[Machine learning] Fine Tune

Fine Tune name suggests, is fine-tuning. In machine learning, learning is generally used in the migration, the number of layer by layer adjusted to achieve the purpose of controlling the migration of some study. So it can use the existing parameters, some of slight changes to adapt to new learning tasks. So, the task of fine-tuning can not adapt to change too much migration.

https://blog.csdn.net/u013841196/article/details/80919857