PLEASE NOTE:

This event is part of the NVIDIA DLI University Ambassador Program, a free workshop exclusively for current academic students, staff and researchers with an active e-mail address from a university (any university, not just Universitat Politècnica de Catalunya - UPC). Attendees must use this e-mail address for the below workshop registration. Industry members can express their interest by providing an e-mail address in this form and be alerted when an industry-oriented DLI workshop is organized in Barcelona.

Description

NVIDIA Deep Learning Institute (DLI) offers hands-on training for developers, data scientists, and researchers looking to solve challenging problems with deep learning. DLI, UPC TelecomBCN and the Barcelona Supercomputing Center (BSC) are excited to announce this half-day practical Deep Learning workshop in Campus Nord UPC on September 12th, 2017.

In this half-day Deep Learning Fundamentals workshop, you will learn to

• Understand general terms and background of deep learning
• Leverage deep neural networks (DNN) within the deep learning workflow to solve a real-world image classification problem using DIGITS.
• Measure object detection approaches in relation to three metrics: model training time, model accuracy and speed of detection during deployment.
• Take three approaches for neural network deployment: DIGITS and Caffe, Caffe, through its Python API and NVIDIA TensorRT™

Agenda (*)

14:00 Deep Learning Demystified and Applied Deep Learning (lecture)

14:30 Image Classification with DIGITS (lab)

15:30 Coffee Break

15:45 Object Detection with DIGITS (lab)

17:00 Neural Network Deployment with DIGITS and TensorRT (lab)

18:00 Closing Comments & Questions

18:30 End of the Workshop

(*) Agenda timing is subject to instructor preference

Attendee profile

Content level: beginner

Pre-Requisites: technical background and basic understanding of deep learning concepts.

Lab Contents

Image Classification with DIGITS

Learn how to leverage deep neural networks (DNN) within the deep learning workflow to solve a real-world image classification problem using NVIDIA DIGITS. You will walk through the process of data preparation, model definition, model training and troubleshooting. You will use validation data to test and try different strategies for improving model performance using GPUs. On completion of this lab, you will be able to use DIGITS to train a DNN on your own image classification application.

Object Detection with DIGITS

This lab explores three approaches to identify a specific feature within an image. Each approach is measured in relation to three metrics: model training time, model accuracy and speed of detection during deployment. On completion of this lab, you will understand the merits of each approach and learn how to detect objects using neural networks trained on NVIDIA DIGITS using real-world datasets.

Neural Network Deployment with DIGITS and TensorRT

This lab will introduce three approaches for neural network deployment. The first approach teaches you to use inference functionality directly within a deep learning framework (NVIDIA DIGITS and Caffe). The second approach teaches you how to integrate inference within a custom application by using a deep learning framework API (Caffe, through its Python API). The final approach teaches you to use TensorRT, which will automatically create an optimized inference run-time from a trained Caffe model and network description file. As you explore these approaches, you will learn about the role of batch size in inference performance, as well as various optimizations that can be made in the inference process. You will also explore inference for a variety of different DNN architectures trained in other DLI labs.

This workshop is brought to you by: