2026年11月11日 技术热点总结

项目简介：the ultimate vocal synthesizer – sing, speak, beatbox, and more

技术栈：Rust

核心介绍：A physical-modeling voice synthesizer in the browser. Every sound comes from a simulated airway (lungs, vocal folds or syrinx, a branched vocal tract, lips, nose), not from samples. The engine is Rust, compiled to WebAssembly for an AudioWorklet; the same code runs natively for tests and the chimera CLI. 1. One instrument: a physically modeled voice (lungs, vocal folds, throat, mouth). No samples, no neural audio.

项目数据：⭐ 1,680 Stars，🍴 191 Forks

项目简介：A native music player you extend with plugins

技术栈：Go

核心介绍：A native desktop music player written in Go with MyGo, no WebView. It plays your local files out of the box, and everything else comes from plugins. go tool mygo dev # run in development go tool mygo build # package into build/ go test ./internal/…

关键特性：Local files (MP3, FLAC, Ogg Vorbis, WAV, and ALAC, FLAC or AAC in .m4a and .aac) with tags, covers and .lrc lyrics；Playlists, liked songs, search, albums and artists；Now Playing with lyric stage styles, cover wall, queue, shuffle and repeat；Notch player…

项目简介：GIF 工厂：按风格制作像素、人物与宠物、粘土吉祥物、光栅卡与动态海报的 GIF Skill。

技术栈：Python

核心介绍：像素与游戏 · 插画与二次元 · 真人与宠物 · 材质与吉祥物 · 卡片与变画 · 海报与文字 · 品牌短片 · 动态绘本 像素小人、人物与宠物、粘土吉祥物、光栅变画、漫画动态海报。 先看效果，再换成你自己的角色和主题；拿不准就先做一张。 浏览风格 · 完整作品库 · 怎么使用 · Skill · 下载项目 人物、背景和大字一起设计。黑白打底，每个角色保留一种主色；从平静切到登场，字体也跟着展开，6 秒循环。 ▶ 播放九宫格 GIF · 使用这个风格 · 九张独立 GIF 每个分类都放了多个角色、动作或场景。点击动图看原尺寸，点击风格名看制作方法；各风格详情页共有 75 个动态案例。100 张完整系列继续放在作品库，默认收起。 像素轻动效 像素轻动效 像素轻动效 像素完整动作

项目数据：⭐ 1,088 Stars，🍴 133 Forks

项目简介：Ocarina of Time running natively on Windows, no emulator. Built on your PC from your own copy. No game files included.

技术栈：C

核心介绍：Read in English O código do jogo é traduzido para C e compilado como um programa do Windows. Os gráficos do Switch 2 são convertidos para Vulkan. Tudo isso acontece no seu PC, a partir do seu próprio arquivo do jogo. > Mudança de rumo (novembro de 2026). Para jogar, hoje recomendamos o Hookshot: ele também compila o jogo no seu PC, a partir da sua cópia, e o instalador dele tem português. > Comparamos os dois no mesmo PC (RTX 5090), com…

项目简介：Accelerate high-latency GEO inflight Wi-Fi via warm BBR TCP multiplexing + 1-RTT Ground-TLS + zlib (zero dependencies)

技术栈：Python

核心介绍：TCP multiplexer and 1-RTT HTTP/TLS relay for geostationary (GEO) inflight Wi-Fi (Inmarsat GX, Viasat, Panasonic, SITA). Pure Python 3 stdlib, zero external dependencies. Measured inflight over Inmarsat GX Ka-band (~620 ms RTT, SITA AS206433 Amsterdam ground station): ORIGIN (6× PARALLEL) RAW CABIN WI-FI (:443) 1-RTT BBR + ZLIB MUX SPEEDUP GitHub Repo ████████████ TIMEOUT ███░░░░░░░░░ 4.66s FIXED ✓

项目数据：⭐ 1,032…

In this work, we study whether frontier agent can continually improve real-world AI infrastructure over long horizons. We introduce Infrathon, a benchmark of diverse, long-horizon AI infrastructure optimization tasks across different levels. We evaluate six models across 150 runs with 24-hour budgets, totaling over 10,000 GPU-hours. Our results show that model rankings vary with the evaluation horizon, while 135 of the 150 runs achieve further improvements during the second half of their budgets. These findings indicate that shorter evaluations can overlook both optimization gains and chang…

Lloyd's k-means algorithm is a long-running standard for clustering, with applications in data analysis, retrieval, quantization and data curation. In this paper, we introduce an exponential moving average optimization procedure for k-means with mini-batch updates, powered by modern deep learning machinery and new insights. We also establish scaling laws on how decay rate and batch size impact final performance. This gives us a formulaic rule to fix them for multiple problem sizes and embedding datasets, without the need for any hyperparameter sweep. The resulting clustering algorithm, coin…

Large language models can generate motion graphics from static images and natural language requests, but it remains difficult to evaluate how well models can follow requests that combine spatial, appearance, and temporal requirements. We introduce MotionGraphicsBench, an expert-curated benchmark of 100 instances for animating an input static SVG with a motion edit prompt. By supplying all required visual elements, our editing task reflects a designer-like workflow and isolates motion generation from graphic generation. Derived from real-world Lottie motion designs, our benchmark contains di…

We introduce FraudBench, an evaluation of large language models' (LLMs) propensities to recommend, assist, or directly take actions that would constitute fraud under U.S. law. Our evaluation contains a novel dataset of 159 scenarios based on real-world U.S. court cases and validated by legal experts. We present each scenario to LLMs via three tasks at varying levels of autonomy: advising users, drafting messages for users to send, and acting directly on behalf of users (477 tasks total). In each scenario, FraudBench re-enacts the facts of a court case to give LLMs an opportunity, but no exp…

Weight tying reduces the parameter cost of language models, but the usual tied output head is limited to a single softmax. We introduce onóna, a two-table parameterization that increases output expressiveness while preserving input-output weight sharing. The Transformer takes the average of two embedding tables as input, and the output head combines the softmax distributions defined by those same tables. A gate conditioned only on the current input token determines the mixture weight. Its values can be precomputed for inference, and the Transformer body remains unchanged. We evaluate onóna …

RidgeFill shortens prefill by not computing most of a model's upper layers for the prompt. The lower layers run exactly up to a split; from the split layer's state a per-layer map predicts every later layer's input, and each late layer writes its cache from that prediction. The design follows YOCO and DeepSeek-V4.1-Flash, where the projection is trained into the model. RidgeFill retrofits it onto models that were never trained for it: the map is a ridge regression fitted once in closed form from one pass over calibration text, the layer's own projections run unchanged, and a chosen set of r…