The definition of Precision in 02_pytorch_classification.ipynb is wrong.

[danli349]

Hello;

The definition of Precision: Proportion of true positives over total number of samples. Higher precision leads to less false positives (model predicts 1 when it should've been 0). in "02_pytorch_classification.ipynb" is wrong.
Precision is the proportion of true positives over positive predictions (not the total number of samples).

Thanks

[37Lime]

Hi,

I have created a PR for this Issue.

To clarify, precision is calculated as TP / (TP + FP), where TP = True Positives and FP = False Positives.

Best wishes.

[danli349]

@37Lime The positive predictions include true positives and false postives.

[37Lime]

@danli349 Sorry, I misinterpreted your sentence.