Files
zhuangang/topics/ai/llm-serving/index.md
T

45 lines
1.5 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
---
id: ai.llm-serving.index
title: 模型侧基础(了解级)
type: index
domain: ai
stack: [llm-serving]
tags: [多模型路由, vLLM, LoRA, 微调, 模型网关]
week: [10, 13]
---
# 模型侧基础(了解级)
BOSS直聘 2026-10-23 的调研中约 7 个岗位提到 LoRA / 微调,也有岗位要求多模型接入和 vLLM 部署。本专题定位**了解级**:会多模型接入和降级(这是应用开发每天要做的),能讲清楚 vLLM 和 LoRA 的原理与取舍,不展开模型训练,避免偏离主线。
## 学习路线
| 顺序 | 课程 | 级别 | 预计 |
| --- | --- | --- | --- |
| 1 | [模型侧基础:多模型接入和路由、vLLM 部署、LoRA / QLoRA 概念](01-routing-vllm-lora.md) | 进阶 | 3 小时 |
## 企业场景
- [公司级大模型网关:多厂商路由、配额、敏感数据走内网模型](scenarios/multi-provider-gateway.md)
## 面试题
- [模型侧面试题(了解级)](interview.md)
## 示例代码
- [examples/model-router](examples/model-router/):按任务路由、限流退避重试、超时切换、熔断、备用模型降级、成本台账(只用标准库)
## 在 16 周计划中
- [第 10 周](../../../plan/week10-agent-m1-tools-mcp.md)、[第 13 周](../../../plan/week13-agent-m4-eval-lowcode.md)
## 官方资料
以官方文档为准:
- vLLM:https://docs.vllm.ai/
- LoRA 论文:https://arxiv.org/abs/2106.09685 ;QLoRA 论文:https://arxiv.org/abs/2305.14314
- Hugging Face PEFT:https://huggingface.co/docs/peft
- LiteLLM:https://docs.litellm.ai/