Annotation weight is an importance coefficient set for different annotation results or labels (e.g., the annotation weight of key targets is higher than that of secondary targets), used to adjust the influence degree of different annotation information during model training.
