Back to Artificial Intelligence & LLMs

Training LLMs: Data, Objectives & Optimization

Understanding the massive undertaking of pre-training a large language model.

5 days

Topics in this Chapter

1

Pre-training Objectives (CLM, MLM)

The self-supervised tasks used to train LLMs from scratch.

2

Data Collection & Curation

The process of assembling and cleaning massive text corpora.

3

Optimization Algorithms (AdamW)

The optimizers used to train models with billions of parameters.

4

Distributed Training Concepts

High-level overview of how training is scaled across many GPUs.

GeekDost - Roadmaps & Snippets for Developers