- 使用 PyTorch 或 TensorFlow 等深度学习框架的经验
- 熟悉大型语言模型及其应用程式
- 了解分散式计算概念
观众
- 机器学习工程师
- 云 AI 专家
优化大型模型以进行微调对于使高级 AI 应用程式可行且具有成本效益至关重要。本课程重点介绍降低计算成本的策略，包括分散式训练、模型量化和硬体优化，使参与者能够高效地部署和微调大型模型。
这种由讲师指导的现场培训（在线或现场）面向希望掌握优化大型模型的技术，以便在实际场景中进行经济高效的微调的高级专业人员。
在本次培训结束时，参与者将能够：
- 了解微调大型模型的挑战。
- 将分散式训练技术应用于大型模型。
- 利用模型量化和修剪提高效率。
- 优化微调任务的硬体利用率。
- 在生产环境中有效地部署微调的模型。
课程形式
- 互动讲座和讨论。
- 大量的练习和练习。
- 在即时实验室环境中动手实施。
课程自定义选项
- 要申请本课程的定制培训，请联系我们进行安排。
优化大型模型简介
- 大型模型体系结构概述
- 微调大型模型的挑战
- 成本效益优化的重要性
分散式训练技术
- 数据和模型并行性简介
- 分散式训练框架：PyTorch 和 TensorFlow
- 跨多个 GPU 和节点扩展
模型量化和修剪
- 了解量化技术
- 应用修剪以减小模型大小
- 准确性和效率之间的权衡
硬体优化
- 为微调任务选择合适的硬体
- 优化 GPU 和 TPU 利用率
- 对大型模型使用专用加速器
高效 Data Management
- 管理大型数据集的策略
- 性能的预处理和批处理
- 数据增强技术
部署优化的模型
- 部署微调模型的技术
- 监控和维护模型性能
- 优化模型部署的真实示例
高级优化技术
- 探索低秩适应 （LoRA）
- 使用适配器进行模组化微调
- 模型优化的未来趋势
总结和后续步骤
United Arab Emirates - Optimizing Large Models for Cost-Effective Fine-Tuning
Qatar - Optimizing Large Models for Cost-Effective Fine-Tuning
Egypt - Optimizing Large Models for Cost-Effective Fine-Tuning
Saudi Arabia - Optimizing Large Models for Cost-Effective Fine-Tuning
South Africa - Optimizing Large Models for Cost-Effective Fine-Tuning
Brasil - Optimizing Large Models for Cost-Effective Fine-Tuning
Canada - Optimizing Large Models for Cost-Effective Fine-Tuning
中国 - Optimizing Large Models for Cost-Effective Fine-Tuning
香港 - Optimizing Large Models for Cost-Effective Fine-Tuning
澳門 - Optimizing Large Models for Cost-Effective Fine-Tuning
台灣 - Optimizing Large Models for Cost-Effective Fine-Tuning
USA - Optimizing Large Models for Cost-Effective Fine-Tuning
Österreich - Optimizing Large Models for Cost-Effective Fine-Tuning
Schweiz - Optimizing Large Models for Cost-Effective Fine-Tuning
Deutschland - Optimizing Large Models for Cost-Effective Fine-Tuning
Czech Republic - Optimizing Large Models for Cost-Effective Fine-Tuning
Denmark - Optimizing Large Models for Cost-Effective Fine-Tuning
Estonia - Optimizing Large Models for Cost-Effective Fine-Tuning
Finland - Optimizing Large Models for Cost-Effective Fine-Tuning
Greece - Optimizing Large Models for Cost-Effective Fine-Tuning
Magyarország - Optimizing Large Models for Cost-Effective Fine-Tuning
Ireland - Optimizing Large Models for Cost-Effective Fine-Tuning
Luxembourg - Optimizing Large Models for Cost-Effective Fine-Tuning
Latvia - Optimizing Large Models for Cost-Effective Fine-Tuning
España - Optimizing Large Models for Cost-Effective Fine-Tuning
Italia - Optimizing Large Models for Cost-Effective Fine-Tuning
Lithuania - Optimizing Large Models for Cost-Effective Fine-Tuning
Nederland - Optimizing Large Models for Cost-Effective Fine-Tuning
Norway - Optimizing Large Models for Cost-Effective Fine-Tuning
Portugal - Optimizing Large Models for Cost-Effective Fine-Tuning
România - Optimizing Large Models for Cost-Effective Fine-Tuning
Sverige - Optimizing Large Models for Cost-Effective Fine-Tuning
Türkiye - Optimizing Large Models for Cost-Effective Fine-Tuning
Malta - Optimizing Large Models for Cost-Effective Fine-Tuning
Belgique - Optimizing Large Models for Cost-Effective Fine-Tuning
France - Optimizing Large Models for Cost-Effective Fine-Tuning
日本 - Optimizing Large Models for Cost-Effective Fine-Tuning
Australia - Optimizing Large Models for Cost-Effective Fine-Tuning
Malaysia - Optimizing Large Models for Cost-Effective Fine-Tuning
New Zealand - Optimizing Large Models for Cost-Effective Fine-Tuning
Philippines - Optimizing Large Models for Cost-Effective Fine-Tuning
Singapore - Optimizing Large Models for Cost-Effective Fine-Tuning
Thailand - Optimizing Large Models for Cost-Effective Fine-Tuning
Vietnam - Optimizing Large Models for Cost-Effective Fine-Tuning
India - Optimizing Large Models for Cost-Effective Fine-Tuning
Argentina - Optimizing Large Models for Cost-Effective Fine-Tuning
Chile - Optimizing Large Models for Cost-Effective Fine-Tuning
Costa Rica - Optimizing Large Models for Cost-Effective Fine-Tuning
Ecuador - Optimizing Large Models for Cost-Effective Fine-Tuning
Guatemala - Optimizing Large Models for Cost-Effective Fine-Tuning
Colombia - Optimizing Large Models for Cost-Effective Fine-Tuning
México - Optimizing Large Models for Cost-Effective Fine-Tuning
Panama - Optimizing Large Models for Cost-Effective Fine-Tuning
Peru - Optimizing Large Models for Cost-Effective Fine-Tuning
Uruguay - Optimizing Large Models for Cost-Effective Fine-Tuning
Venezuela - Optimizing Large Models for Cost-Effective Fine-Tuning
Polska - Optimizing Large Models for Cost-Effective Fine-Tuning
United Kingdom - Optimizing Large Models for Cost-Effective Fine-Tuning
South Korea - Optimizing Large Models for Cost-Effective Fine-Tuning
Pakistan - Optimizing Large Models for Cost-Effective Fine-Tuning
Sri Lanka - Optimizing Large Models for Cost-Effective Fine-Tuning
Bulgaria - Optimizing Large Models for Cost-Effective Fine-Tuning
Bolivia - Optimizing Large Models for Cost-Effective Fine-Tuning
Indonesia - Optimizing Large Models for Cost-Effective Fine-Tuning
Kazakhstan - Optimizing Large Models for Cost-Effective Fine-Tuning
Moldova - Optimizing Large Models for Cost-Effective Fine-Tuning
Morocco - Optimizing Large Models for Cost-Effective Fine-Tuning
Tunisia - Optimizing Large Models for Cost-Effective Fine-Tuning
Kuwait - Optimizing Large Models for Cost-Effective Fine-Tuning
Oman - Optimizing Large Models for Cost-Effective Fine-Tuning
Slovakia - Optimizing Large Models for Cost-Effective Fine-Tuning
Kenya - Optimizing Large Models for Cost-Effective Fine-Tuning
Nigeria - Optimizing Large Models for Cost-Effective Fine-Tuning
Botswana - Optimizing Large Models for Cost-Effective Fine-Tuning
Slovenia - Optimizing Large Models for Cost-Effective Fine-Tuning
Croatia - Optimizing Large Models for Cost-Effective Fine-Tuning
Serbia - Optimizing Large Models for Cost-Effective Fine-Tuning
Bhutan - Optimizing Large Models for Cost-Effective Fine-Tuning
Nepal - Optimizing Large Models for Cost-Effective Fine-Tuning
Uzbekistan - Optimizing Large Models for Cost-Effective Fine-Tuning