-
Notifications
You must be signed in to change notification settings - Fork 1
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
feat: added optimizer n evolutions #126
base: dev
Are you sure you want to change the base?
Conversation
results = ctx.optimization_info.dump_evaluation_results() | ||
decision_metric = results["metrics"]["decision"] | ||
|
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.
сделать опцию выбирать на каком сплите считать метрику (потому что подгоняться под тест плохо)
dataset = load_dataset(input_path) | ||
generator = UtteranceEvolver(Generator(), evolutions, seed) | ||
generator.augment(dataset, split_name=split, n_evolutions=n) |
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.
вычислительно не эффективно, лучше на каждой итерации по одной эволюции добавлять
def _choose_search_space(search_space: str | None) -> list[dict[str, Any]] | Path | str: | ||
if search_space is None: | ||
return SEARCH_SPACE | ||
return search_space |
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.
для случая когда на входе строка тут кажется пропущена конвертация в список словариков
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.
в Pipeline.from_search_space(search_space) можно подавать search_space как файл\строку
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.
надо сконвертить весь этот функционал в класс наподобие UtteranceEvolver
, только назвать его IncrementalUtteranceEvolver
, чтобы у него был метод augment(dataset)
наподобие того который у UtteranceEvolver
потому что текущая версия ограничена в использовании - только через CLI - а хотелось бы главным образом иметь питоновский апи
No description provided.