Error in kaggle notebook kernel

[pplonski]

There is problem with pyarrow verion in kaggle notebook. There is an error:

 [Errno 2] No such file or directory: 'AutoML_1/y.parquet'

See kaggle comment for details: https://www.kaggle.com/mt77pp/mljar-autoeda-automl-prediction/comments#1197484

[pplonski]

• removed fastparquet dependency
• updated pandas package
• remove support for Python 3.6

[hokmingkwan]

Hi, I've encountered the same issue and tried the solution you mentioned. However, it still doesn't work. Would you give me your thoughts?

[pplonski]

@hokmingkwan please start with fresh notebook (not copy or fork of existing one) and try to install mljar directly from the github:

!pip install -q -U git+https://github.com/mljar/mljar-supervised.git@master

[hokmingkwan]

@hokmingkwan please start with fresh notebook (not copy or fork of existing one) and try to install mljar directly from the github:

!pip install -q -U git+https://github.com/mljar/mljar-supervised.git@master

@pplonski I have already tried that start with fresh notebook, but still keeps failing with same error

FileNotFoundError: [Errno 2] No such file or directory: 'AutoML_1/y.parquet'

[pplonski]

@hokmingkwan I've just checked: https://www.kaggle.com/mt77pp/let-me-check

In the fresh notebook it will work. For forked/copy notebooks there might be a problem with numpy versions -> that's why errors.

[pplonski]

@hokmingkwan do you have link for your notebook? What dataset are you trying to analyze?

[hokmingkwan]

@hokmingkwan do you have link for your notebook? What dataset are you trying to analyze?

@pplonski I found out what causes the error. For example, you have a large csv dataset with all numeric data and you read the csv by pandas. So you have a dataframe and would like to reduce the memory usage, you may change the data type to np.float16, ie. df = df.astype(np.float16)
if you run model.fit(df.iloc[:, :-1], df.iloc[:, -1:])
Then it will show this error:
FileNotFoundError: [Errno 2] No such file or directory: 'AutoML_1/y.parquet'