Confusing Results

[YemoGardner]

Hello, I am trying to reproduce your code, but I have found that the generated results are quite confusing and not accurate. I don't know if it's because when trying to convert SVG to NPZ files, a warning appears: Visible Depreciation Warning:
Creating an ndarray from torn nested sequences (which is a list or tuple of lists or triples or ndarrays with different lengths or shapes) is deprecated If you meant to do this, you must specify 'dtype=object' when creating the ndarray.Val=np. anyarray (val)
But if I try to put it in an array, the same unsatisfactory output results will still occur.

[MarkMoHR]

Hi, what's the image you showed? Your input vector image or the output? Could you show your inputs? The image above is very terrible and confusing

[YemoGardner]

Thank you for your reply. This is the result of my output SVG, which was opened with Adobe Illustrator, and the input I used was example 23 from the ‘sample_inputs’ file you provided. I have attached my running steps, maybe you can identify where the problem occurred.

[MarkMoHR]

Hi, from the logs, everything looks fine. But from the results, it looks like the trained weights have not been loaded correctly, as the model seems not able to predict correct strokes at all.