Back to news

AI · Official

Amazon Adds Full Fine-Tuning to SageMaker AI Serverless Model Customization

Amazon AWS Machine Learning blog ·

At a glance

Amazon has added Full Fine-Tuning (FFT) to SageMaker AI Serverless Model Customization, letting users retrain AI models more fully without managing servers. The post argues FFT can beat the lighter LoRA method on complex tasks when large, high-quality datasets are available. It also reports tests showing FFT reaching 90.0 percent on a semantic correctness score for code generation with 100K+ training examples.

Key points

Why it matters

Companies can get more accurate AI models for their own tasks without buying or managing complex computing infrastructure.

Original source

Full Fine-Tuning (FFT) on SageMaker AI Serverless Model Customization

Amazon AWS Machine Learning blog · Nov 19, 09:15 KST

Read on Amazon AWS Machine Learning blog ↗

AI summarized the original article. Check the original for accuracy. Copyright belongs to Amazon AWS Machine Learning blog.

AI