Skip to content

Added output_prr_curve flag#3

Open
silvimica wants to merge 4 commits into
mainfrom
experimental
Open

Added output_prr_curve flag#3
silvimica wants to merge 4 commits into
mainfrom
experimental

Conversation

@silvimica
Copy link
Copy Markdown
Owner

Based on feedback for previous pull

Comment thread scripts/polygraph_eval Outdated
os.chdir(hydra.utils.get_original_cwd())

save_path = args.save_path if "save_path" in args else save_path
output_prr_curves = getattr(args, "output_prr_curves", False)
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Лучше это прям при создании менеджера ниже по коду делать, а то большой промежуток получается между местом где объявляется это и где потом используется

if self.max_rejection == 1:
return "prr"
return f"prr_{self.max_rejection}"
return "prr"
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Вот это наверное опечатка


value = []
for i in range(num_iter):
prr_values = [] # To store PRR scores across iterations
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Посмотри мои комментарии насчет нейминга в этой функции из прошлого ПРа

Comment thread src/lm_polygraph/utils/manager.py Outdated
'cpu' otherwise. Default: None.
language (str): Language to test in claim-level benchmark, one of 'en', 'zh', 'ar', 'ru'. Default: 'en'.
verbose (bool): If set, will print useful info during batch processing. Default: True.
verbose for (e_level, e_name), estimator_values in self.estimations.items():
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Это походу лишнее

Comment thread src/lm_polygraph/utils/manager.py Outdated
s
for e in self.ensemble_estimators
for s in e.stats_dependencies
for s in e.stats_dependenciesgenerate_prr_curve
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Артефакт какой-то

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