[Day 133] Gathering data for the Scottish dataset project + Factor analysis + Grokking ML + MLxFundamentals Day 4 (2)

 Hello :)
Today is Day 133!


A quick summary of today:
  • started putting together some audio clips + transcription for the Scottish dataset project
  • saw how eigenvalues play a role in factor analysis
  • started reading Grokking Machine Learning on Manning.com
  • finished MLx Fundamentals' final session

Firstly, about the audio data

My Scottish partner for this project has recorded various phrases in Glaswegian in the past and uploaded them to youtube. Today I did 4 of the 10 videos. 

To cut the clips I ended up using an app called VideoPad, and even though it is a paid app, it allows me to just cut an audio clips in smaller pieces and save them as new files. 

This is a sample audio waveform of one of the 4 videos

What I did was, make short clips around each expression. I am not sure what these waves are called when there is speech. So for example from the above, I ended up with 36 clips, and uploaded them all (along with the other 3 videos' audio clips) to our project's drive. And the total amount we have so far is 4.75 minutes.  

Secondly, today I read about eigenvalues' role in factor analysis

In my stats class at uni, we learned about factor analysis, and at the end of the chapter I saw the word eigenvalues, and I am glad because once again I will see their real world impact (after my dive into multicollinearity). 

Firstly, about factor analysis, here are the results after using unstandardized variables

And after standardization

Why standardize?

  • result interpretability
  • helps with linearity
  • treats variables equally
Where are the eigenvalues? The overall under each loading. 1.981 and 1.008. Which are sum of the square of each of the 3 values above it. 

To interpret this, we take the 1st row and F1: 0.00089222 - this is F1 accounts for 0.0089% of the variance in Y1 which is Finance, while F2 accounts for 99.90%. And in total, the 2 factor space accounts for 99.99% of the variance in Y1 Finance. 

 The eigenvalues can help to determine which factors to keep (i.e. using scree plots).

Love it when I see the math I studied for ML being used in practice, and *where* it is used.


Thirdly, about Grokking ML

I decided to subscribe to manning.com and the 1st book I decided to read was Grokking ML (as it is one of the most recommended and popular ones). Today I managed to read the 1st 4 chapters(what is ML, types of ML, linear regression, optimization), and I can definitely see why it is popular for beginners, and I am excited to keep reading. 


Finally, the last session from MLxFundamentals was delivered by Wenhan Han a PhD candidate from TU Eindhoven.

It was about loading and using an LLM, and a diffusion model

Some interesting bits from both parts are:

Question to an LLM:

How many kinds of human beings are there in the history?

We saw the top answers from the model. 

Various prompt strategies

Zero-shot

Few shot

Chain-of-thought

Then, how to finetune an LLM with 'unsloth'

We add LoRA adapters
Prepare the data
Data was from huggingface

Train the model
And then inference

For finetuning I need to try unsloth by myself. For some part of the tutorial, an openai api key was required which was unfortunate because I do not have one, so I just watched that part. 

As for diffusion models


I saw the library 'diffusers' was used which I did not know.

And I played around with it a bit, a picture I tried to get and failed was 'a horse riding an astronaut'


That is all for today!

See you tomorrow :)

Popular posts from this blog

[Day 198] Transactions Data Streaming Pipeline Porject [v1 completed]

Image
 Hello :) Today is Day 198! A quick summary of today: data streaming pipeline project [v1 done] Here is a link to the project's repo. Well ... I did not know I can do it in a day (~14 hours) after yesterday's issues but here we are. Turns out in order to insert the full (~70 variables with nested/list structure), I need the proper pyspark schema. And yesterday I did not have that and that is why when I was reading data in the kafka producer I was getting NULL in the columns - my schema was wrong. Well today I not only fixed the schema for the 4 variables I had yesterday, but included *all* the variables that come from the Stripe API ~ 70 (for completeness).  When I run docker-compose, the data streams and is input into the postgres db (and is still running). Unfortunately, the free Stripe API for creating realistic transactions has a limit of 25, so every 3 seconds, 25 new transactions are sent to the db. It has been running half the day (since I got that set up) and as I am wr
Read more

[Day 107] Transforming natural language to charts

Image
 Hello :) Today is Day 107! A quick summary of today: created text2chart - transforming natural language to charts [ github ] [ webapp ] I am super sick, but in the around 3 hours of my fever not being high, I managed to study a bit. I was again thinking about creating some nlp project but was stuck on the idea of how to host a model. After looking around I found this article where they create a similar app called chat2vis  and they use codellama. I know codellama is an open source model that I can get from huggingface and I wondered, how did they host it. Turns out - they ask the user to input a huggingface api key, which then allows for querying and generating content. (AWESOME! this is applicable to some of the other PDF chat apps that I have, but I will look to make an app of them later).  My fever eased a bit, and I sat on my chair to do some work ^^ Here is a breakdown of the app. I used streamlit. Also, 2 default datasets during dev: Financial Statements of Major Companies(2009-
Read more

[Day 54] I became a backprop ninja! (woohoo)

Image
 Hello :) Today is Day 54! A quick summary of today: Read through cs231n's neural network lectures [1] [2] [3] Spent 8 hours going over Andrej Karpathy's backprop ninja tutorial After yesterday's confidence boost in doing backprop, I came in today with the intention of becoming of, as Andrej Karpathy calls it, a backprop ninja.  The cs231n NN lectures were the usual about how NN became to be, param initialization, activation functions, loss functions, regularization and optimizers. Nothing out of the ordinary, but still was good to reassure myself of that knowledge. But that was definitely not the main even of today. As I mentioned, I wanted to become (or at least attempt to become) a backprop ninja today. I have watched that particular tutorial twice, the 1st time I had no idea what is going on, the 2nd time I could follow, and now it was the 3rd time - and now, I wanted to participate and actually see the why and the how of the things we have to do, when we do manual back
Read more