Closed shuangw98 closed 3 years ago
In TFA/FSRW, FSOD fine-tune on both base and novel classes, using the same training examples provided by FSRW. And G-FSOD train for multiple runs on different random samples to obtain averages.
In TFA/FSRW, FSOD fine-tune on both base and novel classes, using the same training examples provided by FSRW. And G-FSOD train for multiple runs on different random samples to obtain averages.