Skip to main content
Practice

What are the Benefits of Fine-tuning?

As we explored in the previous lesson, fine-tuning is the process of retraining a pre-trained model to specialize it for specific tasks.

The key advantage of fine-tuning is that it allows you to create an AI model that is specialized for a specific domain while incorporating up-to-date information.

So how can you practically apply a fine-tuned AI model in the workplace?


1. Creating AI Models Optimized for Specific Domains​

Fine-tuning an existing AI model with financial analysis data can generate a model more specialized in tasks like stock market predictions, risk analysis, and summarizing financial reports.

Such fine-tuned models help financial institutions make faster and more accurate decisions.


2. Improving Customer Service Chatbots with User Feedback​

By continually fine-tuning customer service chatbots using a company's customer interaction data, you can increase customer satisfaction with answers tailored to your services.

Additionally, incorporating customers' language patterns or specific demands can result in more natural conversation flows.


To incorporate rapidly changing market trends into an AI model, continuous fine-tuning is essential.

Fine-tuned models that reflect the latest information can regularly analyze market data or integrate current customer interests into your business.


4. Providing Personalized User Experiences through Recommendation Systems​

By fine-tuning an AI model based on users' service usage data, you can build a more personalized recommendation system.

For instance, by taking into account a user's viewing history, you can offer services that recommend personalized movies or TV programs.


Important Information Regarding Fine-tuning​

Due to OpenAI's policies, connecting external services to OpenAI for fine-tuning is restricted due to policy/technical issues.

This is because the training data uploaded by users for fine-tuning may contain sensitive personal and confidential information.

Moreover, OpenAI strictly limits the number of simultaneous fine-tuning jobs that can be run on external platforms to three, preventing CodeFriends from providing fine-tuning services separately.

The purpose of the comprehensive fine-tuning practice environment you'll experience is to provide the essential knowledge needed for actually performing fine-tuning on the OpenAI platform and to support the creation of JSONL datasets required for actual fine-tuning.


Practice​

In the practice environment on the right, compare the General GPT Model with the fine-tuned Grandma Model.