kevin931/poetic

A machine-learning Python package for poetry.

poetic's Issues

InputLengthError message incorrect with custom model

Describe the bug
The error message of poetic.exceptions.InputLengthError as raised from the poetic.Predictor class is incorrect when the custom model used to initialized the class is of a different input length. The incorrect message reads: poetic.exceptions.InputLengthError: Input length out of bound: must be between 1 and 465.

To Reproduce

    import poetic

    model = poetic.util.Initializer.load_model(model_path=<PATH>, weights_path=<PATH>)
    pred = poetic.Predictor(model=model)
    pred.predict("")

Expected behavior
The error message should reflect the actual maximum input length of the model such as between 1 and 100 instead of 1 and 465.

Desktop (please complete the following information):

• OS: Windows 10
• Architecture: x86-64
• Version: 20H2

poetic information

• Version: 1.1.0
• Build: Dev
• Installation source: GitHub

Environment and dependencies

• Python: 3.7.7
• Environment: conda

Notes
This issue is a note to myself as well as a PA for everyone. I will get to this in a while, but if anyone would like to fix this in the mean time, it will be totally appreciated! PR should be made to the maintenance/v1.1.x branch which will then be merged into dev since v1.1.0 is the earliest version affected by this bug.

Consider creating GUI using guietta?

I'd like to recommand guietta (to everyone!) to create simple GUI interactive apps, which is simple but very powerful!

Development Status Update

This is just a quick update: the project is not abandoned!! I will get to it at some point! If you're reading this, I said hi!!