Skip to content

Fuji with Orbax Emergency Checkpointer #1210

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 29 commits into
base: main
Choose a base branch
from

Conversation

samos123
Copy link
Contributor

@samos123 samos123 commented May 23, 2025

do not merge, just to align on what to test.

Command to test on 2 x v6e-16:

axlearn gcp launch run --cluster=$CLUSTER \
        --runner_name gke_tpu_single \
        --name=$USER-2 \
        --instance_type=tpu-v6e-16 \
        --host_mount_spec=name=tmp,host_path=/tmp,mount_path=/host-tmp \
        --num_replicas=2 \
        --bundler_spec=allow_dirty=True \
        --bundler_type=artifactregistry --bundler_spec=image=tpu \
        --bundler_spec=dockerfile=Dockerfile --bundler_spec=target=tpu \
        -- python3 -m axlearn.common.launch_trainer_main \
          --init_module=axlearn.common.checkpointer_orbax_emergency:local_ckpt_dir=/host-tmp/checkpoints \
          --module=text.gpt.c4_trainer --config=fuji-7B-v2-flash-orbaxem \
          --trainer_dir=gs://$PROJECT_ID-axlearn/$USER-v6e-7b-orbax-2/ \
          --data_dir=gs://axlearn-public/tensorflow_datasets  \
          --jax_backend=tpu \
          --mesh_selector=tpu-v6e-16 \
          --trace_at_steps=3

70b on 4 x v6e-256:

axlearn gcp launch run --cluster=$CLUSTER \
        --runner_name gke_tpu_single \
        --name=$USER \
        --instance_type=tpu-v6e-256 \
        --priority_class=very-high \
        --host_mount_spec=name=tmp,host_path=/tmp,mount_path=/host-tmp \
        --num_replicas=4 \
        --bundler_spec=allow_dirty=True \
        --bundler_type=artifactregistry --bundler_spec=image=tpu \
        --bundler_spec=dockerfile=Dockerfile --bundler_spec=target=tpu \
        -- python3 -m axlearn.common.launch_trainer_main \
          --init_module=axlearn.common.checkpointer_orbax_emergency:local_ckpt_dir=/host-tmp/checkpoints \
          --module=text.gpt.c4_trainer \
          --config=fuji-70B-v2-flash-orbaxem \
          --trainer_dir=gs://tess-checkpoints-us-west1/stoelinga-axlearn-v6e-4k-orbax-1/ \
          --data_dir=gs://axlearn-public/tensorflow_datasets  \
          --jax_backend=tpu \
          --mesh_selector=tpu-v6e-256-4 \
          --trace_at_steps=3

@samos123 samos123 changed the title Fuji Config with Orbax Emergency Checkpointer Fuji with Orbax Emergency Checkpointer May 23, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants