Tutorial 1 - Learn How To Use Modern Convnets - Neuromatch Academy - Deep Learning

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

Tutorial 1: Learn how to use modern convnets

Contents
Tutorial 1: Learn how to use modern convnets
Tutorial Objectives
Setup
Section 1: Modern CNNs and Transfer Learning
Section 2: The History of Convnets
Section 3: Big and Deep Convnets
Section 4: Convnets After AlexNet
Section 5: Inception + ResNeXt
Section 6: Depthwise separable convolutions
Section 7: Transfer Learning
Summary
Daily survey
Bonus: Speed-Accuracy Trade-Off / Different Backbones

Open in Colab Open in Kaggle

Week 2, Day 3: Modern Convnets

By Neuromatch Academy

Content creators: Laura Pede, Richard Vogg, Marissa Weis, Timo Lüddecke, Alexander Ecker

Content reviewers: Arush Tagade, Polina Turishcheva, Yu-Fang Yang, Bettina Hein, Melvin Selim Atay, Kelson Shilling-
Scrivo

Content editors: Gagana B, Roberto Guidotti, Spiros Chavlis

Production editors: Anoop Kulkarni, Roberto Guidotti, Cary Murray, Gagana B, Spiros Chavlis

Tutorial notebook is based on an initial version by Ben Heil

In this tutorial we are going to learn more about Convnets. More specifically, we will:

1. Learn about modern CNNs and Transfer Learning.


2. Understand how architectures incorporate ideas we have about the world.
3. Understand the operating principles underlying the basic building blocks of modern CNNs.
4. Understand the concept of transfer learning and learn to recognize opportunities for applying it.
5. (Bonus) Understand the speed vs. accuracy trade-off.

If you want to download the slides: https://osf.io/download/tzfsn/

https://deeplearning.neuromatch.io/tutorials/W2D3_ModernConvnets/student/W2D3_Tutorial1.html 04/06/24, 2 58 PM
Page 1 of 35
:
Install dependencies
Show code cell source

Install and import feedback gadget


Show code cell source

# Import libraries
import os
import time
import tqdm
import torch
import IPython
import torchvision

import numpy as np
import matplotlib.pyplot as plt

import torch.nn as nn
import torch.nn.functional as F

from torchvision import transforms


from torchvision.models import AlexNet
from torchvision.utils import make_grid
from torchvision.datasets import ImageFolder

from PIL import Image


from io import BytesIO

Figure settings
Show code cell source

Set random seed

https://deeplearning.neuromatch.io/tutorials/W2D3_ModernConvnets/student/W2D3_Tutorial1.html 04/06/24, 2 58 PM
Page 2 of 35
:

You might also like