Open-source alternatives to NeMo

Sglang is a high-performance inference engine and serving system designed for large language and multimodal models. It provides a programmable interface for orchestrating complex generation workflows, enabling developers to coordinate multi-turn dialogues, tool invocations, and reasoning chains through a domain-specific language. The platform is built to support production-scale deployments, offering an OpenAI-compatible API that allows for integration with existing application ecosystems. The system distinguishes itself through a disaggregated architecture that separates compute-intensive pr

gpt-neox is a distributed training system and framework for building large-scale autoregressive language models. It implements the transformer architecture and provides a toolkit for training models with billions of parameters by distributing weights across compute clusters. The framework distinguishes itself through extensive support for distributed model parallelism, including pipeline and sequence parallelism, to overcome single-device memory limits. It further supports sparse model architectures using a mixture of experts system with Sinkhorn-based routing. The project covers a broad ran

DeepSpeed is a distributed deep learning optimization library and framework designed for the training and inference of massive AI models. It serves as a model parallelism orchestrator and a toolkit for scaling large language models across multiple GPUs and compute nodes. The project distinguishes itself through 3D parallelism orchestration, which combines data, pipeline, and tensor parallelism. It utilizes ZeRO-based memory partitioning to eliminate redundant storage and employs CPU-offload memory management to move weights and optimizer states to system RAM. Additionally, it provides special

zero_nlp is a distributed framework for training and fine-tuning large language models and multimodal architectures. It provides a specialized toolkit for distributed model parallelism, allowing neural network layers and weights to be partitioned across multiple GPU devices to train models that exceed the memory capacity of a single processor. The project distinguishes itself through a combination of high-throughput data pipelines and parameter-efficient tuning. It utilizes multi-threading and memory mapping to preprocess and stream datasets exceeding 100GB and implements memory-saving adapta

AudioGPT is an LLM-driven audio framework and processing suite that uses large language models to orchestrate neural audio pipelines. It functions as a multimodal audio generator and processing system, integrating a collection of pretrained models to handle speech synthesis, sound generation, and audio manipulation. The system is distinguished by its ability to generate audio from diverse inputs, including text and images, and its capacity to produce synchronized talking head videos. It also operates as a neural speech translator, converting spoken language between different tongues while pre

ZenML is an orchestration platform designed for building, deploying, and monitoring reproducible machine learning pipelines and agentic workflows. It provides a unified framework that manages the entire lifecycle of machine learning assets, from data processing and model training to the deployment of persistent inference services. By decoupling pipeline logic from underlying compute and storage, the platform enables teams to transition workflows seamlessly from local development environments to production-grade cloud infrastructure. The platform distinguishes itself through a service-oriented

LiveKit is a comprehensive framework for building and orchestrating real-time, multimodal AI agents that interact with users through voice, video, and text. It provides a centralized, event-driven architecture to manage the entire lifecycle of automated participants, from initialization and session state management to graceful shutdown. By utilizing a selective forwarding unit, the platform efficiently routes media streams between participants and agents, ensuring low-latency communication and secure, token-based authentication for all connections. The platform distinguishes itself through it

SpeechBrain is an all-in-one deep learning toolkit designed for speech and audio processing. Built as a modular library, it provides a structured environment for developing, training, and deploying neural network models across a wide range of tasks, including automatic speech recognition, speaker identification, and audio enhancement. The framework distinguishes itself through a configuration-driven approach that separates model architecture and training hyperparameters from application logic. By utilizing externalized configuration files and standardized recipes, it enables reproducible rese

PocketFlow is a graph-based framework for designing and executing large language model operations and reasoning patterns. It serves as an orchestrator for building goal-oriented autonomous agents, multi-agent systems, and retrieval-augmented generation pipelines. The system is distinguished by its ability to coordinate autonomous AI agents that use shared memory and tools to solve complex goals, supported by a structured output engine that enforces schema-consistent responses. It utilizes graph-based workflow orchestration to manage sequences of model operations and supports supervisor-based

This project is a Python framework for building autonomous, event-driven agent systems. It provides a unified runtime for orchestrating multi-agent workflows, managing persistent conversation state, and executing code within secure, isolated sandbox environments. The framework is designed to handle complex task delegation, allowing agents to invoke other agents as tools while maintaining context across multi-turn interactions. The framework distinguishes itself through its deep integration with the Model Context Protocol, enabling agents to connect to external data sources and remote services

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

Metaseq is a transformer sequence modeling toolkit designed for training, fine-tuning, and deploying sequence-to-sequence models using open pre-trained weights. It provides a comprehensive framework for large language model training, including dedicated tools for sequence dataset processing and a standalone inference server for generating text via API requests. The project features specialized utilities for model quantization to reduce parameter precision to eight bits, which lowers memory usage and increases inference speed. It also includes a checkpoint conversion pipeline to transform mode

Megatron-LM is a distributed transformer training library and large language model training framework designed to scale models across thousands of GPUs. It functions as a GPU-optimized deep learning toolkit and a scaling engine for mixture-of-experts architectures, enabling the training of models with hundreds of billions of parameters. The project implements multi-dimensional model parallelism, combining tensor, pipeline, data, expert, and context-based workload distribution. It specifically optimizes mixture-of-experts architectures through integrated memory and communication improvements t

PaddleSpeech is a comprehensive toolkit of neural models for speech recognition, synthesis, and translation built on the PaddlePaddle deep learning framework. It provides a collection of frameworks and tools for converting spoken audio into written text, synthesizing natural audio from text, and performing direct speech translation. The toolkit includes specialized capabilities for keyword spotting to detect trigger words and speaker verification systems that extract unique voiceprints to identify and distinguish between individuals. It also features end-to-end translation tools that map audi

This project is an implementation of the Transformer-XL language model, a neural network architecture designed for long-context language modeling. It provides frameworks for training and deploying models that capture long-term dependencies and relationships in text sequences that extend beyond a fixed context window. The implementation supports both PyTorch and TensorFlow, allowing for distributed training across multiple GPUs and host nodes. It employs a recurrent mechanism to maintain coherence in extended sequences, utilizing segment-level recurrence and state-based memory reuse. The code

Pipecat is a framework and software development kit for building real-time multimodal AI agents and speech-to-speech systems. It utilizes a frame-based data pipeline to route audio, video, and text through a modular sequence of processors, enabling the orchestration of low-latency conversational AI. The project is distinguished by its ability to coordinate complex multimodal services, including speech-to-text, language models, and text-to-speech, within a single pipeline. It features semantic voice activity detection for natural turn-taking, state-machine conversation flows for dialogue manag

ZenML is an extensible machine learning orchestration framework designed to manage the end-to-end lifecycle of data pipelines and AI agent workflows. It functions as a durable orchestrator that executes machine learning tasks as directed acyclic graphs, ensuring that every step is containerized for consistent performance across local, cloud, and hybrid infrastructure. By decoupling pipeline code from underlying compute and storage backends, the platform allows developers to define infrastructure-agnostic stacks that remain portable across diverse environments. The project distinguishes itself

Intel XPU LLM Acceleration Library is a toolkit designed to accelerate large language model inference and finetuning on Intel CPUs, GPUs, and NPUs. It provides a distributed inference engine for scaling models across multiple accelerators, a multimodal model runtime for vision and speech tasks, and a low-bit model quantization tool for converting weights into INT4, FP8, and GGUF formats. The project features a parameter-efficient finetuning framework that enables model adaptation using QLoRA and DPO on Intel hardware. It distinguishes itself by providing specialized optimizations for Intel XP

AISystem is a comprehensive AI full-stack infrastructure project covering the entire pipeline from AI chip architecture to high-level training frameworks. It encompasses the development of AI compiler frameworks, inference engines, and distributed training orchestrators designed to coordinate workloads across a heterogeneous compute stack of CPUs, GPUs, and NPUs. The project focuses on the deep integration of software and hardware, employing software-hardware co-design to align tensor layouts with physical memory structures. It provides specialized capabilities for accelerating Transformer mo

Pythia is a multimodal research framework and distributed training system designed for building, training, and evaluating large models that combine visual and linguistic data. It provides a modular environment for developing vision-language models, focusing on the integration of image and text inputs into shared feature representations. The framework utilizes a modular architecture that decouples model building blocks into interchangeable components, allowing for flexible configuration of vision and language modules. It includes a benchmark suite for executing reference models against standar

ERNIE is a development toolkit for training, fine-tuning, and deploying large language models built on the PaddlePaddle deep learning platform. It provides a comprehensive suite of core components, including an inference server for vision and language models, a training and fine-tuning toolkit, and a framework for building retrieval-augmented generation systems using private knowledge bases. The project features multimodal AI models capable of reasoning across text, images, and video to perform complex visual understanding and information extraction. It distinguishes itself through specialize

This project is an AI-powered IDE extension and LLM coding assistant that provides a conversational interface for generating, refactoring, and debugging code. It functions as an AI agent framework and a Model Context Protocol client, connecting AI models to external data sources and tools to automate complex development tasks. The system is distinguished by its use of autonomous AI agents capable of multi-step task execution, including the ability to read files, modify code, and run terminal commands iteratively. It supports recursive agent orchestration through subagent delegation and employ

This project is a comprehensive educational curriculum and structured learning path covering the full lifecycle of large language models. It provides a guided progression through the theory, architecture, training, and deployment of these models. The curriculum includes specialized guides on transformer architecture, model training tutorials, and frameworks for designing autonomous agents. It also provides dedicated resources for studying model safety and ethics. The material covers a wide range of technical capabilities, including distributed training strategies, parameter-efficient fine-tu

DeepSpeedExamples is a collection of reference implementations and scripts for training, fine-tuning, and executing inference on large-scale AI models using DeepSpeed optimization. It provides a distributed model training guide and practical workflows for adapting large language models through memory-efficient techniques. The repository includes specialized implementations for pipeline parallelism to handle models exceeding single GPU memory and a suite of examples for ZeRO memory optimization to reduce per-device overhead. It also features standardized test suites for benchmarking the throug

TTS-WebUI is a web interface and speech synthesis manager designed to convert written text into spoken audio files. It serves as a self-hosted audio AI suite that allows users to configure speech synthesis models, manage speaker profiles, and generate audio through a graphical dashboard. The system functions as both a visual manager and a generative audio API, providing standardized endpoints and OpenAI-compatible request formats for external applications to trigger synthesis programmatically. It includes a plugin-based extension system that allows new tools and models to be added via externa

Mastra is an orchestration framework designed for building, deploying, and managing autonomous AI agents and multi-agent systems. It provides a comprehensive suite of primitives for creating resilient AI applications, including durable workflow orchestration, event-driven agent loops, and semantic memory management. By integrating these core components, the platform enables developers to build complex, multi-step processes that can reason about goals and execute tasks without manual intervention. The framework distinguishes itself through its focus on observability and secure, isolated execut

WhisperX is an automated speech recognition toolkit designed to convert spoken audio into text while maintaining precise synchronization with the original media. It functions as an integrated pipeline that combines transcription, phoneme-based alignment, and speaker diarization to produce structured, attributed transcripts. The project distinguishes itself through its use of forced alignment, which matches existing text to audio signals at the phoneme level to generate accurate word-level timestamps. It also incorporates speaker diarization to identify and label unique voices within a recordi