تشغيل أي مهارة في Manus بنقرة واحدة

ابدأ الآن

mlx-performance-audit

النجوم٠

التفرعات٠

آخر تحديث١٢ يونيو ٢٠٢٦ في ١١:٣٣

Audit Python MLX repos for lazy-eval, synchronization, compile, dtype, memory, progress, and benchmark issues.

التثبيت

التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.

تشغيل في Manus

المصدر

sealad886

sealad886/mlx-optimizer-plugin

فتح مستودع GitHub عرض مستودعات المنشئ

تنزيل

تشغيل في Manus

المهن ذات الصلةSOC

استنادا إلى تصنيف SOC المهني

مطوّرو البرمجياتمهن الحاسوب والرياضيات·SOC 15-1252

SKILL.md

readonly

name	mlx-performance-audit
description	Audit Python MLX repos for lazy-eval, synchronization, compile, dtype, memory, progress, and benchmark issues.

MLX Performance Audit

Use this skill for repo-wide MLX performance reviews and unknown bottlenecks.

Python Environment

Before any Python execution, use the target repo's .venv. Never install Python packages globally.

Required References

../../references/mlx-core-concepts.md
../../references/eval-and-synchronization.md
../../references/compile-and-transforms.md
../../references/memory-and-dtypes.md
../../references/reporting-format.md

Workflow

Inspect repo state with git --no-pager status --short.
Locate dependency files, MLX imports, training loops, inference loops, dataloaders, benchmark scripts, and progress-reporting patterns.
Run ../../scripts/mlx_audit.py from this plugin repo when the target repo is local and scanning is useful.
Keep candidate findings separate from verified findings.
Recommend measurement before changes: warmup, synchronization, repeated runs, memory telemetry, correctness checks, and representative input sizes.

Report Shape

Use ../../templates/optimization-report.md as the report structure. Every finding needs file, line, evidence, candidate impact, confidence, verification, and residual risk.

المزيد من هذا المستودع

نفس المستودع

mlx-inference-optimizer

sealad886/mlx-optimizer-plugin

Optimize Python MLX inference and generation loops with warmup, batching, cache handling, synchronization, quantization, and memory checks.

2026-06-120

mlx-metal-kernels

sealad886/mlx-optimizer-plugin

Guide MLX Metal profiling, mx.fast escalation, custom Metal kernels, and C++ extensions when profiling proves kernel-level bottlenecks.

2026-06-120

mlx-optimizer

sealad886/mlx-optimizer-plugin

Route Python-first MLX optimization work on Apple Silicon to focused audit, training, inference, Metal, or bridge workflows.

2026-06-120

mlx-portability-bridges

sealad886/mlx-optimizer-plugin

Advise on Python-first MLX integration with Swift, C, C++, and non-native language boundaries.

2026-06-120

mlx-training-optimizer

sealad886/mlx-optimizer-plugin

Optimize Python MLX training loops with value_and_grad, accumulation, checkpointing, dtype, validation cadence, memory telemetry, and progress reporting.

2026-06-120

name	mlx-performance-audit
description	Audit Python MLX repos for lazy-eval, synchronization, compile, dtype, memory, progress, and benchmark issues.

MLX Performance Audit

Use this skill for repo-wide MLX performance reviews and unknown bottlenecks.

Python Environment

Before any Python execution, use the target repo's .venv. Never install Python packages globally.

Required References

../../references/mlx-core-concepts.md
../../references/eval-and-synchronization.md
../../references/compile-and-transforms.md
../../references/memory-and-dtypes.md
../../references/reporting-format.md

Workflow

Inspect repo state with git --no-pager status --short.
Locate dependency files, MLX imports, training loops, inference loops, dataloaders, benchmark scripts, and progress-reporting patterns.
Run ../../scripts/mlx_audit.py from this plugin repo when the target repo is local and scanning is useful.
Keep candidate findings separate from verified findings.
Recommend measurement before changes: warmup, synchronization, repeated runs, memory telemetry, correctness checks, and representative input sizes.

Report Shape

Use ../../templates/optimization-report.md as the report structure. Every finding needs file, line, evidence, candidate impact, confidence, verification, and residual risk.