Random Thoughts

From Noise to Numbers: A Hands-On Guide to Generating MNIST Digits with Diffusion Models (Part 2)

python

dl

In Part 2 of our diffusion model journey, we supercharge our MNIST digit generator! Building on our basic UNet from Part 1, we introduce the powerful diffusers library and its UNet2DModel. See how this enhanced architecture significantly improves denoising quality through direct image prediction, bringing us closer to clearer, sharper digit generation.

From Noise to Numbers: A Hands-On Guide to Generating MNIST Digits with Diffusion Models (Part 1)

python

dl

This post walks you through building MNIST digit generators with diffusion models, starting from the basics. We introduce diffusion models, prepare the MNIST dataset, and train a simple convolutional UNet for direct image prediction. Finally, we evaluate its denoising performance as a foundation for more advanced models.

Building a Web Server From Scratch in Pure Python

python

Building a web server from scratch in Python? Join me as I journey from simple http.server to raw sockets, threading, selectors, and asyncio, exploring different concurrency models and benchmarking performance along the way.

Unleash Your Creativity: A Rapid Prototyping App with Amazon Nova Canvas

python

aws

Explore the power of Amazon Nova Canvas with a hands-on Streamlit app. This rapid prototyping tool offers five powerful image processing features, including text-to-image, inpainting, and outpainting. Unleash your creativity and transform your ideas into stunning visuals.

My Streamlit Deployment Journey on AWS : What Worked, What Didn’t

python

aws

Deploying Streamlit on AWS? I faced WebSocket issues with Lambda and App Runner, found a solution with Lightsail, and built a custom Fargate solution with AWS CDK. This post shares my lessons and cost insights.

Streaming Responses in FastAPI

python

In this blog post, I explore how to stream responses in FastAPI using Server-Sent Events, StreamingResponse, and WebSockets. Through simple examples that simulate LLM outputs, I demonstrate how you can efficiently stream real-time data in your applications.

Tips to Register for an AWS Certification Exam with Pearson VUE from Pakistan

aws

This blog post provides essential tips for individuals in Pakistan facing challenges with AWS certification exam registration through Pearson VUE. It outlines workarounds for the payment limitations, guides readers through the registration process, and offers insights to successfully navigate the system, ensuring a smooth experience.

AWS Certified Data Analytics Specialty DAS-C01 Notes

aws

I am sharing the notes I used to prepare for the exam.

Build your own Generative AI Art Studio with Amazon SageMaker JumpStart

aws

ml

In this notebook, I demonstrate how to use the SageMaker Jumpstart to generate images from text using state-of-the-art Stable Diffusion models.

Creating an Object Detection Model using Amazon Rekognition Custom Labels

aws

ml

Amazon Rekognition Custom Labels is a feature of Amazon Rekognition that enables customers to build specialized image analysis capabilities to detect unique objects and scenes integral to their specific use case. In this post, we will use this service to train a custom model with a small set of labeled images and use it to analyze new images via an API. This service uses AutoML to train models to find objects, scenes, concepts, object locations, and brand locations in images.

Working with partition indexes in AWS Glue

aws

Glue table partition indexes can significantly improve query execution. This post signifies the importance of partition indexes and shows a query performance comparison between tables with partition indexes and without indexes.

Implementing AutoEncoder with PyTorch

pytorch

This is a practice notebook to implement AutoEncoder in PyTorch. An autoencoder takes an image as input, stores it in a lower dimension (term encoder), and tries to reproduce the same image as output, hence the term auto. Autoencoders come in handy to identify and group similar images.

Implementing Word2Vec with PyTorch

pytorch

nlp

This is a practice notebook to implement word2vec in PyTorch.

Generating Text with Recurrent Neural Networks in PyTorch

pytorch

lstm

This is a practice notebook to build a character-level language model with LSTM using PyTorch. We will train a model on an input text, and our goal will be to generate some new text.

Predicting the Sentiment of IMDB Movie Reviews using LSTM in PyTorch

pytorch

lstm

This is a practice notebook to work with a dataset of 50,000 movie reviews from the Internet Movie Database (IMDB) and build an LSTM predictor to distinguish between positive and negative reviews.

Build Temporal Models for Univariate Time series Data with RNN, GRU, LSTM, CNN using PyTorch

pytorch

dl

This is a practice notebook to understand and build models for time series data. We will explore some popular neural network architectures including RNN, GRU, LSTM, and 1D CNN.

Detect Vanishing Gradients in Deep Neural Networks by Plotting Gradient Distributions

pytorch

dl

In this notebook, we will explore how vanishing gradients can affect the training of a deep neural network. We will visualize the gradient flow from the deeper to starting layers during the backpropagation for two popular activation functions, Sigmoid and ReLU.

Convolutional Neural Networks Filters and Feature Maps with PyTorch

pytorch

This is a practice notebook for implementing a convolutional neural network (CNN) on the MNIST dataset with PyTorch. We will implement the now famous LeNet-5 from Yann LeCun, a 7-layer CNN from 1989. Then we will explore and visualize the layers learned by our network including filters, feature maps, and output layers.

Two Class (Binary) Logistic Regression in Pytorch

pytorch

This is a practice notebook for implementing a two class logistic regression model in PyTorch. We will start by generating some synthetic data and then build an end-to-end pipeline to train a model. We will also see two ways to implement logistic regression models.

Linear Regression with PyTorch

pytorch

This is a practice notebook for implementing a linear regression model in PyTorch. We will start by generating some synthetic linear data and then load it into DataLoader class for creating mini-batches. Then build the complete pipeline to train the model and visualize its loss progress in TensorBoard.

Train an Image Classifier using Fastai (Deep Dive Analysis)

aws

ml

sagemaker

fastai

This is a practice notebook using the fastai library to build a simple image classifier for cricket, tennis, basketball, and soccer.

Building a Feature Repository with SageMaker Feature Store

aws

ml

sagemaker

This notebook demonstrates how to build a central feature repository using Amazon SageMaker Feature Store. Feature Store is used to store, retrieve, and share machine learning features.

Deploy Scikit-learn Models to Amazon SageMaker with the SageMaker Python SDK using Script mode

aws

ml

sagemaker

The aim of this notebook is to demonstrate how to train and deploy a scikit-learn model in Amazon SageMaker using script mode.

Loading SageMaker Linear Learner Model with Apache MXNet in Python

aws

ml

sagemaker

How to load SageMaker builtin Linear Learner model with Apache MXNet in Python.

Serverless Inference with SageMaker Serverless Endpoints

aws

ml

sagemaker

How to call an ML model endpoint hosted by SageMaker using serverless technology.

Demystifying Amazon SageMaker Training for scikit-learn Lovers

aws

ml

sagemaker

How to build, train, and deploy a machine learning model with Amazon SageMaker.

storemagic - Don’t lose your variables in Jupyter Notebook

jupyter

python

A post on IPython extension ‘storemagic’ to keep your important variables in persistent storage. Use this trick to keep them safe and stop pulling your hair (if there are any left)

Data Preparation with SageMaker Data Wrangler (Part 5)

aws

ml

sagemaker

A detailed guide on AWS SageMaker Data Wrangler to prepare data for machine learning models. This is a five parts series where we will prepare, import, explore, process, and export data using AWS Data Wrangler. You are reading Part 5:Export data for ML training.

Data Preparation with SageMaker Data Wrangler (Part 4)

aws

ml

sagemaker

A detailed guide on AWS SageMaker Data Wrangler to prepare data for machine learning models. This is a five parts series where we will prepare, import, explore, process, and export data using AWS Data Wrangler. You are reading Part 4:Preprocess data using Data Wrangler.

Data Preparation with SageMaker Data Wrangler (Part 3)

aws

ml

sagemaker

A detailed guide on AWS SageMaker Data Wrangler to prepare data for machine learning models. This is a five parts series where we will prepare, import, explore, process, and export data using AWS Data Wrangler. You are reading Part 3:Explore data with Data Wrangler visualizations.

Data Preparation with SageMaker Data Wrangler (Part 2)

aws

ml

sagemaker

A detailed guide on AWS SageMaker Data Wrangler to prepare data for machine learning models. This is a five parts series where we will prepare, import, explore, process, and export data using AWS Data Wrangler. You are reading Part 2:Import data from multiple sources using Data Wrangler.

Collecting metrics and logs from on-premises servers with the CloudWatch agent

aws

cloudwatch

A detailed guide on using cloudwatch agent to collect logs and metrics from an on-premises Ubuntu server.

Data Preparation with SageMaker Data Wrangler (Part 1)

aws

ml

sagemaker

A detailed guide on AWS SageMaker Data Wrangler to prepare data for machine learning models. This is a five parts series where we will prepare, import, explore, process, and export data using AWS Data Wrangler. You are reading Part 1:Prepare synthetic data and place it on multiple sources.

AWS Machine Learning Certification Notes (MLS-C01)

aws

ml

My notes for AWS ML speciality exam passed on May 14, 2022.

Kaggle - Blue Book for Bulldozers

ml

Docker - Send Container Logs to AWS CloudWatch

docker

python

aws

cloudwatch

A tutorial on sending docker application logs to aws cloudwatch.

Machine Learning Nomenclature

ml

A collection of machine learning terminologies

AWS EFS Sync to S3 Using DataSync

aws

lambda

efs

s3

synchonization

datasync

A tutorial to synchronize EFS with S3 bucket using DataSync service.

AWS EFS Sync to S3 Using Lambda

aws

lambda

efs

s3

synchonization

A tutorial on synchronizing EFS with S3 bucket using a Lambda function.

AWS Lambda - Test and Debug Locally in Visual Studio Code

aws

lambda

docker

A tutorial on testing and debugging AWS Lambda function from Visual Studio Code.

Docker - Debugging Python Application

docker

python

A tutorial on debugging Python application running on Docker inside WSL2.

Docker - Accessing Python Application Logs

docker

python

A tutorial on getting Python application logs running on Docker inside WSL2.

Sklearn Pipeline and Transformers Deep Dive

ml

sklearn

A detailed tutorial on Sklearn Pipeline, ColumnTransformer, FunctionTransformer, and a custom transformer.

AWS CloudFormation Template, Functions, and Commands

aws

Some useful notes on AWS CloudFormation template sections, intrinsic functions, and other tips.

AWS IAM Policy Types

aws

A summary of different AWS IAM policies.

AWS DNS Records - A, CNAME, ALIAS, & MX

aws

A summary of differences between AWS Route53 DNS Records

Python - A collection of output formatting tips

python

Some handy Python output tips

Python - Getting more information from Tracebacks

python

A tutorial to get more information from Python exception stack traceback.

Python Dictionary - Multiple ways to get items

python

A tutorial on multiple ways to get items from a Python dictionary.

My First Blog Post from Jupyter Notebook

jupyter

fastpages

Trying to check if everything is working as intended