|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/llm-fine-tune/
课程评论：没有评论
课程名称：LLM - 使用自定义数据进行微调 课程概述： 欢迎来到“LLM - 使用自定义数据进行微调”课程！如果你热衷于提升你的机器学习技能，这门课程专为你量身定制。准备好启程一段学习之旅，赋予你微调语言模型的能力，利用自定义数据集解锁创新和创造的可能性。 课程内容： 1. **LLM和微调介绍**：本节将介绍课程结构和目标，探讨微调在提升语言模型中的重要性，并深入研究基础模型，为自定义打下基础。了解微调的必要性，探索各种策略，包括关键模型参数的理解以及人工智能和语言建模的基本原则与高级概念。 2. **使用GPT模型进行微调**：本节专注于实践应用，调查可用模型及其应用案例，介绍准备和格式化样本数据的基本步骤。理解令牌计数，并应对可能的警告和成本管理等问题。全面理解微调过程，区分训练数据和验证数据。学习如何将数据上传至OpenAI，创建微调作业，并确保模型的质量保证。 3. **利用Gradient平台快速微调**：Gradient AI平台是唯一支持微调、RAG开发及按需构建LLM的平台。Gradient提供了针对行业的领域特定AI，例如医疗和金融服务，旨在理解领域上下文。利用该平台上传并训练基础模型，使用自定义数据集。 4. **创建埃隆·马斯克推特生成器**：训练基础模型，以埃隆·马斯克的样本推文为基础，并使用“新微调模型”生成风格相似的推文。创建一个Streamlit应用，展示由OpenAI生成的普通推文与您自己模型生成的推文的对比。 5. **数据提取微调模型**：学习如何从原始文本中提取“有价值的信息”。了解如何使用问题与答案的样本数据集，然后处理任何原始文本以获得有价值的信息。通过真实世界的例子，例如从原始开支交易中识别个人、消费金额及商品等。 立即报名，学习如何使用自己的数据微调大型语言模型，解锁个性化应用和机器学习领域的创新潜力！
Welcome to LLM - Fine Tune with Custom Data! If you're passionate about taking your machine learning skills to the next level, this course is tailor-made for you. Get ready to embark on a learning journey that will empower you to fine-tune language models with custom datasets, unlocking a realm of possibilities for innovation and creativity.Introduction to LLM and Fine TuningIn this opening section, you'll be introduced to the course structure and objectives. We'll explore the significance of fine-tuning in enhancing language models and delve into the foundational models that set the stage for customization. Discover the reasons behind the need for fine-tuning and explore various strategies, including an understanding of critical model parameters. Gain a comprehensive understanding of the fundamental principles and advanced concepts in artificial intelligence and language modeling.Fine Tune Using GPT ModelsThis section focuses on practical applications. Survey available models and their use cases, followed by essential steps in preparing and formatting sample data. Understand token counting and navigate potential pitfalls like warnings and cost management. Gain a comprehensive understanding of the fine-tuning process, differentiating between training and validation data. Learn to upload data to OpenAI, create a fine-tune job, and ensure quality assurance for your model.Use Gradient Platform to quickly fine tuneGradient AI Platform: The only AI Agent platform that supports fine-tuning, RAG development, and purpose built LLMs out-of-the-box. Pre-tuned, Domain Expert AI i.e. Gradient offers domain-specific AI designed for your industry. From healthcare to financial services, we've built AI from the ground up to understand domain context. Use the platform to upload and train base foundations models with your own dataset.Create a Elon Musk Tweet Generator Train a foundation model with Elon Mush sample tweets, and then used the 'New Fine Tune Model' to create Elon Mush style tweets. Create a streamlit app to demonstrate side-by-side a normal tweet generated by OpenAI vs your very own model.Data Extraction fine-tune modelLearn how to extract 'valuable information' from a raw text. Learn how to pass sample datasets with question and answers, and then pass any raw text to get valuable information. Use real-world example of identifying person, amount spend and item from raw expense transactions and much more.Enroll now to learn how to fine-tune large language models with your own data, and unlock the potential of personalized applications and innovations in the world of machine learning!
