# FastGRNN: A Fast, Accurate, Stable and Tiny Kilobyte Sized Gated Recurrent Neural Network

arXiv: https://arxiv.org/abs/1901.02358 · HTML: https://mlsanity.com/paper/1901.02358/

## Key facts

- Cited by 38 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Jan 2019.

## Built on by (most cited first, 30 of 38)

- [Matryoshka Representation Learning](https://mlsanity.com/paper/2205.13147/) (2022)
- [UnICORNN: A recurrent model for learning very long time dependencies](https://mlsanity.com/paper/2103.05487/) (2021)
- [Long Expressive Memory for Sequence Modeling](https://mlsanity.com/paper/2110.04744/) (2021)
- [Edge-Cloud Polarization and Collaboration: A Comprehensive Survey for AI](https://mlsanity.com/paper/2111.06061/) (2021)
- [Continual Learning for Recurrent Neural Networks: an Empirical Evaluation](https://mlsanity.com/paper/2103.07492/) (2021)
- [SIRNN: A Math Library for Secure RNN Inference](https://mlsanity.com/paper/2105.04236/) (2021)
- [Machine Learning for Microcontroller-Class Hardware: A Review](https://mlsanity.com/paper/2205.14550/) (2022)
- [Is Complexity Required for Neural Network Pruning? A Case Study on Global Magnitude Pruning](https://mlsanity.com/paper/2209.14624/) (2022)
- [SyReNN: A Tool for Analyzing Deep Neural Networks](https://mlsanity.com/paper/2101.03263/) (2021)
- [Sound Event Detection with Binary Neural Networks on Tightly Power-Constrained IoT Devices](https://mlsanity.com/paper/2101.04446/) (2021)
- [Efficient Neural Networks for Tiny Machine Learning: A Comprehensive Review](https://mlsanity.com/paper/2311.11883/) (2023)
- [Recognition of handwritten MNIST digits on low-memory 2 Kb RAM Arduino board using LogNNet reservoir neural network](https://mlsanity.com/paper/2105.02953/) (2021)
- [Revisiting IoT Device Identification](https://mlsanity.com/paper/2107.07818/) (2021)
- [Doping: A technique for efficient compression of LSTM models using sparse structured additive matrices](https://mlsanity.com/paper/2102.07071/) (2021)
- [T-RECX: Tiny-Resource Efficient Convolutional neural networks with early-eXit](https://mlsanity.com/paper/2207.06613/) (2022)
- [Random orthogonal additive filters: a solution to the vanishing/exploding gradient of deep neural networks](https://mlsanity.com/paper/2210.01245/) (2022)
- [Gated Recurrent Neural Networks with Weighted Time-Delay Feedback](https://mlsanity.com/paper/2212.00228/) (2022)
- [The Expressive Leaky Memory Neuron: an Efficient and Expressive Phenomenological Neuron Model Can Solve Long-Horizon Tasks](https://mlsanity.com/paper/2306.16922/) (2023)
- [Recurrent Neural Networks for Learning Long-term Temporal Dependencies with Reanalysis of Time Scale Representation](https://mlsanity.com/paper/2111.03282/) (2021)
- [YONO: Modeling Multiple Heterogeneous Neural Networks on Microcontrollers](https://mlsanity.com/paper/2203.03794/) (2022)
- [A Method for Medical Data Analysis Using the LogNNet for Clinical Decision Support Systems and Edge Computing in Healthcare](https://mlsanity.com/paper/2108.02428/) (2021)
- [MinUn: Accurate ML Inference on Microcontrollers](https://mlsanity.com/paper/2210.16556/) (2022)
- [An Open-Source Tool for Classification Models in Resource-Constrained Hardware](https://mlsanity.com/paper/2105.05983/) (2021)
- [MAFIA: Machine Learning Acceleration on FPGAs for IoT Applications](https://mlsanity.com/paper/2107.03653/) (2021)
- [Efficient Visual Recognition with Deep Neural Networks: A Survey on Recent Advances and New Directions](https://mlsanity.com/paper/2108.13055/) (2021)
- [Oscillatory Fourier Neural Network: A Compact and Efficient Architecture for Sequential Processing](https://mlsanity.com/paper/2109.13090/) (2021)
- [REDS: Resource-Efficient Deep Subnetworks for Dynamic Resource Constraints](https://mlsanity.com/paper/2311.13349/) (2023)
- [Quantized Approximately Orthogonal Recurrent Neural Networks](https://mlsanity.com/paper/2402.04012/) (2024)
- [Energy-Based Model for Accurate Estimation of Shapley Values in Feature Attribution](https://mlsanity.com/paper/2404.01078/) (2024)
- [Fast-ULCNet: A fast and ultra low complexity network for single-channel speech enhancement](https://mlsanity.com/paper/2601.14925/) (2026)

## Abstract (arXiv)

This paper develops the FastRNN and FastGRNN algorithms to address the twin RNN limitations of inaccurate training and inefficient prediction. Previous approaches have improved accuracy at the expense of prediction costs making them infeasible for resource-constrained and real-time applications. Unitary RNNs have increased accuracy somewhat by restricting the range of the state transition matrix's singular values but have also increased the model size as they require a larger number of hidden units to make up for the loss in expressive power. Gated RNNs have obtained state-of-the-art accuracies by adding extra parameters thereby resulting in even larger models. FastRNN addresses these limitations by adding a residual connection that does not constrain the range of the singular values explicitly and has only two extra scalar parameters. FastGRNN then extends the residual connection to a gate by reusing the RNN matrices to match state-of-the-art gated RNN accuracies but with a 2-4x smaller model. Enforcing FastGRNN's matrices to be low-rank, sparse and quantized resulted in accurate models that could be up to 35x smaller than leading gated and unitary RNNs. This allowed FastGRNN to accurately recognize the "Hey Cortana" wakeword with a 1 KB model and to be deployed on severely resource-constrained IoT microcontrollers too tiny to store other RNN models. FastGRNN's code is available at https://github.com/Microsoft/EdgeML/.