Skip to content
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

get_steps_values & steps #1

Open
wants to merge 2 commits into
base: ricliu-ray-maxtext
Choose a base branch
from
Open

Conversation

tdigangi
Copy link
Collaborator

Retrieve get_steps_values & steps from argv and some minor cleanup.

Verified with ray job submit below.

ray job submit --working-dir "." --runtime-env-json='{"excludes": [".github", "MaxText/test_assets", ".vscode", "test_assets", "getting_started", "end_to_end", "venv", ".git", "pedagogical_examples"],"includes":["assets/"]}' -- python3 MaxText/ray_trainer.py MaxText/configs/base.yml steps=10 steps_per_loop=60  base_output_directory=$OUTPUT_PATH dataset_path=$DATASET_PATH model_name=llama2-7b per_device_batch_size=2  max_target_length=8192 tokenizer_path="assets/tokenizer.llama2" reuse_example_batch=1 ici_fsdp_parallelism=-1 attention='flash' enable_checkpointing=false sa_block_q_dq=2048 sa_block_q_dkv=2048 sa_block_q=1024 profiler="xplane" use_iota_embed=true remat_policy=full gcs_metrics=false

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.

1 participant