Stars
The code used to generate the Wake Vision dataset for person detection
Windows Agent Arena (WAA) 🪟 is a scalable OS platform for testing and benchmarking of multi-modal AI agents.
[TPAMI 2025 & CVPR 2023] IGEV++: Iterative Multi-range Geometry Encoding Volumes for Stereo Matching
OTOv1-v3, NeurIPS, ICLR, TMLR, DNN Training, Compression, Structured Pruning, Erasing Operators, CNN, Diffusion, LLM
Machine Learning Systems: Foundations, Scaling, Agentic AI, and Physical AI (Vols I–IV) • Harvard CS249r | https://mlsysbook.ai
A robust and efficient TinyML inference engine.
Extracts the historic word occurrence of a search term in academic papers
On-device Speech AI for Apple Silicon
Croissant is a high-level format for machine learning datasets that brings together four rich layers.
EfficientNet, MobileNetV3, MobileNetV2, MixNet, etc in JAX w/ Flax Linen and Objax
A playbook for systematically maximizing the performance of deep learning models.
Example workflow for our data-centric speech benchmark
Official codebase used to develop Vision Transformer, SigLIP, MLP-Mixer, LiT and more.
Infrastructure to enable deployment of ML models to low-power resource-constrained embedded targets (including microcontrollers and digital signal processors).
Demonstration of the NEON approach for explainable clustering.
Few-shot Keyword Spotting in Any Language and Multilingual Spoken Word Corpus
attentional sequence-to-sequence model (based on LSTMs) in TFLite Micro, tested on Arduino Nano 33 BLE
This is a list of interesting papers and projects about TinyML.
Interactive Visualization to Build, Train and Test an Autoencoder with Tensorflow.js
MLPerf® Tiny is an ML benchmark suite for extremely low-power systems such as microcontrollers
Simulator + benchmark suite for Micro Aerial Vehicle design.
ginh is not a histogram: visually evaluate your shell usage patterns
scan github projects and automatically create pull requests for typos in readme files