1 个仓库
Awesome GitHub RepositoriesModel Training Diagnostics
Tools for analyzing training stability to identify underfitting and overfitting in neural networks.
Distinct from Performance Diagnostics: Distinct from general performance diagnostics: specifically targets neural network training stability (underfitting/overfitting) rather than system execution speed.
Explore 1 awesome GitHub repository matching testing & quality assurance · Model Training Diagnostics. Refine with filters or upvote what's useful.
Awesome Model Training Diagnostics GitHub Repositories
用 AI 发现最棒的仓库。我们将通过 AI 为您搜索最匹配的仓库。
- udacity/deep-learning
udacity/deep-learning
4,058在 GitHub 上查看该项目是一个深度学习教育课程和实现指南，专为构建和训练神经网络而设计。它提供了一个开发模型以解决模式识别和生成任务的课程。 该材料包括用于计算机视觉训练、自然语言处理和生成式 AI 的专门模块。它涵盖了迁移学习在分类新数据中的实际应用以及合成媒体的创建。 该项目包含网络架构设计、机器学习数据流水线的构建，以及使用模型性能诊断来识别欠拟合或过拟合。 内容通过 Jupyter Notebook 交付。
Provides techniques for analyzing training stability and visualizing network graphs to optimize model performance.
Jupyter Notebook
