",
    "open_instruct/dataset_transformation.py",
    "open_instruct/data_loader.py",
    "open_instruct/data_types.py",
    "open_instruct/dpo_config.py",
    "open_instruct/dpo_utils.py",
    "open_instruct/grpo.py",
    "open_instruct/grpo_callbacks.py",
    "open_instruct/grpo_olmo_core_actor.py",
    "open_instruct/grpo_utils.py",
    "open_instruct/launch_utils.py",
    "open_instruct/logger_utils.py",
    "open_instruct/mix_data.py",
    "open_instruct/mix_data_preferences.py",
    "open_instruct/olmo_core_train_modules.py",
    "open_instruct/padding_free_collator.py",
    "open_instruct/rejection_sampling/prompt_templates.py",
    "open_instruct/rl_utils.py",
    "open_instruct/distillkit/compression/bitpack.py",
    "open_instruct/distillkit/compression/compressor.py",
    "open_instruct/distillkit/compression/config.py",
    "open_instruct/distillkit/compression/monotonic_logprobs.py",
    "open_instruct/distillkit/sample_common.py",
    "open_instruct/rubrics/evolving_rubric_step.py",
    "open_instruct/rubrics/metrics.py",
    "open_instruct/rubrics/prompts.py",
    "open_instruct/rubrics/rubric_utils.py",
    "open_instruct/rubrics/run_utils.py",
    "open_instruct/environments/base.py",
    "open_instruct/environments/examples.py",
    "open_instruct/environments/tools/generic_mcp.py",
    "open_instruct/environments/tools/parsers.py",
    "open_instruct/environments/tools/tools.py",
    "open_instruct/environments/tools/utils.py",
]
exclude = ["scripts/", "decontamination/", "human_eval/"]