NVIDIA / Model-OptimizerLinks
A unified library of SOTA model optimization techniques like quantization, pruning, distillation, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorRT-LLM, TensorRT, vLLM, etc. to optimize inference speed.
☆1,744Updated this week
Alternatives and similar repositories for Model-Optimizer
Users that are interested in Model-Optimizer are comparing it to the libraries listed below
Sorting:
- A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on H…☆3,043Updated last week
- PyTorch native quantization and sparsity for training and inference☆2,601Updated this week
- A pytorch quantization backend for optimum☆1,020Updated last month
- The Triton TensorRT-LLM Backend☆911Updated this week
- [ICML 2023] SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models☆1,576Updated last year
- FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.☆973Updated last year
- [MLSys'25] QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving; [MLSys'25] LServe: Efficient Long-sequence LLM Se…☆799Updated 9 months ago
- This repository contains tutorials and examples for Triton Inference Server☆813Updated 3 weeks ago
- FlashInfer: Kernel Library for LLM Serving☆4,400Updated this week
- Transformers-compatible library for applying various compression algorithms to LLMs for optimized deployment with vLLM☆2,514Updated this week
- A throughput-oriented high-performance serving framework for LLMs☆929Updated 2 months ago
- Distributed Compiler based on Triton for Parallel Systems☆1,303Updated last week
- Mirage Persistent Kernel: Compiling LLMs into a MegaKernel☆2,015Updated 2 weeks ago
- Flash Attention in ~100 lines of CUDA (forward pass only)☆1,036Updated last year
- FlagGems is an operator library for large language models implemented in the Triton Language.☆819Updated this week
- LLM model quantization (compression) toolkit with hw acceleration support for Nvidia CUDA, AMD ROCm, Intel XPU and Intel/AMD/Apple CPU vi…☆952Updated this week
- A fast communication-overlapping library for tensor/expert parallelism on GPUs.☆1,217Updated 4 months ago
- BitBLAS is a library to support mixed-precision matrix multiplications, especially for quantized LLM deployment.☆738Updated 4 months ago
- SOTA low-bit LLM quantization (INT8/FP8/MXFP8/INT4/MXFP4/NVFP4) & sparsity; leading model compression techniques on PyTorch, TensorFlow, …☆2,558Updated last week
- Advanced quantization toolkit for LLMs and VLMs. Support for WOQ, MXFP4, NVFP4, GGUF, Adaptive Schemes and seamless integration with Tra…☆785Updated last week
- depyf is a tool to help you understand and adapt to PyTorch compiler torch.compile.☆772Updated 2 months ago
- cudnn_frontend provides a c++ wrapper for the cudnn backend API and samples on how to use it☆665Updated 2 weeks ago
- PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT☆2,915Updated this week
- Pipeline Parallelism for PyTorch☆784Updated last year
- [EMNLP 2024 & AAAI 2026] A powerful toolkit for compressing large models including LLMs, VLMs, and video generative models.☆647Updated last month
- [MLSys 2024 Best Paper Award] AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration☆3,408Updated 5 months ago
- ONNX Optimizer☆784Updated this week
- Tile primitives for speedy kernels☆3,028Updated 3 weeks ago
- Fast CUDA matrix multiplication from scratch☆996Updated 4 months ago
- A Quirky Assortment of CuTe Kernels☆732Updated this week