Does it actually change my system settings?

No, it is a non-destructive skill that only reads system metadata and generates a JSON report to inform your computational strategy.

What dependencies does it require?

The script primarily requires the psutil library to be installed, which can be done via 'uv pip install psutil'.

Why should I run this before model training?

It identifies the best backend (e.g., MPS for Mac, CUDA for Linux) and ensures your system has enough VRAM and memory for the specific model.

Which GPUs does this skill support?

It supports NVIDIA (via CUDA), AMD (via ROCm), and Apple Silicon (via Metal) across Linux, Windows, and macOS.

Resource Detector & Strategy Advisor

Name: Resource Detector & Strategy Advisor
Author: Activer007

byActiver007

数据科学与机器学习

Identifies system hardware capabilities and provides strategic recommendations for optimal computational performance in scientific tasks.

关于

This skill empowers developers and data scientists to maximize hardware efficiency by automatically detecting CPU cores, GPU accelerators (NVIDIA, AMD, Apple Silicon), memory availability, and disk space. By generating a standardized resource report and strategic recommendations, it helps Claude make informed decisions about whether to employ parallel processing, out-of-core computing, or GPU acceleration before starting resource-intensive tasks like model training or large-scale data analysis. It essentially serves as a pre-flight check for any high-performance computing or data-heavy project.

主要功能

Strategic recommendations for Dask, Joblib, and Multiprocessing
Automated JSON resource reporting for project documentation
0 GitHub stars
Memory-aware data loading strategies for large datasets
CPU architecture and core optimization profiling
Cross-platform detection of NVIDIA, AMD, and Apple Silicon GPUs

使用场景

Identifying available GPU backends (CUDA or MPS) for neural network training
Configuring the optimal number of workers for parallel data processing
Determining if a dataset should be loaded in-memory or via out-of-core libraries like Dask

关于

主要功能

Strategic recommendations for Dask, Joblib, and Multiprocessing
Automated JSON resource reporting for project documentation
0 GitHub stars
Memory-aware data loading strategies for large datasets
CPU architecture and core optimization profiling
Cross-platform detection of NVIDIA, AMD, and Apple Silicon GPUs

使用场景

Identifying available GPU backends (CUDA or MPS) for neural network training
Configuring the optimal number of workers for parallel data processing
Determining if a dataset should be loaded in-memory or via out-of-core libraries like Dask