Javier jrzaurin
This file contains hidden or 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
| Dataset | n_rows | n_cols | objective | neg_pos_ratio | |
|---|---|---|---|---|---|
| adult | 45222 | 15 | binary_classification | 0.3295 | |
| bank_marketing | 41188 | 20 | binary_classification | 0.127 | |
| nyc_taxi | 1458644 | 26 | regression | NA | |
| facebook_comments_vol | 199029 | 54 | regression | NA |
This file contains hidden or 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
| trainer = Trainer(model, objective="binary", metrics=[(Accuracy)]) | |
| trainer.fit(X_tab=X_tab, target=target, n_epochs=5, batch_size=256, val_split=0.2) |
This file contains hidden or 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
| from pytorch_widedeep.models import TabTransformer | |
| tabtransformer = TabTransformer( | |
| column_idx=tab_preprocessor.column_idx, | |
| embed_input=tab_preprocessor.embeddings_input, | |
| continuous_cols=cont_cols, | |
| shared_embed=True, | |
| num_blocks=3, | |
| ) | |
| model = WideDeep(deeptabular=tabtransformer) |
jrzaurin
/ tabtransformer_preprocessing.py
Last active
February 22, 2021 08:57
TabTransformer preprocessing
This file contains hidden or 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
| embed_cols = [ | |
| 'workclass', | |
| 'education', | |
| 'marital_status', | |
| 'occupation', | |
| 'relationship', | |
| 'race' | |
| ] | |
| tab_preprocessor = TabPreprocessor( | |
| embed_cols=embed_cols, |
This file contains hidden or 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
| trainer = Trainer(model, objective="binary", metrics=[(Accuracy)]) | |
| trainer.fit(X_tab=X_tab, target=target, n_epochs=5, batch_size=256, val_split=0.2) |
This file contains hidden or 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
| from pytorch_widedeep.models import TabResnet | |
| tabresnet = TabResnet( | |
| column_idx=tab_preprocessor.column_idx, | |
| embed_input=tab_preprocessor.embeddings_input, | |
| continuous_cols=cont_cols, | |
| batchnorm_cont=True, | |
| blocks_dims=[200, 100, 100], | |
| mlp_hidden_dims=[100, 50], | |
| ) |
This file contains hidden or 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
| from pytorch_widedeep import Trainer | |
| from pytorch_widedeep.metrics import Accuracy | |
| trainer = Trainer(model, objective="binary", metrics=[(Accuracy)]) | |
| trainer.fit(X_tab=X_tab, target=target, n_epochs=5, batch_size=256, val_split=0.2) |
This file contains hidden or 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
| from pytorch_widedeep.models import TabMlp, WideDeep | |
| tabmlp = TabMlp( | |
| mlp_hidden_dims=[200, 100], | |
| column_idx=tab_preprocessor.column_idx, | |
| embed_input=tab_preprocessor.embeddings_input, | |
| continuous_cols=cont_cols, | |
| batchnorm_cont=True, | |
| ) | |
| model = WideDeep(deeptabular=tabmlp) |
This file contains hidden or 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
| from pytorch_widedeep.preprocessing import TabPreprocessor | |
| # define the embedding and continuous columns, and target | |
| embed_cols = [ | |
| ('workclass', 6), | |
| ('education', 8), | |
| ('marital_status', 6), | |
| ('occupation',8), | |
| ('relationship', 6), | |
| ('race', 6)] |
jrzaurin
/ adult_census_data_preparation.py
Last active
February 19, 2021 22:55
Adult census data preparation
This file contains hidden or 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
| import pandas as pd | |
| import numpy as np | |
| from sklearn.model_selection import train_test_split | |
| adult = pd.read_csv("data/adult/adult.csv.zip") | |
| adult.columns = [c.replace("-", "_") for c in adult.columns] | |
| adult["income_label"] = (adult["income"].apply(lambda x: ">50K" in x)).astype(int) | |
| adult.drop("income", axis=1, inplace=True) | |
| for c in adult.columns: |