MarkTechPost
PulseAugur coverage of MarkTechPost — every cluster mentioning MarkTechPost across labs, papers, and developer communities, ranked by signal.
4 天有情绪数据
-
Langfuse 教程展示用于追踪和评估的 LLM 流水线
本教程演示如何使用 Langfuse（一个开源平台）为 LLM 应用程序构建完整的可观测性和评估流水线。该指南涵盖追踪、提示管理、评分和实验执行，提供了一个实用的工作流程。它支持与 OpenAI 或确定性模拟 LLM 集成，允许用户在无需付费模型访问权限的情况下探索 Langfuse 功能。
-
SHAP 指南详细介绍机器学习模型可解释性工作流
本指南提供了一个使用 SHAP 可解释性工作流解释机器学习模型的实用框架。它详细介绍了如何训练基于树的模型，并比较了各种 SHAP Explainer，如 Tree、Exact、Permutation 和 Kernel 方法。本教程还考虑了模型感知和模型无关的技术，探讨了不同方法对准确性和运行时的影响。
-
Django-Unfold 教程展示了如何构建高级管理仪表板
本教程详细介绍了如何使用 Django 和 Django-Unfold 库构建高级管理仪表板。它指导用户完成 Unfold 的安装、项目设置和配置，包括自定义主题、导航以及过滤器和操作等功能。该过程涉及定义电子商务模型、使用示例数据填充数据库，并通过 Web 浏览器访问管理面板。
-
FLARE-FLOSS tutorial shows advanced malware string recovery
This tutorial demonstrates how to use FLARE-FLOSS to extract hidden malware indicators of compromise (IOCs) from Windows executables, going beyond traditional string analysis. It guides users through setting up FLOSS an…
-
Momentum smooths gradient descent's zigzag convergence, accelerating ML training
Gradient descent, a core optimization algorithm, often struggles with uneven loss surfaces, leading to inefficient "zigzagging" convergence. This issue arises from the surface's curvature, where steepness in one directi…
-
ZenML tutorial shows building end-to-end production ML pipelines
This tutorial details the creation of a production-ready machine learning pipeline using ZenML. It covers setting up a ZenML project, defining a custom materializer for specific dataset objects, and building a modular p…
-
Developers' guide tackles AI prompting for production reliability
A new guide addresses the critical need for reliable prompting as AI integrates into production systems. It outlines five techniques: role-specific prompting, negative prompting, JSON prompting, Attentive Reasoning Quer…
-
What is Tokenization Drift and How to Fix It?
Tokenization drift occurs when minor formatting changes in input text, such as spacing or line breaks, lead to different token IDs being generated by a model. This can cause unpredictable shifts in model behavior becaus…
-
Talkie-1930: New 13B LLM trained on pre-1931 English for historical research
Researchers have developed Talkie-1930, a new open-weight language model with 13 billion parameters. This model was trained exclusively on English text published before 1931. Its primary purpose is to facilitate contami…