ModuleNotFoundError：没有名为“sklearn.linear_model.base”的模块

Question

Tried to deploy a ML model in AWS and getting this error on the same. Have updated scikit-learn library but still having this issue.

Answer 1

I had the exact same problem when moving to a new Python environment. I couldn't easily bring my conda env with me to a new OS, so had to hack it together. If you can bring your python environment with you, that should be your first bet--see the conda page: Managing Conda Environments .

The problem is that namespace linear_model.base of sklearn appears to have been removed somewhere between the build where I built the model and the latest build--so if you want your model built with an older version to work, you'll need to create (or recreate) an environment that has the same version of scikit-learn. Updating the module becomes the problem, not the solution . For me, scikit-learn 0.19.1 was what I used to build the model, so I needed to revert to that, which required me to build a new environment based on python 3.6. Not ideal... but it worked.

Make sure when you create the new environment, you install the specific version of scikit-learn first , so all subsequent dependencies will have to be compatible. If you need a specific version of Python, you will create a new environment by:

conda create --name youNameHere python=X.Y

and then ensure you have the right version of scikit-learn with something like this:

conda install scikit-learn=A.BC.D

where "A.BC.D" is your version number.

I'm not sure how much help that is--you may want to install conda on the AWS instance if you have the ability to do so. Here's a good resource on how to do it from the command line if you don't want to just move to another AMI that already has it loaded: https://dev.to/waylonwalker/installing-miniconda-on-linux-from-the-command-line-4ad7

Answer 2

I had similar, but fixed as below:

No longer getting "No module named 'sklearn.linear_model.base'" I am using Python 3.9

You need to edit the following / replace the 2 lines commented out with with these 2 lines and it works - original code left in below for completeness.

from sklearn.linear_model import LinearRegression                               
regressor = LinearRegression()                                                  
                                                                              
#from sklearn.tree import DecisionTreeRegression                                
#regressor = DecisionTreeRegression()

Answer 3

Cause:You have files from a different source and now you have them in a new environment Just do this:Go to your.py file that has the scikit library and note the version Go go your directory and pip install scikit-learn==x.xx.x where x.xx.x is the exact number existing in the.py file See a real example I had when testing some app from github Below is a real example in cmd

C:\todel>python app.py
Traceback (most recent call last):
  File "app.py", line 6, in <module>
    model = pickle.load(open('model.pkl', 'rb'))
ModuleNotFoundError: No module named 'sklearn.linear_model.base'

C:\todel>pip install scikit-learn=0.23.2
ERROR: Invalid requirement: 'scikit-learn=0.23.2'
Hint: = is not a valid operator. Did you mean == ?
WARNING: You are using pip version 20.1.1; however, version 22.0.3 is available.
You should consider upgrading via the 'c:\pyapp\python.exe -m pip install --upgrade pip' command.

C:\todel>pip install scikit-learn==0.23.2
Collecting scikit-learn==0.23.2
  Downloading scikit_learn-0.23.2-cp37-cp37m-win_amd64.whl (6.8 MB)
     |████████████████████████████████| 6.8 MB 504 kB/s
Requirement already satisfied: joblib>=0.11 in c:\pyapp\lib\site-packages (from scikit-learn==0.23.2) (1.1.0)
Requirement already satisfied: scipy>=0.19.1 in c:\pyapp\lib\site-packages (from scikit-learn==0.23.2) (1.4.1)
Requirement already satisfied: threadpoolctl>=2.0.0 in c:\pyapp\lib\site-packages (from scikit-learn==0.23.2) (3.0.0)
Requirement already satisfied: numpy>=1.13.3 in c:\pyapp\lib\site-packages (from scikit-learn==0.23.2) (1.19.2)
Installing collected packages: scikit-learn
  Attempting uninstall: scikit-learn
    Found existing installation: scikit-learn 1.0.2
    Uninstalling scikit-learn-1.0.2:
      Successfully uninstalled scikit-learn-1.0.2
Successfully installed scikit-learn-0.23.2
WARNING: You are using pip version 20.1.1; however, version 22.0.3 is available.
You should consider upgrading via the 'c:\pyapp\python.exe -m pip install --upgrade pip' command.

C:\todel>python app.py
C:\pyapp\lib\site-packages\sklearn\utils\deprecation.py:143: FutureWarning: The sklearn.linear_model.base module is  deprecated in version 0.22 and will be removed in version 0.24. The corresponding classes / functions should instead be imported from sklearn.linear_model. Anything that cannot be imported from sklearn.linear_model is now part of the private API.
  warnings.warn(message, FutureWarning)
C:\pyapp\lib\site-packages\sklearn\base.py:334: UserWarning: Trying to unpickle estimator LinearRegression from version 0.20.2 when using version 0.23.2. This might lead to breaking code or invalid results. Use at your own risk.
  UserWarning)
 * Serving Flask app 'app' (lazy loading)
 * Environment: production
   WARNING: This is a development server. Do not use it in a production deployment.
   Use a production WSGI server instead.
 * Debug mode: on
 * Restarting with stat
C:\pyapp\lib\site-packages\sklearn\utils\deprecation.py:143: FutureWarning: The sklearn.linear_model.base module is  deprecated in version 0.22 and will be removed in version 0.24. The corresponding classes / functions should instead be imported from sklearn.linear_model. Anything that cannot be imported from sklearn.linear_model is now part of the private API.
  warnings.warn(message, FutureWarning)
C:\pyapp\lib\site-packages\sklearn\base.py:334: UserWarning: Trying to unpickle estimator LinearRegression from version 0.20.2 when using version 0.23.2. This might lead to breaking code or invalid results. Use at your own risk.
  UserWarning)
 * Debugger is active!
 * Debugger PIN: 137-011-919
 * Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)