#4343rd of 56 in ML frameworks and libraries
ggml
C/C++ tensor library for ML inference with quantization and many backends
- Stars
- 15.5k
- License
- MIT
- Last commit
- Oct 2026
- Last release
- Oct 2026
- Language
- C++
Missed the latest GitHub check; these facts may be older.
Overview
ggml is a dependency-free C/C++ tensor library that runs machine learning models on CPU, GPU, NPU and in the browser. It provides SIMD-optimized kernels and 2- to 8-bit integer quantization, plus MXFP4 and NVFP4, and it defines the GGUF file format. Changes to the core library are directed to llama.cpp rather than this repository.
Who it is for: C/C++ developers building on-device inference engines
Strengths
- Plain C/C++ with no dependencies; builds with CMake
- Targets x86, ARM, RISC-V, LoongArch, PowerPC, s390x and WebAssembly
- 2- to 8-bit integer quantization, plus MXFP4 and NVFP4 formats
- Zero memory allocations at runtime
Weaknesses
- Low-level library, not a ready-to-use model server or app
- Core changes are submitted via llama.cpp, not this repo
- README gives no hardware requirements or list of supported models
- Documentation is a GGUF spec plus a few external links
What it needs
- GPU optional
- Models: GGUF
Also in ML frameworks and libraries
See all 56| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 95 out of 100 |
| 2 | pytorchPython tensor library with GPU support and tape-based autograd for deep learning | 94 out of 100 |
| 3 | ultralyticsPython library and CLI for training and running YOLO vision models | 92 out of 100 |
| #4 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 89 out of 100 |
| #5 | pytorch-image-modelsPyTorch library of image models, pretrained weights and reference training scripts | 89 out of 100 |