Concepts and Guides
Release Notes

Getting Started

Overview
Setup and installation
Pre-train your first model
Fine-tune your first model
Current release highlights

Concepts

Cerebras Wafer-Scale Cluster
Weight Streaming Execution

Model Zoo

Model Zoo Overview
Trainer Overview
Trainer Configuration Overview
- Correspondance from Legacy to Trainer
Core Workflows
Tutorials
Components
Migration
API
- cerebras.modelzoo
- Trainer API

CS Torch

Cerebras PyTorch API
Writing a custom training loop

Cluster Monitoring

Cerebras job scheduling and monitoring

Fundamentals

Launch your job
- Cerebras cluster settings
Kernel autogeneration with AutoGen
Define environment variables for input workers
Import user dependencies in Cerebras
Special considerations for CV dataloaders
Measure throughput of your model
Managing cluster access controls

Support

Release Notes
Troubleshooting
Glossary

Data preprocessing scripts#

Using HF datasets for auto-regressive LM
Creating HDF5 dataset for GPT models
SlimPajama dataset pre-processing

previous

Visualization and Debugging

next

Using HF datasets for auto-regressive LM

© Copyright 2024, Cerebras Systems.

Created using Sphinx 4.4.0.