Awesome GitHub RepositoriesInteractive Model Interfaces

ChatGLM-6B is an open-source bilingual large language model designed for natural dialogue and text generation in both English and Chinese. It is structured as a dialogue model capable of tasks such as role-playing and information extraction. The project provides implementations for quantized language models, using low-precision weights to reduce GPU memory requirements for local inference. It also supports parameter-efficient fine-tuning, allowing model behavior to be optimized for specific tasks without requiring full retraining. The model includes capabilities for local execution on GPUs a

Dive into LLMs is a framework designed for fine-tuning large language models and constructing modular machine learning pipelines. It provides a structured environment for adjusting pre-trained models on custom datasets while optimizing computational efficiency and training time. The project distinguishes itself by offering an interactive web interface that allows for the deployment and publication of trained models directly to a browser. This enables users to test and interact with model results through a standardized web-based environment. The platform supports the creation of flexible work

This project is a comprehensive toolkit for adapting large language models to the Chinese language, providing a specialized framework for fine-tuning, inference, and local deployment. It serves as a coordinated suite for language-specific adaptation, including tools for expanding tokenizers and implementing retrieval-augmented generation. The project distinguishes itself through a complete pipeline for model adaptation, featuring multilingual tokenizer expansion and a fine-tuning framework that supports instruction-based supervised training and adapter merging. It also includes a dedicated de

ChatGLM3 is a comprehensive framework for deploying, fine-tuning, and serving large language models. It functions as a high-performance inference engine designed to support conversational AI, enabling developers to build interactive agents capable of multi-turn dialogue, autonomous code execution, and structured tool invocation. The project distinguishes itself through its focus on hardware-agnostic deployment and resource optimization. It supports distributed model parallelism across multiple graphics cards, paged key-value caching for concurrent request processing, and weight quantization t

LitGPT is a training and deployment framework for large language models, providing a suite of tools for pretraining, finetuning, quantizing, evaluating, and serving models within a production environment. It includes a dedicated training pipeline for adapting pretrained models to specific tasks, a quantization tool for reducing weight precision, and an inference server for hosting models via web interfaces. The framework supports high-performance model development through custom architecture implementation and the use of predefined recipes to standardize pretraining and finetuning. It enables

Qwen3-TTS is a large language model text-to-speech engine designed to convert written text into natural-sounding human speech. It functions as an audio tokenizer and a generative system for speech synthesis. The project features a promptable voice designer for creating synthetic vocal personas based on natural language descriptions. It also includes a zero-shot voice cloning tool that mimics a target speaker using a short reference audio clip and a transcript. The system provides a framework for speech model fine-tuning to improve speaker likeness and quality through supervised training. Add

ParlAI is a conversational AI research framework designed for training, evaluating, and sharing dialogue models using a unified interface for datasets and agents. It functions as a PyTorch-based training platform and a dialogue data collection system, providing a centralized model zoo for the distribution of versioned pretrained agents. The project distinguishes itself through a knowledge-grounded retrieval system that combines dense and sparse indexing to ground responses in external information. It also provides a comprehensive infrastructure for gathering human-AI interaction data via inte

This project is a comprehensive Node.js software development kit designed for integrating large language models into applications. It serves as a foundational client for interacting with REST and WebSocket services, enabling developers to implement chat functionality, multimodal content generation, and autonomous agent orchestration. The library provides a structured framework for defining executable tools and enforcing JSON schemas, ensuring that model outputs remain programmatically compatible with downstream systems. The SDK distinguishes itself through its robust request orchestration and

Moshi is a real-time voice foundation model and speech-to-speech framework designed for bidirectional, low-latency conversations. It functions as a full-duplex voice interface that processes audio and text concurrently in a single stream, enabling natural human-machine dialogue without sequential processing delays. The system utilizes a neural audio codec to compress high-fidelity audio into low-bitrate tokens for efficient transmission. To manage complex responses and reasoning, it employs internal monologue modeling, which generates a hidden stream of thought tokens alongside audible speech

This project is a web-based user interface for interacting with large language models via API keys. It functions as an OpenAI API client and a general LLM web chat interface, allowing users to send prompts and receive responses through a private web portal. The application features a security layer with password-based access control to restrict public usage. It supports custom request routing and proxy configurations to bypass network restrictions, and it is available as a progressive web app for native-like installation on mobile devices. The interface includes rich text rendering for Markd

This project provides a Chinese large language model based on the LLaMA architecture. It is an instruction-tuned model optimized for natural language processing and multi-turn conversations in Chinese. The system includes a framework for parameter-efficient fine-tuning using low-rank adaptation and quantization to reduce memory requirements. It also implements retrieval augmented generation for local document question answering and supports long-context processing for sequences up to 64K tokens. The project covers a broad set of capabilities including supervised instruction tuning, reinforce

llm-zoomcamp is a comprehensive educational program and course for building real-life AI systems using large language models. It serves as a structured curriculum and implementation guide for developing AI applications and retrieval techniques. The project provides instructional material on building retrieval augmented generation pipelines to ground model responses in custom knowledge bases. It includes training on vector database implementation, semantic search, and the use of function calling to create autonomous agentic workflows. The curriculum covers a broad range of system development

Ce projet est une série de tutoriels de deep learning et un programme éducatif conçu pour enseigner les fondamentaux de PyTorch. Il sert de guide de formation structuré pour maîtriser l'architecture des réseaux de neurones, la différenciation automatique et l'utilisation des tenseurs et des graphes de calcul dynamiques. Le programme se concentre sur des implémentations pratiques, guidant spécifiquement le développement de systèmes de recommandation, de modèles publicitaires et de réseaux d'intérêt pour prédire les préférences des utilisateurs. Il fournit également un contenu pédagogique pour la prévision de séries temporelles et le traitement de données séquentielles. Le matériel couvre un large éventail de capacités de deep learning, incluant la construction de modèles pour la classification d'images et de texte ainsi que pour les données structurées. Il intègre des workflows pour l'accélération GPU, la visualisation des métriques d'entraînement et la création d'interfaces web pour tester les prédictions des modèles. Le projet est livré sous forme de collection de Jupyter Notebooks.

Ce projet est un cursus éducatif en machine learning et une plateforme d'apprentissage délivrée via des Jupyter Notebooks interactifs. Il sert de guide complet pour maîtriser le toolkit de science des données Python, fournissant des tutoriels structurés pour le calcul numérique, la manipulation de données tabulaires et la visualisation statistique. Le cursus inclut des guides d'implémentation spécifiques pour Scikit-Learn et un cours pratique sur TensorFlow pour construire, entraîner et déployer des réseaux de neurones et des modèles de vision par ordinateur. Il couvre le processus de bout en bout de la construction de modèles prédictifs, de la formulation initiale du problème et de la catégorisation des tâches au déploiement des modèles via des interfaces web interactives. Le projet couvre une large surface de capacités incluant le calcul numérique avec des tableaux multidimensionnels, l'analyse exploratoire des données et les routines de prétraitement des données. Il fournit des flux de travail détaillés pour l'apprentissage supervisé et non supervisé, les pipelines de machine learning automatisés, l'optimisation des hyperparamètres et l'évaluation des modèles utilisant des métriques de classification et la validation croisée. Le contenu éducatif est organisé sous forme d'une série de notebooks qui entremêlent code Python et explications narratives pour documenter les flux de travail en science des données.

freegpt-webui est une interface web auto-hébergée pour interagir avec des grands modèles de langage. Il fournit un frontend basé sur le chat conçu pour communiquer avec les modèles GPT 3.5 et GPT 4. L'application permet un chat sans clé API, permettant aux utilisateurs d'accéder à l'IA conversationnelle pour la génération de texte et la récupération d'informations sans fournir ou gérer de clés d'authentification personnelles. Le système gère l'intégration des modèles via une passerelle de proxy inverse et prend en charge le traitement de flux asynchrone pour la génération de texte en temps réel. Les préférences utilisateur et l'historique des conversations sont persistés via le stockage de session côté client.

InfiniteTalk is an open-source system for generating talking head videos driven by audio input. It synthesizes realistic lip movements, head poses, and facial expressions synchronized to a spoken audio track, using either a single still image or a small set of reference video frames as the visual source. The system can produce videos of arbitrary length while maintaining temporal coherence, and it supports animating multiple subjects in a single scene. A key differentiator is the ability to coordinate multiple talking subjects through a structured JSON description, giving each independent lip

Ramalama is a containerized runtime and management tool for large language models. It functions as an OCI AI model manager and registry client, allowing users to package, distribute, and execute AI models as standardized container images. The project differentiates itself by using OCI-compliant distribution for models and retrieval augmented generation assets, enabling the packaging of vector databases into immutable container images. It features hardware-aware image selection that automatically detects GPU or CPU capabilities to pull the most optimized image for the host environment. The sy

Ce projet est une interface web auto-hébergée et une application de bureau conçue pour interagir avec des modèles de langage. Il fournit une plateforme privée pour gérer les sessions de conversation, permettant aux utilisateurs de se connecter à des services d'IA externes tout en gardant le contrôle sur leur historique d'interaction et leurs paramètres de configuration. L'application se distingue en offrant une interface unifiée qui prend en charge les entrées et sorties multimodales, incluant le traitement de l'interaction vocale et la création d'images génératives. Elle sécurise les identifiants sensibles en acheminant les requêtes via un proxy backend et garantit la confidentialité des données en stockant les journaux de conversation et l'historique des sessions localement sur l'appareil de l'utilisateur. Au-delà de la fonctionnalité de chat de base, la plateforme inclut des outils pour le streaming en temps réel des réponses, la gestion de l'historique des conversations et la possibilité d'exporter des journaux pour archivage. Le logiciel est packagé sous forme d'exécutable de bureau autonome, permettant aux utilisateurs d'accéder à ces services d'IA indépendamment d'un navigateur web.

This project provides a web-based integrated development environment for defining, documenting, and simulating software interface specifications. It serves as a browser-based modeling tool that enables teams to create structured API contracts using the RAML modeling language. The environment distinguishes itself through its modular design, which allows the modeling interface to be embedded directly into existing web applications and developer portals. It supports a plugin architecture that enables the integration of custom persistence layers and metadata handlers, allowing teams to attach pro