Tensorflow Lite

TensorFlow Lite is a lightweight machine learning framework designed for deploying models on resource-constrained devices like microcontrollers and mobile phones. Current research emphasizes optimizing inference speed and power efficiency through techniques such as variable-precision arithmetic, sparse matrix computations, and ultra-low bit quantization, often targeting specific hardware architectures like ARM CPUs. This focus on efficient on-device execution is crucial for expanding the reach of AI to applications in areas like IoT and embedded systems, particularly in healthcare and edge computing where data privacy and low latency are paramount.

Papers

April 17, 2023

Dynamically Reconfigurable Variable-precision Sparse-Dense Matrix Acceleration in Tensorflow Lite
Jose Nunez-Yanez, Andres Otero, Eduardo de la Torre
Hardware Accelerator Self Reconfiguration Systolic Array Single Core Tensorflow Lite

August 25, 2022

Deep Learning-based ECG Classification on Raspberry PI using a Tensorflow Lite Model based on PTB-XL Dataset
Kushagra Sharma, Rasit Eskicioglu
Machine Learning Model Internet of Thing Device Electrocardiogram Classification ECG Classification Raspberry Pi TXL PBC Dataset Tensorflow Lite

July 18, 2022

Accelerating Deep Learning Model Inference on Arm CPUs with Ultra-Low Bit Quantization and Runtime
Saad Ashfaq, MohammadHossein AskariHemmat, Sudhakar Sah, Ehsan Saboori, Olivier Mastropietro, Alexander Hoffman
Deep Learning Deep Learning Model Runtime Analysis Low Bit Quantization Deep Learning Inference Virtual ARM Tensorflow Lite

November 30, 2021

TinyML Platforms Benchmarking
Anas Osman, Usman Abid, Luca Gemma, Matteo Perotto, Davide Brunelli
Machine Learning Model TinyML Model Low Power Tensorflow Lite

Tensorflow Lite

Papers

Dynamically Reconfigurable Variable-precision Sparse-Dense Matrix Acceleration in Tensorflow Lite

Deep Learning-based ECG Classification on Raspberry PI using a Tensorflow Lite Model based on PTB-XL Dataset

Accelerating Deep Learning Model Inference on Arm CPUs with Ultra-Low Bit Quantization and Runtime

TinyML Platforms Benchmarking