Awesome GitHub RepositoriesBatch Input Processing

Pkl is a configuration-as-code language used to define, validate, and generate structured configuration files. It functions as a type-safe configuration generator that enforces data integrity through a strongly-typed schema, ensuring configuration values meet defined constraints and types during evaluation. The project distinguishes itself by acting as both a configuration file generator and a binding generator. It transforms high-level programmable definitions into static formats such as JSON, YAML, or XML, and produces language-specific source code to synchronize settings and provide type s

jc is a tool that transforms plain-text results from command-line utilities, system tools, log formats, and text tables into structured JSON data. It functions as a structured data transformer capable of converting various file formats, including CSV, INI, XML, and YAML, into JSON representations for programmatic use. The project includes a collection of specific parsers for Unix commands and system tools such as df, blkid, and various package managers. It also features specialized converters for web server logs, Common Log Format, and Common Event Format strings. The tool covers broad capab

Kreuzberg is a document extraction engine that converts PDFs, Office files, images, and over 90 other formats into clean, structured text and metadata. It is built around a compiled Rust core that can be used as a native library, a command-line tool, a REST API server, or a WebAssembly module for browser-based processing. The system is designed to run entirely on self-hosted infrastructure, with no data leaving the user's environment. What distinguishes Kreuzberg is its breadth of integration surfaces and its pipeline architecture. It exposes extraction capabilities through native bindings fo

Depth-Anything is a monocular depth estimation foundation model that produces dense per-pixel depth maps from a single RGB image. It is built on a DINOv2 Vision Transformer encoder backbone and trained on 62 million unlabeled images using a teacher-student pseudo-labeling framework, enabling robust generalization across diverse scenes without task-specific training. The model outputs both relative depth maps, which capture the ordering of scene points, and metric depth maps with real-world units after fine-tuning on datasets like NYUv2 or KITTI. The project distinguishes itself through its ab

Stanza is a Python natural language processing library designed for tokenization, lemmatization, and dependency parsing across many human languages using neural models. It provides a neural processing pipeline that converts raw text into structured linguistic data objects, alongside a specialized analyzer for extracting medical insights from clinical and biomedical language. The project includes a wrapper that connects Python scripts to Java-based natural language processing tools and remote annotation servers. This enables a bridge for extracting linguistic annotations and analysis data from

ailab is a deep learning tool designed to upscale anime-style images, increasing their resolution while preserving fine details. It is built around a cascade U-Net architecture, a multi-stage neural network model that refines image quality through successive stages, and uses PyTorch for inference. The tool specializes in enhancing anime and cartoon-style artwork, applying super-resolution techniques to boost pixel dimensions without sacrificing visual fidelity. It processes images through a pipeline that includes tensor preprocessing, model inference, and post-processing pixel reconstruction,

Valhalla is an open-source routing engine that calculates optimal paths and travel times using OpenStreetMap data. It is built around a tiled routing graph framework, allowing map data to be organized into small geographic tiles for efficient regional updates and offline routing capability. The project distinguishes itself through a multimodal routing server that combines automobile, pedestrian, bicycle, and public transit modes into single journeys. It includes a GPS trace matching engine to align noisy coordinates to the most probable road network paths and an isochrone and matrix generator

Trafilatura is a Python library and command-line tool for extracting clean, structured text and metadata from web pages. It downloads HTML content, identifies the main body of text, and strips away navigation, ads, and other boilerplate, returning the core article content along with fields like title, author, date, and URL. The tool can also extract user comments and test whether a page contains extractable text, making it a general-purpose web text extraction library. What distinguishes Trafilatura from simpler extractors is its configurable extraction pipeline, which offers high-speed, high

Dieses Projekt ist ein spezialisierter TikTok-API-Scraper und Datenextraktor. Es fungiert als Proxy-basierter Web-Scraper, der Metadaten von Benutzern, Videobeiträge und Trend-Feeds sammelt und eine Webhook-Datenpipeline bereitstellt, um die gescrapten Informationen via HTTP-Requests an externe URLs weiterzuleiten. Das Tool enthält einen Downloader für Videos ohne Wasserzeichen, der hochauflösende Inhalte im lokalen Speicher ablegt. Es verwendet kryptografische Request-Signierung zur Server-Authentifizierung und nutzt Session-Cookie-Authentifizierung in Kombination mit Proxy-Rotation, um den Netzwerkverkehr zu steuern und Ratenbegrenzungen zu umgehen. Die Funktionen umfassen die Massenextraktion von Inhalten durch Batch-Processing-Pipelines sowie den Abruf von Metadaten für Profile, Hashtags und Musiktitel. Extrahierte Daten werden in strukturierte JSON- oder CSV-Formate exportiert; zudem wird der lokale Status nachverfolgt, um den Download-Verlauf zu protokollieren und doppelte Anfragen zu verhindern.

Text Embeddings Inference ist ein hochperformanter Inferenz-Server, der darauf ausgelegt ist, Text-Embedding- und Sequenz-Klassifizierungsmodelle als skalierbare API-Endpunkte zu hosten. Er bietet eine Vektor-Embedding-API zur Konvertierung von Text in dichte Repräsentationen sowie einen Cross-Encoder-Reranking-Server zur Bewertung der Relevanz von Dokumentsequenzen gegenüber einer Abfrage. Das Projekt bietet eine GPU-beschleunigte Inferenz-Engine, die dynamisches Batching und spezialisierte Kernels nutzt, um den Durchsatz zu maximieren. Es bietet eine hochperformante binäre Schnittstelle via gRPC als Alternative zu Standard-HTTP, um Netzwerklatenz und Serialisierungs-Overhead zu reduzieren. Das System deckt ein breites Spektrum an Funktionen ab, einschließlich Dokument-Ähnlichkeits-Ranking, mehrsprachigem Text-Reranking und Sequenz-Klassifizierung zur Vorhersage von Kategorien oder Sentiment. Es unterstützt diverse Deployment-Umgebungen, von serverlosen Auto-Scaling-Containern bis hin zu Air-Gapped-Installationen. Hardwarebeschleunigung ist für NVIDIA-GPUs, AMD-GPUs und Apple Metal verfügbar.

Dieses Projekt ist ein PyTorch-Framework für das Model-Serving, das darauf ausgelegt ist, Machine-Learning-Modelle in der Produktion über skalierbare Netzwerk-Endpunkte bereitzustellen. Es fungiert als leistungsstarker Inference-Server, Optimierer und Modell-Lifecycle-Manager, der das Laden von Modellen, Request-Batching und Hardware-Beschleunigung übernimmt. Das System zeichnet sich durch fortschrittliche Orchestrierungs- und Optimierungsfunktionen aus, wie etwa das Verketten mehrerer Modelle zu sequenziellen Workflows mittels Ausführungsgraphen und den Einsatz von Dynamic Batching zur Verbesserung von Durchsatz und Latenz. Es bietet spezialisierte Unterstützung für generative KI und Large Language Models durch Continuous Batching und Tensor-Parallelität. Zu den breiten Funktionsbereichen gehören GPU-Ressourcenmanagement für diverse Hardware wie NVIDIA, AMD und Apple Silicon sowie ein umfassendes Lifecycle-Management für Registrierung, Versionierung und Worker-Skalierung. Zudem integriert es Observability-Tools zur Überwachung des Systemzustands und der Modellleistung über Prometheus-kompatible Metriken. Der Server wird über eine Kommandozeilenschnittstelle verwaltet, die zur Steuerung des Lifecycles und zur Konfiguration von Laufzeitparametern dient.

CTranslate2 is a C++ inference engine and runtime for Transformer models, designed to execute models on both CPU and GPU with optimizations for speed and memory efficiency. It functions as a model format converter, quantization tool, and REST API server, enabling deployment of neural machine translation, automatic speech recognition, and text generation models. The engine distinguishes itself through a suite of runtime optimizations including layer fusion, weight-matrix quantization, batch-by-length grouping, and a caching allocator that reuses GPU memory. It supports tensor-parallel model di

GAM is a command-line tool for administering Google Workspace and Cloud Identity. It translates command-line arguments into structured API calls, enabling administrators to manage users, groups, organizational units, and domain settings across a Google Workspace environment. The tool handles authentication through OAuth2 flows, service accounts, and workload identity federation, and supports multi-tenant configurations for managing multiple domains or cloud projects from a single installation. GAM distinguishes itself through its batch processing and automation capabilities. It can process la

Segment Anything Fast is a high-performance computer vision inference engine and image segmentation framework built for PyTorch. It provides a specialized environment for automated object isolation and mask generation, designed to process large-scale visual datasets with increased throughput. The project distinguishes itself through a suite of system-level optimization strategies that accelerate deep learning model performance. By utilizing graph-based model compilation, just-in-time kernel fusion, and hardware-aware quantization, it reduces computational latency and memory footprint. These t