DavanagereWorkshop

Workshop on Deep Learning and its applications. Speakers - Saneem Chemmengath, Sneha Mondal, and Abhirut Gupta.

Contact Information

• Saneem Chemmengath (saneem.cg@in.ibm.com)
• Sneha Mondal (snmondal@in.ibm.com)
• Abhirut Gupta (abhirutgupta@in.ibm.com)

Pre-requisites

1. Please download the following datasets -

• Housing Prices https://www.kaggle.com/c/house-prices-advanced-regression-techniques/data

You will need to create a Kaggle account, and accept the terms of the competition. (Follow the instructions from the Download All button.

• MNIST http://yann.lecun.com/exdb/mnist/ Download all 4 files

2. You will need Python for the workshop, we suggest installing Anaconda for Python 3.7 Installer links for different OS can be found at -

Mac - https://repo.anaconda.com/archive/Anaconda3-2018.12-MacOSX-x86_64.pkg
Windows - https://repo.anaconda.com/archive/Anaconda3-2018.12-Windows-x86_64.exe
Linux - https://repo.anaconda.com/archive/Anaconda3-2018.12-Linux-x86_64.sh and power machines - https://repo.anaconda.com/archive/Anaconda3-2018.12-Linux-ppc64le.sh

Anaconda installation guide is at - https://docs.anaconda.com/anaconda/install/

3. After installing anaconda, you can install the following packages

• nltk https://anaconda.org/anaconda/nltk
• gensim https://anaconda.org/anaconda/gensim
• keras https://anaconda.org/anaconda/keras
• tensorflow https://anaconda.org/anaconda/tensorflow
• pytorch https://anaconda.org/pytorch/pytorch
• torchtext https://anaconda.org/derickl/torchtext

The link for each package (https://anaconda.org/anaconda/nltk) contains instructions on how to install it. For example -

conda install -c anaconda nltk

4. Download the spacy English model -

python -m spacy download en

and stopwords from nltk -

python -m nltk.downloader stopwords

Also download the IMDB sentiment dataset, and pre-trained Glove word embeddings on torchtext with the following python code -

import torch
from torchtext import data
from torchtext import datasets

TEXT = data.Field(tokenize='spacy')
LABEL = data.LabelField(dtype=torch.float)

train_data, test_data = datasets.IMDB.splits(TEXT, LABEL)
TEXT.build_vocab(train_data, max_size=25000, vectors="glove.6B.100d")
LABEL.build_vocab(train_data)

This is 84 MB and 862 MB respectively.