-
Notifications
You must be signed in to change notification settings - Fork 657
Updated 1B Configs #866
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
Updated 1B Configs #866
Conversation
eval_interval: 1000 | ||
eval_subset_num_batches: -1 | ||
device_eval_batch_size: ${device_train_microbatch_size} | ||
device_eval_batch_size: 8 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why is this needed?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Just matching all the overwritten comments from this bash script.
README.md
Outdated
torchrun --nproc_per_node=8 scripts/train.py {path_to_train_config} | ||
torchrun --nproc_per_node=8 scripts/train.py {path_to_train_config} --force_save_unsharded=True | ||
``` | ||
> Please use `--force_save_unsharded=True` argument to save checkpoints in unsharded format. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think setting save_num_checkpoints_to_keep = 0
achieves a similar end result and also doesn't save sharded checkpoints through training.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Updated all the configs.
Updated 1B configs wrt to peteish1-anneal.sh to avoid confusion for users.