ValueError: node array from the pickle has an incompatible dtype:

🚀 Deployment
1

I got this error, anyone know why?

Screen Shot 2023-07-06 at 12.09.32 PM
Screen Shot 2023-07-06 at 12.09.32 PM856×482 39.3 KB

Here’s the code to access my model:

def read_model(url='output_model/model.joblib'):

    if ENV == 'AWS':
        file_key = url
        obj = s3.get_object(Bucket=bucket_name, Key=file_key)
        model = joblib.load(io.BytesIO(obj['Body'].read()), 'rb')
    else:
        model = joblib.load(url)
    return model
2

hallo missarah,
for your error, you must ensure

  1. your streamlit version, please check on your requirements.txt with your streamlit
  2. when you activate streamlit, make sure you have same version, your scikit-learn version with scikit-learn version on your requirements.txt
3

thank you, it worked when I specified the right version of scikit-learn which I used to create the model

1 Like