Xcloud #27
Merged
Xcloud #27
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
kensens
approved these changes
Jan 27, 2023
| @@ -48,8 +48,8 @@ def main(argv) -> None: | |||
| uname = getpass.getuser() | |||
|
|
|||
| with xm_abc.create_experiment(experiment_title=_EXP_NAME.value) as experiment: | |||
| job_requirements = xm.JobRequirements(ram=26 * FLAGS.num_gpus * xm.GiB, | |||
| cpu=7 * FLAGS.num_gpus, | |||
| job_requirements = xm.JobRequirements(ram=8 * FLAGS.num_gpus * xm.GiB, | |||
There was a problem hiding this comment.
Is it needed to reduce our resource consumption here? We can request the amount we were using before: https://g3doc.corp.google.com/company/teams/brain-frameworks/xcloud/guide.md?cl=head
crossbeam/experiment/train_eval.py
Outdated
| @@ -191,7 +191,8 @@ def _gather_eval_info(rank, device, local_acc, local_num): | |||
|
|
|||
|
|
|||
| def train_eval_loop(args, device, model, train_files, eval_tasks, | |||
| task_gen, trace_gen): | |||
| task_gen, trace_gen, checkpoint): | |||
| random.shuffle(train_files) | |||
There was a problem hiding this comment.
I think we shouldn't shuffle train_files so that training runs are more consistent/reproducible
| if checkpoint is not None: | ||
| model.load_state_dict(checkpoint['model']) | ||
| optimizer.load_state_dict(checkpoint['optimizer']) | ||
| starting_step = checkpoint['step'] |
There was a problem hiding this comment.
Can we skip starting_step batches from the training dataset, so that we don't train on the same data again upon restart? I think this is just:
for _ in range(starting_step):
next(train_gen)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
stablize jobs