feat: add num_records as a parameter to evaluate method #39
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
malhotra18
requested changes
Oct 17, 2023
| @@ -27,6 +27,8 @@ def get_dataset(config: DataConfig) -> ray.data.Dataset: | |||
| data_loader_config = _get_data_loader_config(data_source, config) | |||
| data_loader = _get_data_loader(config.dataset_mime_type) | |||
| data = data_loader.load_dataset(data_loader_config).materialize() | |||
| if num_records: # pragma: no branch | |||
| data = data.random_sample(num_records/data.count(), seed=1234) | |||
There was a problem hiding this comment.
- please define this seed as a constant
- Update docstring too
| @@ -107,6 +106,7 @@ def evaluate( | |||
| dataset_config: Optional[DataConfig] = None, | |||
| prompt_template: Optional[str] = None, | |||
| save: bool = False, | |||
| num_records=100, | |||
There was a problem hiding this comment.
we decided to have algorithm specific default num_records, right? 100 is too low a number for default. Did we check with science team on this?
There was a problem hiding this comment.
I didn't have defaults from then, and didn't want to wait - I need this test.
But, now that we have it, I will update it.
| self, | ||
| model: ModelRunner, | ||
| model: ModelRunner = None, |
There was a problem hiding this comment.
why is None as a default for this needed? That is an invalid use case for this eval algo.
There was a problem hiding this comment.
I think this happened by an automatic refactor. Will update. Good catch.
|
Created this issue in Ray repository |
pinaraws
reviewed
Oct 17, 2023
| @@ -27,6 +27,8 @@ def get_dataset(config: DataConfig) -> ray.data.Dataset: | |||
| data_loader_config = _get_data_loader_config(data_source, config) | |||
| data_loader = _get_data_loader(config.dataset_mime_type) | |||
| data = data_loader.load_dataset(data_loader_config).materialize() | |||
| if num_records: # pragma: no branch | |||
keerthanvasist
force-pushed
the
ps
branch
2 times, most recently
from
63be5de
to
dbc20eb
Compare
pinaraws
reviewed
Oct 17, 2023
| """ | ||
| with timed_block(f"Loading dataset {config.dataset_name}", logger): | ||
| data_source = get_data_source(config.dataset_uri) | ||
| data_loader_config = _get_data_loader_config(data_source, config) | ||
| data_loader = _get_data_loader(config.dataset_mime_type) | ||
| data = data_loader.load_dataset(data_loader_config).materialize() | ||
| count = data.count() | ||
| util.require(count > 0, "Data has to have atleast one record") |
pinaraws
previously approved these changes
Oct 17, 2023
malhotra18
reviewed
Oct 17, 2023
| self, | ||
| model: ModelRunner, | ||
| dataset_config: Optional[DataConfig] = None, | ||
| prompt_template: Optional[str] = None, | ||
| save: bool = False, | ||
| num_records=100, |
There was a problem hiding this comment.
Note: Merged SemanticRobustness sum accuracy, could you please include this in that too?
pinaraws
previously approved these changes
Oct 17, 2023
malhotra18
previously approved these changes
Oct 17, 2023
xiaoyi-cheng
previously approved these changes
Oct 17, 2023
keerthanvasist
dismissed stale reviews from xiaoyi-cheng, malhotra18, and pinaraws
via
6dd5099
pinaraws
previously approved these changes
Oct 17, 2023
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.
Description of changes:
Add num_records as a parameter to evaluate method.
However, a concerning finding of this PR is that Ray's
random_sampleproduces non-deterministic results even after setting seed parameter.Update:
The sampling is now deterministic. We are using
Dataset.from_pandas(Dataset.to_pandas().sample())to use pandas's sampling ability to get a deterministic sampling. This is a temporary workaround till Ray's issue is fixed.By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.