SFT on multi-host TPUs

SFT on multi-host TPUs#

Supervised fine-tuning (SFT) is a process where a pre-trained large language model is fine-tuned on a labeled dataset to adapt the model to perform better on specific tasks.

This tutorial demonstrates step-by-step instructions for setting up the multi-host TPU environment and then training the model on the Hugging Face dataset using SFT. In this tutorial we use a multi-host TPU such as v6e-256.

We use Tunix, a JAX-based library designed for post-training tasks, to perform SFT.

Let’s get started!

Prerequisites#

Before starting, ensure you have:

  • Access to a Google Cloud Project with TPU quotas.

  • IAM Roles required:

    • Kubernetes Engine Developer (roles/container.developer) to submit and manage workloads on GKE.

    • Artifact Registry Writer (roles/artifactregistry.writer) to upload Docker images.

    • Storage Admin (roles/storage.admin) or Storage Object Admin (roles/storage.objectAdmin) combined with Storage Legacy Bucket Reader (roles/storage.legacyBucketReader) on your GCS bucket to read/write checkpoints and logs. (Note: A bucket-level read permission like storage.buckets.get is required by JAX/TensorStore to verify bucket existence and metadata; using roles/storage.objectAdmin alone will cause a misleading “bucket not found” error).

  • A Hugging Face account with an access token for downloading models.

  • Cluster Toolkit installed and configured. Follow Running MaxText with Cluster Toolkit for gcluster setup.

    • Important: Modern GKE clusters require the GKE auth plugin. If you encounter gke-gcloud-auth-plugin not found when running kubectl commands, you must install it locally (e.g., sudo apt-get install google-cloud-cli-gke-gcloud-auth-plugin for apt installations, or gcloud components install gke-gcloud-auth-plugin for standalone archive installations).

  • A GKE cluster configured for Cluster Toolkit, including healthy Kueue and JobSet components (see Cluster Toolkit documentation, or create a GKE cluster with Pathways when running with Pathways).

  • Docker installed and configured for sudoless use. Follow the steps to configure sudoless Docker.

Build and upload MaxText Docker image#

For instructions on building and uploading the MaxText Docker image with post-training dependencies, please refer to the official documentation.

Environment configuration#

Set up the following environment variables to configure your training run. Replace placeholders with your actual values.

# -- Model configuration --
# The MaxText model name. See `src/maxtext/configs/types.py` for `ModelName` for a
# full list of supported models.
export MODEL=<MODEL_NAME> # e.g., deepseek3-671b

# Your Hugging Face access token. Required to download gated models like Llama.
# You can generate one at https://huggingface.co/settings/tokens.
export HF_TOKEN=<HF_TOKEN>

# -- MaxText configuration --
# Use a GCS bucket you own to store logs and checkpoints. Ideally in the same
# region as your TPUs to minimize latency and costs.
# You can list your buckets and their locations in the
# [Cloud Console](https://console.cloud.google.com/storage/browser) or via
# `gcloud storage buckets list --format="table(name, location)"`.
export BASE_OUTPUT_DIRECTORY=<GCS_BUCKET> # e.g., gs://my-bucket/maxtext-runs

# An arbitrary string to identify this specific run.
# We recommend to include the model, user, and timestamp.
# Note: Workload names cannot exceed 28 characters (or 22 characters when using Pathways due to Kubernetes 63-byte coordinator label limits) and must be valid DNS labels (lowercase alphanumeric and hyphens).
export RUN_NAME=<RUN_NAME>

# -- Workload configuration --
# Your GCP project ID. Find it on the [Cloud Console Dashboard](https://console.cloud.google.com/home/dashboard).
# If you've already set it in your local config, you can retrieve it via:
# gcloud config get-value project
export PROJECT_ID=<PROJECT_ID>

# The GCP location (region or zone) and name of your
# TPU-enabled GKE cluster. Both can be found on the
# [Cloud Console](https://console.cloud.google.com/kubernetes/list).
export LOCATION=<LOCATION> # e.g., 'europe-west4' (region) or 'us-central1-a' (zone)
export GKE_CLUSTER=<CLUSTER_NAME>

# Number of TPU slices for your workload.
export NUM_SLICES=<NUM_SLICES>

# Cluster Toolkit workload placement. Specify the compute type (machine type)
# and topology matching your TPU node pool.
# For example:
#   - v5p-128: COMPUTE_TYPE='ct5p-hightpu-4t', TOPOLOGY='4x4x4'
#   - v6e-256: COMPUTE_TYPE='ct6e-hightpu-4t', TOPOLOGY='16x16'
# To inspect the accelerator and topology labels on your GKE cluster nodes:
# kubectl get nodes -l cloud.google.com/gke-tpu-accelerator -o custom-columns=NAME:.metadata.name,ACCELERATOR:.metadata.labels.cloud\\.google\\.com/gke-tpu-accelerator,TOPOLOGY:.metadata.labels.cloud\\.google\\.com/gke-tpu-topology
export COMPUTE_TYPE=<COMPUTE_TYPE>
export TOPOLOGY=<TOPOLOGY>

# The Docker image you pushed in the previous step
export DOCKER_IMAGE=<IMAGE_NAME>

# -- Fine-Tuning configuration --
export STEPS=<STEPS> # e.g., 1000

# -- Dataset configuration --
export DATASET_NAME=<DATASET_NAME> # e.g., HuggingFaceH4/ultrachat_200k
export TRAIN_SPLIT=<TRAIN_SPLIT> # e.g., train_sft
export TRAIN_DATA_COLUMNS=<DATA_COLUMNS> # e.g., ['messages']

Configure GKE cluster#

Configure access to the target cluster with gcloud, then configure the project, cluster, and location with gcluster as described in Running MaxText with Cluster Toolkit:

gcloud container clusters get-credentials ${GKE_CLUSTER?} \
  --location ${LOCATION?} \
  --project ${PROJECT_ID?}
gcluster job config set project ${PROJECT_ID?}
gcluster job config set cluster ${GKE_CLUSTER?}
gcluster job config set location ${LOCATION?}

Get MaxText model checkpoint#

This section explains how to prepare your model checkpoint for use with MaxText. You have two options: using an existing MaxText checkpoint or converting a Hugging Face checkpoint.

Option 1: Using an existing MaxText checkpoint#

If you already have a MaxText-compatible model checkpoint, simply set the following environment variable and move on to the next section.

export MAXTEXT_CKPT_PATH=<CKPT_PATH> # e.g., gs://my-bucket/my-model-checkpoint/0/items

Note: Make sure that MAXTEXT_CKPT_PATH has the checkpoints created using the correct storage flags:

export USE_PATHWAYS=0  # Set to 1 for Pathways, 0 for McJAX.
checkpoint_storage_use_zarr3=$((1 - USE_PATHWAYS))
checkpoint_storage_use_ocdbt=$((1 - USE_PATHWAYS))

Option 2: Converting a Hugging Face checkpoint#

Refer the steps in Hugging Face to MaxText to convert a hugging face checkpoint to MaxText. Make sure you have correct checkpoint files converted and saved. Similar as Option 1, you can set the following environment and move on.

export MAXTEXT_CKPT_PATH=<CKPT_PATH> # gs://my-bucket/my-checkpoint-directory/0/items

[!IMPORTANT] Automatic scan_layers Resolution: MaxText automatically loads scan_layers from the checkpoint’s saved metadata when resuming (via load_parameters_path) if you do not explicitly specify it on the command-line.

  • You do not need to manually supply scan_layers=False (or scan_layers=True) when loading checkpoints; MaxText will configure this automatically.

  • If you do explicitly provide a scan_layers argument, it must match the checkpoint’s saved setting or a ValueError mismatch error will be raised. See the Checkpoints concept guide for more details.

Submit workload on GKE cluster#

This section provides the command to run SFT on a GKE cluster.

SFT with Multi-Controller JAX (McJAX)#

gcluster job submit \
  --image=${DOCKER_IMAGE?} \
  --name=${RUN_NAME?} \
  --compute-type=${COMPUTE_TYPE?} \
  --topology=${TOPOLOGY?} \
  --num-slices=${NUM_SLICES?} \
  --command="python3 -m maxtext.trainers.post_train.sft.train_sft \
    run_name=${RUN_NAME?} \
    base_output_directory=${BASE_OUTPUT_DIRECTORY?} \
    model_name=${MODEL?} \
    load_parameters_path=${MAXTEXT_CKPT_PATH?} \
    hf_access_token=${HF_TOKEN?} \
    per_device_batch_size=1 \
    steps=${STEPS?} \
    profiler=xplane \
    hf_path=${DATASET_NAME?} \
    train_split=${TRAIN_SPLIT?} \
    train_data_columns=${TRAIN_DATA_COLUMNS?}"

Once the fine-tuning is completed, you can access your model checkpoints at ${BASE_OUTPUT_DIRECTORY}/${RUN_NAME}/checkpoints.

SFT with Pathways#

To submit an SFT workload with Pathways using Cluster Toolkit, use gcluster job submit with the --pathways flag:

export USE_PATHWAYS=1

gcluster job submit \
  --image=${DOCKER_IMAGE?} \
  --name=${RUN_NAME?} \
  --pathways \
  --compute-type=${COMPUTE_TYPE?} \
  --topology=${TOPOLOGY?} \
  --num-slices=${NUM_SLICES?} \
  --pathways-gcs-location=${BASE_OUTPUT_DIRECTORY?} \
  --command="ENABLE_PATHWAYS_PERSISTENCE=1 \
python3 -m maxtext.trainers.post_train.sft.train_sft \
    run_name=${RUN_NAME?} \
    base_output_directory=${BASE_OUTPUT_DIRECTORY?} \
    model_name=${MODEL?} \
    load_parameters_path=${MAXTEXT_CKPT_PATH?} \
    hf_access_token=${HF_TOKEN?} \
    per_device_batch_size=1 \
    steps=${STEPS?} \
    profiler=xplane \
    checkpoint_storage_use_zarr3=$((1 - USE_PATHWAYS)) \
    checkpoint_storage_use_ocdbt=$((1 - USE_PATHWAYS)) \
    enable_single_controller=True"

Once the fine-tuning is completed, you can access your model checkpoints at ${BASE_OUTPUT_DIRECTORY}/${RUN_NAME}/checkpoints.

Monitor and clean up#

gcluster job list
# Note: For Pathways workloads (> 5 pods), specify --main-only=false to retrieve logs from all pods:
gcluster job logs ${RUN_NAME?} --main-only=false
gcluster job cancel ${RUN_NAME?}

You can also inspect the Kubernetes resources directly:

kubectl get jobset -l gcluster.google.com/workload=${RUN_NAME?}
# In Pathways workloads, use the jobset-name label to select all pods (both pathways-head and worker pods):
kubectl get pods -l jobset.sigs.k8s.io/jobset-name=${RUN_NAME?}