New Conference Paper Accepted in ICLR 2020
Our paper entitled “Dynamic Sparse Training: Find Efficient Sparse Network From Scratch With Trainable Masked Layers” has been accepted in the 8th International Conference on Learning Representations (ICLR).