菜单
Model compression techniques: pruning, knowledge distillation, and quantization. Covers compression ordering (P-KD-Q), tools, and evaluation metrics.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
基于 SOC 职业分类
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/mkurman/zorai --skill model-compression命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
Use when provisioning Prime Intellect GPU compute, managing pods/disks/sandboxes, running hosted RL training via prime lab, installing or publishing RL environments, or exposing local services via Prime Tunnel. Covers the `prime` CLI (PyPI: prime) for all Prime Intellect platform operations.
Shared Python package for Science Skills, currently containing http_client -- a unified HTTP client with rate limiting, retries, and exponential backoff. Not a standalone agent skill. Do not invoke directly.
Retrieve and analyze AlphaFold predicted structures for a protein. Use when the user provides a specific UniProt Accession ID and wants structural confidence metrics (pLDDT), domain boundary analysis, or disorder assessment. Do not use if the user only has a protein name, gene name, or amino acid sequence — ask for a UniProt ID first.
| name | model-compression |
| description | Model compression techniques: pruning, knowledge distillation, and quantization. Covers compression ordering (P-KD-Q), tools, and evaluation metrics. |
| tags | ["model-compression","pruning","distillation","quantization","llm","optimization","zorai"] |
Model compression reduces LLM size and inference cost through three main techniques: pruning (removing parameters), knowledge distillation (training smaller student models), and quantization (lower precision weights). A 2025 study on Qwen2.5 3B found the optimal ordering: Pruning → Knowledge Distillation → Quantization (P-KD-Q) achieves 3.68x compression while preserving instruction-following and language understanding.
| Technique | What it does | Compression | Quality Impact |
|---|---|---|---|
| Quantization | Lower precision (FP16→INT4/FP8) | Highest standalone | Low with PTQ, recoverable with QAT |
| Structured Pruning | Remove attention heads, layers, neurons | Moderate | Moderate degradation, recoverable with KD |
| Knowledge Distillation | Train smaller model on larger model outputs | Architecture-dependent | Best quality recovery post-pruning |
The paper shows ordering matters critically:
Applying quantization early causes irreversible information loss that impairs subsequent training.
| Toolkit | Features | Use Case |
|---|---|---|
angelslim (Tencent) | PTQ/QAT, pruning, Eagle3 speculative decoding | Full compression pipeline, 1.25-bit to FP8 |
nvidia-model-optimizer | PTQ, QAT, pruning, distillation, speculative decoding | NVIDIA ecosystem, TensorRT-LLM deployment |
intel-neural-compressor | INT8/FP8/INT4 quantization, pruning, distillation | Intel hardware, ONNX Runtime |
peft + unsloth | QLoRA fine-tuning | Training-time efficiency, adapter-based |