[模型微调] Understanding which scenarios and business use cases prompt engineering, RAG knowledge bases & plugins, and large model fine-tuning are each suitable for.

When it comes to handling data input and output, we have 3 methods: prompt engineering (prompt), RAG knowledge base & plugins, and large model fine-tuning. They have different roles. When developing an application that fits actual business needs, it is often necessary to use these processing methods in combination, and you can hardly find such answers online.

Next, I will introduce these 3 ways of processing data input and output respectively.

Prompt Engineering
It is actually the prompt: using input text to guide the large model to output content.
RAG
It is an upgraded content processing method. It is a combined form of large model + knowledge base + plugins. We can either preset answers or attach external plugins such as weather APIs and real-time news APIs. We can also use various functions to flexibly process and call data.
Large Model Fine-Tuning
Its role is to learn the specific content, tone, emotion, word style, and other aspects we provide.

Once we understand these functions, we can easily address our business logic. We can choose a single processing method, or a suitable combination of multiple methods.
If you want to learn more, welcome to sign up for our training course and become a first-generation AI master~!