Balanced Meta-Softmax for Long-Tailed Visual Recognition

最新推荐文章于 2023-11-10 13:40:08 发布

江水之南

最新推荐文章于 2023-11-10 13:40:08 发布

阅读量328

点赞数

分类专栏：论文文章标签：深度学习

本文链接：https://blog.csdn.net/qq_43845774/article/details/129200208

版权

前人之述备矣

def balanced_softmax_loss(labels, logits, sample_per_class, reduction):
    """Compute the Balanced Softmax Loss between `logits` and the ground truth `labels`.
    Args:
      labels: A int tensor of size [batch].
      logits: A float tensor of size [batch, no_of_classes].
      sample_per_class: A int tensor of size [no of classes].
      reduction: string. One of "none", "mean", "sum"
    Returns:
      loss: A float tensor. Balance