How-to guides#
Explore our how-to guides for optimizing, debugging, and managing your MaxText workloads.
Techniques for maximizing performance, including sharding strategies, Pallas kernels, and benchmarking.
Configure input pipelines using Grain (recommended for determinism), HuggingFace, or TFDS.
Manage GCS checkpoints, handle preemption with emergency checkpointing, and configure multi-tier storage.
Tools for observability: goodput monitoring, hung job debugging, and Vertex AI TensorBoard integration.
Interactive development guides for running MaxText on Google Colab or local JupyterLab environments.
A step-by-step guide for the community to help expand MaxTextβs model library.
How online distillation works in MaxText: loss anatomy, Ξ± / Ξ² / temperature schedule tuning, layer indices, monitoring metrics, and troubleshooting.
Run benchmark evaluation (lm-eval, evalchemy, custom datasets) against MaxText checkpoints using the vLLM-native eval framework.