v0.1.0

Enhances classification performance by computing weights based on the proximity between training and test samples, making the model more effective in scenarios with limited labeled data, class imbalance, noisy labels, or domain shift. The core functionality is provided through components like WeightComputer, WeightedLoss, and WeightedTrainer, which collectively implement the distance-based weighting mechanism using various distance metrics (euclidean, cosine, hamming, jaccard). The repository includes the implementation of the algorithm, experiment utilities, and example code.