Share this page!

Enter Twitter Thread URL to Unroll

Needs to be the full URL like: https://twitter.com/threadreaderapp/status/1644127596119195649

How to get URL link on Twitter App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Akshay 🚀 Profile picture
@akshay_pachaar
Aug 12 • 12 tweets • 5 min read Twitter logo Read on Twitter
I started my career in Data Science back in 2016 ⏳

Here's a detailed roadmap for those starting out today!

What's covered:

- Python
- Machine Learning
- Maths for ML
- ML Books
- MLOps
- LLMs

Let's go! 🚀 Image
1️⃣ Python

If you are new to programming and just getting started.

There isn't a better place to learn Python than David J Malan's CS50p.

Beautiful explanations and great projects.
It's a complete package ⚡️

Check this out 👇
edx.org/course/cs50s-i…
2️⃣ Machine Learning

If you are already into programming and want to start with machine learning.

@AndrewYNg 's Machine learning specialisation has been tested by time and taken by millions.

Check this out 👇
coursera.org/specialization…
3️⃣ Deep Learning Fundamentals

A free course on deep learning using a modern open-source stack.

Taught by @rasbt, best-selling author, professor & AI educator.

Arguably the best course on DL today!

Check this out👇
lightning.ai/pages/courses/…
4️⃣ NLP

NLP Specialization Coursera

A comprehensive curriculum for those interested in Natural Language processing.

Check this out 👇
in.coursera.org/specialization…
5️⃣ Deep Learning for Coders

Looking for some hand on experience in modeling, best practices for training, evaluating your models.

Covers everything from building a classifier to Stable Diffusion! 🚀

Check this out 👇
fast.ai
6️⃣ Mathematics for ML

You start with this journey and at any point of time, you feel like Maths is holding you back.

This course by Imperial college London should help.

Check this out 👇
coursera.org/specialization…
7️⃣ Statistics

If you are a ML practitioner, knowing Statistics can be a great tool in your arsenal.

Sal Khan lectures from Khan Academy are pure gold 🥇

Check this out 👇
khanacademy.org/math/statistic…
8️⃣ Books 📚

Some good books for those who like to read ⬇️

- Intro to Statistical Learning
- Approaching almost any ML prob.
- Deep Learning by Goodfellow
- Deep Learning with Keras
- NLP with Transformers by L. Tunstall
- MLOps by Chip Huyen
9️⃣ MLOps

A model is useless, unless you create a service out of it.

MLOps is a practice that blends Machine Learning, DevOps, & Data Engineering, enabling streamlined delivery of ML systems

One of the best resource to get started!

Check this out👇
madewithml.com
🔟 Large Language Models

Building AI apps using LLMs is going to be a high leverage skill!

@LightningAI offers some of the best resources to masters the LLM landscape today!

Their Lit-GPT is licensed under Apache-2.0! 🔥

Comprehensive LLM tutorials👇
https://t.co/dLL6YlToOxlightning.ai/pages/blog/
Image
That's a wrap!

If you’re interested in:

- Python 🐍
- ML/MLOps 🛠
- CV/NLP 🗣
- LLMs 🧠

Find me → @akshay_pachaar ✔️
Newsletter → ✔️

Everyday, I share tutorials on above topics!

Check my tutorial on self-attention👇
https://t.co/sQVIgeYljHmlspring.beehiiv.com/subscribe

• • •

Missing some Tweet in this thread? You can try to force a refresh
　

Keep Current with Akshay 🚀

Akshay 🚀 Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @akshay_pachaar

Akshay 🚀 Profile picture
Aug 4
One of the most fundamental & profound concept in Statistics is Central Limit Theorem! 📊

Let me break it down for you today!

-- explained with python code --

Let's go!🚀 https://t.co/NH5Kzl6nJUtwitter.com/i/web/status/1…
Image
Simply put, CLT states that the distribution of sample means will approach a normal distribution as the sample size increases.

This is regardless of the population's distribution.

Sounds confusing❓🤔

Let's break it down with an example in next tweet!🎉 Image
Let's say we want to study the heights of people in a town.

We collect a sample of 10 people and calculate the average height.

Now, repeat this 1000 times, and plot the distribution of these averages.

This distribution will tend to be normal!

We simulated this using code👇 Image
Read 6 tweets
Akshay 🚀 Profile picture
Jul 29
A Data Scientist often needs to strike a balance between Precision & Recall!

The two are competing metrics & this is where a PR-curve comes to the rescue!

Today, I'll clearly explain Precision-Recall Curve & how to create one for your model!

Let's go! 🚀 Image
If you already understand the concepts of precision and recall, feel free to skip to the next tweet.

But if not, don't worry—I've got you covered!

You can check my previous post on understanding precision & recall 👇
PR-curve is a graphical representation that shows the trade-off between precision & recall for different threshold values.

Here are 5 simple steps to create one! Image
Read 7 tweets
Akshay 🚀 Profile picture
Jul 28
Everyone should learn how to fine-tune LLMs! 🔥

Introducing Lit-GPT! ⚡️

- NanoGPT based💡
- Supports Llama-2 🦙
- With pre-training/fine-tuning script 📜
- Powered by LightningAI's Lightning Fabric!

I'll show you how to do it in 5️⃣ easy step! ...👇 Image
Lightning AI recently launched Lit-GPT!

A clean, solid & optimised LLM implementation with pre-training & fine-tuning support using LoRA & Adapter.

We'll go step-by-step:
- Installations
- model download
- data prep.
- fine-tuning

1️⃣ Installation & Setup:

Check this 👇 Image
2️⃣ Downloading the model weights

Lit-GPT supports following pretrained weights:

- StableLM
- Pythia
- RedPajama-INCITE

Here we gonna use RedPajama-INCITE 3B weights!

And doing this doesn't get easier!

Check this out 👇 Image
Read 9 tweets
Akshay 🚀 Profile picture
Jul 26
Microsoft is offering FREE courses in following areas:

- AI
- IOT
- Data Science
- Machine Learning

A project-based pedagogy that allows you to learn while building! 🚀

Read More ...👇 Image
1️⃣ AI for beginners

A 12-week, 24-lesson curriculum all about Artificial Intelligence.

You'll learn about:
- Neural Nets
- Deep Learning
- Knowledge representation
- Genetic Algorithms & multi-agent systems

Check this out 👇
https://t.co/rZ3Jdw2Le8microsoft.github.io/AI-For-Beginne…
Image
2️⃣ IOT

Learn about IOT by doing project that cover the journey of food from farm to table.

This includes farming, logistics, manufacturing, retail and consumer - all popular industry areas for IoT devices.

Check this out 👇
https://t.co/o4GJM8gHzmmicrosoft.github.io/IoT-For-Beginn…
Image
Read 6 tweets
Akshay 🚀 Profile picture
Jul 25
Better prompts lead to better responses from LLMs!

But how do you decide which prompt is better❓🤔

Let me show you how it's done: Image
So, how prompt selection is done ?

An LLM is tested on a certain test dataset using different prompt templates & the performance is evaluated.

The challenge with real world data is that it can be noisy.

Check this out👇 Image
Today, there are various ways to prompt LLMs.

Here are a few standard techniques:

- Few Shot prompting
- Instruction prompting
- Templatized prompting
- Chain of thoughts prompting

Images below provide example for each 👇


Image
Image
Image
Image
Read 8 tweets
Akshay 🚀 Profile picture
Jul 24
Two of the fundamental evaluation metrics in machine learning are:

- Precision
- Recall

But, their formal definitions can be a bit confusing.

In next 2 minutes I'll help you clearly understand them!

Let's go!🚀 Image
Let's say there are 10 people in a Town.

2 of them have committed a crime, so in reality:

- 8 are innocent
- 2 are guilty

This is how it looks 👇 Image
Now, we hire a detective to catch the guilty parties!

The detective accuses 3 people of being guilty, while 7 are deemed innocent.

Out of the 3 accused, only 1 is actually guilty.

Out of the 7 predicted innocent, 1 is guilty in reality.

This is how the scenario looks so far👇 Image
Read 8 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

Send Email!

:(