make tensorflow optional about featurewiz HOT 4 CLOSED

I can install featurewiz on Colab with zero requirements of tensorflow - so I am not sure where you are getting that error from. See the screenshot of my install in featurewiz on Colab. On Thursday, March 7, 2024 at 06:05:49 AM EST, Emryn Hofmann ***@***.***> wrote: I've got v 0.5.7 and it's still telling me tensorflow is required? I'm just trying to use SULOV if that helps at all? Also getting this error: (maybe because I can't install tf): { "name": "NameError", "message": "name 'layers' is not defined", "stack": "--------------------------------------------------------------------------- NameError Traceback (most recent call last) Cell In[55], line 1 ----> 1 import featurewiz File \\Lib\\site-packages\\featurewiz\\__init__.py:10 1 # -*- coding: utf-8 -*- 2 ################################################################################ 3 # featurewiz - advanced feature engineering and best features selection in single line of code (...) 7 ################################################################################ 8 # Version 9 from .__version__ import __version__ ---> 10 from .featurewiz import featurewiz 11 from .featurewiz import FE_split_one_field_into_many, FE_add_groupby_features_aggregated_to_dataframe 12 from .featurewiz import FE_start_end_date_time_features File \\Lib\\site-packages\\featurewiz\\featurewiz.py:49 47 from .my_encoders import Rare_Class_Combiner, Rare_Class_Combiner_Pipe, FE_create_time_series_features 48 from .my_encoders import Column_Names_Transformer ---> 49 from .auto_encoders import DenoisingAutoEncoder, VariationalAutoEncoder, GANAugmenter, GAN 50 from .auto_encoders import dae_hyperparam_selection, vae_hyperparam_selection, CNNAutoEncoder 51 from .stacking_models import get_class_distribution File \\Lib\\site-packages\\featurewiz\\auto_encoders.py:449 446 import pandas as pd 447 from sklearn.base import BaseEstimator, TransformerMixin --> 449 class VariationalAutoEncoder(BaseEstimator, TransformerMixin): 450 \"\"\" 451 Variational Autoencoder (VAE) for feature extraction in multi-class classification problems. 452 (...) 538 539 \"\"\" 540 try: File \\Lib\\site-packages\\featurewiz\\auto_encoders.py:545, in VariationalAutoEncoder() 542 except: 543 print('tensorflow >= 2.5 not installed in machine. Please install and try again. ') --> 545 class Sampling(layers.Layer): 546 \"\"\"Uses (z_mean, z_log_var) to sample z, the vector encoding a digit.\"\"\" 547 def call(self, inputs): NameError: name 'layers' is not defined" } — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you modified the open/close state.Message ID: ***@***.***>