Alex Emelyanov

Уточнение по методу GET /accounts/filter/ premium

2

https://highloadcup.ru/media/condition/accounts_rules_201218_4.html Непонятно как должен работать предикат `premium_null`. Если значение `1` должно отфильтровать всех у кого есть премиум на текущую дату, то это то же самое что `premium_now`, если всех у...

Add keys evaluation with Uber gem Add :if option Do not use cache in case of cache_store_expire_time is equal or less than 0

Add to experiments sessions metadata next fields: - description - test scoring metric - user-experienced accuracy can be added via API call just set `metadata` as a key-value dictionary for...

tasks `score_model_performance_daily_task` and `predict_by_model_task` Extra parameters: * `model_ids` - list of model ids * `multipredict_filter` - ROI-syntax expression with multi-predict rule

For example import data task https://app.auger.ai/admin/cluster_tasks/295252 with a big dataset causes SIGKILL: ``` [2021-02-24 10:34:41,824: INFO/MainProcess] Received task: a2ml.tasks_queue.tasks_hub_api.import_data_task[a211b108-2684-40e6-8ee6-dd82520e46ae] [2021-02-24 10:34:43,864: WARNING/ForkPoolWorker-346] [warning] SOCKS support in urllib3 requires the installation...

We can count FP and FN for multi class problems per class

Iill-defined empirical covariance in build diverge model task

1

Sometimes build divergence model fails with error ``` Fitting the mixture model failed because some components have ill-defined empirical covariance (for instance caused by singleton or collapsed samples). Try to...

https://github.com/slundberg/shap can be applied to many model types and can give cool charts like this

Numeric types as a string https://app-staging.auger.ai/admin/cluster_tasks/30759 - failed https://app-staging.auger.ai/admin/cluster_tasks/30759 - succeeded