diff --git "a/PyTorchConference2025_GithubRepos.json" "b/PyTorchConference2025_GithubRepos.json" --- "a/PyTorchConference2025_GithubRepos.json" +++ "b/PyTorchConference2025_GithubRepos.json" @@ -1,195 +1,4 @@ [ - { - "repo_name": "flashinfer-bench", - "repo_link": "https://github.com/flashinfer-ai/flashinfer-bench", - "category": "benchmark", - "github_about_section": "Building the Virtuous Cycle for AI-driven LLM Systems", - "homepage_link": "https://bench.flashinfer.ai", - "github_topic_closest_fit": "benchmark", - "contributors_all": 12, - "contributors_2025": 11, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 1.0, - "180-day-contributor-retention-rate": "not-enough-data" - }, - { - "repo_name": "cuJSON", - "repo_link": "https://github.com/AutomataLab/cuJSON", - "category": "library leveraging parallel compute", - "github_about_section": "cuJSON: A Highly Parallel JSON Parser for GPUs", - "homepage_link": "https://dl.acm.org/doi/10.1145/3760250.3762222", - "github_topic_closest_fit": "json-parser", - "contributors_all": 2, - "contributors_2025": 2, - "contributors_2024": 2, - "contributors_2023": 0, - "growth_2025_percent": 0, - "90-day-contributor-retention-rate": 1.0, - "180-day-contributor-retention-rate": 1.0 - }, - { - "repo_name": "triton-runner", - "repo_link": "https://github.com/toyaix/triton-runner", - "github_about_section": "Multi-Level Triton Runner supporting Python, IR, PTX, and cubin.", - "homepage_link": "https://triton-runner.org", - "contributors_all": 1, - "contributors_2025": 1, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 1.0, - "180-day-contributor-retention-rate": "not-enough-data" - }, - { - "repo_name": "reference-kernels", - "repo_link": "https://github.com/gpu-mode/reference-kernels", - "category": "kernel examples", - "github_about_section": "Official Problem Sets / Reference Kernels for the GPU MODE Leaderboard!", - "homepage_link": "https://gpumode.com", - "contributors_all": 16, - "contributors_2025": 16, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.8, - "180-day-contributor-retention-rate": 0.6666666666666666 - }, - { - "repo_name": "intelliperf", - "repo_link": "https://github.com/AMDResearch/intelliperf", - "category": "performance testing", - "github_about_section": "Automated bottleneck detection and solution orchestration", - "homepage_link": "https://arxiv.org/html/2508.20258v1", - "github_topic_closest_fit": "profiling", - "contributors_all": 7, - "contributors_2025": 7, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.8, - "180-day-contributor-retention-rate": 0.75 - }, - { - "repo_name": "TritonBench", - "repo_link": "https://github.com/thunlp/TritonBench", - "category": "benchmark", - "github_about_section": "TritonBench: Benchmarking Large Language Model Capabilities for Generating Triton Operators", - "homepage_link": "https://arxiv.org/abs/2502.14752", - "github_topic_closest_fit": "benchmark", - "contributors_all": 3, - "contributors_2025": 3, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.6666666666666666, - "180-day-contributor-retention-rate": 0.0 - }, - { - "repo_name": "CU2CL", - "repo_link": "https://github.com/vtsynergy/CU2CL", - "github_about_section": "A prototype CUDA-to-OpenCL source-to-source translator, built on the Clang compiler framework", - "homepage_link": "http://chrec.cs.vt.edu/cu2cl", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 3, - "contributors_2025": 0, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.6666666666666666, - "180-day-contributor-retention-rate": 0.6666666666666666 - }, - { - "repo_name": "FTorch", - "repo_link": "https://github.com/Cambridge-ICCS/FTorch", - "category": "wrapper", - "github_about_section": "A library for directly calling PyTorch ML models from Fortran.", - "homepage_link": "https://cambridge-iccs.github.io/FTorch", - "github_topic_closest_fit": "machine-learning", - "contributors_all": 20, - "contributors_2025": 11, - "contributors_2024": 8, - "contributors_2023": 9, - "growth_2025_percent": 37, - "90-day-contributor-retention-rate": 0.6470588235294118, - "180-day-contributor-retention-rate": 0.5625 - }, - { - "repo_name": "composable_kernel", - "repo_link": "https://github.com/ROCm/composable_kernel", - "category": "gpu kernels", - "github_about_section": "Composable Kernel: Performance Portable Programming Model for Machine Learning Tensor Operators", - "homepage_link": "https://rocm.docs.amd.com/projects/composable_kernel", - "contributors_all": 190, - "contributors_2025": 140, - "contributors_2024": 58, - "contributors_2023": 33, - "growth_2025_percent": 141, - "90-day-contributor-retention-rate": 0.6180555555555556, - "180-day-contributor-retention-rate": 0.5043478260869565 - }, - { - "repo_name": "monarch", - "repo_link": "https://github.com/meta-pytorch/monarch", - "github_about_section": "PyTorch Single Controller", - "homepage_link": "https://meta-pytorch.org/monarch", - "contributors_all": 85, - "contributors_2025": 85, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.6097560975609756, - "180-day-contributor-retention-rate": "not-enough-data" - }, - { - "repo_name": "oneDPL", - "repo_link": "https://github.com/uxlfoundation/oneDPL", - "github_about_section": "oneAPI DPC++ Library (oneDPL)", - "homepage_link": "https://software.intel.com/content/www/us/en/develop/tools/oneapi/components/dpc-library.html", - "contributors_all": 67, - "contributors_2025": 17, - "contributors_2024": 29, - "contributors_2023": 28, - "growth_2025_percent": -41, - "90-day-contributor-retention-rate": 0.5970149253731343, - "180-day-contributor-retention-rate": 0.5151515151515151 - }, - { - "repo_name": "nixl", - "repo_link": "https://github.com/ai-dynamo/nixl", - "github_about_section": "NVIDIA Inference Xfer Library (NIXL)", - "contributors_all": 78, - "contributors_2025": 78, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.5957446808510638, - "180-day-contributor-retention-rate": 0.391304347826087 - }, - { - "repo_name": "lightning-thunder", - "repo_link": "https://github.com/Lightning-AI/lightning-thunder", - "github_about_section": "PyTorch compiler that accelerates training and inference. Get built-in optimizations for performance, memory, parallelism, and easily write your own.", - "contributors_all": 76, - "contributors_2025": 44, - "contributors_2024": 47, - "contributors_2023": 29, - "growth_2025_percent": -6, - "90-day-contributor-retention-rate": 0.582089552238806, - "180-day-contributor-retention-rate": 0.5 - }, - { - "repo_name": "Primus-Turbo", - "repo_link": "https://github.com/AMD-AGI/Primus-Turbo", - "contributors_all": 12, - "contributors_2025": 12, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.5714285714285714, - "180-day-contributor-retention-rate": "not-enough-data" - }, { "repo_name": "llvm-project", "repo_link": "https://github.com/llvm/llvm-project", @@ -200,564 +9,342 @@ "contributors_all": 6680, "contributors_2025": 2378, "contributors_2024": 2130, - "contributors_2023": 1920, - "growth_2025_percent": 11, - "90-day-contributor-retention-rate": 0.543317230273752, - "180-day-contributor-retention-rate": 0.49890369370888854 - }, - { - "repo_name": "aiter", - "repo_link": "https://github.com/ROCm/aiter", - "github_about_section": "AI Tensor Engine for ROCm", - "homepage_link": "https://rocm.blogs.amd.com/software-tools-optimization/aiter-ai-tensor-engine/README.html", - "contributors_all": 151, - "contributors_2025": 145, - "contributors_2024": 10, - "contributors_2023": 0, - "growth_2025_percent": 1350, - "90-day-contributor-retention-rate": 0.5352112676056338, - "180-day-contributor-retention-rate": 0.3902439024390244 - }, - { - "repo_name": "Triton-distributed", - "repo_link": "https://github.com/ByteDance-Seed/Triton-distributed", - "github_about_section": "Distributed Compiler based on Triton for Parallel Systems", - "homepage_link": "https://triton-distributed.readthedocs.io", - "contributors_all": 30, - "contributors_2025": 30, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.5238095238095238, - "180-day-contributor-retention-rate": 0.6 + "contributors_2023": 1920 }, { - "repo_name": "SYCL-Docs", - "repo_link": "https://github.com/KhronosGroup/SYCL-Docs", - "github_about_section": "SYCL Open Source Specification", - "homepage_link": "https://khronos.org/sycl", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 67, - "contributors_2025": 13, - "contributors_2024": 20, - "contributors_2023": 27, - "growth_2025_percent": -35, - "90-day-contributor-retention-rate": 0.5223880597014925, - "180-day-contributor-retention-rate": 0.45454545454545453 - }, - { - "repo_name": "rocm-systems", - "repo_link": "https://github.com/ROCm/rocm-systems", - "github_about_section": "super repo for rocm systems projects", - "contributors_all": 1032, - "contributors_2025": 440, - "contributors_2024": 323, - "contributors_2023": 204, - "growth_2025_percent": 36, - "90-day-contributor-retention-rate": 0.5213675213675214, - "180-day-contributor-retention-rate": 0.44954128440366975 - }, - { - "repo_name": "rocSOLVER", - "repo_link": "https://github.com/ROCm/rocSOLVER", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-libraries repo", - "homepage_link": "https://github.com/ROCm/rocm-libraries", - "contributors_all": 59, - "contributors_2025": 20, - "contributors_2024": 23, - "contributors_2023": 15, - "growth_2025_percent": -13, - "90-day-contributor-retention-rate": 0.5172413793103449, - "180-day-contributor-retention-rate": 0.5094339622641509 - }, - { - "repo_name": "cuda-python", - "repo_link": "https://github.com/NVIDIA/cuda-python", - "github_about_section": "CUDA Python: Performance meets Productivity", - "homepage_link": "https://nvidia.github.io/cuda-python", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 48, - "contributors_2025": 41, - "contributors_2024": 12, - "contributors_2023": 1, - "growth_2025_percent": 241, - "90-day-contributor-retention-rate": 0.5, - "180-day-contributor-retention-rate": 0.375 - }, - { - "repo_name": "RaBitQ", - "repo_link": "https://github.com/gaoj0017/RaBitQ", - "github_about_section": "[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", - "github_topic_closest_fit": "nearest-neighbor-search", - "contributors_all": 2, - "contributors_2025": 2, - "contributors_2024": 1, - "contributors_2023": 0, - "growth_2025_percent": 100, - "90-day-contributor-retention-rate": 0.5, - "180-day-contributor-retention-rate": 0.5 - }, - { - "repo_name": "MIOpen", - "repo_link": "https://github.com/ROCm/MIOpen", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-libraries repo", - "homepage_link": "https://github.com/ROCm/rocm-libraries", - "contributors_all": 204, - "contributors_2025": 47, - "contributors_2024": 62, - "contributors_2023": 44, - "growth_2025_percent": -24, - "90-day-contributor-retention-rate": 0.49019607843137253, - "180-day-contributor-retention-rate": 0.44148936170212766 - }, - { - "repo_name": "modular", - "repo_link": "https://github.com/modular/modular", - "category": "parallel computing", - "github_about_section": "The Modular Platform (includes MAX & Mojo)", - "homepage_link": "https://docs.modular.com", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 366, - "contributors_2025": 222, - "contributors_2024": 205, - "contributors_2023": 99, - "growth_2025_percent": 8, - "90-day-contributor-retention-rate": 0.4854368932038835, - "180-day-contributor-retention-rate": 0.4681647940074906 - }, - { - "repo_name": "hatchet", - "repo_link": "https://github.com/LLNL/hatchet", - "category": "performance testing", - "github_about_section": "Graph-indexed Pandas DataFrames for analyzing hierarchical performance data", - "homepage_link": "https://llnl-hatchet.readthedocs.io", - "github_topic_closest_fit": "profiling", - "contributors_all": 25, - "contributors_2025": 3, - "contributors_2024": 6, - "contributors_2023": 8, - "growth_2025_percent": -50, - "90-day-contributor-retention-rate": 0.48, - "180-day-contributor-retention-rate": 0.44 - }, - { - "repo_name": "hipBLASLt", - "repo_link": "https://github.com/AMD-AGI/hipBLASLt", - "category": "Basic Linear Algebra Subprograms (BLAS)", - "github_about_section": "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", - "github_topic_closest_fit": "matrix-multiplication", - "contributors_all": 111, - "contributors_2025": 69, - "contributors_2024": 70, - "contributors_2023": 35, - "growth_2025_percent": -1, - "90-day-contributor-retention-rate": 0.4774774774774775, - "180-day-contributor-retention-rate": 0.3577981651376147 - }, - { - "repo_name": "truss", - "repo_link": "https://github.com/basetenlabs/truss", + "repo_name": "vllm", + "repo_link": "https://github.com/vllm-project/vllm", "category": "inference engine", - "github_about_section": "The simplest way to serve AI/ML models in production", - "homepage_link": "https://truss.baseten.co", + "github_about_section": "A high-throughput and memory-efficient inference and serving engine for LLMs", + "homepage_link": "https://docs.vllm.ai", "github_topic_closest_fit": "inference", - "contributors_all": 72, - "contributors_2025": 44, - "contributors_2024": 30, - "contributors_2023": 21, - "growth_2025_percent": 46, - "90-day-contributor-retention-rate": 0.47540983606557374, - "180-day-contributor-retention-rate": 0.4423076923076923 + "contributors_all": 1885, + "contributors_2025": 1369, + "contributors_2024": 579, + "contributors_2023": 145 }, { - "repo_name": "rocRAND", - "repo_link": "https://github.com/ROCm/rocRAND", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-libraries repo", - "homepage_link": "https://github.com/ROCm/rocm-libraries", - "contributors_all": 85, - "contributors_2025": 10, - "contributors_2024": 32, - "contributors_2023": 26, - "growth_2025_percent": -68, - "90-day-contributor-retention-rate": 0.47058823529411764, - "180-day-contributor-retention-rate": 0.42857142857142855 + "repo_name": "pytorch", + "repo_link": "https://github.com/pytorch/pytorch", + "category": "machine learning framework", + "github_about_section": "Tensors and Dynamic neural networks in Python with strong GPU acceleration", + "homepage_link": "https://pytorch.org", + "github_topic_closest_fit": "machine-learning", + "contributors_all": 5434, + "contributors_2025": 1187, + "contributors_2024": 1090, + "contributors_2023": 1024 }, { - "repo_name": "rdma-core", - "repo_link": "https://github.com/linux-rdma/rdma-core", - "github_about_section": "RDMA core userspace libraries and daemons", - "contributors_all": 437, - "contributors_2025": 58, - "contributors_2024": 61, - "contributors_2023": 66, - "growth_2025_percent": -4, - "90-day-contributor-retention-rate": 0.4696261682242991, - "180-day-contributor-retention-rate": 0.4262295081967213 + "repo_name": "transformers", + "repo_link": "https://github.com/huggingface/transformers", + "category": "multi-purpose library", + "github_about_section": "Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.", + "homepage_link": "https://huggingface.co/transformers", + "github_topic_closest_fit": "machine-learning", + "contributors_all": 3582, + "contributors_2025": 860, + "contributors_2024": 769, + "contributors_2023": 758 }, { - "repo_name": "Tensile", - "repo_link": "https://github.com/ROCm/Tensile", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-libraries repo", - "homepage_link": "https://github.com/ROCm/rocm-libraries", - "contributors_all": 137, - "contributors_2025": 16, - "contributors_2024": 25, - "contributors_2023": 22, - "growth_2025_percent": -36, - "90-day-contributor-retention-rate": 0.46715328467153283, - "180-day-contributor-retention-rate": 0.4090909090909091 + "repo_name": "sglang", + "repo_link": "https://github.com/sgl-project/sglang", + "category": "inference engine", + "github_about_section": "SGLang is a fast serving framework for large language models and vision language models.", + "homepage_link": "https://docs.sglang.ai", + "github_topic_closest_fit": "inference", + "contributors_all": 937, + "contributors_2025": 796, + "contributors_2024": 189, + "contributors_2023": 1 }, { - "repo_name": "executorch", - "repo_link": "https://github.com/pytorch/executorch", - "category": "model compiler", - "github_about_section": "On-device AI across mobile, embedded and edge for PyTorch", - "homepage_link": "https://executorch.ai", - "contributors_all": 437, - "contributors_2025": 267, - "contributors_2024": 243, - "contributors_2023": 77, - "growth_2025_percent": 9, - "90-day-contributor-retention-rate": 0.4631578947368421, - "180-day-contributor-retention-rate": 0.3905325443786982 + "repo_name": "hhvm", + "repo_link": "https://github.com/facebook/hhvm", + "category": "virtual machine", + "github_about_section": "A virtual machine for executing programs written in Hack.", + "homepage_link": "https://hhvm.com", + "github_topic_closest_fit": "virtual-machine", + "contributors_all": 2624, + "contributors_2025": 692, + "contributors_2024": 648, + "contributors_2023": 604 }, { - "repo_name": "hipCUB", - "repo_link": "https://github.com/ROCm/hipCUB", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-libraries repo", - "homepage_link": "https://github.com/ROCm/rocm-libraries", - "contributors_all": 54, - "contributors_2025": 10, - "contributors_2024": 19, - "contributors_2023": 13, - "growth_2025_percent": -47, - "90-day-contributor-retention-rate": 0.46296296296296297, - "180-day-contributor-retention-rate": 0.46296296296296297 + "repo_name": "llama.cpp", + "repo_link": "https://github.com/ggml-org/llama.cpp", + "category": "inference engine", + "github_about_section": "LLM inference in C/C++", + "homepage_link": "https://ggml.ai", + "github_topic_closest_fit": "inference", + "contributors_all": 1374, + "contributors_2025": 535, + "contributors_2024": 575, + "contributors_2023": 461 }, { - "repo_name": "ROCm", - "repo_link": "https://github.com/ROCm/ROCm", - "github_about_section": "AMD ROCm Software - GitHub Home", - "homepage_link": "https://rocm.docs.amd.com", - "contributors_all": 166, - "contributors_2025": 67, - "contributors_2024": 61, - "contributors_2023": 44, - "growth_2025_percent": 9, - "90-day-contributor-retention-rate": 0.4605263157894737, - "180-day-contributor-retention-rate": 0.3880597014925373 + "repo_name": "kubernetes", + "repo_link": "https://github.com/kubernetes/kubernetes", + "category": "container orchestration", + "github_about_section": "Production-Grade Container Scheduling and Management", + "homepage_link": "https://kubernetes.io", + "github_topic_closest_fit": "kubernetes", + "contributors_all": 5041, + "contributors_2025": 509, + "contributors_2024": 498, + "contributors_2023": 565 }, { - "repo_name": "rocPRIM", - "repo_link": "https://github.com/ROCm/rocPRIM", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-libraries repo", - "homepage_link": "https://github.com/ROCm/rocm-libraries", - "contributors_all": 76, - "contributors_2025": 12, - "contributors_2024": 28, - "contributors_2023": 15, - "growth_2025_percent": -57, - "90-day-contributor-retention-rate": 0.4605263157894737, - "180-day-contributor-retention-rate": 0.4605263157894737 + "repo_name": "tensorflow", + "repo_link": "https://github.com/tensorflow/tensorflow", + "category": "machine learning framework", + "github_about_section": "An Open Source Machine Learning Framework for Everyone", + "homepage_link": "https://tensorflow.org", + "github_topic_closest_fit": "machine-learning", + "contributors_all": 4618, + "contributors_2025": 500, + "contributors_2024": 523, + "contributors_2023": 630 }, { - "repo_name": "ondemand", - "repo_link": "https://github.com/OSC/ondemand", - "github_about_section": "Supercomputing. Seamlessly. Open, Interactive HPC Via the Web", - "homepage_link": "https://openondemand.org", - "github_topic_closest_fit": "hpc", - "contributors_all": 117, - "contributors_2025": 43, - "contributors_2024": 23, - "contributors_2023": 21, - "growth_2025_percent": 86, - "90-day-contributor-retention-rate": 0.45544554455445546, - "180-day-contributor-retention-rate": 0.4166666666666667 + "repo_name": "verl", + "repo_link": "https://github.com/volcengine/verl", + "category": "reinforcement learning", + "github_about_section": "verl: Volcano Engine Reinforcement Learning for LLMs", + "homepage_link": "https://verl.readthedocs.io", + "github_topic_closest_fit": "deep-reinforcement-learning", + "contributors_all": 462, + "contributors_2025": 454, + "contributors_2024": 10, + "contributors_2023": 0 }, { - "repo_name": "hip", - "repo_link": "https://github.com/ROCm/hip", - "github_about_section": "HIP: C++ Heterogeneous-Compute Interface for Portability", - "homepage_link": "https://rocmdocs.amd.com/projects/HIP", - "contributors_all": 288, - "contributors_2025": 46, - "contributors_2024": 31, - "contributors_2023": 25, - "growth_2025_percent": 48, - "90-day-contributor-retention-rate": 0.4423791821561338, - "180-day-contributor-retention-rate": 0.36923076923076925 + "repo_name": "rocm-systems", + "repo_link": "https://github.com/ROCm/rocm-systems", + "category": "multi-purpose library", + "github_about_section": "super repo for rocm systems projects", + "homepage_link": "https://amd.com/en/products/software/rocm.html", + "github_topic_closest_fit": "amd", + "contributors_all": 1032, + "contributors_2025": 440, + "contributors_2024": 323, + "contributors_2023": 204 }, { - "repo_name": "tilelang", - "repo_link": "https://github.com/tile-ai/tilelang", - "category": "parallel computing dsl", - "github_about_section": "Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels", - "homepage_link": "https://tilelang.com", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 90, - "contributors_2025": 89, - "contributors_2024": 1, - "contributors_2023": 0, - "growth_2025_percent": 8800, - "90-day-contributor-retention-rate": 0.4423076923076923, - "180-day-contributor-retention-rate": 0.3611111111111111 + "repo_name": "ray", + "repo_link": "https://github.com/ray-project/ray", + "category": "multi-purpose library", + "github_about_section": "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", + "github_topic_closest_fit": "machine-learning", + "contributors_all": 1381, + "contributors_2025": 397, + "contributors_2024": 223, + "contributors_2023": 230 }, { - "repo_name": "rocFFT", - "repo_link": "https://github.com/ROCm/rocFFT", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-libraries repo", - "homepage_link": "https://github.com/ROCm/rocm-libraries", - "contributors_all": 81, - "contributors_2025": 17, - "contributors_2024": 20, - "contributors_2023": 19, - "growth_2025_percent": -15, - "90-day-contributor-retention-rate": 0.43209876543209874, - "180-day-contributor-retention-rate": 0.42857142857142855 + "repo_name": "spark", + "repo_link": "https://github.com/apache/spark", + "github_about_section": "Apache Spark - A unified analytics engine for large-scale data processing", + "homepage_link": "https://spark.apache.org", + "github_topic_closest_fit": "big-data", + "contributors_all": 3083, + "contributors_2025": 322, + "contributors_2024": 300, + "contributors_2023": 336 }, { - "repo_name": "lean4", - "repo_link": "https://github.com/leanprover/lean4", - "category": "theorem prover", - "github_about_section": "Lean 4 programming language and theorem prover", - "homepage_link": "https://lean-lang.org", - "github_topic_closest_fit": "lean", - "contributors_all": 278, - "contributors_2025": 110, - "contributors_2024": 85, - "contributors_2023": 64, - "growth_2025_percent": 29, - "90-day-contributor-retention-rate": 0.42857142857142855, - "180-day-contributor-retention-rate": 0.3905579399141631 + "repo_name": "goose", + "repo_link": "https://github.com/block/goose", + "category": "agent", + "github_about_section": "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", + "github_topic_closest_fit": "ai-agents", + "contributors_all": 332, + "contributors_2025": 319, + "contributors_2024": 32, + "contributors_2023": 0 }, { - "repo_name": "tritonparse", - "repo_link": "https://github.com/meta-pytorch/tritonparse", - "github_about_section": "TritonParse: A Compiler Tracer, Visualizer, and Reproducer for Triton Kernels", - "homepage_link": "https://meta-pytorch.org/tritonparse", - "contributors_all": 15, - "contributors_2025": 15, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.42857142857142855, - "180-day-contributor-retention-rate": "not-enough-data" + "repo_name": "elasticsearch", + "repo_link": "https://github.com/elastic/elasticsearch", + "category": "search engine", + "github_about_section": "Free and Open Source, Distributed, RESTful Search Engine", + "homepage_link": "https://elastic.co/products/elasticsearch", + "github_topic_closest_fit": "search-engine", + "contributors_all": 2297, + "contributors_2025": 316, + "contributors_2024": 284, + "contributors_2023": 270 }, { - "repo_name": "hipBLAS", - "repo_link": "https://github.com/ROCm/hipBLAS", - "category": "Basic Linear Algebra Subprograms (BLAS)", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-libraries repo", - "github_topic_closest_fit": "matrix-multiplication", - "contributors_all": 72, - "contributors_2025": 21, - "contributors_2024": 24, - "contributors_2023": 14, - "growth_2025_percent": -12, - "90-day-contributor-retention-rate": 0.4027777777777778, - "180-day-contributor-retention-rate": 0.4 + "repo_name": "jax", + "repo_link": "https://github.com/jax-ml/jax", + "category": "scientific computing", + "github_about_section": "Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more", + "homepage_link": "https://docs.jax.dev", + "github_topic_closest_fit": "scientific-computing", + "contributors_all": 997, + "contributors_2025": 312, + "contributors_2024": 280, + "contributors_2023": 202 }, { - "repo_name": "rocSPARSE", - "repo_link": "https://github.com/ROCm/rocSPARSE", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-libraries repo", - "homepage_link": "https://github.com/ROCm/rocm-libraries", - "contributors_all": 65, - "contributors_2025": 19, - "contributors_2024": 24, - "contributors_2023": 18, - "growth_2025_percent": -20, - "90-day-contributor-retention-rate": 0.4, - "180-day-contributor-retention-rate": 0.41379310344827586 + "repo_name": "modelcontextprotocol", + "repo_link": "https://github.com/modelcontextprotocol/modelcontextprotocol", + "category": "mcp", + "github_about_section": "Specification and documentation for the Model Context Protocol", + "homepage_link": "https://modelcontextprotocol.io", + "github_topic_closest_fit": "mcp", + "contributors_all": 327, + "contributors_2025": 298, + "contributors_2024": 42, + "contributors_2023": 0 }, { - "repo_name": "triSYCL", - "repo_link": "https://github.com/triSYCL/triSYCL", - "github_about_section": "Generic system-wide modern C++ for heterogeneous platforms with SYCL from Khronos Group", - "homepage_link": "https://trisycl.github.io/triSYCL/Doxygen/triSYCL/html/index.html", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 31, - "contributors_2025": 0, - "contributors_2024": 1, - "contributors_2023": 3, - "growth_2025_percent": -100, - "90-day-contributor-retention-rate": 0.3870967741935484, - "180-day-contributor-retention-rate": 0.22580645161290322 + "repo_name": "executorch", + "repo_link": "https://github.com/pytorch/executorch", + "category": "model compiler", + "github_about_section": "On-device AI across mobile, embedded and edge for PyTorch", + "homepage_link": "https://executorch.ai", + "contributors_all": 437, + "contributors_2025": 267, + "contributors_2024": 243, + "contributors_2023": 77 }, { - "repo_name": "hhvm", - "repo_link": "https://github.com/facebook/hhvm", - "github_about_section": "A virtual machine for executing programs written in Hack.", - "homepage_link": "https://hhvm.com", - "contributors_all": 2624, - "contributors_2025": 692, - "contributors_2024": 648, - "contributors_2023": 604, - "growth_2025_percent": 6, - "90-day-contributor-retention-rate": 0.38206785137318255, - "180-day-contributor-retention-rate": 0.3457513604018418 + "repo_name": "numpy", + "repo_link": "https://github.com/numpy/numpy", + "category": "scientific computing", + "github_about_section": "The fundamental package for scientific computing with Python.", + "homepage_link": "https://numpy.org", + "github_topic_closest_fit": "scientific-computing", + "contributors_all": 2172, + "contributors_2025": 235, + "contributors_2024": 233, + "contributors_2023": 252 }, { - "repo_name": "tflite-micro", - "repo_link": "https://github.com/tensorflow/tflite-micro", - "github_about_section": "Infrastructure to enable deployment of ML models to low-power resource-constrained embedded targets (including microcontrollers and digital signal processors).", - "contributors_all": 111, - "contributors_2025": 19, - "contributors_2024": 25, - "contributors_2023": 31, - "growth_2025_percent": -24, - "90-day-contributor-retention-rate": 0.38095238095238093, - "180-day-contributor-retention-rate": 0.3269230769230769 + "repo_name": "triton", + "repo_link": "https://github.com/triton-lang/triton", + "category": "parallel computing dsl", + "github_about_section": "Development repository for the Triton language and compiler", + "homepage_link": "https://triton-lang.org", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 522, + "contributors_2025": 233, + "contributors_2024": 206, + "contributors_2023": 159 }, { - "repo_name": "cudnn-frontend", - "repo_link": "https://github.com/NVIDIA/cudnn-frontend", + "repo_name": "modular", + "repo_link": "https://github.com/modular/modular", "category": "parallel computing", - "github_about_section": "cudnn_frontend provides a c++ wrapper for the cudnn backend API and samples on how to use it", - "homepage_link": "https://developer.nvidia.com/cudnn", + "github_about_section": "The Modular Platform (includes MAX & Mojo)", + "homepage_link": "https://docs.modular.com", "github_topic_closest_fit": "parallel-programming", - "contributors_all": 12, - "contributors_2025": 6, - "contributors_2024": 5, - "contributors_2023": 1, - "growth_2025_percent": 20, - "90-day-contributor-retention-rate": 0.375, - "180-day-contributor-retention-rate": 0.2857142857142857 - }, - { - "repo_name": "sglang", - "repo_link": "https://github.com/sgl-project/sglang", - "category": "inference engine", - "github_about_section": "SGLang is a fast serving framework for large language models and vision language models.", - "homepage_link": "https://docs.sglang.ai", - "github_topic_closest_fit": "inference", - "contributors_all": 937, - "contributors_2025": 796, - "contributors_2024": 189, - "contributors_2023": 1, - "growth_2025_percent": 321, - "90-day-contributor-retention-rate": 0.36363636363636365, - "180-day-contributor-retention-rate": 0.2436548223350254 + "contributors_all": 366, + "contributors_2025": 222, + "contributors_2024": 205, + "contributors_2023": 99 }, { - "repo_name": "milvus", - "repo_link": "https://github.com/milvus-io/milvus", - "category": "vector database", - "github_about_section": "Milvus is a high-performance, cloud-native vector database built for scalable vector ANN search", - "homepage_link": "https://milvus.io", - "github_topic_closest_fit": "vector-search", - "contributors_all": 387, - "contributors_2025": 95, - "contributors_2024": 84, - "contributors_2023": 72, - "growth_2025_percent": 13, - "90-day-contributor-retention-rate": 0.36363636363636365, - "180-day-contributor-retention-rate": 0.29859154929577464 + "repo_name": "scipy", + "repo_link": "https://github.com/scipy/scipy", + "category": "scientific computing", + "github_about_section": "SciPy library main repository", + "homepage_link": "https://scipy.org", + "github_topic_closest_fit": "scientific-computing", + "contributors_all": 1973, + "contributors_2025": 210, + "contributors_2024": 251, + "contributors_2023": 245 }, { - "repo_name": "tensorflow", - "repo_link": "https://github.com/tensorflow/tensorflow", - "category": "machine learning framework", - "github_about_section": "An Open Source Machine Learning Framework for Everyone", - "homepage_link": "https://tensorflow.org", - "github_topic_closest_fit": "machine-learning", - "contributors_all": 4618, - "contributors_2025": 500, - "contributors_2024": 523, - "contributors_2023": 630, - "growth_2025_percent": -4, - "90-day-contributor-retention-rate": 0.3572371315442147, - "180-day-contributor-retention-rate": 0.29946642952423297 + "repo_name": "ollama", + "repo_link": "https://github.com/ollama/ollama", + "category": "inference engine", + "github_about_section": "Get up and running with OpenAI gpt-oss, DeepSeek-R1, Gemma 3 and other models.", + "homepage_link": "https://ollama.com", + "github_topic_closest_fit": "inference", + "contributors_all": 574, + "contributors_2025": 202, + "contributors_2024": 314, + "contributors_2023": 97 }, { - "repo_name": "helion", - "repo_link": "https://github.com/pytorch/helion", - "category": "parallel computing dsl", - "github_about_section": "A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.", - "homepage_link": "https://helionlang.com", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 49, - "contributors_2025": 49, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.35714285714285715, - "180-day-contributor-retention-rate": 1.0 + "repo_name": "trl", + "repo_link": "https://github.com/huggingface/trl", + "github_about_section": "Train transformer language models with reinforcement learning.", + "homepage_link": "http://hf.co/docs/trl", + "contributors_all": 433, + "contributors_2025": 189, + "contributors_2024": 154, + "contributors_2023": 122 }, { - "repo_name": "jax", - "repo_link": "https://github.com/jax-ml/jax", - "category": "scientific computing", - "github_about_section": "Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more", - "homepage_link": "https://docs.jax.dev", - "github_topic_closest_fit": "scientific-computing", - "contributors_all": 997, - "contributors_2025": 312, - "contributors_2024": 280, - "contributors_2023": 202, - "growth_2025_percent": 11, - "90-day-contributor-retention-rate": 0.35538954108858056, - "180-day-contributor-retention-rate": 0.3171007927519819 + "repo_name": "flashinfer", + "repo_link": "https://github.com/flashinfer-ai/flashinfer", + "category": "gpu kernels", + "github_about_section": "FlashInfer: Kernel Library for LLM Serving", + "homepage_link": "https://flashinfer.ai", + "github_topic_closest_fit": "attention", + "contributors_all": 205, + "contributors_2025": 158, + "contributors_2024": 50, + "contributors_2023": 11 }, { - "repo_name": "kubernetes", - "repo_link": "https://github.com/kubernetes/kubernetes", - "category": "container orchestration", - "github_about_section": "Production-Grade Container Scheduling and Management", - "homepage_link": "https://kubernetes.io", - "github_topic_closest_fit": "kubernetes", - "contributors_all": 5041, - "contributors_2025": 509, - "contributors_2024": 498, - "contributors_2023": 565, - "growth_2025_percent": 2, - "90-day-contributor-retention-rate": 0.35439449169704335, - "180-day-contributor-retention-rate": 0.2948559670781893 + "repo_name": "aiter", + "repo_link": "https://github.com/ROCm/aiter", + "github_about_section": "AI Tensor Engine for ROCm", + "homepage_link": "https://rocm.blogs.amd.com/software-tools-optimization/aiter-ai-tensor-engine/README.html", + "contributors_all": 151, + "contributors_2025": 145, + "contributors_2024": 10, + "contributors_2023": 0 }, { - "repo_name": "pocl", - "repo_link": "https://github.com/pocl/pocl", - "github_about_section": "pocl - Portable Computing Language", - "homepage_link": "https://portablecl.org", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 166, - "contributors_2025": 26, - "contributors_2024": 27, - "contributors_2023": 21, - "growth_2025_percent": -3, - "90-day-contributor-retention-rate": 0.3496932515337423, - "180-day-contributor-retention-rate": 0.30625 + "repo_name": "LMCache", + "repo_link": "https://github.com/LMCache/LMCache", + "github_about_section": "Supercharge Your LLM with the Fastest KV Cache Layer", + "homepage_link": "https://lmcache.ai", + "contributors_all": 152, + "contributors_2025": 144, + "contributors_2024": 18, + "contributors_2023": 0 }, { - "repo_name": "server", - "repo_link": "https://github.com/triton-inference-server/server", - "github_about_section": "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", + "repo_name": "composable_kernel", + "repo_link": "https://github.com/ROCm/composable_kernel", + "category": "gpu kernels", + "github_about_section": "Composable Kernel: Performance Portable Programming Model for Machine Learning Tensor Operators", + "homepage_link": "https://rocm.docs.amd.com/projects/composable_kernel", + "contributors_all": 190, + "contributors_2025": 140, + "contributors_2024": 58, + "contributors_2023": 33 + }, + { + "repo_name": "Mooncake", + "repo_link": "https://github.com/kvcache-ai/Mooncake", + "github_about_section": "Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.", + "homepage_link": "https://kvcache-ai.github.io/Mooncake", "github_topic_closest_fit": "inference", - "contributors_all": 147, - "contributors_2025": 24, - "contributors_2024": 36, - "contributors_2023": 34, - "growth_2025_percent": -33, - "90-day-contributor-retention-rate": 0.3401360544217687, - "180-day-contributor-retention-rate": 0.3129251700680272 + "contributors_all": 138, + "contributors_2025": 133, + "contributors_2024": 13, + "contributors_2023": 0 }, { - "repo_name": "Vulkan-Tools", - "repo_link": "https://github.com/KhronosGroup/Vulkan-Tools", - "category": "graphics api", - "github_about_section": "Vulkan Development Tools", - "homepage_link": "https://vulkan.org", - "github_topic_closest_fit": "vulkan", - "contributors_all": 248, - "contributors_2025": 20, - "contributors_2024": 24, - "contributors_2023": 24, - "growth_2025_percent": -16, - "90-day-contributor-retention-rate": 0.33884297520661155, - "180-day-contributor-retention-rate": 0.3181818181818182 + "repo_name": "torchtitan", + "repo_link": "https://github.com/pytorch/torchtitan", + "github_about_section": "A PyTorch native platform for training generative AI models", + "contributors_all": 145, + "contributors_2025": 119, + "contributors_2024": 43, + "contributors_2023": 1 }, { "repo_name": "ao", @@ -768,277 +355,217 @@ "contributors_all": 178, "contributors_2025": 114, "contributors_2024": 100, - "contributors_2023": 5, - "growth_2025_percent": 14, - "90-day-contributor-retention-rate": 0.33774834437086093, - "180-day-contributor-retention-rate": 0.3 + "contributors_2023": 5 }, { - "repo_name": "ThunderKittens", - "repo_link": "https://github.com/HazyResearch/ThunderKittens", - "category": "parallel computing", - "github_about_section": "Tile primitives for speedy kernels", - "homepage_link": "https://hazyresearch.stanford.edu/blog/2024-10-29-tk2", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 34, - "contributors_2025": 29, - "contributors_2024": 13, - "contributors_2023": 0, - "growth_2025_percent": 123, - "90-day-contributor-retention-rate": 0.3333333333333333, - "180-day-contributor-retention-rate": 0.3333333333333333 + "repo_name": "lean4", + "repo_link": "https://github.com/leanprover/lean4", + "category": "theorem prover", + "github_about_section": "Lean 4 programming language and theorem prover", + "homepage_link": "https://lean-lang.org", + "github_topic_closest_fit": "lean", + "contributors_all": 278, + "contributors_2025": 110, + "contributors_2024": 85, + "contributors_2023": 64 }, { - "repo_name": "OpenCL-SDK", - "repo_link": "https://github.com/KhronosGroup/OpenCL-SDK", - "github_about_section": "OpenCL SDK", - "homepage_link": "https://khronos.org/opencl", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 25, - "contributors_2025": 8, - "contributors_2024": 6, - "contributors_2023": 9, - "growth_2025_percent": 33, - "90-day-contributor-retention-rate": 0.3333333333333333, - "180-day-contributor-retention-rate": 0.34782608695652173 + "repo_name": "ComfyUI", + "repo_link": "https://github.com/comfyanonymous/ComfyUI", + "category": "user interface", + "github_about_section": "The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.", + "homepage_link": "https://comfy.org", + "github_topic_closest_fit": "stable-diffusion", + "contributors_all": 278, + "contributors_2025": 108, + "contributors_2024": 119, + "contributors_2023": 94 }, { - "repo_name": "Self-Forcing", - "repo_link": "https://github.com/guandeh17/Self-Forcing", - "category": "video generation", - "github_about_section": "Official codebase for \"Self Forcing: Bridging Training and Inference in Autoregressive Video Diffusion\" (NeurIPS 2025 Spotlight)", - "homepage_link": "https://self-forcing.github.io", - "github_topic_closest_fit": "diffusion-models", - "contributors_all": 4, - "contributors_2025": 4, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.3333333333333333, - "180-day-contributor-retention-rate": "not-enough-data" + "repo_name": "unsloth", + "repo_link": "https://github.com/unslothai/unsloth", + "category": "fine tuning", + "github_about_section": "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", + "github_topic_closest_fit": "fine-tuning", + "contributors_all": 127, + "contributors_2025": 102, + "contributors_2024": 27, + "contributors_2023": 3 }, { - "repo_name": "streamv2v", - "repo_link": "https://github.com/Jeff-LiangF/streamv2v", - "category": "video generation", - "github_about_section": "Official Pytorch implementation of StreamV2V.", - "homepage_link": "https://jeff-liangf.github.io/projects/streamv2v", - "github_topic_closest_fit": "diffusion-models", - "contributors_all": 7, - "contributors_2025": 3, - "contributors_2024": 6, - "contributors_2023": 0, - "growth_2025_percent": -50, - "90-day-contributor-retention-rate": 0.3333333333333333, - "180-day-contributor-retention-rate": 0.16666666666666666 + "repo_name": "burn", + "repo_link": "https://github.com/tracel-ai/burn", + "github_about_section": "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", + "contributors_all": 237, + "contributors_2025": 99, + "contributors_2024": 104, + "contributors_2023": 62 }, { - "repo_name": "pytorch", - "repo_link": "https://github.com/pytorch/pytorch", - "category": "machine learning framework", - "github_about_section": "Tensors and Dynamic neural networks in Python with strong GPU acceleration", - "homepage_link": "https://pytorch.org", - "github_topic_closest_fit": "machine-learning", - "contributors_all": 5434, - "contributors_2025": 1187, - "contributors_2024": 1090, - "contributors_2023": 1024, - "growth_2025_percent": 8, - "90-day-contributor-retention-rate": 0.33145302470336513, - "180-day-contributor-retention-rate": 0.2921212121212121 + "repo_name": "accelerate", + "repo_link": "https://github.com/huggingface/accelerate", + "github_about_section": "A simple way to launch, train, and use PyTorch models on almost any device and distributed configuration, automatic mixed precision (including fp8), and easy-to-configure FSDP and DeepSpeed support.", + "homepage_link": "https://huggingface.co/docs/accelerate", + "contributors_all": 392, + "contributors_2025": 97, + "contributors_2024": 124, + "contributors_2023": 149 }, { - "repo_name": "warp", - "repo_link": "https://github.com/NVIDIA/warp", - "category": "spatial computing", - "github_about_section": "A Python framework for accelerated simulation, data generation and spatial computing.", - "homepage_link": "https://nvidia.github.io/warp", - "github_topic_closest_fit": "physics-simulation", - "contributors_all": 79, - "contributors_2025": 40, - "contributors_2024": 29, - "contributors_2023": 17, - "growth_2025_percent": 37, - "90-day-contributor-retention-rate": 0.32857142857142857, - "180-day-contributor-retention-rate": 0.2833333333333333 + "repo_name": "terminal-bench", + "repo_link": "https://github.com/laude-institute/terminal-bench", + "category": "benchmark", + "github_about_section": "A benchmark for LLMs on complicated tasks in the terminal", + "homepage_link": "https://tbench.ai", + "github_topic_closest_fit": "benchmark", + "contributors_all": 96, + "contributors_2025": 96, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "triton", - "repo_link": "https://github.com/triton-lang/triton", - "category": "parallel computing dsl", - "github_about_section": "Development repository for the Triton language and compiler", - "homepage_link": "https://triton-lang.org", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 522, - "contributors_2025": 233, - "contributors_2024": 206, - "contributors_2023": 159, - "growth_2025_percent": 13, - "90-day-contributor-retention-rate": 0.3236607142857143, - "180-day-contributor-retention-rate": 0.27791563275434245 + "repo_name": "DeepSpeed", + "repo_link": "https://github.com/deepspeedai/DeepSpeed", + "github_about_section": "DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.", + "homepage_link": "https://deepspeed.ai", + "contributors_all": 442, + "contributors_2025": 96, + "contributors_2024": 134, + "contributors_2023": 165 }, { - "repo_name": "torchtitan", - "repo_link": "https://github.com/pytorch/torchtitan", - "github_about_section": "A PyTorch native platform for training generative AI models", - "contributors_all": 145, - "contributors_2025": 119, - "contributors_2024": 43, - "contributors_2023": 1, - "growth_2025_percent": 176, - "90-day-contributor-retention-rate": 0.3229166666666667, - "180-day-contributor-retention-rate": 0.3424657534246575 + "repo_name": "milvus", + "repo_link": "https://github.com/milvus-io/milvus", + "category": "vector database", + "github_about_section": "Milvus is a high-performance, cloud-native vector database built for scalable vector ANN search", + "homepage_link": "https://milvus.io", + "github_topic_closest_fit": "vector-search", + "contributors_all": 387, + "contributors_2025": 95, + "contributors_2024": 84, + "contributors_2023": 72 }, { - "repo_name": "nccl", - "repo_link": "https://github.com/NVIDIA/nccl", - "github_about_section": "Optimized primitives for collective multi-GPU communication", - "homepage_link": "https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/index.html", - "contributors_all": 51, - "contributors_2025": 7, - "contributors_2024": 5, - "contributors_2023": 6, - "growth_2025_percent": 40, - "90-day-contributor-retention-rate": 0.32, - "180-day-contributor-retention-rate": 0.24489795918367346 + "repo_name": "cutlass", + "repo_link": "https://github.com/NVIDIA/cutlass", + "category": "parallel computing", + "github_about_section": "CUDA Templates and Python DSLs for High-Performance Linear Algebra", + "homepage_link": "https://docs.nvidia.com/cutlass/index.html", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 238, + "contributors_2025": 94, + "contributors_2024": 64, + "contributors_2023": 66 }, { - "repo_name": "metaflow", - "repo_link": "https://github.com/Netflix/metaflow", - "github_about_section": "Build, Manage and Deploy AI/ML Systems", - "homepage_link": "https://metaflow.org", - "contributors_all": 121, - "contributors_2025": 37, - "contributors_2024": 35, - "contributors_2023": 28, - "growth_2025_percent": 5, - "90-day-contributor-retention-rate": 0.3185840707964602, - "180-day-contributor-retention-rate": 0.2761904761904762 + "repo_name": "tilelang", + "repo_link": "https://github.com/tile-ai/tilelang", + "category": "parallel computing dsl", + "github_about_section": "Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels", + "homepage_link": "https://tilelang.com", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 90, + "contributors_2025": 89, + "contributors_2024": 1, + "contributors_2023": 0 }, { - "repo_name": "terminal-bench", - "repo_link": "https://github.com/laude-institute/terminal-bench", - "category": "benchmark", - "github_about_section": "A benchmark for LLMs on complicated tasks in the terminal", - "homepage_link": "https://tbench.ai", - "github_topic_closest_fit": "benchmark", - "contributors_all": 96, - "contributors_2025": 96, + "repo_name": "monarch", + "repo_link": "https://github.com/meta-pytorch/monarch", + "github_about_section": "PyTorch Single Controller", + "homepage_link": "https://meta-pytorch.org/monarch", + "contributors_all": 85, + "contributors_2025": 85, "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.3181818181818182, - "180-day-contributor-retention-rate": 0.29411764705882354 + "contributors_2023": 0 }, { - "repo_name": "ray", - "repo_link": "https://github.com/ray-project/ray", - "github_about_section": "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", - "contributors_all": 1381, - "contributors_2025": 397, - "contributors_2024": 223, - "contributors_2023": 230, - "growth_2025_percent": 78, - "90-day-contributor-retention-rate": 0.31763766959297685, - "180-day-contributor-retention-rate": 0.25989672977624784 + "repo_name": "Liger-Kernel", + "repo_link": "https://github.com/linkedin/Liger-Kernel", + "category": "kernel examples", + "github_about_section": "Efficient Triton Kernels for LLM Training", + "homepage_link": "https://openreview.net/pdf?id=36SjAIT42G", + "github_topic_closest_fit": "triton", + "contributors_all": 120, + "contributors_2025": 78, + "contributors_2024": 61, + "contributors_2023": 0 }, { - "repo_name": "TensorRT", - "repo_link": "https://github.com/NVIDIA/TensorRT", - "github_about_section": "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", - "contributors_all": 104, - "contributors_2025": 10, - "contributors_2024": 18, - "contributors_2023": 19, - "growth_2025_percent": -44, - "90-day-contributor-retention-rate": 0.31683168316831684, - "180-day-contributor-retention-rate": 0.26 + "repo_name": "nixl", + "repo_link": "https://github.com/ai-dynamo/nixl", + "github_about_section": "NVIDIA Inference Xfer Library (NIXL)", + "contributors_all": 78, + "contributors_2025": 78, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "spark", - "repo_link": "https://github.com/apache/spark", - "github_about_section": "Apache Spark - A unified analytics engine for large-scale data processing", - "homepage_link": "https://spark.apache.org", - "github_topic_closest_fit": "big-data", - "contributors_all": 3083, - "contributors_2025": 322, - "contributors_2024": 300, - "contributors_2023": 336, - "growth_2025_percent": 7, - "90-day-contributor-retention-rate": 0.3159986750579662, - "180-day-contributor-retention-rate": 0.2586496472959355 + "repo_name": "jupyterlab", + "repo_link": "https://github.com/jupyterlab/jupyterlab", + "category": "user interface", + "github_about_section": "JupyterLab computational environment.", + "homepage_link": "https://jupyterlab.readthedocs.io", + "github_topic_closest_fit": "jupyter", + "contributors_all": 698, + "contributors_2025": 77, + "contributors_2024": 85, + "contributors_2023": 100 }, { - "repo_name": "AdaptiveCpp", - "repo_link": "https://github.com/AdaptiveCpp/AdaptiveCpp", - "github_about_section": "Compiler for multiple programming models (SYCL, C++ standard parallelism, HIP/CUDA) for CPUs and GPUs from all vendors: The independent, community-driven compiler for C++-based heterogeneous programming models. Lets applications adapt themselves to all the hardware in the system - even at runtime!", - "homepage_link": "https://adaptivecpp.github.io", - "contributors_all": 93, - "contributors_2025": 32, - "contributors_2024": 32, - "contributors_2023": 24, - "growth_2025_percent": 0, - "90-day-contributor-retention-rate": 0.3146067415730337, - "180-day-contributor-retention-rate": 0.3023255813953488 + "repo_name": "hipBLASLt", + "repo_link": "https://github.com/AMD-AGI/hipBLASLt", + "category": "Basic Linear Algebra Subprograms (BLAS)", + "github_about_section": "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", + "github_topic_closest_fit": "matrix-multiplication", + "contributors_all": 111, + "contributors_2025": 69, + "contributors_2024": 70, + "contributors_2023": 35 }, { - "repo_name": "vllm", - "repo_link": "https://github.com/vllm-project/vllm", - "category": "inference engine", - "github_about_section": "A high-throughput and memory-efficient inference and serving engine for LLMs", - "homepage_link": "https://docs.vllm.ai", - "github_topic_closest_fit": "inference", - "contributors_all": 1885, - "contributors_2025": 1369, - "contributors_2024": 579, - "contributors_2023": 145, - "growth_2025_percent": 136, - "90-day-contributor-retention-rate": 0.30577223088923555, - "180-day-contributor-retention-rate": 0.2512315270935961 + "repo_name": "peft", + "repo_link": "https://github.com/huggingface/peft", + "github_about_section": "PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.", + "homepage_link": "https://huggingface.co/docs/peft", + "github_topic_closest_fit": "lora", + "contributors_all": 272, + "contributors_2025": 69, + "contributors_2024": 111, + "contributors_2023": 115 }, { - "repo_name": "LMCache", - "repo_link": "https://github.com/LMCache/LMCache", - "github_about_section": "Supercharge Your LLM with the Fastest KV Cache Layer", - "homepage_link": "https://lmcache.ai", - "contributors_all": 152, - "contributors_2025": 144, - "contributors_2024": 18, - "contributors_2023": 0, - "growth_2025_percent": 700, - "90-day-contributor-retention-rate": 0.30526315789473685, - "180-day-contributor-retention-rate": 0.3235294117647059 + "repo_name": "ROCm", + "repo_link": "https://github.com/ROCm/ROCm", + "github_about_section": "AMD ROCm Software - GitHub Home", + "homepage_link": "https://rocm.docs.amd.com", + "contributors_all": 166, + "contributors_2025": 67, + "contributors_2024": 61, + "contributors_2023": 44 }, { - "repo_name": "lapack", - "repo_link": "https://github.com/Reference-LAPACK/lapack", - "category": "linear algebra", - "github_about_section": "LAPACK is a library of Fortran subroutines for solving the most commonly occurring problems in numerical linear algebra.", - "homepage_link": "https://netlib.org/lapack", - "github_topic_closest_fit": "linear-algebra", - "contributors_all": 178, - "contributors_2025": 20, - "contributors_2024": 24, - "contributors_2023": 42, - "growth_2025_percent": -16, - "90-day-contributor-retention-rate": 0.29310344827586204, - "180-day-contributor-retention-rate": 0.23837209302325582 + "repo_name": "mcp-agent", + "repo_link": "https://github.com/lastmile-ai/mcp-agent", + "category": "mcp", + "github_about_section": "Build effective agents using Model Context Protocol and simple workflow patterns", + "github_topic_closest_fit": "mcp", + "contributors_all": 63, + "contributors_2025": 63, + "contributors_2024": 1, + "contributors_2023": 0 }, { - "repo_name": "Mooncake", - "repo_link": "https://github.com/kvcache-ai/Mooncake", - "github_about_section": "Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.", - "homepage_link": "https://kvcache-ai.github.io/Mooncake", - "github_topic_closest_fit": "inference", - "contributors_all": 138, - "contributors_2025": 133, - "contributors_2024": 13, - "contributors_2023": 0, - "growth_2025_percent": 923, - "90-day-contributor-retention-rate": 0.2894736842105263, - "180-day-contributor-retention-rate": 0.2631578947368421 + "repo_name": "rdma-core", + "repo_link": "https://github.com/linux-rdma/rdma-core", + "github_about_section": "RDMA core userspace libraries and daemons", + "contributors_all": 437, + "contributors_2025": 58, + "contributors_2024": 61, + "contributors_2023": 66 }, { "repo_name": "onnx", @@ -1050,95 +577,127 @@ "contributors_all": 370, "contributors_2025": 56, "contributors_2024": 45, - "contributors_2023": 61, - "growth_2025_percent": 24, - "90-day-contributor-retention-rate": 0.28888888888888886, - "180-day-contributor-retention-rate": 0.2507204610951009 + "contributors_2023": 61 }, { - "repo_name": "elasticsearch", - "repo_link": "https://github.com/elastic/elasticsearch", - "category": "search engine", - "github_about_section": "Free and Open Source, Distributed, RESTful Search Engine", - "homepage_link": "https://elastic.co/products/elasticsearch", - "github_topic_closest_fit": "search-engine", - "contributors_all": 2297, - "contributors_2025": 316, - "contributors_2024": 284, - "contributors_2023": 270, - "growth_2025_percent": 11, - "90-day-contributor-retention-rate": 0.2855227882037534, - "180-day-contributor-retention-rate": 0.2478399272396544 + "repo_name": "letta", + "repo_link": "https://github.com/letta-ai/letta", + "category": "agent", + "github_about_section": "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", + "github_topic_closest_fit": "ai-agents", + "contributors_all": 157, + "contributors_2025": 56, + "contributors_2024": 75, + "contributors_2023": 47 }, { - "repo_name": "ome", - "repo_link": "https://github.com/sgl-project/ome", - "github_about_section": "OME is a Kubernetes operator for enterprise-grade management and serving of Large Language Models (LLMs)", - "homepage_link": "http://docs.sglang.ai/ome", - "github_topic_closest_fit": "k8s", - "contributors_all": 28, - "contributors_2025": 28, + "repo_name": "helion", + "repo_link": "https://github.com/pytorch/helion", + "category": "parallel computing dsl", + "github_about_section": "A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.", + "homepage_link": "https://helionlang.com", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 49, + "contributors_2025": 49, "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.2777777777777778, - "180-day-contributor-retention-rate": "not-enough-data" + "contributors_2023": 0 }, { - "repo_name": "Liger-Kernel", - "repo_link": "https://github.com/linkedin/Liger-Kernel", - "category": "kernel examples", - "github_about_section": "Efficient Triton Kernels for LLM Training", - "homepage_link": "https://openreview.net/pdf?id=36SjAIT42G", - "github_topic_closest_fit": "triton", - "contributors_all": 120, - "contributors_2025": 78, - "contributors_2024": 61, - "contributors_2023": 0, - "growth_2025_percent": 27, - "90-day-contributor-retention-rate": 0.2765957446808511, - "180-day-contributor-retention-rate": 0.18421052631578946 + "repo_name": "hip", + "repo_link": "https://github.com/ROCm/hip", + "github_about_section": "HIP: C++ Heterogeneous-Compute Interface for Portability", + "homepage_link": "https://rocmdocs.amd.com/projects/HIP", + "contributors_all": 288, + "contributors_2025": 46, + "contributors_2024": 31, + "contributors_2023": 25 }, { - "repo_name": "roctracer", - "repo_link": "https://github.com/ROCm/roctracer", - "github_about_section": "[DEPRECATED] Moved to ROCm/rocm-systems repo", - "homepage_link": "https://github.com/ROCm/rocm-systems", - "contributors_all": 45, - "contributors_2025": 8, - "contributors_2024": 11, - "contributors_2023": 6, - "growth_2025_percent": -27, - "90-day-contributor-retention-rate": 0.2727272727272727, - "180-day-contributor-retention-rate": 0.2682926829268293 + "repo_name": "openevolve", + "repo_link": "https://github.com/codelion/openevolve", + "github_about_section": "Open-source implementation of AlphaEvolve", + "github_topic_closest_fit": "genetic-algorithm", + "contributors_all": 46, + "contributors_2025": 46, + "contributors_2024": 0, + "contributors_2023": 0 + }, + { + "repo_name": "lightning-thunder", + "repo_link": "https://github.com/Lightning-AI/lightning-thunder", + "github_about_section": "PyTorch compiler that accelerates training and inference. Get built-in optimizations for performance, memory, parallelism, and easily write your own.", + "contributors_all": 76, + "contributors_2025": 44, + "contributors_2024": 47, + "contributors_2023": 29 + }, + { + "repo_name": "truss", + "repo_link": "https://github.com/basetenlabs/truss", + "category": "inference engine", + "github_about_section": "The simplest way to serve AI/ML models in production", + "homepage_link": "https://truss.baseten.co", + "github_topic_closest_fit": "inference", + "contributors_all": 72, + "contributors_2025": 44, + "contributors_2024": 30, + "contributors_2023": 21 + }, + { + "repo_name": "ondemand", + "repo_link": "https://github.com/OSC/ondemand", + "github_about_section": "Supercomputing. Seamlessly. Open, Interactive HPC Via the Web", + "homepage_link": "https://openondemand.org", + "github_topic_closest_fit": "hpc", + "contributors_all": 117, + "contributors_2025": 43, + "contributors_2024": 23, + "contributors_2023": 21 + }, + { + "repo_name": "pybind11", + "repo_link": "https://github.com/pybind/pybind11", + "github_about_section": "Seamless operability between C++11 and Python", + "homepage_link": "https://pybind11.readthedocs.io", + "github_topic_closest_fit": "bindings", + "contributors_all": 404, + "contributors_2025": 43, + "contributors_2024": 45, + "contributors_2023": 42 + }, + { + "repo_name": "cuda-python", + "repo_link": "https://github.com/NVIDIA/cuda-python", + "github_about_section": "CUDA Python: Performance meets Productivity", + "homepage_link": "https://nvidia.github.io/cuda-python", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 48, + "contributors_2025": 41, + "contributors_2024": 12, + "contributors_2023": 1 }, { - "repo_name": "Vulkan-Docs", - "repo_link": "https://github.com/KhronosGroup/Vulkan-Docs", - "category": "graphics api", - "github_about_section": "The Vulkan API Specification and related tools", - "homepage_link": "https://vulkan.org", - "github_topic_closest_fit": "vulkan", - "contributors_all": 141, - "contributors_2025": 18, - "contributors_2024": 21, - "contributors_2023": 34, - "growth_2025_percent": -14, - "90-day-contributor-retention-rate": 0.27205882352941174, - "180-day-contributor-retention-rate": 0.19402985074626866 + "repo_name": "warp", + "repo_link": "https://github.com/NVIDIA/warp", + "category": "spatial computing", + "github_about_section": "A Python framework for accelerated simulation, data generation and spatial computing.", + "homepage_link": "https://nvidia.github.io/warp", + "github_topic_closest_fit": "physics-simulation", + "contributors_all": 79, + "contributors_2025": 40, + "contributors_2024": 29, + "contributors_2023": 17 }, { - "repo_name": "DeepSpeed", - "repo_link": "https://github.com/deepspeedai/DeepSpeed", - "github_about_section": "DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.", - "homepage_link": "https://deepspeed.ai", - "contributors_all": 442, - "contributors_2025": 96, - "contributors_2024": 134, - "contributors_2023": 165, - "growth_2025_percent": -28, - "90-day-contributor-retention-rate": 0.2621359223300971, - "180-day-contributor-retention-rate": 0.21212121212121213 + "repo_name": "metaflow", + "repo_link": "https://github.com/Netflix/metaflow", + "github_about_section": "Build, Manage and Deploy AI/ML Systems", + "homepage_link": "https://metaflow.org", + "contributors_all": 121, + "contributors_2025": 37, + "contributors_2024": 35, + "contributors_2023": 28 }, { "repo_name": "numba", @@ -1148,94 +707,51 @@ "contributors_all": 430, "contributors_2025": 36, "contributors_2024": 32, - "contributors_2023": 55, - "growth_2025_percent": 12, - "90-day-contributor-retention-rate": 0.2589073634204275, - "180-day-contributor-retention-rate": 0.2028985507246377 - }, - { - "repo_name": "jupyterlab", - "repo_link": "https://github.com/jupyterlab/jupyterlab", - "category": "user interface", - "github_about_section": "JupyterLab computational environment.", - "homepage_link": "https://jupyterlab.readthedocs.io", - "github_topic_closest_fit": "jupyter", - "contributors_all": 698, - "contributors_2025": 77, - "contributors_2024": 85, - "contributors_2023": 100, - "growth_2025_percent": -9, - "90-day-contributor-retention-rate": 0.25735294117647056, - "180-day-contributor-retention-rate": 0.22388059701492538 + "contributors_2023": 55 }, { - "repo_name": "ort", - "repo_link": "https://github.com/pytorch/ort", - "github_about_section": "Accelerate PyTorch models with ONNX Runtime", - "contributors_all": 47, - "contributors_2025": 0, - "contributors_2024": 7, - "contributors_2023": 9, - "growth_2025_percent": -100, - "90-day-contributor-retention-rate": 0.2553191489361702, - "180-day-contributor-retention-rate": 0.1702127659574468 + "repo_name": "SWE-bench", + "repo_link": "https://github.com/SWE-bench/SWE-bench", + "category": "benchmark", + "github_about_section": "SWE-bench: Can Language Models Resolve Real-world Github Issues?", + "homepage_link": "https://swebench.com", + "github_topic_closest_fit": "benchmark", + "contributors_all": 66, + "contributors_2025": 33, + "contributors_2024": 37, + "contributors_2023": 9 }, { - "repo_name": "scipy", - "repo_link": "https://github.com/scipy/scipy", - "category": "scientific computing", - "github_about_section": "SciPy library main repository", - "homepage_link": "https://scipy.org", - "github_topic_closest_fit": "scientific-computing", - "contributors_all": 1973, - "contributors_2025": 210, - "contributors_2024": 251, - "contributors_2023": 245, - "growth_2025_percent": -16, - "90-day-contributor-retention-rate": 0.2542901716068643, - "180-day-contributor-retention-rate": 0.2216931216931217 + "repo_name": "AdaptiveCpp", + "repo_link": "https://github.com/AdaptiveCpp/AdaptiveCpp", + "github_about_section": "Compiler for multiple programming models (SYCL, C++ standard parallelism, HIP/CUDA) for CPUs and GPUs from all vendors: The independent, community-driven compiler for C++-based heterogeneous programming models. Lets applications adapt themselves to all the hardware in the system - even at runtime!", + "homepage_link": "https://adaptivecpp.github.io", + "contributors_all": 93, + "contributors_2025": 32, + "contributors_2024": 32, + "contributors_2023": 24 }, { - "repo_name": "torchdynamo", - "repo_link": "https://github.com/pytorch/torchdynamo", - "github_about_section": "A Python-level JIT compiler designed to make unmodified PyTorch programs faster.", - "contributors_all": 63, - "contributors_2025": 0, - "contributors_2024": 1, - "contributors_2023": 4, - "growth_2025_percent": -100, - "90-day-contributor-retention-rate": 0.25396825396825395, - "180-day-contributor-retention-rate": 0.09523809523809523 + "repo_name": "Triton-distributed", + "repo_link": "https://github.com/ByteDance-Seed/Triton-distributed", + "github_about_section": "Distributed Compiler based on Triton for Parallel Systems", + "homepage_link": "https://triton-distributed.readthedocs.io", + "contributors_all": 30, + "contributors_2025": 30, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "cutlass", - "repo_link": "https://github.com/NVIDIA/cutlass", + "repo_name": "ThunderKittens", + "repo_link": "https://github.com/HazyResearch/ThunderKittens", "category": "parallel computing", - "github_about_section": "CUDA Templates and Python DSLs for High-Performance Linear Algebra", - "homepage_link": "https://docs.nvidia.com/cutlass/index.html", + "github_about_section": "Tile primitives for speedy kernels", + "homepage_link": "https://hazyresearch.stanford.edu/blog/2024-10-29-tk2", "github_topic_closest_fit": "parallel-programming", - "contributors_all": 238, - "contributors_2025": 94, - "contributors_2024": 64, - "contributors_2023": 66, - "growth_2025_percent": 46, - "90-day-contributor-retention-rate": 0.25118483412322273, - "180-day-contributor-retention-rate": 0.2393617021276596 - }, - { - "repo_name": "goose", - "repo_link": "https://github.com/block/goose", - "category": "agent", - "github_about_section": "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", - "github_topic_closest_fit": "ai-agents", - "contributors_all": 332, - "contributors_2025": 319, - "contributors_2024": 32, - "contributors_2023": 0, - "growth_2025_percent": 896, - "90-day-contributor-retention-rate": 0.25, - "180-day-contributor-retention-rate": 0.2540983606557377 + "contributors_all": 34, + "contributors_2025": 29, + "contributors_2024": 13, + "contributors_2023": 0 }, { "repo_name": "dstack", @@ -1247,211 +763,182 @@ "contributors_all": 69, "contributors_2025": 28, "contributors_2024": 42, - "contributors_2023": 14, - "growth_2025_percent": -33, - "90-day-contributor-retention-rate": 0.25, - "180-day-contributor-retention-rate": 0.2545454545454545 + "contributors_2023": 14 }, { - "repo_name": "KernelBench", - "repo_link": "https://github.com/ScalingIntelligence/KernelBench", - "category": "benchmark", - "github_about_section": "KernelBench: Can LLMs Write GPU Kernels? - Benchmark with Torch -> CUDA problems", - "homepage_link": "https://scalingintelligence.stanford.edu/blogs/kernelbench", - "github_topic_closest_fit": "benchmark", - "contributors_all": 19, - "contributors_2025": 16, - "contributors_2024": 3, - "contributors_2023": 0, - "growth_2025_percent": 433, - "90-day-contributor-retention-rate": 0.25, - "180-day-contributor-retention-rate": 0.2 + "repo_name": "ome", + "repo_link": "https://github.com/sgl-project/ome", + "github_about_section": "OME is a Kubernetes operator for enterprise-grade management and serving of Large Language Models (LLMs)", + "homepage_link": "http://docs.sglang.ai/ome", + "github_topic_closest_fit": "k8s", + "contributors_all": 28, + "contributors_2025": 28, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "kernels", - "repo_link": "https://github.com/huggingface/kernels", - "category": "gpu kernels", - "github_about_section": "Load compute kernels from the Hub", - "contributors_all": 15, - "contributors_2025": 14, - "contributors_2024": 2, - "contributors_2023": 0, - "growth_2025_percent": 600, - "90-day-contributor-retention-rate": 0.25, - "180-day-contributor-retention-rate": 0.3333333333333333 + "repo_name": "pocl", + "repo_link": "https://github.com/pocl/pocl", + "github_about_section": "pocl - Portable Computing Language", + "homepage_link": "https://portablecl.org", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 166, + "contributors_2025": 26, + "contributors_2024": 27, + "contributors_2023": 21 }, { - "repo_name": "ZLUDA", - "repo_link": "https://github.com/vosen/ZLUDA", - "github_about_section": "CUDA on non-NVIDIA GPUs", - "homepage_link": "https://vosen.github.io/ZLUDA", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 15, - "contributors_2025": 8, - "contributors_2024": 4, - "contributors_2023": 0, - "growth_2025_percent": 100, - "90-day-contributor-retention-rate": 0.25, - "180-day-contributor-retention-rate": 0.2222222222222222 + "repo_name": "server", + "repo_link": "https://github.com/triton-inference-server/server", + "github_about_section": "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", + "github_topic_closest_fit": "inference", + "contributors_all": 147, + "contributors_2025": 24, + "contributors_2024": 36, + "contributors_2023": 34 }, { - "repo_name": "omnitrace", - "repo_link": "https://github.com/ROCm/omnitrace", - "category": "performance testing", - "github_about_section": "Omnitrace: Application Profiling, Tracing, and Analysis", - "homepage_link": "https://rocm.docs.amd.com/projects/omnitrace", - "github_topic_closest_fit": "profiling", - "contributors_all": 16, - "contributors_2025": 2, - "contributors_2024": 12, - "contributors_2023": 2, - "growth_2025_percent": -83, - "90-day-contributor-retention-rate": 0.25, - "180-day-contributor-retention-rate": 0.1875 + "repo_name": "Vulkan-Hpp", + "repo_link": "https://github.com/KhronosGroup/Vulkan-Hpp", + "category": "graphics api", + "github_about_section": "Open-Source Vulkan C++ API", + "homepage_link": "https://vulkan.org", + "github_topic_closest_fit": "vulkan", + "contributors_all": 102, + "contributors_2025": 21, + "contributors_2024": 15, + "contributors_2023": 15 }, { - "repo_name": "BitBLAS", - "repo_link": "https://github.com/microsoft/BitBLAS", - "category": "Basic Linear Algebra Subprograms (BLAS)", - "github_about_section": "BitBLAS is a library to support mixed-precision matrix multiplications, especially for quantized LLM deployment.", - "github_topic_closest_fit": "matrix-multiplication", - "contributors_all": 17, - "contributors_2025": 5, - "contributors_2024": 14, - "contributors_2023": 0, - "growth_2025_percent": -64, - "90-day-contributor-retention-rate": 0.23529411764705882, - "180-day-contributor-retention-rate": 0.125 + "repo_name": "ccache", + "repo_link": "https://github.com/ccache/ccache", + "github_about_section": "ccache - a fast compiler cache", + "homepage_link": "https://ccache.dev", + "contributors_all": 218, + "contributors_2025": 20, + "contributors_2024": 28, + "contributors_2023": 22 }, { - "repo_name": "flashinfer", - "repo_link": "https://github.com/flashinfer-ai/flashinfer", - "category": "gpu kernels", - "github_about_section": "FlashInfer: Kernel Library for LLM Serving", - "homepage_link": "https://flashinfer.ai", - "github_topic_closest_fit": "attention", - "contributors_all": 205, - "contributors_2025": 158, - "contributors_2024": 50, - "contributors_2023": 11, - "growth_2025_percent": 216, - "90-day-contributor-retention-rate": 0.2265625, - "180-day-contributor-retention-rate": 0.15217391304347827 + "repo_name": "lapack", + "repo_link": "https://github.com/Reference-LAPACK/lapack", + "category": "linear algebra", + "github_about_section": "LAPACK is a library of Fortran subroutines for solving the most commonly occurring problems in numerical linear algebra.", + "homepage_link": "https://netlib.org/lapack", + "github_topic_closest_fit": "linear-algebra", + "contributors_all": 178, + "contributors_2025": 20, + "contributors_2024": 24, + "contributors_2023": 42 }, { - "repo_name": "nvcc4jupyter", - "repo_link": "https://github.com/andreinechaev/nvcc4jupyter", - "github_about_section": "A plugin for Jupyter Notebook to run CUDA C/C++ code", - "homepage_link": "https://nvcc4jupyter.readthedocs.io", - "contributors_all": 9, - "contributors_2025": 0, - "contributors_2024": 3, - "contributors_2023": 3, - "growth_2025_percent": -100, - "90-day-contributor-retention-rate": 0.2222222222222222, - "180-day-contributor-retention-rate": 0.1111111111111111 + "repo_name": "Vulkan-Tools", + "repo_link": "https://github.com/KhronosGroup/Vulkan-Tools", + "category": "graphics api", + "github_about_section": "Vulkan Development Tools", + "homepage_link": "https://vulkan.org", + "github_topic_closest_fit": "vulkan", + "contributors_all": 248, + "contributors_2025": 20, + "contributors_2024": 24, + "contributors_2023": 24 + }, + { + "repo_name": "tflite-micro", + "repo_link": "https://github.com/tensorflow/tflite-micro", + "github_about_section": "Infrastructure to enable deployment of ML models to low-power resource-constrained embedded targets (including microcontrollers and digital signal processors).", + "contributors_all": 111, + "contributors_2025": 19, + "contributors_2024": 25, + "contributors_2023": 31 + }, + { + "repo_name": "Vulkan-Docs", + "repo_link": "https://github.com/KhronosGroup/Vulkan-Docs", + "category": "graphics api", + "github_about_section": "The Vulkan API Specification and related tools", + "homepage_link": "https://vulkan.org", + "github_topic_closest_fit": "vulkan", + "contributors_all": 141, + "contributors_2025": 18, + "contributors_2024": 21, + "contributors_2023": 34 + }, + { + "repo_name": "quack", + "repo_link": "https://github.com/Dao-AILab/quack", + "category": "kernel examples", + "github_about_section": "A Quirky Assortment of CuTe Kernels", + "contributors_all": 17, + "contributors_2025": 17, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "numpy", - "repo_link": "https://github.com/numpy/numpy", - "category": "scientific computing", - "github_about_section": "The fundamental package for scientific computing with Python.", - "homepage_link": "https://numpy.org", - "github_topic_closest_fit": "scientific-computing", - "contributors_all": 2172, - "contributors_2025": 235, - "contributors_2024": 233, - "contributors_2023": 252, - "growth_2025_percent": 0, - "90-day-contributor-retention-rate": 0.21983786361468766, - "180-day-contributor-retention-rate": 0.18807561803199224 + "repo_name": "oneDPL", + "repo_link": "https://github.com/uxlfoundation/oneDPL", + "github_about_section": "oneAPI DPC++ Library (oneDPL)", + "homepage_link": "https://software.intel.com/content/www/us/en/develop/tools/oneapi/components/dpc-library.html", + "contributors_all": 67, + "contributors_2025": 17, + "contributors_2024": 29, + "contributors_2023": 28 }, { - "repo_name": "SWE-bench", - "repo_link": "https://github.com/SWE-bench/SWE-bench", + "repo_name": "KernelBench", + "repo_link": "https://github.com/ScalingIntelligence/KernelBench", "category": "benchmark", - "github_about_section": "SWE-bench: Can Language Models Resolve Real-world Github Issues?", - "homepage_link": "https://swebench.com", + "github_about_section": "KernelBench: Can LLMs Write GPU Kernels? - Benchmark with Torch -> CUDA problems", + "homepage_link": "https://scalingintelligence.stanford.edu/blogs/kernelbench", "github_topic_closest_fit": "benchmark", - "contributors_all": 66, - "contributors_2025": 33, - "contributors_2024": 37, - "contributors_2023": 9, - "growth_2025_percent": -10, - "90-day-contributor-retention-rate": 0.21428571428571427, - "180-day-contributor-retention-rate": 0.14583333333333334 - }, - { - "repo_name": "burn", - "repo_link": "https://github.com/tracel-ai/burn", - "github_about_section": "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", - "contributors_all": 237, - "contributors_2025": 99, - "contributors_2024": 104, - "contributors_2023": 62, - "growth_2025_percent": -4, - "90-day-contributor-retention-rate": 0.21359223300970873, - "180-day-contributor-retention-rate": 0.13043478260869565 + "contributors_all": 19, + "contributors_2025": 16, + "contributors_2024": 3, + "contributors_2023": 0 }, { - "repo_name": "llama.cpp", - "repo_link": "https://github.com/ggml-org/llama.cpp", - "category": "inference engine", - "github_about_section": "LLM inference in C/C++", - "homepage_link": "https://ggml.ai", - "github_topic_closest_fit": "inference", - "contributors_all": 1374, - "contributors_2025": 535, - "contributors_2024": 575, - "contributors_2023": 461, - "growth_2025_percent": -6, - "90-day-contributor-retention-rate": 0.21308724832214765, - "180-day-contributor-retention-rate": 0.16818181818181818 + "repo_name": "reference-kernels", + "repo_link": "https://github.com/gpu-mode/reference-kernels", + "category": "kernel examples", + "github_about_section": "Official Problem Sets / Reference Kernels for the GPU MODE Leaderboard!", + "homepage_link": "https://gpumode.com", + "contributors_all": 16, + "contributors_2025": 16, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "Vulkan-Hpp", - "repo_link": "https://github.com/KhronosGroup/Vulkan-Hpp", - "category": "graphics api", - "github_about_section": "Open-Source Vulkan C++ API", - "homepage_link": "https://vulkan.org", - "github_topic_closest_fit": "vulkan", - "contributors_all": 102, - "contributors_2025": 21, - "contributors_2024": 15, - "contributors_2023": 15, - "growth_2025_percent": 40, - "90-day-contributor-retention-rate": 0.20833333333333334, - "180-day-contributor-retention-rate": 0.18085106382978725 + "repo_name": "synthetic-data-kit", + "repo_link": "https://github.com/meta-llama/synthetic-data-kit", + "category": "synthetic data generation", + "github_about_section": "Tool for generating high quality Synthetic datasets", + "homepage_link": "https://pypi.org/project/synthetic-data-kit", + "github_topic_closest_fit": "synthetic-dataset-generation", + "contributors_all": 15, + "contributors_2025": 15, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "pybind11", - "repo_link": "https://github.com/pybind/pybind11", - "github_about_section": "Seamless operability between C++11 and Python", - "homepage_link": "https://pybind11.readthedocs.io", - "github_topic_closest_fit": "bindings", - "contributors_all": 404, - "contributors_2025": 43, - "contributors_2024": 45, - "contributors_2023": 42, - "growth_2025_percent": -4, - "90-day-contributor-retention-rate": 0.2071611253196931, - "180-day-contributor-retention-rate": 0.18441558441558442 + "repo_name": "tritonparse", + "repo_link": "https://github.com/meta-pytorch/tritonparse", + "github_about_section": "TritonParse: A Compiler Tracer, Visualizer, and Reproducer for Triton Kernels", + "homepage_link": "https://meta-pytorch.org/tritonparse", + "contributors_all": 15, + "contributors_2025": 15, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "modelcontextprotocol", - "repo_link": "https://github.com/modelcontextprotocol/modelcontextprotocol", - "category": "mcp", - "github_about_section": "Specification and documentation for the Model Context Protocol", - "homepage_link": "https://modelcontextprotocol.io", - "github_topic_closest_fit": "mcp", - "contributors_all": 327, - "contributors_2025": 298, - "contributors_2024": 42, - "contributors_2023": 0, - "growth_2025_percent": 609, - "90-day-contributor-retention-rate": 0.20512820512820512, - "180-day-contributor-retention-rate": 0.12987012987012986 + "repo_name": "kernels", + "repo_link": "https://github.com/huggingface/kernels", + "category": "gpu kernels", + "github_about_section": "Load compute kernels from the Hub", + "contributors_all": 15, + "contributors_2025": 14, + "contributors_2024": 2, + "contributors_2023": 0 }, { "repo_name": "Wan2.2", @@ -1463,246 +950,252 @@ "contributors_all": 14, "contributors_2025": 14, "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.2, - "180-day-contributor-retention-rate": "not-enough-data" + "contributors_2023": 0 }, { - "repo_name": "ComfyUI", - "repo_link": "https://github.com/comfyanonymous/ComfyUI", - "category": "user interface", - "github_about_section": "The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.", - "homepage_link": "https://comfy.org", - "github_topic_closest_fit": "stable-diffusion", - "contributors_all": 278, - "contributors_2025": 108, - "contributors_2024": 119, - "contributors_2023": 94, - "growth_2025_percent": -9, - "90-day-contributor-retention-rate": 0.19753086419753085, - "180-day-contributor-retention-rate": 0.16216216216216217 + "repo_name": "SYCL-Docs", + "repo_link": "https://github.com/KhronosGroup/SYCL-Docs", + "github_about_section": "SYCL Open Source Specification", + "homepage_link": "https://khronos.org/sycl", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 67, + "contributors_2025": 13, + "contributors_2024": 20, + "contributors_2023": 27 }, { - "repo_name": "ccache", - "repo_link": "https://github.com/ccache/ccache", - "github_about_section": "ccache - a fast compiler cache", - "homepage_link": "https://ccache.dev", - "contributors_all": 218, - "contributors_2025": 20, - "contributors_2024": 28, - "contributors_2023": 22, - "growth_2025_percent": -28, - "90-day-contributor-retention-rate": 0.18396226415094338, - "180-day-contributor-retention-rate": 0.15384615384615385 + "repo_name": "Primus-Turbo", + "repo_link": "https://github.com/AMD-AGI/Primus-Turbo", + "contributors_all": 12, + "contributors_2025": 12, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "transformers", - "repo_link": "https://github.com/huggingface/transformers", - "github_about_section": "Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.", - "homepage_link": "https://huggingface.co/transformers", - "contributors_all": 3582, - "contributors_2025": 860, - "contributors_2024": 769, - "contributors_2023": 758, - "growth_2025_percent": 11, - "90-day-contributor-retention-rate": 0.1778975741239892, - "180-day-contributor-retention-rate": 0.146606914212548 + "repo_name": "flashinfer-bench", + "repo_link": "https://github.com/flashinfer-ai/flashinfer-bench", + "category": "benchmark", + "github_about_section": "Building the Virtuous Cycle for AI-driven LLM Systems", + "homepage_link": "https://bench.flashinfer.ai", + "github_topic_closest_fit": "benchmark", + "contributors_all": 12, + "contributors_2025": 11, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "mistral-inference", - "repo_link": "https://github.com/mistralai/mistral-inference", - "category": "inference engine", - "github_about_section": "Official inference library for Mistral models", - "homepage_link": "https://mistral.ai", - "github_topic_closest_fit": "inference", - "contributors_all": 29, - "contributors_2025": 2, - "contributors_2024": 17, - "contributors_2023": 14, - "growth_2025_percent": -88, - "90-day-contributor-retention-rate": 0.1724137931034483, - "180-day-contributor-retention-rate": 0.13793103448275862 + "repo_name": "FTorch", + "repo_link": "https://github.com/Cambridge-ICCS/FTorch", + "category": "wrapper", + "github_about_section": "A library for directly calling PyTorch ML models from Fortran.", + "homepage_link": "https://cambridge-iccs.github.io/FTorch", + "github_topic_closest_fit": "machine-learning", + "contributors_all": 20, + "contributors_2025": 11, + "contributors_2024": 8, + "contributors_2023": 9 }, { - "repo_name": "synthetic-data-kit", - "repo_link": "https://github.com/meta-llama/synthetic-data-kit", - "category": "synthetic data generation", - "github_about_section": "Tool for generating high quality Synthetic datasets", - "homepage_link": "https://pypi.org/project/synthetic-data-kit", - "github_topic_closest_fit": "synthetic-dataset-generation", - "contributors_all": 15, - "contributors_2025": 15, + "repo_name": "TensorRT", + "repo_link": "https://github.com/NVIDIA/TensorRT", + "github_about_section": "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", + "contributors_all": 104, + "contributors_2025": 10, + "contributors_2024": 18, + "contributors_2023": 19 + }, + { + "repo_name": "TileIR", + "repo_link": "https://github.com/microsoft/TileIR", + "category": "parallel computing dsl", + "github_about_section": "TileIR (tile-ir) is a concise domain-specific IR designed to streamline the development of high-performance GPU/CPU kernels (e.g., GEMM, Dequant GEMM, FlashAttention, LinearAttention). By employing a Pythonic syntax with an underlying compiler infrastructure on top of TVM, TileIR allows developers to focus on productivity without sacrificing the low-level optimizations necessary for state-of-the-art performance.", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 10, + "contributors_2025": 10, + "contributors_2024": 1, + "contributors_2023": 0 + }, + { + "repo_name": "kernels-community", + "repo_link": "https://github.com/huggingface/kernels-community", + "category": "gpu kernels", + "homepage_link": "https://huggingface.co/kernels-community", + "github_about_section": "Kernel sources for https://huggingface.co/kernels-community", + "contributors_all": 9, + "contributors_2025": 9, "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.16666666666666666, - "180-day-contributor-retention-rate": 0.0 + "contributors_2023": 0 }, { - "repo_name": "accelerate", - "repo_link": "https://github.com/huggingface/accelerate", - "github_about_section": "A simple way to launch, train, and use PyTorch models on almost any device and distributed configuration, automatic mixed precision (including fp8), and easy-to-configure FSDP and DeepSpeed support.", - "homepage_link": "https://huggingface.co/docs/accelerate", - "contributors_all": 392, - "contributors_2025": 97, - "contributors_2024": 124, - "contributors_2023": 149, - "growth_2025_percent": -21, - "90-day-contributor-retention-rate": 0.16442048517520216, - "180-day-contributor-retention-rate": 0.14772727272727273 + "repo_name": "GEAK-agent", + "repo_link": "https://github.com/AMD-AGI/GEAK-agent", + "category": "agent", + "github_about_section": "It is an LLM-based AI agent, which can write correct and efficient gpu kernels automatically.", + "github_topic_closest_fit": "ai-agents", + "contributors_all": 9, + "contributors_2025": 9, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "mcp-agent", - "repo_link": "https://github.com/lastmile-ai/mcp-agent", - "category": "mcp", - "github_about_section": "Build effective agents using Model Context Protocol and simple workflow patterns", - "github_topic_closest_fit": "mcp", - "contributors_all": 63, - "contributors_2025": 63, - "contributors_2024": 1, - "contributors_2023": 0, - "growth_2025_percent": 6200, - "90-day-contributor-retention-rate": 0.1509433962264151, - "180-day-contributor-retention-rate": 0.15625 + "repo_name": "neuronx-distributed-inference", + "repo_link": "https://github.com/aws-neuron/neuronx-distributed-inference", + "contributors_all": 11, + "contributors_2025": 9, + "contributors_2024": 3, + "contributors_2023": 0 }, { - "repo_name": "trl", - "repo_link": "https://github.com/huggingface/trl", - "github_about_section": "Train transformer language models with reinforcement learning.", - "homepage_link": "http://hf.co/docs/trl", - "contributors_all": 433, - "contributors_2025": 189, - "contributors_2024": 154, - "contributors_2023": 122, - "growth_2025_percent": 22, - "90-day-contributor-retention-rate": 0.14705882352941177, - "180-day-contributor-retention-rate": 0.11904761904761904 + "repo_name": "OpenCL-SDK", + "repo_link": "https://github.com/KhronosGroup/OpenCL-SDK", + "github_about_section": "OpenCL SDK", + "homepage_link": "https://khronos.org/opencl", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 25, + "contributors_2025": 8, + "contributors_2024": 6, + "contributors_2023": 9 + }, + { + "repo_name": "ZLUDA", + "repo_link": "https://github.com/vosen/ZLUDA", + "github_about_section": "CUDA on non-NVIDIA GPUs", + "homepage_link": "https://vosen.github.io/ZLUDA", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 15, + "contributors_2025": 8, + "contributors_2024": 4, + "contributors_2023": 0 }, { - "repo_name": "neuronx-distributed-inference", - "repo_link": "https://github.com/aws-neuron/neuronx-distributed-inference", - "contributors_all": 11, - "contributors_2025": 9, - "contributors_2024": 3, - "contributors_2023": 0, - "growth_2025_percent": 200, - "90-day-contributor-retention-rate": 0.14285714285714285, - "180-day-contributor-retention-rate": 0.2 + "repo_name": "intelliperf", + "repo_link": "https://github.com/AMDResearch/intelliperf", + "category": "performance testing", + "github_about_section": "Automated bottleneck detection and solution orchestration", + "homepage_link": "https://arxiv.org/html/2508.20258v1", + "github_topic_closest_fit": "profiling", + "contributors_all": 7, + "contributors_2025": 7, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "peft", - "repo_link": "https://github.com/huggingface/peft", - "github_about_section": "PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.", - "homepage_link": "https://huggingface.co/docs/peft", - "github_topic_closest_fit": "lora", - "contributors_all": 272, - "contributors_2025": 69, - "contributors_2024": 111, - "contributors_2023": 115, - "growth_2025_percent": -37, - "90-day-contributor-retention-rate": 0.13877551020408163, - "180-day-contributor-retention-rate": 0.09051724137931035 + "repo_name": "nccl", + "repo_link": "https://github.com/NVIDIA/nccl", + "github_about_section": "Optimized primitives for collective multi-GPU communication", + "homepage_link": "https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/index.html", + "contributors_all": 51, + "contributors_2025": 7, + "contributors_2024": 5, + "contributors_2023": 6 }, { - "repo_name": "letta", - "repo_link": "https://github.com/letta-ai/letta", - "category": "agent", - "github_about_section": "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", - "github_topic_closest_fit": "ai-agents", - "contributors_all": 157, - "contributors_2025": 56, - "contributors_2024": 75, - "contributors_2023": 47, - "growth_2025_percent": -25, - "90-day-contributor-retention-rate": 0.13793103448275862, - "180-day-contributor-retention-rate": 0.0948905109489051 + "repo_name": "cudnn-frontend", + "repo_link": "https://github.com/NVIDIA/cudnn-frontend", + "category": "parallel computing", + "github_about_section": "cudnn_frontend provides a c++ wrapper for the cudnn backend API and samples on how to use it", + "homepage_link": "https://developer.nvidia.com/cudnn", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 12, + "contributors_2025": 6, + "contributors_2024": 5, + "contributors_2023": 1 }, { - "repo_name": "unsloth", - "repo_link": "https://github.com/unslothai/unsloth", - "category": "fine tuning", - "github_about_section": "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", - "github_topic_closest_fit": "fine-tuning", - "contributors_all": 127, - "contributors_2025": 102, - "contributors_2024": 27, - "contributors_2023": 3, - "growth_2025_percent": 277, - "90-day-contributor-retention-rate": 0.1348314606741573, - "180-day-contributor-retention-rate": 0.11864406779661017 + "repo_name": "BitBLAS", + "repo_link": "https://github.com/microsoft/BitBLAS", + "category": "Basic Linear Algebra Subprograms (BLAS)", + "github_about_section": "BitBLAS is a library to support mixed-precision matrix multiplications, especially for quantized LLM deployment.", + "github_topic_closest_fit": "matrix-multiplication", + "contributors_all": 17, + "contributors_2025": 5, + "contributors_2024": 14, + "contributors_2023": 0 }, { - "repo_name": "openevolve", - "repo_link": "https://github.com/codelion/openevolve", - "github_about_section": "Open-source implementation of AlphaEvolve", - "github_topic_closest_fit": "genetic-algorithm", - "contributors_all": 46, - "contributors_2025": 46, + "repo_name": "Self-Forcing", + "repo_link": "https://github.com/guandeh17/Self-Forcing", + "category": "video generation", + "github_about_section": "Official codebase for \"Self Forcing: Bridging Training and Inference in Autoregressive Video Diffusion\" (NeurIPS 2025 Spotlight)", + "homepage_link": "https://self-forcing.github.io", + "github_topic_closest_fit": "diffusion-models", + "contributors_all": 4, + "contributors_2025": 4, "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.12, - "180-day-contributor-retention-rate": "not-enough-data" + "contributors_2023": 0 }, { - "repo_name": "StreamDiffusion", - "repo_link": "https://github.com/cumulo-autumn/StreamDiffusion", - "category": "image generation", - "github_about_section": "StreamDiffusion: A Pipeline-Level Solution for Real-Time Interactive Generation", - "homepage_link": "https://arxiv.org/abs/2312.12491", - "github_topic_closest_fit": "diffusion-models", - "contributors_all": 29, - "contributors_2025": 0, - "contributors_2024": 9, - "contributors_2023": 25, - "growth_2025_percent": -100, - "90-day-contributor-retention-rate": 0.10344827586206896, - "180-day-contributor-retention-rate": 0.06896551724137931 + "repo_name": "TritonBench", + "repo_link": "https://github.com/thunlp/TritonBench", + "category": "benchmark", + "github_about_section": "TritonBench: Benchmarking Large Language Model Capabilities for Generating Triton Operators", + "homepage_link": "https://arxiv.org/abs/2502.14752", + "github_topic_closest_fit": "benchmark", + "contributors_all": 3, + "contributors_2025": 3, + "contributors_2024": 0, + "contributors_2023": 0 }, { - "repo_name": "TileIR", - "repo_link": "https://github.com/microsoft/TileIR", - "category": "parallel computing dsl", - "github_about_section": "TileIR (tile-ir) is a concise domain-specific IR designed to streamline the development of high-performance GPU/CPU kernels (e.g., GEMM, Dequant GEMM, FlashAttention, LinearAttention). By employing a Pythonic syntax with an underlying compiler infrastructure on top of TVM, TileIR allows developers to focus on productivity without sacrificing the low-level optimizations necessary for state-of-the-art performance.", - "github_topic_closest_fit": "parallel-programming", - "contributors_all": 10, - "contributors_2025": 10, - "contributors_2024": 1, - "contributors_2023": 0, - "growth_2025_percent": 900, - "90-day-contributor-retention-rate": 0.1, - "180-day-contributor-retention-rate": 0.0 + "repo_name": "hatchet", + "repo_link": "https://github.com/LLNL/hatchet", + "category": "performance testing", + "github_about_section": "Graph-indexed Pandas DataFrames for analyzing hierarchical performance data", + "homepage_link": "https://llnl-hatchet.readthedocs.io", + "github_topic_closest_fit": "profiling", + "contributors_all": 25, + "contributors_2025": 3, + "contributors_2024": 6, + "contributors_2023": 8 }, { - "repo_name": "quack", - "repo_link": "https://github.com/Dao-AILab/quack", - "category": "kernel examples", - "github_about_section": "A Quirky Assortment of CuTe Kernels", - "contributors_all": 17, - "contributors_2025": 17, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.09090909090909091, - "180-day-contributor-retention-rate": "not-enough-data" + "repo_name": "streamv2v", + "repo_link": "https://github.com/Jeff-LiangF/streamv2v", + "category": "video generation", + "github_about_section": "Official Pytorch implementation of StreamV2V.", + "homepage_link": "https://jeff-liangf.github.io/projects/streamv2v", + "github_topic_closest_fit": "diffusion-models", + "contributors_all": 7, + "contributors_2025": 3, + "contributors_2024": 6, + "contributors_2023": 0 }, { - "repo_name": "ollama", - "repo_link": "https://github.com/ollama/ollama", + "repo_name": "mistral-inference", + "repo_link": "https://github.com/mistralai/mistral-inference", "category": "inference engine", - "github_about_section": "Get up and running with OpenAI gpt-oss, DeepSeek-R1, Gemma 3 and other models.", - "homepage_link": "https://ollama.com", + "github_about_section": "Official inference library for Mistral models", + "homepage_link": "https://mistral.ai", "github_topic_closest_fit": "inference", - "contributors_all": 574, - "contributors_2025": 202, - "contributors_2024": 314, - "contributors_2023": 97, - "growth_2025_percent": -35, - "90-day-contributor-retention-rate": 0.08414872798434442, - "180-day-contributor-retention-rate": 0.06458333333333334 + "contributors_all": 29, + "contributors_2025": 2, + "contributors_2024": 17, + "contributors_2023": 14 + }, + { + "repo_name": "omnitrace", + "repo_link": "https://github.com/ROCm/omnitrace", + "category": "performance testing", + "github_about_section": "Omnitrace: Application Profiling, Tracing, and Analysis", + "homepage_link": "https://rocm.docs.amd.com/projects/omnitrace", + "github_topic_closest_fit": "profiling", + "contributors_all": 16, + "contributors_2025": 2, + "contributors_2024": 12, + "contributors_2023": 2 + }, + { + "repo_name": "cuJSON", + "repo_link": "https://github.com/AutomataLab/cuJSON", + "category": "library leveraging parallel compute", + "github_about_section": "cuJSON: A Highly Parallel JSON Parser for GPUs", + "homepage_link": "https://dl.acm.org/doi/10.1145/3760250.3762222", + "github_topic_closest_fit": "json-parser", + "contributors_all": 2, + "contributors_2025": 2, + "contributors_2024": 2, + "contributors_2023": 0 }, { "repo_name": "IMO2025", @@ -1714,10 +1207,18 @@ "contributors_all": 2, "contributors_2025": 2, "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.0, - "180-day-contributor-retention-rate": "not-enough-data" + "contributors_2023": 0 + }, + { + "repo_name": "RaBitQ", + "repo_link": "https://github.com/gaoj0017/RaBitQ", + "github_about_section": "[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", + "github_topic_closest_fit": "nearest-neighbor-search", + "contributors_all": 2, + "contributors_2025": 2, + "contributors_2024": 1, + "contributors_2023": 0 }, { "repo_name": "torchdendrite", @@ -1727,10 +1228,49 @@ "contributors_all": 2, "contributors_2025": 1, "contributors_2024": 1, - "contributors_2023": 0, - "growth_2025_percent": 0, - "90-day-contributor-retention-rate": 0.0, - "180-day-contributor-retention-rate": 0.0 + "contributors_2023": 0 + }, + { + "repo_name": "triton-runner", + "repo_link": "https://github.com/toyaix/triton-runner", + "github_about_section": "Multi-Level Triton Runner supporting Python, IR, PTX, and cubin.", + "homepage_link": "https://triton-runner.org", + "contributors_all": 1, + "contributors_2025": 1, + "contributors_2024": 0, + "contributors_2023": 0 + }, + { + "repo_name": "nvcc4jupyter", + "repo_link": "https://github.com/andreinechaev/nvcc4jupyter", + "github_about_section": "A plugin for Jupyter Notebook to run CUDA C/C++ code", + "homepage_link": "https://nvcc4jupyter.readthedocs.io", + "contributors_all": 9, + "contributors_2025": 0, + "contributors_2024": 3, + "contributors_2023": 3 + }, + { + "repo_name": "CU2CL", + "repo_link": "https://github.com/vtsynergy/CU2CL", + "github_about_section": "A prototype CUDA-to-OpenCL source-to-source translator, built on the Clang compiler framework", + "homepage_link": "http://chrec.cs.vt.edu/cu2cl", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 3, + "contributors_2025": 0, + "contributors_2024": 0, + "contributors_2023": 0 + }, + { + "repo_name": "triSYCL", + "repo_link": "https://github.com/triSYCL/triSYCL", + "github_about_section": "Generic system-wide modern C++ for heterogeneous platforms with SYCL from Khronos Group", + "homepage_link": "https://trisycl.github.io/triSYCL/Doxygen/triSYCL/html/index.html", + "github_topic_closest_fit": "parallel-programming", + "contributors_all": 31, + "contributors_2025": 0, + "contributors_2024": 1, + "contributors_2023": 3 }, { "repo_name": "cupti", @@ -1741,38 +1281,37 @@ "contributors_all": 1, "contributors_2025": 0, "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": 0.0, - "180-day-contributor-retention-rate": 0.0 + "contributors_2023": 0 }, { - "repo_name": "kernels-community", - "repo_link": "https://github.com/huggingface/kernels-community", - "category": "gpu kernels", - "homepage_link": "https://huggingface.co/kernels-community", - "github_about_section": "Kernel sources for https://huggingface.co/kernels-community", - "contributors_all": 9, - "contributors_2025": 9, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": "not-enough-data", - "180-day-contributor-retention-rate": "not-enough-data" + "repo_name": "ort", + "repo_link": "https://github.com/pytorch/ort", + "github_about_section": "Accelerate PyTorch models with ONNX Runtime", + "contributors_all": 47, + "contributors_2025": 0, + "contributors_2024": 7, + "contributors_2023": 9 }, { - "repo_name": "GEAK-agent", - "repo_link": "https://github.com/AMD-AGI/GEAK-agent", - "category": "agent", - "github_about_section": "It is an LLM-based AI agent, which can write correct and efficient gpu kernels automatically.", - "github_topic_closest_fit": "ai-agents", - "contributors_all": 9, - "contributors_2025": 9, - "contributors_2024": 0, - "contributors_2023": 0, - "growth_2025_percent": "No 2024 data", - "90-day-contributor-retention-rate": "not-enough-data", - "180-day-contributor-retention-rate": "not-enough-data" + "repo_name": "StreamDiffusion", + "repo_link": "https://github.com/cumulo-autumn/StreamDiffusion", + "category": "image generation", + "github_about_section": "StreamDiffusion: A Pipeline-Level Solution for Real-Time Interactive Generation", + "homepage_link": "https://arxiv.org/abs/2312.12491", + "github_topic_closest_fit": "diffusion-models", + "contributors_all": 29, + "contributors_2025": 0, + "contributors_2024": 9, + "contributors_2023": 25 + }, + { + "repo_name": "torchdynamo", + "repo_link": "https://github.com/pytorch/torchdynamo", + "github_about_section": "A Python-level JIT compiler designed to make unmodified PyTorch programs faster.", + "contributors_all": 63, + "contributors_2025": 0, + "contributors_2024": 1, + "contributors_2023": 4 }, { "repo_name": "wandb",