NVIDIA DLI offers hands-on training for developers, data scientists, and researchers looking to solve challenging problems with deep learning and accelerated computing.

In this workshop you will learn how to train convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to generate captions from images and video using TensorFlow and the Microsoft Common Objects in Context (COCO) dataset. 

PREREQUISITES

Familiarity with basic Python (functions and variables),

Understanding of fundamentals of convolutional neural networks for computer vision (e.g. see material here or here)

Prior experience training neural networks (in any framework, e.g. see this simple example in keras) 

TOOLS AND FRAMEWORKS: TensorFlow

SCHEDULE

8:30 - 8:45   Breakfast

8:45 - 10:15   Part 1: Segmentation

10:15 - 10:30  Coffee break 

10:30 - 12:00  Part 2: Word Generation

12:00 - 1:00   Lunch

1:00 - 3:00   Part 3: Image & Video Captioning

INSTRUCTOR: 

Dima Lituiev, PhD,

Bakar Computational Health Sciences Institute, UCSF

Certified NVIDIA Deep Learning Institute Ambassador