Pipeline parallelism has been demonstrated to be a remarkable approach t...
We present Rhino, a system for accelerating tensor programs with automat...
This paper presents TAG, an automatic system to derive optimized DNN tra...
This paper proposes DisCo, an automatic deep learning compilation module...
Many recent machine learning models show dynamic shape characteristics.
...
We show in this work that memory intensive computations can result in se...
The last decade has witnessed growth in the computational requirements f...
It is a challenging task to train large DNN models on sophisticated GPU
...