vLLM is a high-throughput, memory-efficient inference and serving engine for large language models (LLMs). It implements PagedAttention for near-zero waste KV cache management, continuous batching, chunked prefill, prefix caching, CUDA/HIP graphs, speculative decoding, and torch.compile. The system serves OpenAI-compatible APIs (chat completions, completions, embeddings, rerank, speech-to-text), Anthropic Messages API, Cohere API, and custom pooling/classification endpoints. It supports 200+ model architectures including dense, MoE, hybrid SSM/attention, and multimodal models (vision, audio, video). The architecture is a hybrid Python/Rust system: a Python core handles model execution, scheduling, and kernel dispatch, while a Rust frontend provides a high-performance HTTP/gRPC server with middleware for auth, metrics, and load management. Key entry points include the `vllm` CLI (serve, bench, launch), the `LLM` class for offline inference, and the OpenAI-compatible API server. External dependencies include PyTorch, HuggingFace Transformers, FlashAttention, CUTLASS, Triton, NCCL/RCCL, and various quantization libraries. The project is production-grade with extensive CI/CD, Docker images, Helm charts, and deployment integrations for Kubernetes, Ray, SageMaker, and more.
Ray is a distributed compute framework that unifies task/actor execution, data processing, model training, hyperparameter tuning, RL, and model serving under a single runtime. The core is written in C++ (raylet, GCS, object store, scheduling) with a Python frontend, communicating via gRPC and Arrow-compatible plasma object storage. Production entry points include `ray.init()` for cluster bootstrap, `@ray.remote` for task/actor offload, Ray Serve for model deployment (FastAPI/DAG/gang-scheduling patterns), Ray Data for streaming ETL, Ray Train/Tune for distributed training and hyperparameter search, and RLlib for reinforcement learning. The architecture is a hybrid: a distributed systems platform (scheduler, object store, autoscaler) layered with high-level AI libraries, all built via Bazel with extensive CI (Docker images, build matrices, release tests). It is not a data research repo, though Ray Data and RLlib contain research-grade data pipeline and RL primitives; the dominant character is a production-grade application framework.
ExLlamaV3 is a high-performance inference library for running large language models on consumer GPUs, with a focus on flexible quantization (EXL3 trellis format) and parallel inference. The architecture is a modular Python library with a compiled C++/CUDA extension providing custom kernels for attention (paged, MLA, DSA, recurrent), MoE routing, quantization, and sampling. Key entry points include the `Model`, `Config`, `Tokenizer`, `Cache`, and `Generator` classes, as well as CLI tools for model conversion (`convert.py`), quantization optimization (`sc_optimize.py`), and evaluation (`eval/`). Data flows: model weights are loaded from safetensors, quantized offline, then at inference time the library manages paged KV caches, supports tensor parallelism across GPUs, and provides streaming generation with filters (Formatron, LLGuidance) and custom samplers. External dependencies include PyTorch, safetensors, tokenizers, Triton, and optional packages like xformers and llguidance. The codebase demonstrates strong engineering with reusable module abstractions, a comprehensive attention dispatch system, and a sophisticated quantization pipeline, though it carries some technical drift from unmaintained dependencies and a large surface of architecture-specific files.
AIBrix is a Kubernetes-native control plane and intelligent gateway for serving production LLM workloads. The repository implements two intertwined planes: a data plane composed of Envoy ext_proc gateway plugins (cmd/plugins, pkg/plugins/gateway) that intercept OpenAI-compatible inference requests and route them to vLLM, SGLang, or TensorRT-LLM replicas using algorithms such as least-load, prefix-cache-aware routing, SLO queueing, token-load tracking, and prefill/decode disaggregation; and a control plane of Kubernetes controllers (cmd/controllers) that reconcile custom resources for autoscaling (PodAutoscaler), LoRA adapters, model warmup/claims, KV caches, RayCluster fleets, and StormService/RoleSet orchestration. Supporting runtime code includes a Python metadata and batch-job service, a distributed KV-cache package, a GPU optimizer, and an OpenAI-compatible frontend.
BentoML is a machine-learning model serving framework and deployment platform. In production it lets developers define inference services via a decorator-based SDK (`@bentoml.service`, `@bentoml.api`), package those services into versioned "Bentos" with locked Python dependencies, and build OCI container images or deploy to Kubernetes, EC2, or BentoCloud. The serving layer supports both HTTP (Starlette/Uvicorn) and gRPC, with adaptive batching, model runners, Prometheus metrics, OpenTelemetry tracing, and pluggable monitoring. A comprehensive CLI (`bentoml serve/build/containerize/deploy`) is the primary operational surface, while the `_bentoml_sdk` and `_bentoml_impl` packages provide the user-facing and internal implementation layers respectively. Architecturally, the project separates concern into model registry, service definition, packaging, server workers, and cloud clients. Model artifacts are persisted in a versioned local store via framework adapters for PyTorch, TensorFlow, Transformers, ONNX, scikit-learn, MLflow, XGBoost, LightGBM, and others. Data flows from trained model artifacts into the registry, get referenced by service code, get bundled by `bentoml build` into self-contained Bento packages, and are served by Circus-managed worker processes. The core request path involves serialization, batching, resource allocation, and process supervision, all implemented in internal modules with generally clean abstractions. The repository is mid-migration from a legacy `bentoml.legacy.Service` API to a newer `_bentoml_sdk`, which creates visible duplication: two HTTP client stacks, multiple protobuf versions, and deprecation shims. This is a real maintenance cost but not a functional blocker; the modern SDK is actively developed, documented, and tested. The repo is clearly an application/framework rather than a data-research project; research-like artifacts are limited to test fixtures, training examples, and ML framework integration tests.
Cog is an open-source CLI and runtime for packaging machine learning models into production-ready Docker containers. It automatically generates HTTP inference servers with OpenAPI schema, handles GPU/CUDA/cuDNN compatibility, and provides a full weight management pipeline for large model artifacts. The project consists of a Go CLI (build, push, serve, predict, doctor, weights management), a Rust/PyO3 prediction server (coglet) with slot-based concurrency, a Python SDK for defining model interfaces (BaseRunner, Input, streaming), and a TypeScript playground UI. Key workflows include building OCI images from cog.yaml, pushing to registries, serving models via HTTP with SSE streaming, and managing model weights via content-addressed storage and lockfiles. The architecture is multi-language but cleanly separated, with well-defined protocols between components.
NanoFlow is a high-throughput LLM serving framework designed to optimize inference through intra-device parallelism and efficient resource utilization. It employs nano-batching (splitting requests at operation granularity) and asynchronous CPU scheduling to overlap compute, memory, and network operations within a single device. The framework supports tensor parallelism for multi-GPU deployment and includes advanced KV-cache management strategies, including offloading to SSDs to avoid recomputation. It is evaluated on models such as Llama2, Llama3, and Qwen2, demonstrating significant throughput improvements over baselines like TensorRT-LLM. The architecture is modular, comprising core components for execution and worker management, model-specific implementations, operation kernels with multiple backends (Torch, CUDA, etc.), KV-cache mechanisms, and utility functions for profiling and resource management. Key entry points include `run_llama3.py` for launching the service and various test scripts for multi-GPU scenarios. Data flows involve request processing through transformer layers, with overlapping of computation and communication via streams and asynchronous execution, managed by a topological executor. The framework relies on a complex dependency stack including NVIDIA CUDA, NCCL, CUTLASS, FlashInfer, MSCCL++, and Gurobi, which introduces maintenance challenges. However, it demonstrates strong engineering practices through reusable components such as the IOWrapper for tensor dependency management, a cross-platform profiler, and a linear programming-based buffer allocator. While active development continues, the repository faces challenges in dependency management and platform fragmentation, indicating a need for careful version control to maintain stability.
fastllm is a high-performance LLM inference engine written in C++ with CUDA/CPU backends, designed for local and server deployment of large language models. In production, it loads models from HuggingFace, GGUF, and custom formats, then executes inference with advanced features: quantization (FP8, INT4, NVFP4, GGUF), speculative decoding, multi-token prediction (MTP), Mixture-of-Experts (MoE), paged KV cache, tensor parallelism, NUMA-aware execution, and disk offloading. It serves models through multiple interfaces: a C++ CLI, a Python CLI (ftllm), an OpenAI-compatible HTTP API server (Chat Completions, Responses, Anthropic Messages), a WebUI, a desktop launcher with agent integrations (DeepSeek Harness, OpenCode, Codex, Claude Code), and an Android app. The system also includes model conversion/export tools, a launcher with hardware auto-recommendation, and a plugin system for UI customization.
Xinference is a production-grade model serving platform that provides OpenAI-compatible REST APIs for a wide range of AI models including LLMs, embeddings, rerank, image generation, video, audio/TTS, and world models. It supports multiple inference backends (vLLM, SGLang, transformers, MLX, llama.cpp) and can be deployed as a single process or a distributed cluster using an actor-based supervisor/worker architecture (xoscar). The system includes a web UI for model management, authentication/authorization with OAuth2 scopes, token-aware routing for load balancing, and monitoring capabilities. It bundles many third-party model implementations for turnkey deployment, which reduces setup friction but introduces maintenance overhead from vendored code. The architecture is a hybrid of a monolithic FastAPI server and a distributed actor system. Key entry points are CLI commands (xinference, xinference-supervisor, xinference-worker) and the REST API. The frontend is a Next.js application statically exported and served by the backend. The codebase demonstrates strong patterns in modular route registration, plugin-style model registries, and streaming response handling, but also carries significant technical debt from heavy vendoring, duplicated code across third-party integrations, and version-sensitive patches to upstream libraries like vLLM.
Made With ML is an educational data science project that implements an end-to-end ML pipeline for classifying GitHub repositories into categories (NLP, computer vision, MLOps, other) based on their titles and descriptions. The core data lineage flows from raw CSV datasets (dataset.csv, holdout.csv) through text cleaning, tokenization via SciBERT, and stratified splitting into train/val/test sets, followed by distributed training with PyTorch and Ray Train, hyperparameter tuning via HyperOpt, evaluation with slice-based metrics using Snorkel, and serving through a FastAPI/Ray Serve deployment. The project demonstrates MLOps patterns including MLflow experiment tracking, data validation with Great Expectations, behavioral model testing, and CI/CD via GitHub Actions with Anyscale for cloud execution. The codebase is well-structured with clear separation of concerns across data.py, train.py, tune.py, evaluate.py, predict.py, serve.py, models.py, and utils.py. It includes comprehensive tests at the code, data, and model levels. However, the project shows signs of staleness with pinned dependencies from 2023 (Ray 2.7.0, transformers 4.28.1, torch 2.0.0) that have known security vulnerabilities and compatibility issues with current Python versions. The data pipeline components—particularly the CustomPreprocessor, stratified splitting, and text cleaning—are genuinely reusable patterns. The evaluation framework with slice-based metrics and behavioral tests (invariance, directional expectations, minimum functionality) represents solid ML engineering practice. The deployment configuration is tightly coupled to Anyscale's proprietary platform, limiting portability.
The solo-cli is a command-line tool designed to deploy and manage Physical AI systems on local hardware. It integrates functionalities for serving large language models (via Ollama, vLLM, or llama.cpp), controlling robotic arms (through the LeRobot framework), authenticating with a custom model hub (Solo Hub), and fine-tuning models using external datasets. The application features a modular command structure with lazy-loaded dependencies to optimize startup time, cross-platform hardware detection for automated backend selection, and a layered configuration system. Additional utilities include Docker management, system information gathering, and filesystem-safe string handling for robotics data operations. The project also includes example implementations of Model Context Protocol (MCP) servers for various domains, demonstrating how to extend the CLI's capabilities for specialized AI workflows.
This repository is a curated, actively-maintained personal knowledge base of academic paper reading notes, focused on systems for machine learning, cloud computing, and related areas. It is organized by conference venue (e.g., OSDI, SOSP, MLSys) and by topic (e.g., LLM systems, GPU sharing, memory disaggregation). The core content consists of markdown files with structured summaries (TL;DR, background, design, evaluation) and cross-linked topic aggregation pages. The repository is the source for a GitBook static site and includes scripts for building and serving that site, as well as utilities for PDF text extraction. While not a data pipeline or software application, it functions as a high-quality research resource with consistent formatting, curation guidelines, and reusable templates for note-taking and paper list organization.