High-efficiency floating-point neural network inference operators for mobile, server, and Web

BladeDISC is an end-to-end DynamIc Shape Compiler project for machine learning workloads.

The Tensor Algebra SuperOptimizer for Deep Learning

Batch normalization fusion for PyTorch

[MLSys 2021] IOS: Inter-Operator Scheduler for CNN Acceleration

Optimize layers structure of Keras model to reduce computation time

A set of tool which would make your life easier with Tensorrt and Onnxruntime. This Repo is designed for YoloV3

Official Repo for SparseLLM: Global Pruning of LLMs (NeurIPS 2024)

The blog, read report and code example for AGI/LLM related knowledge.

cross-platform modular neural network inference library, small and efficient

Learn the ins and outs of efficiently serving Large Language Models (LLMs). Dive into optimization techniques, including KV caching and Low Rank Adapters (LoRA), and gain hands-on experience with Predibase’s LoRAX framework inference server.

Faster inference YOLOv8: Optimize and export YOLOv8 models for faster inference using OpenVINO and Numpy 🔢

[WIP] A template for getting started writing code using GGML

A constrained expectation-maximization algorithm for feasible graph inference.

Modified inference engine for quantized convolution using product quantization

Batch Partitioning for Multi-PE Inference with TVM (2020)

🤖️ Optimized CUDA Kernels for Fast MobileNetV2 Inference

PyTorch Mobile: Android examples of usage in applications

Interface for TensorRT engines inference along with an example of YOLOv4 engine being used.

MLP-Rank: A graph theoretical approach to structured pruning of deep neural networks based on weighted Page Rank centrality as introduced by the related thesis.