g-claude

Ascend NPU hardware entry point — health check, environment setup, shell script template. Starting point for any Ascend workflow.

vLLM-Ascend serving toolchain — install, offline validation, scenario tuning, online serving, contribution guide.

Download models from ModelScope or HuggingFace to local storage before inference, quantization, or evaluation.

Model quantization on Ascend NPUs — W4A8/W8A8/W4A4, one-click and custom YAML, MoE mixed precision, VLM support, accuracy recovery.

AISBench evaluation framework — accuracy benchmarks (GSM8K, MMLU, AIME) and performance benchmarks against vLLM services.

Publish model README to GitCode with auto-inferred YAML frontmatter tags from HuggingFace/ModelScope API and local transformers config.

Review Python code for idiomatic style and suggest improvements. Flags unpythonic patterns and applies refactors.

Structured Git commits with WHAT/WHY/HOW format, optimized as AI context for future sessions.

Set up a Python project with uv — pyproject.toml, Aliyun mirror, ruff linting, pre-commit hooks.

Configure a Neovim plugin from GitHub — fetches docs via context7, writes minimal init.lua/init.vim config, updates README.md in Chinese.

MindStudio-Modeling performance evaluation framework for Ascend NPUs. Provides tensor-cast for model-level analysis and serving-cast-simulation for LLM serving parameter tuning.