| Day 1 | |
|
8.00-9.00 |
Registration |
|
Session 1 9.00-11.00 |
Motivation Foundations and Terminology of Deep Learning |
|
Session 2 11.30-1.30 |
Machine Learning Basics Linear Regression, Logistic Regression Gradient Descent |
| Session 3 2.15-4.00 |
Designing & Optimizing Neural Network Model Building Deep Models and Hyperparameter Tuning |
| Session 4 4.30-6.30 |
Deep Learning Hacks System/project level tricks and regularization strategies |
| Day 2 |
|
| Session 1 9.00-11.00 |
Convolutional Neural Network Computer Vision with CNNs |
| Session 2 11.30-1.30 |
Recurrent Neural Network Sequence Modelling with RNNs, generative modelling in Deep Learning |
| Session 3 2.15-4.00 |
Advanced Learning Topics Hands-On Labs |
| Session 4 4.30-6.30 |
Assessment |
| Day 3 - NVIDIA Deep Learning Institute Hands on Labs The NVIDIA Deep Learning Institute offers hands on training for developers, data scientists and researchers, looking to solve challanging problems with deep learning |
|
| 09:00 - 11.00 | Image Classification with DIGITS (lab) |
| 11:00 - 1:00 | Object Detection with DIGITS (lab) |
| 2:00 - 4.00 | Azure Notebook on CNN Transfer Learning and Auto Encoders (lab) |
- Two day Deep Learning Workshop followed by a One day NVIDIA DLI workshop are completely free for Collaborators (20 pax), Zonal leads (5 pax), Basic partners (2 pax). All such faculty members from these institutions should contact their institution's Single point of contact for coupon codes with 100% discount.
- Apart from those not being covered by point one, For all other academic participants, the fee for the workshop is 6000 INR + 18% GST (Total 7080 INR). This Fees is only for two days of Deep Learning Workshop by leadingindia.ai initiative. NVIDIA DLI workshop will be given free to all such participants.
- For all Industry professionals the fee for the workshop is 6000 INR + 18% GST (Total 7080 INR). This Fees is only for two days of Deep Learning Workshop by leadingindia.ai initiative. They are not allowed to attend NVIDIA DLI workshops as per the Guidelines from our Industry collaborator.
- All those who successfully register in a workshop will get a referral code. This referral code will be available in your dashboard. Once five people use your referral code to do their workshop registration, then you will get 5000 INR credited to your account. This will go on like this for additonal five referrals hereafter. That means you will get 5000INR for every five referrals of your referral code for workshop registration.
- Your referral code can not be combined with the discounted registrations. Even those who themselves have done discounted registration with 100% or 10% discount can use this referral offer to popularize the workshops.
- For IEEE members, ACM Members, CSI Members, UG/PG/PhD students and for persons with a referral code there is a discount of 10% on base fee, so it will come down to 5400+18%= 6372 INR. If you fall in more than one category (e.g You are a student and IEEE member), still the discount will remain the same i.e 10%.
- All participants will get participation certificate, if they attend the respective workshops in full. However, all those who clear the assessment test will receive the "successfully completed certificate". Assessment Test will consist of 30 Multiple Choice Questions and you need to answer atleast 15 questions correctly. NVDIA DLI Workshop Certificate will be separately given by NVIDIA
- Working Lunch and Tea/Snacks in the morning and evening for all three days will be given without any cost.
- Stay arrangements differ at different locations depending upon the facility being given by the respective institution. You should check at the link given against each location.
- All should come with their laptops for the workshop. However, it is not mandatory. Those without laptops may not be able to enjoy the hands on tasks and coding experience. Without laptops participants may not be able to complete the NVIDIA Labs
- For whatsoever reasons if you dont attend the workshop Registration Amount is non-refundable and Coupon can not be reused again for some other workshop.
- Make sure that their laptop is set up prior to the workshop by following these steps:Ensure websockets runs smoothly on their laptop by going to http://websocketstest.com/ Make sure that WebSockets work for the laptop by seeing under Environment, WebSockets is supported and Data Receive, Send and Echo Test all check Yes under WebSockets (Port 80). If there are issues with WebSockets, try updating browser or trying a different browser. The labs will not run without WebSockets support Best browsers for the labs are Chrome, FireFox and Safari. The labs will run in IE but it is not an optimal experience.
Pre-requisite: Workshop Lab Material with Azure Notebooks
Our workshop notebooks based on Tensorflow and Keras can be practiced using free online platform Azure Notebooks.-
Account creation: To get started with Azure notebooks you have to log-in to https://notebooks.azure.com/ using the Microsoft account. If you have an existing Microsoft account you can use it for logging in to Azure Notebooks. Otherwise you can create a Microsoft account in https://account.microsoft.com/account
-
Getting started with Azure Notebook: After logging-in, you can click on “Library” tab to go into your library where you can “Add a new Library”. After creating a library, you can add new file which can be Notebooks of Python, R or F#.
-
Using Existing Repos: If you have the URL of some existing repositories, you can clone them into your library. For example the below library can be cloned by visiting the URL. https://notebooks.azure.com/jakevdp/libraries/pythondatasciencehandbook
After visiting the URL, you can press “Clone” to copy the entire folder/repository into your library. Then any modification can be done by opening, editing and running those .ipynb notebooks. -
Our Notebooks: During the time of workshop, we will share links to our notebook repository for your practice.
If you cannot use Azure Notebook services during the workshop, our same lab material can be used by installing Anaconda Python and necessary packages.
Anaconda 5.1 - Python Version 3.6 You can Download it from the link freely from the given link.
Click here
Python Libraries: numpy, pandas, sklearn, nltk, tensorflow, keras and for plotting matplotlib
Installation of Packages:
For installing the above mentioned python packages please follow the procedure mentioned below.
For installing new packages, system must be connected through internet.
Press start in windows and search for "Anaconda Prompt". When you will click on the application, a prompt window will be open,wait for few seconds.
You can check/ install any package in Anaconda Python by..
pip install package_name
f-ex:
pip install numpy
and press Enter key after typing these commands.
Note. If you are installing numpy like this, it will give requirement already satisfied message on the prompt. that means package is already installed in the anaconda python.
Installing Opencv for Image processing
Start the Anaconda Prompt: Press start in windows and search for "Anaconda Prompt". when you will click on the application, a prompt window will be open, wait for few seconds.
type command:
pip install opencv-python
press Enter.
Installing Tensorflow and Keras:
Start the Anaconda Prompt in administrator mode: Press start in “windows”, type “Anaconda Prompt”, and right click on the application and click “Run as Administrator”.
Repeat step 2
f.ex- pip install tensorflow
pip install keras
Help: You can also take help of YouTube video: https://www.youtube.com/watch?v=z0qhKP2liHs for sample installation of Anaconda + Tensorflow + Keras
Note: If you want to work with regular python installed from (https://www.python.org/downloads/), installation of numpy, pandas, sklearn, ntlk, tensorflow, keras libraries can be still done using “pip install” command similar to step 2 and 3. However, installation of Anaconda Python is highly encouraged rather than naïve python.
Tensorflow using Google Colab
Sign-in using Google Account to access free Online Tensorflow notebooks at http://colab.research.google.com/
GitHub Account
To join GitHub, the world's leading software development platform, visit the following URL. https://github.com/
GitHub is widely used for open-source software development, you can use it to
- Make your project repositories easily accessible from anywhere
- Maintain history and control versions easily
- Collaborate with other developers that helps you to develop a network
- Fork other developer's repositories and can show up your work to them.
- Build a profile to impress the recruiter
Useful links to go through before the workshop
-
GitHub Links:
-
https://github.com/ChristosChristofidis/awesome-deep-learning
-
Blogs:
-
Book and other Resources: ML Basic