Caffe con Troll: Shallow Ideas to Speed Up Deep Learning

Abstract

We present Caffe con Troll (CcT), a fully compatible end-to-end version of the popular framework Caffe with rebuilt internals. We built CcT to examine the performance characteristics of training and deploying general-purpose convolutional neural networks across different hardware architectures. We find that, by employing standard batching optimizations for CPU training, we achieve a 4.5x throughput improvement over Caffe on popular networks like CaffeNet. Moreover, with these improvements, the end-to-end training time for CNNs is directly proportional to the FLOPS delivered by the CPU, which enables us to efficiently train hybrid CPU-GPU systems for CNNs.

References in corpus (3)

• Caffe: Convolutional Architecture for Fast Feature Embedding
• HOGWILD!: A Lock-Free Approach to Parallelizing Stochastic Gradient Descent
• DimmWitted: A Study of Main-Memory Statistical Analytics

Cited by in corpus (9)

• Faster CNNs with Direct Sparse Convolutions and Guided Pruning
• SparkNet: Training Deep Networks in Spark
• Cluster Pruning: An Efficient Filter Pruning Method for Edge AI Vision Applications
• MEC: Memory-efficient Convolution for Deep Neural Network
• A Systematic Approach to Blocking Convolutional Neural Networks
• On-the-fly Operation Batching in Dynamic Computation Graphs
• Distributed Training Large-Scale Deep Architectures
• KeystoneML: Optimizing Pipelines for Large-Scale Advanced Analytics
• Deep Learning Approximation: Zero-Shot Neural Network Speedup