TinyML Projects [NeurIPS’22] On-Device Training Under 256KB Memory [Paper] [Website] [Demo] #On-device Learning, #Memory, #Training, #System, #Compiler [arXiv] BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird’s-Eye View Representation [Paper] [Website] [Demo] #Point-Cloud, #Self-driving, #3D Vision [NeurIPS’21] Delayed Gradient Averaging: Tolerate the Communication Latency for Federated Learning [Paper] [Slides] [Poster] [Website] #On-device Learning, #Latency, #Federated Learning [NeurIPS’21] MCUNetV2: Memory-Efficient Patch-based Inference for Tiny Deep Learning [Paper] [Slides] [Website] [Demo] [Use Cases] #Inference, #Micro-Controller [NeurIPS’20] TinyTL: Reduce Activations, Not Trainable Parameters for Efficient On-Device Learning [Paper] [Slides] [Code] [Website] #On-device Learning, #Memory-Efficient, #Transfer Learning [NeurIPS’20] MCUNet: Tiny Deep Learning on IoT Devices [Paper] [Slides] [Poster] [Code] [Website] [Demo] #Inference, #Micro-Controller [NeurIPS’19] Deep Leakage from Gradients [Paper] [Slides] [Code] [Website] #Training, #Federated Learning, #Privacy [ICLR’18] Deep Gradient Compression: Reducing the Communication Bandwidth for Distributed Training [Paper] [Code] #Distributed Training, #Bandwidth