search

josephjaspers / blackcat_tensors

Matrix-Vector Library Designed for Neural Network Construction. cuda (gpu) support, openmp (multithreaded cpu) support, partial support of BLAS, expression template based implementation PTX code generation identical to hand written kernels, and support for auto-differentiation
12 stars 4 forks source link

About learning_rate Gradually diminishing,is it necessary? #34

Closed xinsuinizhuan closed 4 years ago

xinsuinizhuan commented 4 years ago

A yolo's learning_rate, it decrease at epoche's 80% and 90%, whether it is necessary in our model?

josephjaspers commented 4 years ago

It is not necessary but it is recommended to decrease the step size in gradient descent. (So when the neural network is getting close to the minima it decreases the step-size so as to not jump-over the minima).