2026年6月29日 技术热点总结

项目简介：An extension of the traditional PDF standard, allowing multiple files to be stored in a single file via metadata

技术栈：TypeScript

核心介绍：PDFx is an open, backwards compatible extension of PDF that bundles many documents into a single file, plus a minimal desktop viewer for macOS, Windows, and Linux. A .pdfx file is a fully valid PDF: open it anywhere and every page shows in sequence. Open it in PDFx and it splits back into the original documents. Plain single PDFs work as they are.

项目数据：⭐ 485 Stars，🍴 52 Forks

项目简介：Official inference code for Krea 2

技术栈：Python

核心介绍：Krea 2 – an image generation model from Krea AI. API Docs • Hugging Face (RAW) • Hugging Face (TURBO) • Technical Blog This is the official repository for the open version of Krea 2, an image model trained from scratch focused on creative and stylistic exploration. The repository contains inference code and instructions to run the model.

项目数据：⭐ 365 Stars，🍴 27 Forks

项目简介：Reusable AI video production skills library for creation, recreation, motion design, openers, and QA.

技术栈：Python

核心介绍：这是一个长期维护的 AI 视频制作 skill 精选仓库，用来沉淀我在视频创作、视频复刻、动效生成、片头包装、质检交付等流程中反复使用的方法。 仓库不是只服务于某一个视频项目。后续优秀的视频制作 skill 都会优先收录在这里，并按“视频类型 / 风格 / 工作流环节”持续扩展。当前首批收录的是参考视频复刻质检、暗色 SaaS / AI 产品短片、黑底白字打字开场。 这些 skill 适用于 Codex、Claude Code、Cursor 等支持本地 skill 目录或 Skills CLI 的 AI 编程/创作代理。仓库中的每个一级目录都是一个独立可安装的 skill。 ▶ Watch Black White Text Opener 示例成片：雪踏乌云暗色 SaaS 介绍短片 ▶ Watch Xuetawuyun Dark SaaS Showcase 示例成片：Presenton 复刻 Bitexact 成片 ▶ Watch Presenton Replica Bitexact Show…

ABACUS is a unified vision-language model that handles object counting, crowd counting, referring-expression counting, and count-faithful image generation without any benchmark-specific training required. Our model is built on existing 3B-parameter unified foundation model and is adapted for object localization tasks using three key innovations: density-aware adaptive zooming with objectness maps for spatial grounding; a boundary-aware count policy via GRPO to eliminate crop-boundary errors; and a cycle-consistent GRPO strategy where the understanding branch self-critiques generated outputs…

Process reward models enable fine-grained, step-level evaluation of LLMs, yet building them for agentic settings remains prohibitively difficult: long-horizon interactions, irreversible actions, and stochastic environment feedback make both human annotation and Monte Carlo estimation infeasible at scale. In this work, we show that reinforcement learning (RL) post-training already provides the ingredients for effective step-level scoring, eliminating the need for dedicated reward model training altogether. Concretely, we derive an implicit advantage under a general stochastic Markov decision…

Reasoning capability has advanced rapidly in large language models (LLMs), leading to an increasing size of key-value (KV) cache in both prefilling and decoding stages. Existing KV cache compression methods mainly rely on attention weights to estimate token importance. While attention effectively captures contextual relevance, it overlooks complementary information-theoretic signals related to predictive uncertainty and token informativeness. In this paper, we revisit token importance from a forward-looking perspective and introduce Forward Influence, a metric that measures how compressed t…

Earth Observation (EO) forecasting aims to predict future Earth surface dynamics from satellite observations under changing meteorological conditions. In this paper, we view this task as a partially observed, weather-driven world modeling problem, in which weather acts as a conditioning signal, while forecasting remains uncertain due to sparse observations and unobserved land-surface states. However, existing methods do not fully capture this setting: deterministic models collapse uncertainty into a single future prediction, while diffusion-based methods typically treat weather variables as…

The prevalent dual-branch paradigm, i.e., training a side network to encode visual conditions and fusing its intermediate-layer features to a frozen pretrained main network, has shown remarkable success in visual-condition controllable generation. Despite its widespread adoption, the role of the side branch and its training efficiency remain underexplored. In this paper, we first revisit this mainstream paradigm through the lens of score-based generative modeling: 1) The main network preserves visual perceptual quality by providing a prior unconditional score. 2) The side network steers con…

As agentic systems continue to evolve and are widely deployed in real-world scenarios, there is a growing demand to faithfully evaluate their capabilities. However, current benchmarks are typically built on popular applications with relatively simple tasks and focus on a narrow set of capabilities while overlooking broader dimensions, resulting in saturated performance on modern agents and failing to probe their limitations. To this end, we introduce GauntletBench, a web-based benchmark for evaluating agent generalisation in challenging scenarios, focusing on three underexplored capabilitie…