deadlock and first nf-test #156
Merged
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
…hecking if model learns, made dnafloat into default test
…ult be inside each process subdir instead of home dir
…stimulus into nf-test_pipeline
… still missing gpu info
… what nextflow allocates for the single process
alessiovignoli
added
bug
This was
linked to
issues
Jun 17, 2024
| # shuffle the data | ||
| csv_obj.shuffle_labels() | ||
| # shuffle the data with a default seed. TODO get the seed for the config if and when that is going to be set there. | ||
| csv_obj.shuffle_labels(seed=42) |
There was a problem hiding this comment.
Why not add it directly to?
def main(data_csv, config_json, out_path)
```
with a default value
Co-authored-by: Jose Espinosa-Carrasco <[email protected]>
Co-authored-by: Jose Espinosa-Carrasco <[email protected]>
Co-authored-by: Jose Espinosa-Carrasco <[email protected]>
Co-authored-by: Jose Espinosa-Carrasco <[email protected]>
Co-authored-by: Jose Espinosa-Carrasco <[email protected]>
Co-authored-by: Jose Espinosa-Carrasco <[email protected]>
Co-authored-by: Jose Espinosa-Carrasco <[email protected]>
bin/launch_utils.py
Outdated
| def memory_split_for_ray_init(memory_str: Union[str, None]) -> Tuple[float, float]: | ||
| """ | ||
| compute the memory requirements for ray init. |
There was a problem hiding this comment.
Process the input memory value into the right unit and allocates 30% for overhead and 70% for tuning.
mathysgrapotte
approved these changes
Jul 15, 2024
suzannejin
approved these changes
Jul 15, 2024
Co-authored-by: mathysgrapotte <[email protected]>
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Outdated suggestions cannot be applied.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.
The changes revolve around creating the first nf-test for the workflow of data handling.
All improvements to try not to get deadlock hapenning. WHen that happens ray tune gets stuck in either PENDING or RUNNING mode without any error. To prevent that a number of changes were necessary:
if ray is initialized through tuner.fit() (that calls ray.init()) on the cluster it reads completely wrong the available resources. it sees way more than what is allocated for it. That's way ray is initialized esplicitly with a given set of values.