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

fix eval_harness download cache-dir #163

Merged
merged 2 commits into from
Nov 30, 2023

Conversation

pihao7
Copy link
Contributor

@pihao7 pihao7 commented Nov 27, 2023

This is to fix a issue when run evaluation stage on DGX cloud,

python3 /opt/NeMo-Megatron-Launcher/launcher_scripts/main.py \
             data_dir=/mount_workspace/data  stages=[evaluation] cluster_type=bcp \
             ...

the above script would generate bcprun command like this

bcprun --nnodes 1 --npernode 1 -no_redirect  --env 'TRANSFORMERS_OFFLINE=1' --env 'TORCH_NCCL_AVOID_RECORD_STREAMS=1' --env 'NCCL_NVLS_ENABLE=0' --cmd "
  python3 /opt/NeMo-Megatron-Launcher/launcher_scripts/nemo_launcher/collections/eval_harness/download.py \
  --tasks=all_tasks \
  --cache-dir=/mount_workspace/data/eval_harness_data"

but the eval datasets are not download to the cache-dir, because the download.py instead expected underscore version of cache_dir.

@dimapihtar dimapihtar self-requested a review November 30, 2023 23:58
Copy link
Collaborator

@dimapihtar dimapihtar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. Thank you!

@dimapihtar dimapihtar merged commit d1c88d1 into NVIDIA:master Nov 30, 2023
3 checks passed
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