A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 6, 00:45 UTC
[CI/Docs] Improve aarch64/DGX Spark support for dev setup (#38057)
Signed-off-by: Ben Browning <bbrownin@redhat.com> Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Ben Browning committed
Mar 25, 2026 at 12:24 UTC
1ac2ef2e5335ca0af99aee438998c9305461f563
2 files changed
+16
-14
AGENTS.md
+15
-13
@@ -39,6 +39,8 @@ If work is duplicate/trivial busywork, **do not proceed**. Return a short explan
39
40
## 2. Development Workflow
41
42
+- **Never use system `python3` or bare `pip`/`pip install`.** All Python commands must go through `uv` and `.venv/bin/python`.
43
+
44
### Environment setup
45
46
```bash
@@ -58,33 +60,33 @@ pre-commit install
60
61
```bash
62
# If you are only making Python changes:
61
-VLLM_USE_PRECOMPILED=1 uv pip install -e .
63
+VLLM_USE_PRECOMPILED=1 uv pip install -e . --torch-backend=auto
64
65
# If you are also making C/C++ changes:
64
-uv pip install -e .
66
+uv pip install -e . --torch-backend=auto
67
```
68
69
### Running tests
70
69
-Tests require extra dependencies.
70
-All versions for test dependencies should be read from `requirements/test.txt`
71
+> Requires [Environment setup](#environment-setup) and [Installing dependencies](#installing-dependencies).
72
73
```bash
73
-# Install bare minimum test dependencies:
74
-uv pip install pytest pytest-asyncio tblib
75
-
76
-# Install additional test dependencies as needed, or install them all as follows:
74
+# Install test dependencies.
75
+# requirements/test.txt is pinned to x86_64; on other platforms, use the
76
+# unpinned source file instead:
77
+uv pip install -r requirements/test.in # resolves for current platform
78
+# Or on x86_64:
79
uv pip install -r requirements/test.txt
80
79
-# Run specific test from specific test file
80
-pytest tests/path/to/test.py -v -s -k test_name
81
-
82
-# Run all tests in directory
83
-pytest tests/path/to/dir -v -s
81
+# Run a specific test file (use .venv/bin/python directly;
82
+# `source activate` does not persist in non-interactive shells):
83
+.venv/bin/python -m pytest tests/path/to/test_file.py -v
84
```
85
86
### Running linters
87
88
+> Requires [Environment setup](#environment-setup).
89
+
90
```bash
91
# Run all pre-commit hooks on staged files:
92
pre-commit run
requirements/test.in
+1
-1
@@ -61,7 +61,7 @@ runai-model-streamer[s3,gcs,azure]==0.15.7
61
fastsafetensors>=0.2.2 # 0.2.2 contains important fixes for multi-GPU mem usage
62
instanttensor>=0.1.5
63
pydantic>=2.12 # 2.11 leads to error on python 3.13
64
-decord==0.6.0
64
+decord==0.6.0; platform_machine == "x86_64"
65
terratorch >= 1.2.2 # Required for Prithvi tests
66
imagehash # Required for Prithvi tests
67
segmentation-models-pytorch > 0.4.0 # Required for Prithvi tests