Learn Before
Concept

GPT (Generative Pre-Training)

The GPT (Generative Pre-Training) model utilizes a Transformer decoder as its foundational architecture. It is trained via an autoregressive language modeling objective and consists of 100100 million parameters. Unlike subsequent models, GPT typically requires task-specific fine-tuning to perform effectively on individual downstream tasks.

0

1

Updated 2026-05-15

Tags

Data Science

Ch.2 Generative Models - Foundations of Large Language Models

Foundations of Large Language Models

Foundations of Large Language Models Course

Computing Sciences

D2L

Dive into Deep Learning @ D2L