Overview

The Cerebras documentation provides a thorough overview of core workflows essential for AI model development and deployment:

The Quickstart Guide For Data Preprocessing guide offers a streamlined introduction to preparing data for training.
The Pre-training and fine-tuning workflows are detailed, enabling users to effectively train and optimize their models.
Specialized workflows for downstream tasks, such as the BigCode Eval Harness (BCEH) and Eleuther Eval Harness (EEH), are included for targeted applications.
The Pretraining with Upstream Validation tutorial includes downstream validation for your pre-training run. Additionally, the multi-phase and summary sections offer guidance on managing complex training processes and summarizing results, ensuring comprehensive support for every stage of model development.