ECE 498NSU/598NSG: Deep Learning in Hardware

Homeworks

• 2019/9/5 Homework-1 Assigned

  • Homework files: [PDF] [Python file for Problem 5] [Python helper file] [DNN model for Problem 5]

  • Due date: 2019/9/19

  • Solution: [PDF] [ADALINE Notebook] [Quantization Notebook]

• 2019/9/19 Homework-2 Assigned

  • Homework files: [PDF] [Dynamic ranges and precision offsets for Problem 2]

  • Due date: 2019/10/3

  • Solution: [PDF] [Notebook for Problem 3, 4, 5, 6] [Python script for uniform DNN quantization] [Python script for per-layer DNN quantization]

• 2019/10/4 Homework-3 Assigned

  • Homework files: [PDF]

  • Due date: 2019/10/29

  • Solution: [PDF] [Python files for Problem 4]

• 2019/10/29 Homework-4 Assigned

  • Homework files: [PDF] [Model Files]

  • Due date: 2019/11/12

  • Solution: [PDF]

Scribing

• 10% of your grade will be based on scribing lectures. Each student will scribe three lectures over the course of the semester. A signup sheet is provided below. Please sign up by writing your name in an empty scribe field. You should sign up as soon as you can to avoid running out of options by the end of the semester. Every week, we will provide the topics of the upcoming lectures which may help you decide whether you would like to scribe or not. Each lecture cannot be scribed by more than 6 students. It is the student's responsibility to ensure they are able to finalize three scribed lectures by the end of the semester. The deadline to submit your scribed lectures is exactly one week after the lecture date. Submission is via email to one of the two TAs.

• Please use the following Google Sheet to sign up for Scribing [link]