GitHub

Personalized Federated Fine-Tuning for LLMs via Data-Driven Heterogeneous Model Architectures

This repository contains the official implementation for the work “Personalized Federated Fine-Tuning for LLMs via Data-Driven Heterogeneous Model Architectures”. Large language models (LLMs) are ...
腾讯网

分类数据 EDA 实战：如何发现隐藏的层次结构

点击上方“Deephub Imba”,关注公众号,好文章不错过 !探索性数据分析（EDA）的本质不是画图和算统计量，而是不被自己的数据欺骗。分类列是最容易出问题的地方。city、category、product、department、role、customer_type——这些列看起来很简单，跑个 ...
腾讯网

再见，痛苦学术绘图！谷歌又一款Banana火了，一作是北大博士生

就在2026年2月4日，Google Cloud AI Research联合北京大学的研究团队正式发布了PaperBanana——一个真正意义上的“学术插图Agent框架”。它的目标非常直白：把“自主AI科学家”工作流里最后一个还得靠人肉硬扛的环节— ...
techENT

Google’s “Agentic Vision” Gives Gemini 3 Flash an Investigative Eye

This week, Google introduced a new capability for its Gemini 3 Flash model called “Agentic Vision” that fundamentally changes ...
GitHub

Rationales for Sequential Predictions

Source code for the paper: Rationales for Sequential Predictions by Keyon Vafa, Yuntian Deng, David Blei, and Sasha Rush (EMNLP 2021). Check out our Colab notebook, which generates a sequence with GPT ...
Hacker

Your First Interactive Plot in Python: A Hands-On Plotly Guide

A software engineer and book author with many years of experience, I have dedicated my career to the art of automation. A software engineer and book author with many years of experience, I have ...