[KongchangAI]
ConceptValidation Accuracy

Validation Accuracy

A metric in machine learning that measures the correctness of a model's classification decisions on a validation set. It is commonly used for model selection and checkpoint selection.

Timeline (last 90 days)

Nov 11

Validation loss (such as cross-entropy) measures the difference between the model's predicted probability distribution and the true labels, while validation accuracy (or F1) measures the correctness of hard classification decisions; the two are not always monotonically in sync.

Unverified50%
Nov 11

The checkpoint selected by the lowest-loss criterion and the checkpoint selected by the highest-accuracy criterion may be models from different stages of training, so their test performance may naturally differ.

Unverified50%

All Facts (2)

Source Articles