Projects sharing features with MNBVC

Chinese-Vicuna is a Chinese large language model and instruction-following AI based on the LLaMA architecture. It is specifically designed for natural language understanding and generation in the Chinese language, utilizing an instruction-tuned model to follow complex user prompts across conversations. The project provides a LoRA fine-tuning framework and quantization systems to enable model adaptation and inference on consumer hardware. It implements quantized inference to reduce memory usage on both CPUs and GPUs, supported by a low-level C++ implementation to minimize system resource requi

pdfminer.six is a programmatic tool for extracting text, layout information, and metadata from PDF documents into machine-readable formats. It functions as a document parser that converts internal PDF objects and structures into accessible data objects for analysis. The project includes utilities for decrypting RC4 and AES encrypted files to enable content extraction. It also provides a layout analyzer to identify fonts, colors, and text locations to determine the organizational structure of pages. The system covers a broad range of extraction capabilities, including the retrieval of embedde

python-goose is a Python library for web scraping and content extraction. It functions as an HTML boilerplate remover and article parser designed to isolate primary text and metadata from web pages by stripping away navigation, layout noise, and non-essential elements. The tool features multilingual processing capabilities, utilizing language-specific stop-word analyzers to identify and extract primary content across different languages. It also identifies and collects embedded media, including source URLs and embed codes for lead images and videos associated with an article. The library cov

DataFlow is an agent-based workflow orchestrator and data pipeline designed to synthesize, clean, and augment large-scale datasets for training large language models. It functions as a synthetic data generator and text curation tool, utilizing an intelligent assistant to assemble modular processing operators into functional pipelines based on user requirements. The project distinguishes itself through a low-code approach, providing a web-based visual interface for designing and monitoring multi-stage execution flows. It features an operator-based registry system that allows for the integratio

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

Compromise is a natural language processing library and rule-based text parser designed to analyze unstructured text. It functions as a toolkit for identifying parts of speech, linguistic patterns, and semantic meaning, while providing specialized engines for named entity recognition and the parsing of temporal and numeric data. The project is distinguished by its linguistic morphological engine, which can conjugate verbs across different tenses and inflect nouns and adjectives. It further allows for linguistic model customization through a plugin system that enables the extension of lexicons

This project is a comprehensive framework and toolkit for developing, optimizing, and deploying transformer-based models across multimodal, document intelligence, and natural language processing tasks. It provides a unified neural architecture that processes text, vision, audio, and document layout data through a shared set of weights, enabling researchers and developers to build foundational models that align cross-modal representations. The platform distinguishes itself through advanced training and inference strategies designed for large-scale deep learning. It incorporates specialized mec

Langroid is a multi-agent orchestration framework and tool integration suite designed for building complex AI applications. It serves as a multi-modal integration layer that connects diverse local and remote language models with an agentic retrieval-augmented generation system. The project distinguishes itself through a collaborative message-exchange paradigm, allowing specialized agents to delegate tasks hierarchically and coordinate via structured communication. It features an advanced state management system for conversational AI, including the ability to rewind and prune conversation hist

LlamaIndex is a comprehensive development framework designed to connect private or external data sources to large language models. It functions as a data-centric toolkit that enables the construction of retrieval-augmented generation systems, allowing developers to build applications that provide context-aware answers based on specific organizational information. The project distinguishes itself through a robust agentic orchestration engine that supports the creation of autonomous agents capable of multi-step reasoning, memory management, and complex tool execution. Beyond simple retrieval, i

jetson-inference is a set of libraries and tools for executing optimized deep learning models on embedded GPU hardware. Its primary purpose is to enable real-time computer vision and AI inference at the edge with low latency and high throughput. The project distinguishes itself through high-performance streaming analytics and the ability to execute concurrent AI pipelines on auto-grade silicon. It provides specialized support for multi-sensor stream processing, utilizing zero-copy data transport to load camera frames directly into GPU memory. The codebase covers a broad surface of capabiliti

AutoGluon is an automated machine learning framework and multimodal library designed to automate the end-to-end pipeline from data preprocessing to high-accuracy model training and validation. It functions as an automated model trainer for tabular, image, text, and time series data, as well as a tool for time series forecasting and foundation model finetuning. The project is distinguished by its ability to jointly process and fuse different data types, allowing for the construction of multimodal neural networks that integrate images, text, and structured tables. It supports zero-shot inferenc

UltraChat is a collection of large-scale conversational datasets and instruction-tuning data designed for training and evaluating generative AI models. It provides structured JSON data consisting of complex, multi-round dialogue sequences intended to refine the performance of large language models in chat tasks. The project focuses on improving reasoning and response quality through a diverse set of interactions across multiple sectors. These datasets are used for supervised fine-tuning and instruction tuning workflows to improve how models follow complex directions and maintain context acros

Pandoc is a universal document converter that translates content between a wide range of markup and binary formats. It functions by parsing input documents into a unified intermediate abstract syntax tree, which serves as the foundation for consistent manipulation and transformation across diverse output types. The system is distinguished by its modular reader-writer pipeline, which decouples input parsing from output generation to allow for granular control over document structure. Users can programmatically manipulate this intermediate tree through a robust filter system, supporting both ex

SnowNLP is a Python library for Chinese natural language processing. It provides tools for text segmentation, sentiment analysis, document classification, and phonetic transliteration. The library includes capabilities for training and saving custom machine learning models for tokenization and sentiment analysis using raw training datasets. It covers a range of linguistic processing areas, including parts of speech tagging, sentence splitting, and text similarity measurement. The toolkit also provides utilities for extracting key information through text summarization and calculating word im

RedPajama-Data is a toolset for preprocessing large-scale text datasets used to train large language models. It provides a preprocessing pipeline focused on cleaning, deduplicating, and scoring massive collections of text to ensure data quality and diversity. The project utilizes a document quality scoring framework that employs machine learning and statistical heuristics to evaluate if documents are suitable for training. It includes a dataset filtering pipeline that uses classifiers and blocklists to remove undesirable words or URLs. The system features a text deduplication toolset that el

JioNLP is a Chinese natural language processing toolkit designed for cleaning, normalizing, and extracting structured information from unstructured text. It functions as a linguistic analyzer for Chinese characters and a rule-based named entity extractor, providing a specialized system for sentiment scoring and synthetic data generation for machine learning workflows. The project features a lexicon-based sentiment analysis engine that computes numerical emotional tone scores and a data augmentation library that uses back-translation and synonym replacement to expand training datasets. It incl

Synonyms is a Chinese natural language processing tool focused on semantic analysis. It provides capabilities for Chinese word segmentation, part-of-speech tagging, and the retrieval of synonyms based on semantic proximity. The project converts words and sentences into numerical vector representations to calculate similarity scores. This allows for the determination of semantic proximity between different phrases and the identification of chatbot intent through sentence comparison. The system also includes tools for automated keyword extraction and importance ranking to identify significant

Synonyms is a natural language processing library and semantic similarity engine specifically designed for Chinese text. It functions as a word embedding toolkit and tokenizer that extracts semantic meaning and identifies synonyms by calculating the conceptual closeness between words and sentences. The system provides a toolkit for Chinese word embedding and synonym discovery, allowing for the retrieval of semantically similar words to expand vocabulary. It distinguishes itself through a configuration-driven approach to model loading, which supports the integration of custom word embeddings t

This is a Chinese natural language processing toolkit providing a suite of tools for word segmentation, part-of-speech tagging, and named entity recognition. It includes a neural dependency parser for analyzing syntactic and semantic relationships between words and a machine learning training suite for creating custom linguistic models using annotated datasets. The toolkit distinguishes itself through its deployment flexibility, offering a dockerized server and a web service interface that exposes processing capabilities via API. It supports the use of pretrained models and allows for the int

ConvertX is a web-based file conversion management platform designed to transform documents, images, and video files between various formats. It utilizes system-level binary orchestration to execute conversion tasks, leveraging background worker threads to handle concurrent, high-volume bulk processing without blocking the user interface. The platform distinguishes itself through a comprehensive security and access control framework, which includes multi-user account management, session-based token authentication, and role-based permissions. Users can secure their output files with passwords

LAC is a Chinese lexical analysis engine and toolkit designed for joint word segmentation, part-of-speech tagging, and named entity recognition. It functions as a high-performance system that identifies word boundaries and grammatical categories using trained machine learning models. The project features a lightweight, compiled native runtime that enables on-device natural language processing and embedding into mobile applications. It includes model compression and conversion to optimize for resource-constrained environments and supports multi-threaded parallel execution to increase throughpu

Fengshenbang-LM is a Chinese language model ecosystem and pre-training framework designed for the development and fine-tuning of billion-parameter large language models. It serves as a natural language processing pipeline and cross-modal AI platform capable of generating content across different modalities, including text-to-image generation and protein structure prediction. The project provides a domain-specific model adapter for applying pretrained models to specialized industries such as healthcare, finance, and law. It utilizes a distributed configuration system and data sharding to manag

FileConverter is a batch processing utility designed to automate file format transformations and system-level file management. It functions as a metadata-aware engine that integrates directly into the operating system shell, allowing users to trigger conversion tasks through the file manager context menu. The tool distinguishes itself by combining format conversion with automated post-processing workflows. Users can define reusable conversion presets to ensure consistent quality and settings across batch operations. Furthermore, the system utilizes metadata from source files to dynamically ge

Calibre is a comprehensive suite for digital library management, serving as a centralized hub for organizing, converting, and editing e-book collections. It functions as a multi-purpose platform that combines a relational database for metadata tracking with a powerful processing engine capable of transforming document formats and restructuring internal markup. Beyond local management, the software acts as a content server, enabling users to host their libraries over a network for remote access and reading via standard web browsers. The project distinguishes itself through its deep extensibili

This project is a high-performance markdown-to-HTML parser designed for use in browser, server-side, and command-line environments. It functions as a configurable syntax processor that transforms plain text documents into structured web content, providing a flexible engine for rendering dynamic documentation and web-based text. The parser features a modular, extensible pipeline that allows developers to intercept the document transformation process at multiple stages. Through custom tokenization, rendering overrides, and lifecycle hooks, users can define unique syntax, modify the token stream

This project provides a comprehensive Chinese language corpus designed to support the training and fine-tuning of large language models. It serves as a structured natural language processing resource, offering a collection of text data that includes dialogue, customer service interactions, and creative writing. The dataset is organized into distinct thematic categories, allowing for targeted model development across specific conversational and narrative contexts. By providing information in standardized, schema-agnostic text formats, the collection ensures portability across various machine l

Fairseq is a PyTorch toolkit for sequence-to-sequence modeling, specializing in neural machine translation, automatic speech recognition, and large-scale language model training. It provides a framework for processing and aligning diverse data sources, including text, audio, and video, to support tasks such as speech-to-text conversion and multimodal sequence learning. The project is distinguished by its distributed training capabilities, which utilize parameter sharding, mixed-precision training, and CPU offloading to handle models that exceed single-device memory. It also includes specializ