Rinchin

### Motivation For example, black covers imports sorting, so we don't (may be) need isort darglint is slow ### Proposal Remove redundant python linters

Blending uses weights from last iteration, not best

1

## 🐛 Bug Log from tutorial 1: > [10:39:41] Layer 1 training completed. > > [10:39:41] Blending: optimization starts with equal weights and score 0.7529907972036357 > [10:39:41] Blending: iteration 0:...

5-fold CV causes different number of classes in train/test

2

## 🐛 Bug if train contains less than 5 instances of any class, one or more folds fails with "y_true and y_pred contain different number of classes" error. ### To...

Add forced features as a basis in an iterative selection strategy

1

## 🚀 Feature Request Features force_input must be in out set of features. But features that correlated with forced features are not need in final feature set. I'd like to...

Handle import errors of extras modules

1

## 🚀 Feature Request LightAutoML should inform a user about additional extras that he needs to install to avoid the import problem. Similar to: https://github.com/sb-ai-lab/LightAutoML/blob/f2fa4107b5256e65ace778561ac62037f48d3190/lightautoml/report/report_deco.py#L1134 https://github.com/sb-ai-lab/LightAutoML/blob/f2fa4107b5256e65ace778561ac62037f48d3190/lightautoml/text/tokenizer.py#L21 Or automatically install missing...

TabularAutoML completes its execution too early

1

## 🚀 Feature Request If one of the ML_algo spends more time than it has been allocated, the entire pipeline ends. If https://github.com/sb-ai-lab/LightAutoML/blob/c423c056fccd496cf6876db48692485221459a4e/lightautoml/automl/base.py#L226 ### Proposal Continue the execution while it...

## 🚀 Feature Request Save original column name after feature transformation ### Motivation https://github.com/sb-ai-lab/LightAutoML/pull/58#discussion_r1178909078 ### Proposal Create Column class, that will keep all mappings for column names related to original...

## 🚀 Feature Request Add matching function to map AutoML's predictions for multiclass task and call it before predict method return predictions. `automl.reader.class_mapping` is not convenient.