Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 15 additions & 1 deletion pyproject.toml
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@ readme = "README.md"
description = "a lightweight vLLM implementation built from scratch"
requires-python = ">=3.10,<3.13"
dependencies = [
"torch>=2.4.0",
"torch==2.8.0",
"triton>=3.0.0",
"transformers>=4.51.0",
"flash-attn",
Expand All @@ -25,3 +25,17 @@ Homepage="https://github.com/GeeeekExplorer/nano-vllm"
[tool.setuptools.packages.find]
where = ["."]
include = ["nanovllm*"]

[tool.uv]
# flash-attn imports torch in setup.py, so it must build in the project env (not an isolated one)
no-build-isolation-package = ["flash-attn"]

# Force torch to resolve from PyTorch's cu126 index so we get a CUDA-12 wheel
# (CUDA-13 wheels of torch don't have matching flash-attn precompiled wheels).
[[tool.uv.index]]
name = "pytorch-cu126"
url = "https://download.pytorch.org/whl/cu126"
explicit = true

[tool.uv.sources]
torch = { index = "pytorch-cu126" }
Loading