GGML
Visit ToolGGML is a tensor library for machine learning that enables large models and high performance on commodity hardware. It is used by projects like llama.cpp and whisper.cpp.
At a glance
Trending
GGML is a tensor library for machine learning that enables large models and high performance on commodity hardware. It is used by projects like llama.cpp and whisper.cpp.
Trending
About
GGML is a tensor library designed for machine learning, focusing on enabling large models and high performance on readily available hardware. It boasts a low-level, cross-platform implementation with support for integer quantization, which is crucial for optimizing model size and speed. A key differentiator is its minimal design, aiming for a small and simple codebase, and its commitment to zero memory allocations during runtime for efficiency. The library is open-core, freely available under the MIT license, and encourages community contributions. It was founded by Georgi Gerganov in 2023 and later acquired by Hugging Face in 2026, highlighting its significance in the AI ecosystem.
Capabilities
Pricing & Plans
Open Source
Not publicly disclosed. Check ggml.ai for current pricing.
FAQs
Trending