ETRI-Knowledge Sharing Plaform

State-Of-The-Art Hardware Architecture Artificial Intelligence Transformer-Based Single Chip Parallel Computing Floating-Point Numbers Hardware Accelerator Design Time Hardware Acceleration Low-Precision Parallel Performance Resource Utilization Computer Systems Processor Architecture Radix-4 Outer Product Light-Weight Neural Processing Memory Usage Convolution Neural Network(CNN) Power Optimization And Operation Software Framework Matrix Multiplication Neural Processor Computing Performance Process Technology Silicon Process Processor Performance Rapid Growth Soc System 28 Nm Inference Engine And Frameworks Parallel Operations Floating Point Unit Learning Application Network Algorithm Processing Unit Weight Algorithm Dedicated Software Accuracy-Preserving Computational Algorithm Basic Linear Algebra Subprograms Operating Frequency Deep Learning(DL) Intermediate Representation Practical Design Operation Flow Mixed Precision