Metadata-Version: 2.4
Name: ahu-paimon-toolkit
Version: 0.2.0
Summary: Player-Aware Intelligent Monitoring and Operations Navigator - a decoupled toolkit for video understanding, VLM inference, and LLM-as-Judge evaluation.
Author-email: SiriusAhu <SiriusLLLUUU@gmail.com>
Requires-Python: >=3.12
Requires-Dist: httpx>=0.28.1
Requires-Dist: loguru>=0.7.3
Requires-Dist: matplotlib>=3.10.0
Requires-Dist: mss>=10.1.0
Requires-Dist: numpy>=2.4.2
Requires-Dist: openai>=1.0.0
Requires-Dist: opencv-python-headless>=4.13.0.92
Requires-Dist: pydantic>=2.12.5
Requires-Dist: seaborn>=0.13.0
Provides-Extra: dev
Requires-Dist: pytest-asyncio>=1.3.0; extra == 'dev'
Requires-Dist: pytest>=9.0.2; extra == 'dev'
Provides-Extra: windows
Requires-Dist: pywin32>=311; extra == 'windows'
Description-Content-Type: text/markdown

# Ahu's Paimon Toolkit

This is a toolkit for SiriusAhu's FYP, *PAIMON* (Player-Aware Intelligent Monitoring and Operations Navigator).

## `vLLM` Benchmark

This toolkit provides a series of benchmark functions for `vLLM` server.

**Note**: The `vLLM` benchmark needs the download of `QWEN3-VL-2B-Instruct` model, which could be automatically downloaded by running benchmark functions.

To speed up the download process, China users can set environment variable `HF_ENDPOINT` to `https://hf-mirror.com` before running benchmark functions.

```bash
export HF_ENDPOINT=https://hf-mirror.com
```

## Usage

```bash
python -m ahu_paimon_toolkit.benchmark_vllm
```

## License