0% found this document useful (0 votes)
111 views2 pages

GenAI 20 Weeks Roadmap

Uploaded by

ndeekshith78
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
111 views2 pages

GenAI 20 Weeks Roadmap

Uploaded by

ndeekshith78
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

भारतीय सूचना प्रौद्योगिकी संस्थान, नािपुर

Indian Institute of Information Technology, Nagpur


An Institution of National Importance by an Act of Parliament
Dept. of Computer Science & Engineering
Five Months Online Certification Programme on Generative AI
20 Weeks Learning Roadmap of the Certification (4 hours per week)
Month Week Topic Key Focus Areas Hands-On / Evaluation Activities
01 1 Core Math Linear Algebra (Matrices, • Solve gradient problems and
Foundations Vectors), Probability Basics, matrix operations using Python.
(Condensed) Gradient Descent
2 ML Fundamentals Supervised & Unsupervised • Train a simple classification,
Learning, Key Evaluation regression and clustering model
Metrics using Scikit-learn.
3 Deep Learning Neural Networks, Activation • Build and train a simple NN with
Basics and Loss Functions, TensorFlow/PyTorch.
Optimizations
4 Deep Models, CNN, Sequence Models • Set up a basic LLM pipeline and API
Tools & Platforms Hugging Face, OpenAI API, integration.
Azure ML, Google Vertex AI • Nvidia Certification on
“Fundamentals of Deep
Learning”.
• Quiz 1.
02 5 Autoencoders Autoencoders (Basic & • Implement and train a VAE for
and Variants Variational), Data dimensionality reduction.
Reconstruction Applications
6 Generative GAN Architecture, Variants • Build and train a GAN to generate
Adversarial (StyleGAN, CycleGAN) simple images.
Networks (GANs)
7 Attention Self-Attention, Intuition, and • Implement a self-attention
Mechanisms Applications mechanism in PyTorch.
8 Transformer Transformer Architecture, • Fine-tune a pre-trained
Models BERT, GPT, T5 Transformer on a custom dataset.
• Quiz 2.
03 9 Large Language GPT-3, GPT-4, Tokenization, • Experiment with OpenAI’s GPT
Models (LLMs) Pretraining and Fine-tuning models for text generation.
• Text classification using
Transformers.
• Release of Capstone Project
Problem statements.
10 Deployment & Deploy LLMs, Post-Training • NER using Transformers.
Optimizations Optimizations, Few- and • Deploy a fine-tuned LLM on a
Zero-Shot Learning cloud platform like AWS or Azure.
• Nvidia Certification on “Building
Transformers based NLP
Applications”.
11 Scaling and Data, Model, and Tensor • Optimize model training using
Parallelism Parallelism; Scaling Laws distributed frameworks.
12 Prompt Prompting Techniques, • Experiment with advanced
Engineering and Instruction Fine-Tuning, prompting and instruction fine-
PEFT RLHF tuning on GPT-based models.
• Quiz 3.
04 13 Improving LLM Retrieval-Augmented • Integrate RAG into an LLM for
Capabilities Generation (RAG), Vector knowledge-based augmentation.
Databases, LlamaIndex
14 GenAI Use Cases Azure OpenAI, Google • Build a GenAI application (e.g., SQL
and Platforms Vertex AI, Custom GenAI builder, grammar correction).
Applications
15 Computer Vision Object Detection, Image • Implementation of YOLO
Segmentation, OpenCV
16 GenAI for Vision Diffusion Models (Stable • Perform image segmentation
Diffusion, Classifier-Free using U-Net.
Guidance) • Generate images using Stable
Diffusion tools.
• Quiz 4.
• Mid evaluation of Capstone
Project.
05 17 GenAI for Vision CLIP • Implantation of CLIP and relevant
applications.
• Nvidia Certification on
“Generative AI with Diffusion
Models”.
18 Multimodal Vision-Language Models, • Implement a pipeline for image-
Models Multimodal GenAI text tasks (e.g., visual Q&A).
Techniques • Build an image captioning or visual
Q&A pipeline.
• Implementation of Deepfake and
hateful memes on multimodal
data.
• Nvidia Certification on
“Applications of AI for Anomaly
Detection”.
19 Ethical AI and Fairness, Bias, Copyright • Audit a pre-trained model for
Responsible AI Issues, Model Auditing biases and implement safety
measures.
20 Capstone Project Real-World GenAI • Evaluation of Capstone Project,
Evaluation Application Development Other evaluations, Grading, and
Certifications.

Nvidia Certification Links:

1. https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+C-FX-01+V3
2. https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+C-FX-03+V3
3. https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+S-FX-14+V1
4. https://learn.nvidia.com/courses/course-detail?course_id=course-v1:DLI+C-TC-01+V1

Programme Committee
1. Dr. Tausif Diwan, Associate Dean, IIIT Nagpur (Coordinator)
2. Dr. Amol Bhopale, Assistant professor, Dept. of CSE, IIIT Nagpur (Co-coordinator)
3. Dr. Nileshchandra Pikle, Assistant professor, Dept. of CSE, IIIT Nagpur (Co-coordinator)

You might also like