Social and customizable AI writing assistant! ✍️

DISCONTINUED (July 2025): the maintenance for this repository has been discontinued. Please refer to: https://github.com/STMicroelectronics/st-mems-machine-learning-core/ for the up-to-date HAL repository. Configuration files, examples and tools for the Machine Learning Core feature (MLC) available in STMicroelectronics MEMS sensors.

Structured inference with Llama 2 in your browser

This project has implemented the RAG function on Jetson and supports TXT and PDF document formats. It uses MLC for 4-bit quantization of the Llama2-7b model, utilizes ChromaDB as the vector database, and connects these features with Llama_Index. I hope you like this project.

[ECCV'24] Self-training Room Layout Estimation via Geometry-aware Ray-casting

A collapsed-cone convolution radiotherapy dose calculation algorithm

The proposed method captures local and global correlations using Low Rank label subspace transformation for Multi-label learning with Missing Labels (LRMML). The model considers an auxiliary label matrix which facilitates the missing label information recovery.

《机器学习编译》的学习笔记和练习代码

Varian ESAPI script that plots doserate vs gantry angle for VMAT plans

TVM MLIR interface

Examples, tutorials, and tools for the MLC, a dedicated core for machine learning processing embedded in STMicroelectronics MEMS sensors.

Browser-based AI assistant that runs entirely on your device without requiring an internet connection (after initial model download)

Mathematics in Industry Landscape - A Presentation for MLC Webinar 2021

The code provided is a complete implementation of a diversity gain engine for an embedded system using STM32 microcontrollers. It involves initializing and configuring several peripherals, such as ADC, UART, SPI, and LoRa modules, to implement a diversity gain engine algorithm with three different combining techniques,: EGC, MLC, and SC.

edge_flutter enables seamless on-device Large Language Model inference using Apple's Neural Engine and Core ML, with automatic fallback to cloud models when device performance doesn't meet specified latency/memory targets.