zhanwenchen / relaug

MIT License
0 stars 0 forks source link

Use My Own Training Confusion Matrix #120

Open zhanwenchen opened 1 year ago

zhanwenchen commented 1 year ago

S+V during BPL may not work as well. If I use my own training confusion matrix, maybe that will make this good?

Or dynamically update the confusion matrix? Wait. Can't we just use the one in the inference?