Cross-Entropy on k4i's bloghttps://k4i.top/zh/tags/cross-entropy/Recent content in Cross-Entropy on k4i's blogHugo -- gohugo.iozhsky_io@outlook.com (K4i)sky_io@outlook.com (K4i)All content is subject to the license of <a rel="license noopener" href="https://creativecommons.org/licenses/by-nc-sa/4.0/" target="_blank">CC BY-NC-SA 4.0</a> .Tue, 23 Jun 2026 10:00:00 +0800Loss Function：模型到底在优化什么https://k4i.top/zh/posts/loss-functions-cross-entropy/Tue, 23 Jun 2026 10:00:00 +0800sky_io@outlook.com (K4i)Tue, 23 Jun 2026 10:00:00 +0800https://k4i.top/zh/posts/loss-functions-cross-entropy/<p>前向传播给出预测，反向传播计算梯度，梯度下降更新参数。但中间还有一个非常关键的问题：</p> <blockquote> <p>预测结果到底怎样才算“错”？错多少？</p> </blockquote> <p>这个问题由 <strong>loss function（损失函数）</strong> 回答。它把模型输出和真实标签变成一个标量：</p>K4ifeatured imagedeep-learningloss-functioncross-entropygradient-descentAI