Generative AI with MLFlow
Generative AI with MLFlow
This course teaches you how to build, fine-tune, and deploy Generative AI models using MLflow. You’ll work with text generation (LLMs), image generation (Diffusion models), prompt engineering, and containerized deployment. Through hands-on labs, you’ll master MLflow for experiment tracking, model registry, and performance monitoring.
Find Our Students At
About the course
The Generative AI Using MLflow course is a complete, hands-on guide to building, managing, and deploying cutting-edge AI models. You’ll begin by understanding Generative AI concepts, model types, and real-world use cases, before diving into MLflow fundamentals for experiment tracking and model management. Using Hugging Face, you’ll work with powerful text generation models like GPT-2, fine-tuning them on custom datasets and logging results to MLflow.You’ll then explore image generation with Stable Diffusion, track prompts and generated images, and compare experimental results. The course covers essential prompt engineering techniques, domain-specific dataset creation, and structured experiment logging. You’ll learn to package and serve models via REST APIs, deploy them in Docker containers with Streamlit/FastAPI, and optimize access through NGINX.Finally, you’ll master model evaluation by logging key metrics, monitoring inference performance, and tracking system usage in MLflow. By the end, you’ll have the skills to take Generative AI projects from concept to production with full reproducibility and scalability.
Course details
Course Content
Introduction to Linux
- What is Linux?
- Role of Linux in Artificial Intelligence
- Lab: Installing Red Hat Linux
Linux CLI for AI Use Cases
- Software management in Linux
- Lab: Configuring package repositories
- Lab: Installing and updating packages
- Service management in Linux
- Lab: Managing services with systemctl
Linux Networking for Machine Learning
- Understanding how IP addresses work
- Understanding Linux firewall
- Lab: Allowing and denying port access using firewalld
- Understanding remote connections using SSH
- Lab: Using SSH key-based authentication
- Lab: Using SSH password-based authentication
Python Basics
- Python Installation & Environment Setup (Anaconda / venv)
- Variables, Data Types (str, int, float, bool, list, tuple, dict, set)
- Basic Operators and Expressions
- Control Flow: if-else, for, while loops
- Functions & Return Values
- Lab: Write a Python script to take a sentence and count words, characters, and vowels.
Working with Libraries
- Installing and Importing Libraries (pip, conda)
- Using Popular Libraries (numpy, pandas, matplotlib)
- Reading Documentation & Examples
- Lab: Install numpy and pandas, create a DataFrame from a list of dictionaries, anddisplay basic statistics.
File Handling & Data Processing
- Reading & Writing Text Files
- Working with CSV and JSON
- Handling Exceptions (try-except)
- Lab: Read a CSV file of text prompts and write the first 10 prompts to a new text file.
Python for Machine Learning
- Lists, List Comprehensions, and Iterators
- NumPy Arrays and Operations
- Pandas DataFrames for Dataset Handling
- Data Cleaning & Transformation
- Lab: Load a CSV dataset into Pandas, drop missing values, and save the cleaneddataset.
Python for NLP
- String Manipulation (split, join, slicing)
- Regular Expressions (re module)
- Tokenization Basics (using nltk or transformers)
- Lab: Tokenize a set of text prompts and count token frequencies.
Python for AI/ML Workflows
- Using scikit-learn for Simple Models
- 6.2 Understanding Functions vs Classes in ML Code
- 6.3 Saving and Loading Models (pickle, joblib)
- Lab: Train a simple text classification model with scikit-learn and save it to disk.
Working with Hugging Face Transformers
- Installing transformers & datasets libraries
- Loading Pretrained Models (from_pretrained)
- Generating Text from a Model
- Lab: Load GPT-2 from Hugging Face and generate a short text paragraph.
Python for API & App Deployments
- Basics of FastAPI and Streamlit
- Sending and Receiving JSON via API Calls
- Running Local Servers for Testing
- Lab: Build a simple FastAPI endpoint that returns text generated from a model.
Introduction to Generative AI
- What is Generative AI?
- Types of Generative Models (LLMs, GANs, Diffusion, VAEs)
- Use Cases of Generative AI
- Lab: Explore AI-generated content examples using Hugging Face Datasets
MLflow Fundamentals
- What is MLflow?
- Components of MLflow: Tracking, Projects, Models, Registry
- MLflow Architecture & Workflow
- Lab: Install MLflow on Linux and launch MLflow Tracking UI
- Lab: Log a basic machine learning model to MLflow
Text Generation using LLMs
- Introduction to Transformers and Hugging Face
- Using Pretrained Models like GPT-2
- Fine-tuning GPT-2 on Custom Text Dataset
- Lab: Run inference using GPT-2 with Hugging Face
- Lab: Log training metrics and artifacts to MLflow
Serving Generative Models
- MLflow Model Packaging and Signature
- MLflow Model Registry Workflow
- Lab: Register a model in MLflow
- Lab: Serve a model using mlflow models serve
- Lab: Deploy GPT-2 as a REST API and test with curl
Image Generation using Diffusion Models
- Introduction to Diffusion Models (Stable Diffusion Overview)
- Using Pretrained Diffusers Library
- Lab: Generate images from text using Stable Diffusion
- Lab: Track prompts and images in MLflow
- Lab: Run comparative image generation experiments and track results
Custom Datasets & Prompt Engineering
- UWhat is Prompt Engineering?
- Impact of Prompts on Generated Output
- Creating a Domain-Specific Dataset
- Lab: Prepare a CSV dataset for fine-tuning
- Lab: Create multiple prompt styles and evaluate outputs
- Lab: Log datasets and prompt outputs to MLflow
Containerization & Deployment
- Dockerizing MLflow Models
- Hosting with Streamlit / FastAPI
- Lab: Build Docker image with MLflow model
- LAB: Overriding values with --set
- Lab: Deploy MLflow model container behind NGINX reverse proxy
Monitoring, Evaluation, and Performance
- Metrics to Evaluate Generative Models
- Using MLflow to Track Inference Metrics
- Lab: Log custom metrics like BLEU, ROUGE in MLflow
- Lab: Monitor GPU vs CPU inference times using nvidia-smi
- Lab: Track API response time, error rates, and logs in MLflow
What our students say about us
Access Labs Anytime
Frequently Asked Questions
Is this training live or pre-recorded?
How will I join the live online training?
How many students are there in a single batch?
How can I practice the labs?
Do you offer payment flexibility?
Can I attend the training in classroom?
Hello, here's a modal.
Now click that shiny button below.
Course details