performance-optimization

优化Kaggle竞赛的训练性能。在训练速度慢、出现GPU内存问题时,或用户主动询问性能调整时主动使用。关键词:慢、优化、OOM、CUDA、batch_size、num_workers、加速、性能

📖 中文导读:优化Kaggle竞赛的训练性能。在训练速度慢、出现GPU内存问题时,或用户主动询问性能调整时主动使用。关键词:慢、优化、OOM、CUDA、batch_size、num_workers、加速、性能 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。

name: performance-optimization description: Optimize training performance for Kaggle competitions. Use PROACTIVELY when training is slow, GPU memory issues occur, or user asks about performance tuning. Keywords: 遅い, 最適化, OOM, CUDA, batch_size, num_workers, 高速化, パフォーマンス

When to Use (PROACTIVE)

This skill should be activated automatically when:

  • User reports slow training
  • CUDA Out of Memory (OOM) errors occur
  • User asks about batch_size or num_workers tuning
  • User mentions: "遅い", "最適化", "OOM", "高速化"
  • Training speed is suboptimal

What This Skill Does

Provides GPU-specific optimization strategies:

  • Batch size and num_workers recommendations by GPU
  • Mixed Precision Training implementation
  • Gradient Accumulation patterns
  • Memory optimization techniques
  • Parallel processing patterns

How to Use

Refer to gpu-tuning-guide.md for:

  • GPU-specific hyperparameter recommendations
  • Mixed precision training code
  • Gradient accumulation implementation
  • Memory profiling techniques
  • Common performance bottlenecks and solutions