k-bit quantization library for PyTorch that cuts LLM memory use for inference and training.
- Pricing
- Free
- Category
- Model Efficiency
- Compliance
- Not stated
Verified 2026-10-22Website
Model Efficiency
Hugging Face Optimum
Open-source optimization toolkit for training and running transformer models efficiently.
Verified 2026-10-22
Overview
Hugging Face Optimum is an open-source extension of Transformers, Diffusers, TIMM and Sentence Transformers that provides optimization tools for training and running models on targeted hardware. It covers ONNX and ONNX Runtime, OpenVINO, NVIDIA TensorRT-LLM, AMD Instinct GPUs and Ryzen AI NPUs, AWS Trainium and Inferentia, Intel Gaudi, and FuriosaAI accelerators. Optimum is distributed under Apache 2.0 and installs from PyPI.
Use cases
- Optimize models for target accelerators
- Export models to ONNX
- Quantize and accelerate inference
- Run training on specialized hardware
Sources
More in Model Efficiency
Category page →C/C++ LLM inference with 1.5- to 8-bit quantization and broad hardware backend support.
- Pricing
- Free
- Category
- Model Efficiency
- Compliance
- Not stated
Verified 2026-10-22Website
Open-source library for quantizing and compressing LLMs for efficient vLLM deployment.
- Pricing
- Free
- Category
- Model Efficiency
- Compliance
- Not stated
Verified 2026-10-22Website