[
    {
        "github_repo_link": "https://github.com/pytorch/pytorch",
        "repo_name": "pytorch",
        "repo_description": "Tensors and Dynamic neural networks in Python with strong GPU acceleration",
        "homepage_link": "https://pytorch.org",
        "closest_github_tag": "machine-learning",
        "category": "machine learning framework"
    },
    {
        "github_repo_link": "https://github.com/ggml-org/llama.cpp",
        "repo_name": "llama.cpp",
        "repo_description": "LLM inference in C/C++",
        "closest_github_tag": "ggml",
        "category": "inference engine"
    },
    {
        "github_repo_link": "https://github.com/onnx/onnx",
        "repo_name": "onnx",
        "repo_description": "Open standard for machine learning interoperability",
        "homepage_link": "https://onnx.ai/",
        "closest_github_tag": "deep-learning"
    },
    {
        "github_repo_link": "https://github.com/ray-project/ray",
        "repo_name": "ray",
        "repo_description": "Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.",
        "homepage_link": "https://ray.io",
        "closest_github_tag": "deep-learning"
    },
    {
        "github_repo_link": "https://github.com/vllm-project/vllm",
        "repo_name": "vllm",
        "repo_description": "A high-throughput and memory-efficient inference and serving engine for LLMs",
        "homepage_link": "https://docs.vllm.ai",
        "closest_github_tag": "inference",
        "category": "inference engine"
    },
    {
        "github_repo_link": "https://github.com/ollama/ollama",
        "repo_name": "ollama",
        "repo_description": "Get up and running with OpenAI gpt-oss, DeepSeek-R1, Gemma 3 and other models.",
        "homepage_link": "https://ollama.com",
        "closest_github_tag": "llms",
        "category": "inference engine"
    },
    {
        "github_repo_link": "https://github.com/sgl-project/sglang",
        "repo_name": "sglang",
        "repo_description": "SGLang is a fast serving framework for large language models and vision language models.",
        "homepage_link": "https://docs.sglang.ai/",
        "closest_github_tag": "inference",
        "category": "inference engine"
    },
    {
        "github_repo_link": "https://github.com/modular/modular",
        "repo_name": "modular",
        "repo_description": "The Modular Platform (includes MAX & Mojo)",
        "homepage_link": "https://docs.modular.com/",
        "closest_github_tag": "mojo"
    },
    {
        "github_repo_link": "https://github.com/pytorch/ao",
        "repo_name": "ao",
        "repo_description": "PyTorch native quantization and sparsity for training and inference",
        "homepage_link": "https://pytorch.org/ao/stable/index.html",
        "closest_github_tag": "quantization"
    },
    {
        "github_repo_link": "https://github.com/triton-lang/triton",
        "repo_name": "triton",
        "repo_description": "Development repository for the Triton language and compiler",
        "homepage_link": "https://triton-lang.org/",
        "category": "dsl"
    },
    {
        "github_repo_link": "https://github.com/HazyResearch/ThunderKittens",
        "repo_name": "ThunderKittens",
        "repo_description": "Tile primitives for speedy kernels"
    },
    {
        "github_repo_link": "https://github.com/gpu-mode/reference-kernels",
        "repo_name": "reference-kernels",
        "repo_description": "Official Problem Sets / Reference Kernels for the GPU MODE Leaderboard!",
        "homepage_link": "https://gpumode.com",
        "closest_github_tag": "gpu",
        "category": "kernels"
    },
    {
        "github_repo_link": "https://github.com/pytorch/executorch",
        "repo_name": "executorch",
        "repo_description": "On-device AI across mobile, embedded and edge for PyTorch",
        "homepage_link": "https://executorch.ai",
        "closest_github_tag": "mobile",
        "category": "model compiler"
    },
    {
        "github_repo_link": "https://github.com/guandeh17/Self-Forcing",
        "repo_name": "Self-Forcing",
        "repo_description": "Official codebase for \"Self Forcing: Bridging Training and Inference in Autoregressive Video Diffusion\" (NeurIPS 2025 Spotlight)",

    },
    {
        "github_repo_link": "https://github.com/cumulo-autumn/StreamDiffusion",
        "repo_name": "StreamDiffusion",
        "repo_description": "StreamDiffusion: A Pipeline-Level Solution for Real-Time Interactive Generation"
    },
    {
        "github_repo_link": "https://github.com/comfyanonymous/ComfyUI",
        "repo_name": "ComfyUI",
        "repo_description": "The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.",
        "homepage_link": "https://www.comfy.org/",
        "closest_github_tag": "stable-diffusion"
    },
    {
        "github_repo_link": "https://github.com/Jeff-LiangF/streamv2v",
        "repo_name": "streamv2v",
        "repo_description": "Official Pytorch implementation of StreamV2V. ",
        "homepage_link": "https://jeff-liangf.github.io/projects/streamv2v/"
    },
    {
        "github_repo_link": "https://github.com/letta-ai/letta",
        "repo_name": "letta",
        "repo_description": "Letta is the platform for building stateful agents: open AI with advanced memory that can learn and self-improve over time.",
        "homepage_link": "https://docs.letta.com/",
        "closest_github_tag": "ai-agents"
    },
    {
        "github_repo_link": "https://github.com/jupyterlab/jupyterlab",
        "repo_name": "jupyterlab",
        "repo_description": "JupyterLab computational environment.",
        "homepage_link": "https://jupyterlab.readthedocs.io/",
        "closest_github_tag": "jupyter",
        "category": "ui"
    },
    {
        "github_repo_link": "https://github.com/ROCm/rocm-systems",
        "repo_name": "rocm-systems",
        "repo_description": "super repo for rocm systems projects"
    },
    {
        "github_repo_link": "https://github.com/NVIDIA/cutlass",
        "repo_name": "cutlass",
        "repo_description": "CUDA Templates and Python DSLs for High-Performance Linear Algebra",
        "homepage_link": "https://docs.nvidia.com/cutlass/index.html",
        "closest_github_tag": "cuda"
    },
    {
        "github_repo_link": "https://github.com/pytorch/helion",
        "repo_name": "helion",
        "repo_description": "A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.",
        "category": "dsl"
    },
    {
        "github_repo_link": "https://github.com/jax-ml/jax",
        "repo_name": "jax",
        "repo_description": "Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more",
        "homepage_link": "https://docs.jax.dev",
        "closest_github_tag": "jax"
    },
    {
        "github_repo_link": "https://github.com/tensorflow/tensorflow",
        "repo_name": "tensorflow",
        "repo_description": "An Open Source Machine Learning Framework for Everyone",
        "homepage_link": "https://tensorflow.org",
        "closest_github_tag": "deep-learning",
        "category": "machine learning framework"
    },
    {
        "github_repo_link": "https://github.com/deepspeedai/DeepSpeed",
        "repo_name": "DeepSpeed",
        "repo_description": "DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.",
        "homepage_link": "https://www.deepspeed.ai/",
        "closest_github_tag": "gpu"
    },
    {
        "github_repo_link": "https://github.com/triton-inference-server/server",
        "repo_name": "server",
        "repo_description": "The Triton Inference Server provides an optimized cloud and edge inferencing solution. ",
        "homepage_link": "https://docs.nvidia.com/deeplearning/triton-inference-server/user-guide/docs/index.html",
        "closest_github_tag": "inference"
    },
    {
        "github_repo_link": "https://github.com/ROCm/ROCm",
        "repo_name": "ROCm",
        "repo_description": "AMD ROCm™ Software - GitHub Home",
        "homepage_link": "https://rocm.docs.amd.com",
        "closest_github_tag": "documentation"
    },
    {
        "github_repo_link": "https://github.com/llvm/llvm-project",
        "repo_name": "llvm-project",
        "repo_description": "The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.",
        "homepage_link": "http://llvm.org",
        "category": "compiler"
    },
    {
        "github_repo_link": "https://github.com/cwpearson/cupti",
        "repo_name": "cupti",
        "repo_description": "Profile how CUDA applications create and modify data in memory.",
        "category": "profiler"
    },
    {
        "github_repo_link": "https://github.com/LLNL/hatchet",
        "repo_name": "hatchet",
        "repo_description": "Graph-indexed Pandas DataFrames for analyzing hierarchical performance data",
        "homepage_link": "https://llnl-hatchet.readthedocs.io",
        "closest_github_tag": "performance",
        "category": "profiler"
    },
    {
        "github_repo_link": "https://github.com/toyaix/triton-runner",
        "repo_name": "triton-runner",
        "repo_description": "Multi-Level Triton Runner supporting Python, IR, PTX, and cubin.",
        "homepage_link": "https://triton-runner.org",
        "closest_github_tag": "triton"
    },
    {
        "github_repo_link": "https://github.com/ByteDance-Seed/Triton-distributed",
        "repo_name": "Triton-distributed",
        "repo_description": "Distributed Compiler based on Triton for Parallel Systems",
        "homepage_link": "https://triton-distributed.readthedocs.io/en/latest/",
        "category": "model compiler"
    },
    {
        "github_repo_link": "https://github.com/linkedin/Liger-Kernel",
        "repo_name": "Liger-Kernel",
        "repo_description": "Efficient Triton Kernels for LLM Training",
        "homepage_link": "https://openreview.net/pdf?id=36SjAIT42G",
        "closest_github_tag": "triton",
        "category": "kernels"
    },
    {
        "github_repo_link": "https://github.com/thunlp/TritonBench",
        "repo_name": "TritonBench",
        "repo_description": "TritonBench: Benchmarking Large Language Model Capabilities for Generating Triton Operators",
        "category": "benchmark"
    },
    {
        "github_repo_link": "https://github.com/meta-pytorch/tritonparse",
        "repo_name": "tritonparse",
        "repo_description": "TritonParse: A Compiler Tracer, Visualizer, and Reproducer for Triton Kernels",
        "homepage_link": "https://meta-pytorch.org/tritonparse/",
        "closest_github_tag": "triton"
    },
    {
        "github_repo_link": "https://github.com/elastic/elasticsearch",
        "repo_name": "elasticsearch",
        "repo_description": "Free and Open Source, Distributed, RESTful Search Engine",
        "homepage_link": "https://www.elastic.co/products/elasticsearch",
        "closest_github_tag": "search-engine",
        "category": "search engine"
    },
    {
        "github_repo_link": "https://github.com/kubernetes/kubernetes",
        "repo_name": "kubernetes",
        "repo_description": "Production-Grade Container Scheduling and Management",
        "homepage_link": "https://kubernetes.io",
        "closest_github_tag": "containers"
    },
    {
        "github_repo_link": "https://github.com/modelcontextprotocol/modelcontextprotocol",
        "repo_name": "modelcontextprotocol",
        "repo_description": "Specification and documentation for the Model Context Protocol",
        "homepage_link": "https://modelcontextprotocol.io"
    },
    {
        "github_repo_link": "https://github.com/lastmile-ai/mcp-agent",
        "repo_name": "mcp-agent",
        "repo_description": "Build effective agents using Model Context Protocol and simple workflow patterns",
        "closest_github_tag": "ai-agents"
    },
    {
        "github_repo_link": "https://github.com/milvus-io/milvus",
        "repo_name": "milvus",
        "repo_description": "Milvus is a high-performance, cloud-native vector database built for scalable vector ANN search",
        "homepage_link": "https://milvus.io",
        "closest_github_tag": "vector-search",
        "category": "vector databse"
    },
    {
        "github_repo_link": "https://github.com/gaoj0017/RaBitQ",
        "repo_name": "RaBitQ",
        "repo_description": "[SIGMOD 2024] RaBitQ: Quantizing High-Dimensional Vectors with a Theoretical Error Bound for Approximate Nearest Neighbor Search",
        "homepage_link": "https://github.com/VectorDB-NTU/RaBitQ-Library",
        "closest_github_tag": "nearest-neighbor-search"
    },
    {
        "github_repo_link": "https://github.com/Airtable/airtable.js",
        "repo_name": "airtable.js",
        "repo_description": "Airtable javascript client"
    },
    {
        "github_repo_link": "https://github.com/mistralai/mistral-inference",
        "repo_name": "mistral-inference",
        "repo_description": "Official inference library for Mistral models",
        "homepage_link": "https://mistral.ai/",
        "closest_github_tag": "llm-inference",
        "category": "inference engine"
    },
    {
        "github_repo_link": "https://github.com/dstackai/dstack",
        "repo_name": "dstack",
        "repo_description": "dstack is an open-source control plane for running development, training, and inference jobs on GPUs—across hyperscalers, neoclouds, or on-prem.",
        "homepage_link": "https://dstack.ai",
        "closest_github_tag": "orchestration"
    },
    {
        "github_repo_link": "https://github.com/numpy/numpy",
        "repo_name": "numpy",
        "repo_description": "The fundamental package for scientific computing with Python.",
        "homepage_link": "https://numpy.org",
        "closest_github_tag": "python",
        "category": "python library"
    },
    {
        "github_repo_link": "https://github.com/scipy/scipy",
        "repo_name": "scipy",
        "repo_description": "SciPy library main repository",
        "homepage_link": "https://scipy.org",
        "closest_github_tag": "python",
        "category": "python library"
    },
    {
        "github_repo_link": "https://github.com/numba/numba",
        "repo_name": "numba",
        "repo_description": "NumPy aware dynamic Python compiler using LLVM",
        "homepage_link": "https://numba.pydata.org/",
        "closest_github_tag": "compiler"
    },
    {
        "github_repo_link": "https://github.com/sandialabs/torchdendrite",
        "repo_name": "torchdendrite",
        "repo_description": "Dendrites for PyTorch and SNNTorch neural networks ",
        "closest_github_tag": "scr-3078",
        "category": "machine learning framework"
    },
    {
        "github_repo_link": "https://github.com/Lightning-AI/lightning-thunder",
        "repo_name": "lightning-thunder",
        "repo_description": "PyTorch compiler that accelerates training and inference. Get built-in optimizations for performance, memory, parallelism, and easily write your own."
    },
    {
        "github_repo_link": "https://github.com/pytorch/torchdynamo",
        "repo_name": "torchdynamo",
        "repo_description": "A Python-level JIT compiler designed to make unmodified PyTorch programs faster."
    },
    {
        "github_repo_link": "https://github.com/microsoft/TileIR",
        "repo_name": "TileIR",
        "category": "dsl"
    },
    {
        "github_repo_link": "https://github.com/pytorch/torchtitan",
        "repo_name": "torchtitan",
        "repo_description": "A PyTorch native platform for training generative AI models"
    },
    {
        "github_repo_link": "https://github.com/NVIDIA/cudnn-frontend",
        "repo_name": "cudnn-frontend",
        "repo_description": "cudnn_frontend provides a c++ wrapper for the cudnn backend API and samples on how to use it"
    },
    {
        "github_repo_link": "https://github.com/pytorch/ort",
        "repo_name": "ort",
        "repo_description": "Accelerate PyTorch models with ONNX Runtime"
    },
    {
        "github_repo_link": "https://github.com/NVIDIA/nccl",
        "repo_name": "nccl",
        "repo_description": "Optimized primitives for collective multi-GPU communication",
        "homepage_link": "https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/index.html",
        "closest_github_tag": "cuda"
    },
    {
        "github_repo_link": "https://github.com/sgl-project/ome",
        "repo_name": "ome",
        "repo_description": "OME is a Kubernetes operator for enterprise-grade management and serving of Large Language Models (LLMs)",
        "homepage_link": "http://docs.sglang.ai/ome/",
        "closest_github_tag": "k8s"
    },
    {
        "github_repo_link": "https://github.com/volcengine/verl",
        "repo_name": "verl",
        "repo_description": "verl: Volcano Engine Reinforcement Learning for LLMs",
        "homepage_link": "https://verl.readthedocs.io/en/latest/index.html"
    },
    {
        "github_repo_link": "https://github.com/aws-neuron/neuronx-distributed-inference",
        "repo_name": "neuronx-distributed-inference",
        "category": "inference engine"
    },
    {
        "github_repo_link": "https://github.com/meta-pytorch/monarch",
        "repo_name": "monarch",
        "repo_description": "PyTorch Single Controller",
        "homepage_link": "https://meta-pytorch.org/monarch"
    },
    {
        "github_repo_link": "https://github.com/ai-dynamo/nixl",
        "repo_name": "nixl",
        "repo_description": "NVIDIA Inference Xfer Library (NIXL)"
    },
    {
        "github_repo_link": "https://github.com/LMCache/LMCache",
        "repo_name": "LMCache",
        "repo_description": "Supercharge Your LLM with the Fastest KV Cache Layer",
        "homepage_link": "https://lmcache.ai/",
        "closest_github_tag": "inference"
    },
    {
        "github_repo_link": "https://github.com/linux-rdma/rdma-core",
        "repo_name": "rdma-core",
        "repo_description": "RDMA core userspace libraries and daemons",
        "homepage_link": null,
        "closest_github_tag": "linux-kernel"
    },
    {
        "github_repo_link": "https://github.com/NVIDIA/TensorRT",
        "repo_name": "TensorRT",
        "repo_description": "NVIDIA® TensorRT™ is an SDK for high-performance deep learning inference on NVIDIA GPUs. This repository contains the open source components of TensorRT.",
        "homepage_link": "https://developer.nvidia.com/tensorrt",
        "closest_github_tag": "inference"
    },
    {
        "github_repo_link": "https://github.com/Cambridge-ICCS/FTorch",
        "repo_name": "FTorch",
        "repo_description": "A library for directly calling PyTorch ML models from Fortran.",
        "homepage_link": "https://cambridge-iccs.github.io/FTorch/",
        "closest_github_tag": "deep-learning"
    },
    {
        "github_repo_link": "https://github.com/facebook/hhvm",
        "repo_name": "hhvm",
        "repo_description": "A virtual machine for executing programs written in Hack.",
        "homepage_link": "https://hhvm.com",
        "closest_github_tag": "hack"
    },
    {
        "github_repo_link": "https://github.com/vosen/ZLUDA",
        "repo_name": "ZLUDA",
        "repo_description": "CUDA on non-NVIDIA GPUs",
        "homepage_link": "https://vosen.github.io/ZLUDA/",
        "closest_github_tag": "cuda"
    },
    {
        "github_repo_link": "https://github.com/vtsynergy/CU2CL",
        "repo_name": "CU2CL",
        "repo_description": "A prototype CUDA-to-OpenCL source-to-source translator, built on the Clang compiler framework",
        "homepage_link": "http://chrec.cs.vt.edu/cu2cl"
    },
    {
        "github_repo_link": "https://github.com/pocl/pocl",
        "repo_name": "pocl",
        "repo_description": "pocl - Portable Computing Language",
        "homepage_link": "https://portablecl.org",
        "closest_github_tag": "opencl"
    },
    {
        "github_repo_link": "https://github.com/apache/spark",
        "repo_name": "spark",
        "repo_description": "Apache Spark - A unified analytics engine for large-scale data processing",
        "homepage_link": "https://spark.apache.org/",
        "closest_github_tag": "big-data"
    },
    {
        "github_repo_link": "https://github.com/codelion/openevolve",
        "repo_name": "openevolve",
        "repo_description": "Open-source implementation of AlphaEvolve",
        "homepage_link": "",
        "closest_github_tag": "genetic-algorithm"
    },
    {
        "github_repo_link": "https://github.com/ROCm/hipBLAS",
        "repo_name": "hipBLAS",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-libraries repo",
        "homepage_link": "https://github.com/ROCm/rocm-libraries",
        "closest_github_tag": "hip"
    },
    {
        "github_repo_link": "https://github.com/ROCm/roctracer",
        "repo_name": "roctracer",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-systems repo ",
        "homepage_link": "https://github.com/ROCm/rocm-systems"
    },
    {
        "github_repo_link": "https://github.com/huggingface/peft",
        "repo_name": "peft",
        "repo_description": "🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.",
        "homepage_link": "https://huggingface.co/docs/peft",
        "closest_github_tag": "lora"
    },
    {
        "github_repo_link": "https://github.com/ROCm/hip",
        "repo_name": "hip",
        "repo_description": "HIP: C++ Heterogeneous-Compute Interface for Portability",
        "homepage_link": "https://rocmdocs.amd.com/projects/HIP/",
        "closest_github_tag": "hip"
    },
    {
        "github_repo_link": "https://github.com/ROCm/composable_kernel",
        "repo_name": "composable_kernel",
        "repo_description": "Composable Kernel: Performance Portable Programming Model for Machine Learning Tensor Operators",
        "homepage_link": "https://rocm.docs.amd.com/projects/composable_kernel/en/latest/"
    },
    {
        "github_repo_link": "https://github.com/ROCm/aiter",
        "repo_name": "aiter",
        "repo_description": "AI Tensor Engine for ROCm"
    },
    {
        "github_repo_link": "https://github.com/AMDResearch/intelliperf",
        "repo_name": "intelliperf",
        "repo_description": "Automated bottleneck detection and solution orchestration",
        "closest_github_tag": "performance"
    },
    {
        "github_repo_link": "https://github.com/AMD-AGI/GEAK-agent",
        "repo_name": "GEAK-agent",
        "repo_description": "It is an LLM-based AI agent, which can write correct and efficient gpu kernels automatically."
    },
    {
        "github_repo_link": "https://github.com/AMD-AGI/torchtitan",
        "repo_name": "torchtitan",
        "repo_description": "A PyTorch native platform for training generative AI models"
    },
    {
        "github_repo_link": "https://github.com/AMD-AGI/hipBLASLt",
        "repo_name": "hipBLASLt",
        "repo_description": "hipBLASLt is a library that provides general matrix-matrix operations with a flexible API and extends functionalities beyond a traditional BLAS library",
        "homepage_link": "https://rocm.docs.amd.com/projects/hipBLASLt/en/latest/index.html"
    },
    {
        "github_repo_link": "https://github.com/AMD-AGI/rocm-torchtitan",
        "repo_name": "rocm-torchtitan"
    },
    {
        "github_repo_link": "https://github.com/HazyResearch/Megakernels",
        "repo_name": "Megakernels",
        "repo_description": "kernels, of the mega variety"
    },
    {
        "github_repo_link": "https://github.com/huggingface/kernels",
        "repo_name": "kernels",
        "repo_description": "Load compute kernels from the Hub",
        "category": "kernels"
    },
    {
        "github_repo_link": "https://github.com/tile-ai/tilelang",
        "repo_name": "tilelang",
        "repo_description": " Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels",
        "homepage_link": "https://tilelang.com/",
        "category": "dsl"
    },
    {
        "github_repo_link": "https://github.com/opencv/opencv",
        "repo_name": "opencv",
        "repo_description": "Open Source Computer Vision Library",
        "homepage_link": "https://opencv.org",
        "closest_github_tag": "image-processing"
    },
    {
        "github_repo_link": "https://github.com/Lightning-AI/lightning-thunder",
        "repo_name": "lightning-thunder",
        "repo_description": "PyTorch compiler that accelerates training and inference. Get built-in optimizations for performance, memory, parallelism, and easily write your own."
    },
    {
        "github_repo_link": "https://github.com/tracel-ai/burn",
        "repo_name": "burn",
        "repo_description": "Burn is a next generation tensor library and Deep Learning Framework that doesn't compromise on flexibility, efficiency and portability.",
        "homepage_link": "https://burn.dev",
        "closest_github_tag": "machine-learning"
    },
    {
        "github_repo_link": "https://github.com/huggingface/kernels-community",
        "repo_name": "kernels-community",
        "repo_description": "Kernel sources for https://huggingface.co/kernels-community",
        "category": "kernels"
    },
    {
        "github_repo_link": "https://github.com/flashinfer-ai/flashinfer-bench",
        "repo_name": "flashinfer-bench",
        "repo_description": "Building the Virtuous Cycle for AI-driven LLM Systems",
        "homepage_link": "https://bench.flashinfer.ai",
        "category": "benchmark"
    },
    {
        "github_repo_link": "https://github.com/OSC/ondemand",
        "repo_name": "ondemand",
        "repo_description": "Supercomputing. Seamlessly. Open, Interactive HPC Via the Web",
        "homepage_link": "https://openondemand.org/",
        "closest_github_tag": "hpc"
    },
    {
        "github_repo_link": "https://github.com/flashinfer-ai/flashinfer",
        "repo_name": "flashinfer",
        "repo_description": "FlashInfer: Kernel Library for LLM Serving",
        "homepage_link": "https://flashinfer.ai",
        "closest_github_tag": "attention"
    },
    {
        "github_repo_link": "https://github.com/ScalingIntelligence/KernelBench",
        "repo_name": "KernelBench",
        "repo_description": "KernelBench: Can LLMs Write GPU Kernels? - Benchmark with Torch -> CUDA problems",
        "homepage_link": "https://scalingintelligence.stanford.edu/blogs/kernelbench/",
        "closest_github_tag": "benchmark",
        "category": "benchmark"
    },
    {
        "github_repo_link": "https://github.com/thunlp/TritonBench",
        "repo_name": "TritonBench",
        "repo_description": "TritonBench: Benchmarking Large Language Model Capabilities for Generating Triton Operators",
        "category": "benchmark"
    },
    {
        "github_repo_link": "https://github.com/AutomataLab/cuJSON",
        "repo_name": "cuJSON",
        "repo_description": "cuJSON: A Highly Parallel JSON Parser for GPUs"
    },
    {
        "github_repo_link": "https://github.com/Netflix/metaflow",
        "repo_name": "metaflow",
        "repo_description": "Build, Manage and Deploy AI/ML Systems",
        "homepage_link": "https://metaflow.org",
        "closest_github_tag": "machine-learning"
    },
    {
        "github_repo_link": "https://github.com/harmonic-ai/IMO2025",
        "repo_name": "IMO2025"
    },
    {
        "github_repo_link": "https://github.com/leanprover/lean4",
        "repo_name": "lean4",
        "repo_description": "Lean 4 programming language and theorem prover",
        "homepage_link": "https://lean-lang.org",
        "closest_github_tag": "lean"
    },
    {
        "github_repo_link": "https://github.com/NVIDIA/warp",
        "repo_name": "warp",
        "repo_description": "A Python framework for accelerated simulation, data generation and spatial computing.",
        "homepage_link": "https://nvidia.github.io/warp/",
        "closest_github_tag": "gpu"
    },
    {
        "github_repo_link": "https://github.com/NVIDIA/cuda-python",
        "repo_name": "cuda-python",
        "repo_description": "CUDA Python: Performance meets Productivity",
        "homepage_link": "https://nvidia.github.io/cuda-python/"
    },
    {
        "github_repo_link": "https://github.com/basetenlabs/truss",
        "repo_name": "truss",
        "repo_description": "The simplest way to serve AI/ML models in production",
        "homepage_link": "https://truss.baseten.co",
        "closest_github_tag": "machine-learning"
    },
    {
        "github_repo_link": "https://github.com/laude-institute/terminal-bench",
        "repo_name": "terminal-bench",
        "repo_description": "A benchmark for LLMs on complicated tasks in the terminal",
        "homepage_link": "https://www.tbench.ai",
        "category": "benchmark"
    },
    {
        "github_repo_link": "https://github.com/block/goose",
        "repo_name": "goose",
        "repo_description": "an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM",
        "homepage_link": "https://block.github.io/goose/",
        "closest_github_tag": "mcp",
        "category": "agent"
    },
    {
        "github_repo_link": "https://github.com/kvcache-ai/Mooncake",
        "repo_name": "Mooncake",
        "repo_description": "Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.",
        "homepage_link": "https://kvcache-ai.github.io/Mooncake/",
        "closest_github_tag": "inference"
    },
    {
        "github_repo_link": "https://github.com/SWE-bench/SWE-bench",
        "repo_name": "SWE-bench",
        "repo_description": "SWE-bench: Can Language Models Resolve Real-world Github Issues?",
        "homepage_link": "https://www.swebench.com",
        "closest_github_tag": "benchmark",
        "category": "benchmark"
    },
    {
        "github_repo_link": "https://github.com/Dao-AILab/quack",
        "repo_name": "quack",
        "repo_description": "A Quirky Assortment of CuTe Kernels",
        "category": "kernels"
    },
    {
        "github_repo_link": "https://github.com/KhronosGroup/SYCL-Docs",
        "repo_name": "SYCL-Docs",
        "repo_description": "SYCL Open Source Specification"
    },
    {
        "github_repo_link": "https://github.com/triSYCL/triSYCL",
        "repo_name": "triSYCL",
        "repo_description": " Generic system-wide modern C++ for heterogeneous platforms with SYCL from Khronos Group",
        "closest_github_tag": "opencl"
    },
    {
        "github_repo_link": "https://github.com/pybind/pybind11",
        "repo_name": "pybind11",
        "repo_description": "Seamless operability between C++11 and Python",
        "homepage_link": "https://pybind11.readthedocs.io/",
        "closest_github_tag": "bindings"
    },
    {
        "github_repo_link": "https://github.com/andreinechaev/nvcc4jupyter",
        "repo_name": "nvcc4jupyter",
        "repo_description": "A plugin for Jupyter Notebook to run CUDA C/C++ code",
        "category": "compiler"
    },
    {
        "github_repo_link": "https://github.com/ROCm/rocSOLVER",
        "repo_name": "rocSOLVER",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-libraries repo",
        "homepage_link": "https://github.com/ROCm/rocm-libraries",
        "closest_github_tag": "rocm"
    },
    {
        "github_repo_link": "https://github.com/ROCm/Tensile",
        "repo_name": "Tensile",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-libraries repo",
        "homepage_link": "https://github.com/ROCm/rocm-libraries",
        "closest_github_tag": "gpu"
    },
    {
        "github_repo_link": "https://github.com/ROCm/rocPRIM",
        "repo_name": "rocPRIM",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-libraries repo ",
        "homepage_link": "https://github.com/ROCm/rocm-libraries",
        "closest_github_tag": "hip"
    },
    {
        "github_repo_link": "https://github.com/ROCm/hipCUB",
        "repo_name": "hipCUB",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-libraries repo ",
        "homepage_link": "https://github.com/ROCm/rocm-libraries"
    },
    {
        "github_repo_link": "https://github.com/ROCm/rocFFT",
        "repo_name": "rocFFT",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-libraries repo",
        "homepage_link": "https://github.com/ROCm/rocm-libraries",
        "closest_github_tag": "hip"
    },
    {
        "github_repo_link": "https://github.com/ROCm/rocSPARSE",
        "repo_name": "rocSPARSE",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-libraries repo",
        "homepage_link": "https://github.com/ROCm/rocm-libraries"
    },
    {
        "github_repo_link": "https://github.com/ROCm/rocRAND",
        "repo_name": "rocRAND",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-libraries repo ",
        "homepage_link": "https://github.com/ROCm/rocm-libraries",
        "closest_github_tag": "hip"
    },
    {
        "github_repo_link": "https://github.com/ROCm/MIOpen",
        "repo_name": "MIOpen",
        "repo_description": "[DEPRECATED] Moved to ROCm/rocm-libraries repo",
        "homepage_link": "https://github.com/ROCm/rocm-libraries"
    },
    {
        "github_repo_link": "https://github.com/Reference-LAPACK/lapack",
        "repo_name": "lapack",
        "repo_description": "LAPACK development repository",
        "closest_github_tag": "linear-algebra"
    },
    {
        "github_repo_link": "https://github.com/ccache/ccache",
        "repo_name": "ccache",
        "repo_description": "ccache – a fast compiler cache",
        "homepage_link": "https://ccache.dev",
        "closest_github_tag": "compiler",
        "category": "compiler"
    },
    {
        "github_repo_link": "https://github.com/ROCm/omnitrace",
        "repo_name": "omnitrace",
        "repo_description": "Omnitrace: Application Profiling, Tracing, and Analysis",
        "homepage_link": "https://rocm.docs.amd.com/projects/omnitrace/en/docs-6.2.4/",
        "closest_github_tag": "performance-analysis"
    },
    {
        "github_repo_link": "https://github.com/KhronosGroup/OpenCL-SDK",
        "repo_name": "OpenCL-SDK",
        "repo_description": "OpenCL SDK"
    },
    {
        "github_repo_link": "https://github.com/meta-llama/synthetic-data-kit",
        "repo_name": "synthetic-data-kit",
        "repo_description": "Tool for generating high quality Synthetic datasets",
        "homepage_link": "https://pypi.org/project/synthetic-data-kit/",
        "closest_github_tag": "generation"
    },
    {
        "github_repo_link": "https://github.com/unslothai/unsloth",
        "repo_name": "unsloth",
        "repo_description": "Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek-R1, Qwen3, Gemma 3, TTS 2x faster with 70% less VRAM.",
        "homepage_link": "https://docs.unsloth.ai/",
        "closest_github_tag": "unsloth"
    },
    {
        "github_repo_link": "https://github.com/KhronosGroup/Vulkan-Docs",
        "repo_name": "Vulkan-Docs",
        "repo_description": "The Vulkan API Specification and related tools"
    },
    {
        "github_repo_link": "https://github.com/tensorflow/tflite-micro",
        "repo_name": "tflite-micro",
        "repo_description": "Infrastructure to enable deployment of ML models to low-power resource-constrained embedded targets (including microcontrollers and digital signal processors)."
    },
        {
        "github_repo_link": "https://github.com/Wan-Video/Wan2.2",
        "repo_name": "Wan2.2",
        "repo_description": "Wan: Open and Advanced Large-Scale Video Generative Models",
        "homepage_link": "https://wan.video",
        "closest_github_tag": "video-generation"
    },
    {
        "github_repo_link": "https://github.com/AMD-AGI/Primus-Turbo",
        "repo_name": "Primus-Turbo"
    }
]