10/04/2023, 16:50 ERA Syllabus

Published using Google Docs Learn More Report Abuse

Updated automatically every 5

ERA Syllabus minutes

Announcing ERA (Extensive AI: Reimagined

and Advanced), the most advanced deep AI
 course in the world, designed to give you
unparalleled hands-on experience in cutting-
edge AI  technologies! This is the 9th
iteration of our highly successful training
program, now merging  our two beloved
courses, EVA (Extensive Vision AI) and END
(Extensive NLP via Deep  Learning), along
with key sessions from EMLO (Extensive
Machine Learning Ops).

Registrations open on April 5th and

close on April 18th. The course
commences on April  29th, with classes
held every Saturday at 9:00 AM.
Due to recent breakthroughs in AI
research, the entire course has been re-
written, with an  updated curriculum that
covers both vision and NLP topics, as well
as machine learning  operations. In nearly
every session, students will be developing
applications and deploying  them on the
cloud!
Each weekly session features hands-on
exercises, accessible through our new

https://docs.google.com/document/d/e/2PACX-1vR_9al1F7Cm5B9KkINFfkbUk0duHHxnJgvyTg9qBsRiqSDi8bdU_Yp7ARwT2uXQm… 1/5
10/04/2023, 16:50 ERA Syllabus
Learning  Management System (LMS),
ensuring that students gain practical
Published using Google Docsexperience in addition to Learn More Report Abuse
theoretical knowledge. Every session
starts with solving the previous week's
assignment and  quiz, reinforcing the Updated automatically every 5
ERA Syllabus learning process. minutes

ERA is the culmination of our best

offerings, providing a comprehensive and
unparalleled  deep AI learning experience.
 

Here is the complete Syllabus (updates are

shown in bold):

Session Name Description

Session Fundamentals of   Introduction to

1 Artificial neural network
Intelligence; concepts, data
representation for
 images, text, and
audio, converting
spatial to temporal
data and  vice
versa,
understanding
convolutions, fully
connected layers,
and  forward
propagation
mathematics

Session Exploring Neural   Delving into multi-

2 Network channel convolutions,
Architectures; max pooling, layer
structures,  receptive
fields, attention
spans, and receptive
field calculation.

Session Git and Python   Introducing Git,

3 Essentials; GitHub, and Python
for effective
collaboration and
 programming in AI
projects.

Session Building the First   Understanding

4 Neural Networks; kernels,
channels, word
embeddings,
embedding
 mutation,
embedding
layers,
comparing vision
and language
 models,
positional
encoding,
SoftMax, and
GPU processing.

Session Introduction to Learning

5 PyTorch; PyTorch basics,
tensors,
AutoGrad,
squeezing and
 unsqueezing
operations, and
constructing a
neural network
from  scratch
using PyTorch.

Session Backpropagation Grasping

6 and  Advanced   backpropagation
Architectures; mathematics, fully
connected layers,
 spatial data loss
and recovery,
modern AI
architectures,
negative  log
likelihood,
https://docs.google.com/document/d/e/2PACX-1vR_9al1F7Cm5B9KkINFfkbUk0duHHxnJgvyTg9qBsRiqSDi8bdU_Yp7ARwT2uXQm… 2/5
10/04/2023, 16:50 ERA Syllabus
SoftMax, and
learning rates
Published using Google Docs Learn More Report Abuse
Session In-Depth Coding   Hands-on coding
7 Practice; exercises and
optimization of code
Updated automatically every 5
ERA Syllabus performance minutes
Session Advanced Exploring batch
8 Techniques  and normalization,
Optimizations; dropout, word-
level masking,
 regularization,
convolution types,
dense problems,
and optimizers

Session Data Understanding

9 Augmentation data
 and augmentation,
Visualization; preprocessing,
tokenization,
 class activation
maps,
GradCAM, and
attention
visualization
 techniques

Session PyTorch Learning PyTorch

10 Lightning and  AI Lightning, Fabric,
Application   and creating
Development; Lightning-based AI
 applications

Session Residual Examining ResNet

11 Connections in and the concept of
 CNNs and FC residual
Layers; connections in
neural  networks

Session Building and Creating machine

12 Deploying  AI learning pipelines for
Applications; training and
deployment of  AI
models

Session YOLO and Object Understanding

13   anchor boxes
Detection and the YOLO
Techniques; object
detection
 algorithm

Session Multi-GPU Implementing multi-

14 Training  and GPU training,
Scalable Model dynamic batching,
 Serving; and autoscaling  for
model serving

Session UNETs, Variational Exploring

15   UNETs,
AutoEncoders, variational
and   autoencoders,
and their
Applications; practical
 applications

Session Transformers and Studying

16   transformers, self-
Advanced attention, cross-
Embedding attention, multi-
 Techniques; head  attention,
Word2Vec,
encoding types,
and advanced
tokenization  and
embeddings

Session Encoder Investigating encoder

17 Architectures architectures and
 and BERT; training BERT models

Session Masked Learning about

18 AutoEncoders masked
and Vision   autoencoders,
Transformers; vision
transformers,
their
 architectures,
https://docs.google.com/document/d/e/2PACX-1vR_9al1F7Cm5B9KkINFfkbUk0duHHxnJgvyTg9qBsRiqSDi8bdU_Yp7ARwT2uXQm… 3/5
10/04/2023, 16:50 ERA Syllabus
and training
techniques
Published using Google Docs Learn More Report Abuse
Session Decoders and   Understanding
19 Generative Pre decoder-only
trained architectures,
Updated automatically every 5
ERA Syllabus Transformers; training
GPT/GPT-2 minutes
 models, and
exploring sparse
attention

Session Training and Techniques for

21 Fine training large
tuning Large language models
Language and fine-tuning
 Models; them on a single
GPU &
Chinchilla
Scaling Laws.

Session CLIP Models Studying

22 and   contrastive
Training; language-
image pre-
training,
understanding
 CLIP &
Segment
Anything
models, and
training them
from scratch

Session Generative Exploring the

23 Art and world of
 Stable generative
Diffusion; art and the
concept of
stable
diffusion in AI

Session Automatic Introduction

24 Speech   to Whisper
Recognition  and
Fundamentals; automatic
speech
recognition
(ASR)
 technologies

Session Reinforcement Understanding

25 Learning  Part deep Q
I; learning, A3C,
and DDPG in
reinforcement
 learning

Session Reinforcement Exploring

26 Learning  Part advanced
II; reinforcement
learning
algorithms like
T3D and
 Agent57

Session Reinforcement Investigating

27 Learning  from reinforcement
Human learning from
Feedback; human
feedback
(RLHF)  and
language-
human
alignment in
models like
ChatGPT and
 InstructGPT

Session Training Step-by-step

28 ChatGPT guide to training a
from  Scratch; ChatGPT model
from scratch

Session Training Techniques

29 Multimodal and best
 GPTs; practices for
training

https://docs.google.com/document/d/e/2PACX-1vR_9al1F7Cm5B9KkINFfkbUk0duHHxnJgvyTg9qBsRiqSDi8bdU_Yp7ARwT2uXQm… 4/5
10/04/2023, 16:50 ERA Syllabus
multimodal
generative
Published using Google Docs  pre-trained Learn More Report Abuse
transformers

Session Capstone Comprehensive

Updated automatically every 5
ERA Syllabus 30 Project; capstone
project that minutes
consolidates
and applies the
 knowledge
gained
throughout the
course,
encouraging
students to
 create an
innovative AI
solution.

https://docs.google.com/document/d/e/2PACX-1vR_9al1F7Cm5B9KkINFfkbUk0duHHxnJgvyTg9qBsRiqSDi8bdU_Yp7ARwT2uXQm… 5/5