Search EN Upload Download free for 30 days

 0% (1) · 3K views · 21 pages You might also like

Deep Learning R18 Jntuh Lab Manual

100% (4)
MACHINE LEARNING R23 material
Get instant answers

32 pages
Uploaded by nphotographer99 AI-enhanced title and description

74% (23)
The document discusses applying convolutional neural networks (CNNs) to common computer vision problems like image classification, object detection,
Impact of Artificial Intelligence To The
semantic segmentation, and instance segme… Full description
Academic Performance of Students in
San Luis Senior High School
Save 0% 100% Embed Share Print Report
36 pages

No ratings yet
Ask AI Download now
1 of 21  Search document  AI Module 1, 2 Notes 4 TH Sem Updated
Download as pdf or txt

lOMoARcPSD|34023260

41 pages

100% (1)
CN Lab Manual r22!3!1

57 pages

100% (1)
DL Notes 1 5 Deep Learning

189 pages

100% (1)
Deep Learning Lab Manual

19 pages

100% (7)
Machine Learning Unit 1

112 pages

DEEP Learning R18 LAB Manual 100% (1)

CS3551 DC - Unit - I Qbank Final

60 pages

100% (2)
ML Unit-1

12 pages

100% (2)
Question Bank: Subject Name: Artificial
Intelligence & Machine Learning Subject
Code: 18CS71 Sem: VII
8 pages

No ratings yet
CS8711 - Cloud Computing Laboratory
Record: Department of Computer
Science & Engineering
5 pages

75% (8)
R18 Os Lab Manual PDF
Ad Download to read ad-free

84 pages

No ratings yet
Application: Craig Dean Memorial
lOMoARcPSD|34023260
Scholarship

2 pages
R18 B.Tech. AI&ML DEEP LEARNING LAB
No ratings yet
1. Setting up the Spyder IDE Environment and Executing a Python The Arts Vocabulary For IELTS
Program You're Reading a Preview
AIM: Setting up the Spyder IDE Environment and Executing a Python
Program Upload your documents to download. 19 pages

DESCRIPTION: No ratings yet

Upload to Download NNDL Lab Manual
To set up the Spyder IDE (Integrated Development Environment) and
execute a Python program, follow these steps:
OR
1. Install Download and install Python from official website: 41 pages
Python https://www.python.org/downloads/
2. Install Anaconda,
Become a Scribd a popular
member Python
to read distributionfull
and download fordocuments.
data science, 100% (10)
Spyder includes Spyder and pre-installed packages for scientific Deep Learning Lab Manual
computing and analysis. Download from the official website:
Start your 30 day free trial
https://www.anaconda.com/products/individual and follow
installation instructions. 30 pages
3. Launch Launch Spyder IDE after installing Anaconda.
Spyder No ratings yet
4. Create In Spyder, you can create a new Python file or open an existing Ad3511 Deep Learning Lab Manual III
or open one. To create a new file, go to "File" > "New File" or use the Yearjnn
a keyboard shortcut Ctrl+N (Command+N on Mac).
Python Alternatively, you can open an existing Python file by going to 58 pages
file "File" > "Open" or using the keyboard shortcut Ctrl+O
(Command+O on Mac). No ratings yet
5. Write Write Python code in editor window for simple program. For Unit 1 (Fiot)
your example,
Python print("Hello, World!")
code
38 pages
6. Save the Save your Python file with the ".py" extension, for example,
Python "hello.py". You can do this by going to "File" > "Save" or
No ratings yet
file using the keyboard shortcut Ctrl+S (Command+S on Mac).
DL Lab Manual 2022-23
7. Execute To execute your Python program, go to "Run" > "Run" or press
the the F5 key. The output of your program will appear in the
Python console pane at the bottom of the Spyder IDE.
34 pages
progra
m
No ratings yet
ME P4252-II Semester - MACHINE
That's it! You have set up the Spyder IDE environment and executed a LEARNING
Python program. You can now continue writing and executing more complex
Python code within Spyder. 48 pages

No ratings yet
1 Os Lab Manual

107 pages

Ad Download to read ad-free No ratings yet

AI & DS - AD3351 DAA - 2marks (Unit 1 &
2) Question Bank

7 pages

lOMoARcPSD|34023260
No ratings yet
Unit-I Notes

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

29 pages

VIVA QUESTIONS:

➢
You're Reading a Preview
How can you install Spyder IDE on your computer, and what are the
No ratings yet
Solving XOR Problem Using DNN AIDS
benefits of using Spyder over a standard text editor for Python
programming? Upload your documents to download.

4 pages
➢ How do you execute a Python program
Upload to in the Spyder IDE, and what
Download
are the different ways to run the code? No ratings yet
IT8601-Computational Intelligence PDF
OR

12 pages
Become a Scribd member to read and download full documents.
No ratings yet
CS-605 Data - Analytics - Lab Complete
Start your 30 day free trial
Manual (2) - 1672730238

56 pages

100% (1)
Cs3551 Distributed Computing L T P C

2 pages

100% (1)
Data Science-Lab Manual

15 pages

No ratings yet
NLP Question Paper Solution

27 pages

No ratings yet
Ccs369 - Text and Speech Analysis - Lab
Manual

23 pages
2
No ratings yet
CCS335-cloud Computing Lab

2 pages
Ad Download to read ad-free

No ratings yet
Data Science Lab Manual - CS3361-
Ramprakash S

47 pages
lOMoARcPSD|34023260

100% (1)
Soft Computing Lab Record
R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

2. Installing Keras, Tensorflow and Pytorch libraries and making use of

them You're Reading a Preview 35 pages

AIM: 100% (1)

Upload your documents to download. Traffic Sign Recognition - PPT
Installing Keras, Tensorflow and Pytorch libraries and making use of them
DESCRIPTION: Upload to Download
8 pages
To install Keras, TensorFlow, and PyTorch libraries and make use of them, you
can follow the steps below: 40% (5)
OR
1. Install Ensure Python is installed and download the latest ML OLD Question Paper
Python and version from the official website:
Become
pip (if not a Scribd member to read and download full documents.
https://www.python.org/downloads/
already Python comes pre-installed with pip; install if 2 pages

installed) unavailable via pip website:

Start your 30 day free trial
https://pip.pypa.io/en/stable/installing/ 67% (15)
2. Install Open a command prompt or terminal and run the Seminar Report On "Neural Network
TensorFlow following command to install TensorFlow using pip: and Their Applications"
pip install tensorflow
3. Install Keras Keras integrated into TensorFlow, ensuring automatic 32 pages
installation during installation. However, you can
explicitly install Keras using pip: pip install keras No ratings yet
4. Install Install PyTorch by visiting official website, selecting Syllabus 6th Sem 21cs63
PyTorch appropriate command based on system configuration:
https://pytorch.org/get-started/locally/
For example, to install the CPU-only version of PyTorch 7 pages
using pip, you can run: pip install torch torchvision
5. Verify After installing the libraries, you can verify that 100% (1)
installations everything is set up correctly by launching a Python CS3491 Artificial Intelligence and
interpreter or creating a Python script and importing the Machine Learning Nov Dec 2023
libraries: Question Paper Download
import tensorflow as tf 2 pages
import keras
import torch No ratings yet
print("TensorFlow version:", tf. version ) Machine Learning Techniques:
print("Keras version:", keras. version ) Important Questions Unit-1
print("PyTorch version:", torch. version )
This code will output the versions of the installed 8 pages
libraries, confirming that everything is installed correctly.
6. Using the Install libraries, use for machine learning models No ratings yet
libraries training. Scripting Language
Here's a basic example of how you can create a simple

3
55 pages

No ratings yet
DEEP LEARNING NOTES - Btech

Ad Download to read ad-free

26 pages

No ratings yet
CS3491 Ai & ML Lab Manual

lOMoARcPSD|34023260

57 pages

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

No ratings yet
Web Security Unit 5
neural network using TensorFlow/Keras and PyTorch:
You're Reading a Preview
➢ Using TensorFlow/Keras:

import tensorflow as tf 22 pages

from tensorflow.keras
Upload your documents import layers
to download.
# Create a simple neural network 0% (1)
model = tf.keras.Sequential([ Ccs335-Cloud Lab Manual Complete
Upload to activation='relu',
layers.Dense(64, Download
input_shape=(784,)),
layers.Dense(10, activation='softmax')
61 pages
]) OR
# Compile the model
100% (2)
model.compile(optimizer='adam', IR Question Bank
Become a Scribd member to read and download full documents.
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
# Train the model (example data used here)
Start your 30 day free trial 29 pages
# model.fit(train_data, train_labels, epochs=10,
batch_size=32)
No ratings yet
➢ Using PyTorch:
Data Analytics Lab Manual
import torch
import torch.nn as nn
import torch.optim as optim
23 pages
# Create a simple neural network
class SimpleNet(nn.Module):
No ratings yet
def init (self):
NLP UNIT 2 (Ques Ans Bank)
super(SimpleNet, self). init ()
self.fc1 = nn.Linear(784, 64)
self.fc2 = nn.Linear(64, 10)
def forward(self, x): 26 pages

x = torch.relu(self.fc1(x))
x = self.fc2(x) No ratings yet

return x ML - LAB Record

model = SimpleNet()
# Define loss function and optimizer
criterion = nn.CrossEntropyLoss() 36 pages
optimizer = optim.Adam(model.parameters(), lr=0.001)
# Train the model (example data used here) No ratings yet
# for epoch in range(10): It Web Essential Lab Manual
# running_loss = 0.0
# for data, labels in train_loader:
# optimizer.zero_grad() 33 pages
# outputs = model(data)
# loss = criterion(outputs, labels) No ratings yet
Devops Unit 3 Complete
4

18 pages

No ratings yet
Ad Download to read ad-free Unit 2 Part 1

34 pages

50% (2)
lOMoARcPSD|34023260
Principles That Guide Practice

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad 20 pages

# loss.backward() 100% (2)

# You're Reading a Preview

optimizer.step() Advanced Data Structures Notes
# running_loss += loss.item()
#Upload
print(f"Epoch {epoch+1},
your documents Loss:
to download.
{running_loss/len(train_loader)}") 142 pages

In these examples, we've created simple neural network architectures using

TensorFlow/Keras and PyTorch, Upload to Download
but you can build more complex models 67% (3)

depending on your specific tasks and requirements. Vtu ML Lab Manual

VIVA QUESTIONS: OR
➢ How can you install Keras, TensorFlow, and PyTorch libraries on 47 pages

your system, and what are the differences between these deep
Become a Scribd member to read and download full documents. No ratings yet
learning frameworks?
Deep Learning Laboratory

➢ Explain the process ofStart your 30aday

building free trial
simple neural network using Keras
for image classification, and how do you compile and train the
model? 69 pages

No ratings yet
Image and Video Analytics

3 pages

No ratings yet
Dev Ops Lab Manual

25 pages

No ratings yet
AD3501 Deep Learning Syllabus

1 page

0% (1)
NLP Handwritten Notes

59 pages

5 No ratings yet
Me cp4212 Software Engineering Manual

34 pages

Ad Download to read ad-free

No ratings yet
DeepLearning - Lab Manual

20 pages

lOMoARcPSD|34023260

No ratings yet
DL-Experiments-1 To 5

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

43 pages

You're Reading a Preview No ratings yet

RLDL128
3. Applying the Upload
Convolution Neural Network
your documents on computer vision
to download.
problems

AIM: Applying the Convolution Neural

Upload Network on computer vision problems
to Download 73 pages

DESCRIPTION:
No ratings yet
CNNs (Convolutional Neural Networks) Azeem. 604 New 3
ORrevolutionize computer vision by
learning spatial hierarchies automatically.
Here are some common
Become a Scribd computer
member tovision
readproblems where full
and download CNNs are frequently
documents. 13 pages
used:

Image classification No ratings yet

Image Start yourinvolves assigning
30 day free trial a label or category to
MD Burhan6621documentation1
Classificatio an input image from a predefined set of classes. CNNs excel
n at this task due to their ability to learn hierarchical features
and capture patterns in images.
Object Object detection involves identifying and localizing multiple 23 pages

Detection objects of interest within an image. CNNs can be combined

with techniques like Region Proposal Networks (RPNs) or No ratings yet
Single Shot Multibox Detector (SSD) to accomplish object Technical Seminar Doc (Rupesh)
detection tasks.
Semantic Semantic segmentation involves classifying each pixel of an
Segmentatio image into a specific category. CNNs with fully 24 pages
n convolutional architectures, such as U-Net or DeepLab, are
commonly used for semantic segmentation. No ratings yet
Instance Instance segmentation goes beyond semantic segmentation Azeez Seminar
Segmentatio by not only classifying each pixel but also distinguishing
n individual object instances. CNNs, combined with methods
like Mask R-CNN, are often used for instance segmentation 23 pages
tasks.
Image CNNs can be used for image generation tasks, such as 100% (1)
Generation generating realistic images from scratch. Generative Topic 2 - Report Writing
Adversarial Networks (GANs) and Variational Autoencoders
(VAEs) are popular architectures for image synthesis.
Style Style transfer involves transferring the visual style of one
7 pages
Transfer image to another while preserving the content. CNNs are
used to separate and manipulate the content and style
100% (2)
representations of the images. Diamond Module JPNJ FINAL
Image Image captioning combines computer vision and natural
Captioning language processing to generate textual descriptions for

156 pages
6

No ratings yet
061 HRG - Auckland Council (Larry
Murphy) - Property

Ad Download to read ad-free

28 pages

100% (1)
Aqa Gcse English Literature Coursework
Mark Scheme

lOMoARcPSD|34023260

8 pages

No ratings yet
R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad
Prof Ed Part 3

input images. CNNs are used to extract visual features,

You're Reading a Preview
which are then combined with recurrent neural networks 3 pages
(RNNs) to generate captions.
Super- Super-resolution
Upload your aims to upscale
documents low-resolution images to
to download. No ratings yet
Resolution higher resolutions. CNNs, particularly architectures like Vedanta Lanjigarh - Innovative Award
SRGAN (Super-Resolution Generative Adversarial For Tribal Art and Dhokra Art Projects
Network), are Upload
commonlyto Download
used for super-resolution tasks. For Its "Aajeevika" Initiative at The…
Face CNNs have demonstrated remarkable performance in face
5 pages
Recognition recognition tasks, where the goal is to identify individuals
OR
based on their facial features.
No ratings yet
Image CNNs can be applied to remove noise, artifacts, or restore Aula 8 - Humans and Their Abilities Part
Denoising damaged images, making them useful in image denoising 2
and Become a Scribd
and member
restoration to read and download full documents.
tasks.
Restoration
15 pages
Start your 30 day free trial
To apply CNNs to these computer vision problems, you'll need labeled datasets No ratings yet
for training, and you'll build and train CNN architectures using popular deep Chapter 9
learning libraries like TensorFlow, Keras, or PyTorch. The specific architecture
and training process will vary depending on the problem at hand. It's also
common to use pre-trained CNN models (e.g., VGG, ResNet, MobileNet) and 18 pages

fine-tune them on specific tasks if you have limited data.

No ratings yet
Overall, CNNs have revolutionized computer vision and continue to advance Being 13
the state-of-the-art in a wide range of applications, bringing significant
improvements to the accuracy and robustness of computer vision systems.
2 pages

VIVA QUESTIONS: No ratings yet

Behavioral Finance in A Nutshell
➢ What is the main advantage of using Convolutional Neural Networks
(CNNs) over traditional fully connected neural networks for computer
vision tasks?
1 page
➢ Explain the typical architecture of a Convolutional Neural Network used
for image classification, and what are the functions of different layers in
100% (2)
this architecture?
Sopa de Letras Verbos

2 pages

No ratings yet
Taiwan University Lists

8 pages

No ratings yet
Free SAP C - TS410 - 1809 Questions -
Ad Download to read ad-free Pass SAP C - TS410 - 1809

2 pages

No ratings yet
02 Writing The C1 Part 2
lOMoARcPSD|34023260

25 pages
R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad
No ratings yet
Plan Layout

You're Reading a Preview

4. Image classification
Uploadon MNIST
your dataset
documents to(CNN model with Fully
download. 5 pages

connected layer)
No ratings yet
AIM: Image classification on MNIST dataset
Upload to (CNN model with Fully
Download Classroom Management For Adhd
connected layer)
DESCRIPTION:
OR 21 pages
Image classification on the MNIST dataset is a classic example and a great
starting point for understanding how to build a Convolutional Neural Network No ratings yet
(CNN)Become a Scribd
with a Fully member
Connected to read
Layer and download
for image full The
recognition. documents.
MNIST English 7 Q2 Module 3
dataset consists of 28x28 grayscale images of handwritten digits from 0 to 9.
Each image is associated with a label
Start your representing the digit it represents.
30 day free trial
28 pages
Here's a step-by-step guide to building a CNN model with a Fully Connected
Layer for image classification on the MNIST dataset using TensorFlow and
No ratings yet
Keras: Lesson 1
Step 1: Import the necessary libraries

import numpy as np
11 pages

import tensorflow as tf
No ratings yet
from tensorflow.keras import layers, models, datasets The Eight Royal Trees: Age Group-4-7
Step 2: Load and preprocess the MNIST dataset Yrs

# Load MNIST dataset

2 pages

(train_images, train_labels), (test_images, test_labels) =

datasets.mnist.load_data() No ratings yet
Chapter 1 Mathematics in Our World
# Normalize pixel values to range [0, 1]
train_images, test_images = train_images / 255.0, test_images / 255.0
23 pages
# Expand dimensions to add channel dimen
No ratings yet
sion (for grayscale images)
The Relationship of Educational
train_images = np.expand_dims(train_images, axis=-1) Preparation, Autonomy, and Critical
Thinking To Nursing Job Satisfaction
test_images = np.expand_dims(test_images, axis=-1)
9 pages
Step 3: Build the CNN model
No ratings yet

8 Copy For Students V Enriquez Theory

(20230323082714)

6 pages

No ratings yet
Ad Download to read ad-free
DWH-DM Assignment

5 pages

No ratings yet
Curmudgeonly Advice: Donald R. Kinder
lOMoARcPSD|34023260

2012

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

16 pages

model = models.Sequential()
You're Reading a Preview
No ratings yet
# Convolutional layers Evaluating Speech

model.add(layers.Conv2D(32, (3, documents

Upload your 3), activation='relu', input_shape=(28, 28, 1)))
to download.
model.add(layers.MaxPooling2D((2, 2))) 2 pages

model.add(layers.Conv2D(64, Upload to Download

(3, 3), activation='relu'))
No ratings yet
model.add(layers.MaxPooling2D((2, 2))) Quarter 3 Exam

OR
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
# Flatten the 3D output to 1D 3 pages
Become a Scribd member to read and download full documents.
model.add(layers.Flatten())
No ratings yet
# Fully Connected layers Elegant Rosé Watercolor Stains Student
Start your 30 day free trial Pack For High School by Slidesgo
model.add(layers.Dense(64, activation='relu'))
model.add(layers.Dense(10, activation='softmax')) # 10 output classes (0-9 59 pages

digits)
Ad Download to read ad-free
Step 4: Compile the model
model.compile(optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])

Step 5: Train the model

model.fit(train_images, train_labels, epochs=5, batch_size=64,

validation_split=0.1)
Step 6: Evaluate the model on the test set

test_loss, test_accuracy = model.evaluate(test_images, test_labels)

print("Test accuracy:", test_accuracy)
Step 7: Make predictions on new data

predictions = model.predict(test_images[:5])
predicted_labels = np.argmax(predictions, axis=1)
print("Predicted labels:", predicted_labels)

print("True labels:", test_labels[:5])

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

This code will create a CNN model with three convolutional layers followed by
You're Reading a Preview
a fully connected layer. It will then train the model on the MNIST dataset and
evaluate its accuracy on the test set. Finally, it will make predictions on five test
Upload
images and display the yourlabels
predicted documents to download.
along with the true labels.
You can experiment with the number of layers, units, and other hyperparameters
to improve the model's performance.
UploadAdditionally,
to Download you can use techniques like
data augmentation and dropout to further enhance the model's generalization
capabilities.
OR
VIVA QUESTIONS:

➢ What is the MNIST dataset, and why is it commonly used for testing
Become
image a Scribd member
classification models?to read and download full documents.
➢ Describe the architecture of a CNN model with a fully connected
layer for image classification
Start youron the free
30 day MNISTtrial dataset.

10

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

You're Reading a Preview

5. Applying the Deep Learning Models in the field of Natural Language
Processing
Upload your documents to download.
AIM: Applying the Deep Learning Models in the field of Natural Language
Processing
Upload to Download
DESCRIPTION:
Deep learning models have made significant contributions to the field of
OR
Natural Language Processing (NLP), enabling the development of powerful
language models and applications. Some of the key deep learning models used
in NLPBecome
include:a Scribd member to read and download full documents.
Recurrent Neural Networks (RNNs):
RNNs are designed to handle sequential
Start data,free
your 30 day making
trial them well-suited for
natural language processing tasks where the order of words matters. They
process input data step-by-step while maintaining hidden states to capture
context. However, traditional RNNs suffer from vanishing gradient problems.
Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) are
popular variants of RNNs that address this issue.
Applications:
Text Sentiment Named Entity Language
classification analysis Recognition modeling
(NER)

Transformers:
Transformers introduced the attention mechanism, enabling more efficient and
parallelized processing of sequential data. They have revolutionized NLP by
capturing long-range dependencies effectively and have become the backbone
of modern language models.
Applications:
Machine Translation (e.g., Text generation Question-Answering
Google's Transformer- (e.g., OpenAI's GPT (e.g., Google's BERT-
based model "BERT" for series) based model "BERT" for
NMT) QA)

Bidirectional Encoder Representations from Transformers (BERT):

BERT is a transformer-based language model pre-trained on a large corpus of
text data. It learns contextualized word representations, allowing it to
understand the context in which a word appears in a sentence. BERT's pre-

11

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

trained representations can be fine-tuned for a wide range of NLP tasks, making
You're Reading a Preview
it a versatile and powerful model.
Applications:
Text UploadEntity
Named your documents to download.Question-
Sentiment
classification Recognition analysis Answering
(NER) Upload to Download

Generative Pre-trained Transformer (GPT):

Ad Download to read ad-free
OR
GPT is a family of transformer-based language models developed by OpenAI.
GPT-3, in particular, is one of the largest language models ever created, with
175 billion parameters. It demonstrates impressive capabilities in natural
Become a Scribd member to read and download full documents.
language understanding and generation.
Applications:
Start your 30 day free trial
Text completion Text generation (e.g., Language translation
creative writing, code
generation)

Convolutional Neural Networks for NLP:

Although more commonly used for computer vision, CNNs can be adapted to
NLP tasks. They are often employed for text classification and sentiment
analysis by treating text as a 1D sequence of tokens.
Applications:
Text classification Sentiment analysis

Sequence-to-Sequence Models:
Sequence-to-sequence models use encoder-decoder architectures to handle tasks
that involve transforming one sequence into another, such as machine
translation and summarization.
Applications:
Machine translation Text summarization

Attention Mechanisms:
Attention mechanisms are not models themselves, but they have been
instrumental in improving the performance of various NLP models. They allow
models to focus on specific parts of the input during processing, enhancing their
understanding and performance.
Applications:

Language translation Text generation

12

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

These deep learning models, along with their variants and combinations, have
You're Reading a Preview
driven significant advancements in the field of Natural Language Processing,
enabling the development of more sophisticated and context-aware language
Upload
models and applications. As your documents
research to download.
progresses, we can expect further
improvements and innovations in NLP, ultimately leading to more accurate and
human-like language processing systems.
Upload to Download

VIVA QUESTIONS:
OR
➢ What is the role of Word Embeddings in Natural Language
Processing, and how do they benefit deep learning models for NLP
Become a Scribd member to read and download full documents.
tasks?
➢ How do Recurrent Neural Networks (RNNs) handle sequential data
in Natural LanguageStart
Processing, and
your 30 day what
free trialare some of the challenges
they face?

13

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

6. Train a sentiment analysis model on IMDB dataset, use RNN layers

You're Reading a Preview
with LSTM/GRU notes
AIM: Train a sentiment analysis model on IMDB dataset, use RNN layers with
LSTM/GRU notes Upload your documents to download.
DESCRIPTION:
Upload to Download
To train a sentiment analysis model on the IMDB dataset using RNN layers with
LSTM/GRU units, we'll use TensorFlow and Keras. The IMDB dataset contains
OR
movie reviews labeled as positive or negative sentiments.
Step 1: Import the necessary libraries.
import Become
tensorflowa Scribd
as tf member to read and download full documents.
from tensorflow.keras.datasets import imdb
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow.keras.models Start your Sequential
import 30 day free trial
from tensorflow.keras.layers import Embedding, LSTM, GRU, Dense
Step 2: Load and preprocess the IMDB dataset.
# Load the IMDB dataset, keeping only the top 10,000 most frequent words
num_words = 10000
(train_data, train_labels), (test_data, test_labels) =
imdb.load_data(num_words=num_words)
# Pad the sequences to ensure they all have the same length
max_length = 250
train_data = pad_sequences(train_data, maxlen=max_length)
test_data = pad_sequences(test_data, maxlen=max_length)
Step 3: Build the RNN model with LSTM or GRU layers.
embedding_dim = 100
model = Sequential()
model.add(Embedding(input_dim=num_words, output_dim=embedding_dim,
input_length=max_length))
# Choose LSTM or GRU layer
# LSTM Layer
# model.add(LSTM(units=64, dropout=0.2, recurrent_dropout=0.2))
# GRU Layer
model.add(GRU(units=64, dropout=0.2, recurrent_dropout=0.2))
model.add(Dense(units=1, activation='sigmoid'))
Step 4: Compile and train the model.
model.compile(optimizer='adam', loss='binary_crossentropy',
metrics=['accuracy'])
batch_size = 128

14

Ad Download to read ad-free

lOMoARcPSD|34023260

Ad Download to read ad-free

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

epochs = 5
You're Reading a Preview
model.fit(train_data, train_labels, batch_size=batch_size, epochs=epochs,
validation_split=0.2)
Step 5: Evaluate theUpload
model your documents
on the test set. to download.
test_loss, test_accuracy = model.evaluate(test_data, test_labels)
print("Test accuracy:", test_accuracy)
That's it! With these steps, youUpload to Download
have built and trained a sentiment analysis
model on the IMDB dataset using RNN layers with LSTM/GRU units. The
model will learn to predict whether movie reviews are positive or negative
OR
based on the given text. You can experiment with different hyperparameters,
such as the number of LSTM/GRU units, the embedding dimension, or the
numberBecome a Scribd
of epochs, to seemember
how theytoaffect
read the
andmodel's
download full documents.
performance.

VIVA QUESTIONS:
Start your 30 day free trial

➢ What is the IMDB dataset, and what is the goal of training a

sentiment analysis model on this dataset using RNN layers with
LSTM/GRU?

➢ Explain the role of LSTM/GRU units in the sentiment analysis model

and how they address the vanishing gradient problem in traditional
RNNs?

15

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

7. Applying the Autoencoder algorithms for encoding the real-world data

You're Reading a Preview
AIM: Applying the Autoencoder algorithms for encoding the real-world data
DESCRIPTION:
Upload
Autoencoders are a class your documents
of neural to download.
network architectures used for unsupervised
learning. They are particularly useful for encoding real-world data into a lower-
dimensional representation, also known as the "latent space" or "encoding
space." Autoencoders consist of Upload to Download
two main parts: the encoder, which maps the
input data to the latent space, and the decoder, which reconstructs the data from
the encoded representation. The objective is to minimize the difference between
OR
the original input and the reconstructed output, encouraging the model to learn a
compact representation of the data.
Become
Applying a Scribd for
Autoencoders member to read
encoding and download
real-world full documents.
data involves the following
steps:

Step 1: Data Preparation Start your 30 day free trial

Collect and preprocess the real-world data you want to encode. Ensure that the
data is in a suitable format and normalize it if necessary.

Step 2: Build the Autoencoder Model

Create the Autoencoder model using a neural network library like TensorFlow
or Keras.
import tensorflow as tf
from tensorflow.keras.layers import Input, Dense
from tensorflow.keras.models import Model
# Define the input size
input_size = <your_input_size>
# Define the size of the latent space (encoded representation)
latent_dim = <your_desired_latent_dimension>
# Encoder
input_data = Input(shape=(input_size,))
encoded = Dense(latent_dim, activation='relu')(input_data)
# Decoder
decoded = Dense(input_size, activation='sigmoid')(encoded)
# Autoencoder model
autoencoder = Model(input_data, decoded)

Step 3: Compile and Train the Autoencoder

Compile the model with an appropriate optimizer and loss function, and then
train the Autoencoder on your real-world data.
autoencoder.compile(optimizer='adam', loss='mean_squared_error')
# Assuming you have your data loaded into 'data'

16

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

autoencoder.fit(data, data, epochs=50, batch_size=32)

You're Reading a Preview
Step 4: Encode Real-World Data
Once the Autoencoder Upload your you
is trained, documents to download.
can use the encoder part of the model to
encode your real-world data into the latent space.
encoder = Model(input_data, encoded)
Upload to Download
encoded_data = encoder.predict(data)
The encoded_data variable now contains the encoded representation of your
real-world data in the lower-dimensional latent space.
OR

Step 5: Decode Encoded Data (Optional)

Become
If desired, a Scribd
you can member
use the decodertopart
readofand download
the model full documents.
to reconstruct the data
from the encoded representation.
# Assuming you have the encoder part of the model stored in 'encoder'
Start your 30 day free trial
decoder_input = Input(shape=(latent_dim,))
decoded_output = autoencoder.layers[-1](decoder_input) # Get the last layer of
the decoder
decoder = Model(decoder_input, decoded_output)
# Reconstruct data from encoded representation
reconstructed_data = decoder.predict(encoded_data)

Autoencoders are powerful tools for dimensionality reduction and feature

learning in real-world data. They are widely used in various applications,
including anomaly detection, data compression, and denoising. By training an
Autoencoder, you can obtain a more compact representation of your data, which
can be useful for downstream tasks or visualizations.

VIVA QUESTIONS:

➢ What is the main purpose of using autoencoder algorithms in

encoding real-world data, and how do they work?

➢ Explain the typical architecture of an autoencoder model, and what

are some popular variations of autoencoders used for different real-
world data encoding tasks?

17
Ad Download to read ad-free

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

8. Applying Generative Adversial Networks for image generation and

unsupervised tasks. You're Reading a Preview
AIM: Applying Generative Adversial Networks for image generation and
unsupervised tasks. Upload your documents to download.
DESCRIPTION:
Generative Adversarial Networks (GANs) are a powerful class of deep learning
models used for various tasks, Upload to Download
including image generation and unsupervised
learning. GANs consist of two neural networks: a generator and a discriminator.
The generator generates fake data (e.g., images), while the discriminator tries to
OR
distinguish between real data and fake data generated by the generator. The two
networks are trained simultaneously in a competition, where the generator tries
Become
to produce data athat
Scribd
foolsmember to read and
the discriminator, download
and full documents.
the discriminator tries to get
better at distinguishing real from fake data.
Here's how GANs can be applied for image generation and unsupervised tasks:
Start your 30 day free trial
Image Generation with GANs:
The primary application of GANs is image generation. The generator learns to
create new images that resemble the training data. The discriminator provides
feedback to the generator, guiding it to produce more realistic images over time.

Step 1: Import the necessary libraries.

import numpy as np
import tensorflow as tf
from tensorflow.keras.layers import Dense, LeakyReLU, BatchNormalization,
Reshape, Conv2DTranspose, Conv2D, Flatten
from tensorflow.keras.models import Sequential

Step 2: Build the Generator and Discriminator models.

def build_generator(latent_dim):
model = Sequential()
model.add(Dense(256, input_dim=latent_dim))
model.add(LeakyReLU(alpha=0.01))
model.add(BatchNormalization(momentum=0.8))
model.add(Dense(512))
model.add(LeakyReLU(alpha=0.01))
model.add(BatchNormalization(momentum=0.8))
model.add(Dense(1024))
model.add(LeakyReLU(alpha=0.01))
model.add(BatchNormalization(momentum=0.8))
model.add(Dense(28 * 28 * 1, activation='tanh'))
model.add(Reshape((28, 28, 1)))

18

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

return model
You're Reading a Preview
def build_discriminator(img_shape):
model = Sequential()
Upload your documents to download.
model.add(Flatten(input_shape=img_shape))
model.add(Dense(512))
model.add(LeakyReLU(alpha=0.01))
model.add(Dense(256)) Upload to Download
model.add(LeakyReLU(alpha=0.01))
model.add(Dense(1, activation='sigmoid'))
OR
return model

Step 3:Become
Define the GANmember
a Scribd model by tocombining the Generator
read and download and
full documents.
Discriminator.
def build_gan(generator, discriminator):
Start your 30 day free trial
discriminator.trainable = False
model = Sequential()
model.add(generator)
model.add(discriminator)
return model

Step 4: Train the GAN model.

# Define hyperparameters
latent_dim = 100
img_shape = (28, 28, 1)
# Build and compile the discriminator and generator
generator = build_generator(latent_dim)
discriminator = build_discriminator(img_shape)
gan = build_gan(generator, discriminator)
discriminator.compile(loss='binary_crossentropy',
optimizer=tf.keras.optimizers.Adam(0.0002, 0.5))
gan.compile(loss='binary_crossentropy',
optimizer=tf.keras.optimizers.Adam(0.0002, 0.5))
# Load and preprocess real image data (e.g., from MNIST dataset)
# Training loop
epochs = 10000
batch_size = 64
half_batch = batch_size // 2
for epoch in range(epochs):
# Train the discriminator
idx = np.random.randint(0, X_train.shape[0], half_batch)

19

Ad Download to read ad-free

lOMoARcPSD|34023260

R18 B.Tech. AI&ML DEEP LEARNING LAB JNTU Hyderabad

real_images = X_train[idx]
You're Reading a Preview
noise = np.random.normal(0, 1, (half_batch, latent_dim))
generated_images = generator.predict(noise)
Upload your documents to download.np.ones((half_batch,
d_loss_real = discriminator.train_on_batch(real_images,
1)))
d_loss_fake = discriminator.train_on_batch(generated_images,
np.zeros((half_batch, 1))) Upload to Download
d_loss = 0.5 * np.add(d_loss_real, d_loss_fake)
# Train the generator
OR
noise = np.random.normal(0, 1, (batch_size, latent_dim))
valid_labels = np.ones((batch_size, 1))
Become
g_loss a Scribd member to read
= gan.train_on_batch(noise, and download full documents.
valid_labels)

Unsupervised Tasks with GANs:

Start your 30 day free trial
GANs can be used for various unsupervised learning tasks, such as data
augmentation, feature extraction, and anomaly detection. For example, a pre-
trained GAN generator can be used to generate additional data for training
purposes, which can improve the performance of other models. GANs can also
be used for feature extraction by using the intermediate layers of the generator
as feature representations for downstream tasks.
In summary, GANs are versatile models that have proven to be effective in
image generation and unsupervised learning tasks. They have a wide range of
applications and continue to be an active area of research in the deep learning
community.

VIVA QUESTIONS:

➢ How do Generative Adversarial Networks (GANs) work for image

generation, and what are the main components of a GAN model?
➢ How can GANs be applied to unsupervised tasks, and what are some
examples of unsupervised learning tasks where GANs have been
successfully used?

20

Reward Your Curiosity

Everything you want to read.
Anytime. Anywhere. Any device.

Read free for 30 days

No Commitment. Cancel anytime.

We and our 10 partners store and access information on your device for personalized ads and content. Personal data may be processed, such as cookie identifiers, unique device identifiers, and browser information.
Third parties may store and access information on your device and process this personal data. You may change or withdraw your preferences by clicking on the cookie icon or link; however, as a consequence, you
may not see relevant ads or personalized content.
Our website may use these cookies to:
Share this document
Measure the audience of the advertising on our website, without profiling Customize Your Choices
Display personalized ads based on your navigation and your profile
    
Personalize our editorial content based on your navigation
Accept All
Allow you to share content on social networks or platforms present on our website
Send you advertising based on your location
Ad Continue Without Accepting
Privacy Policy
Third Parties

Storage Targeted Advertising Personalization Analytics

About Suppor t Legal Social Get our fr ee apps

About Scribd Help / FAQ Terms Instagram

Everand: Ebooks & Audiobooks Accessibility Privacy Twitter

SlideShare Purchase help Copyright Facebook

Join our team! AdChoices Cookie Preferences Pinterest

Contact us Do not sell or share my personal

information

Documents
Language: English Copyright © 2025 Scribd Inc.