research/evals/pyproject.toml

[project]
name = "slm-evals"
version = "0.1.0"
description = "SLM agentic benchmark suite (BFCL, tau-bench, GAIA, SWE-bench)"
readme = "README.md"
requires-python = ">=3.12"
dependencies = [
    "accelerate>=0.30.0",
    "bitsandbytes>=0.43.0",
    "datasets>=2.19.0",
    "huggingface-hub>=0.22.0",
    "pandas>=2.0.0",
    "peft>=0.14.0",
    "pyyaml>=6.0",
    "torch>=2.2.0",
    "tqdm>=4.66.0",
    "transformers>=4.41.0",
]

[project.optional-dependencies]
lm-eval = [
    "lm-eval[hf,ifeval]>=0.4.9",
]

[project.scripts]
slm-benchmark = "slm_evals.run_benchmark:main"
slm-lm-eval = "slm_evals.run_lm_eval:main"

[build-system]
requires = ["uv_build>=0.8.13,<0.9.0"]
build-backend = "uv_build"