Abstract: On-device training on resource-constrained hardware, such as microcontrollers with limited memory and fixed-function convolutional accelerators, remains an open challenge in embedded ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results