MOVE37 About Blog Project Links

Game of DL Stacks

Hook blablabla. This article will compare four Machine Learning (ML) libraries that was popular on the Internet. The comparison matrixes is inherited from DLIF(Deep Learning Implementations and Frameworks) tutorial, Benchmarking State-of-the-Art Deep Learning Software Tools, and Caffe: Convolutional Architecture for Fast Feature Embedding. The goal is to understand the unique pros and cons of the tech stacks and what they are most useful for, and then choose one for performing experiment for ML hardware research.

alt text Framework Comparison among various machine learning frameworks from DLIF presentation1

alt text
Speedup Comarison for various machine learning frameworks made by the Dell Lab 2

Caffe

TensorFlow

Torch/PyTorch

MXNet


Share this: