The easiest LLM Fine Tuning UI just Landed! 🚀
Now, ANYONE can fine-tune (almost) any LLM available on Hugging Face Hub by just uploading a CSV and choosing the parameters and by a single click of a button! 💥 Here's how you can do it: 1/N First, you need a huggingface account! If you dont have one, create one: .
Once your account is setup, click this link:
You can choose any name for the space 2/N hf.co
huggingface.co/new-space?temp…

Here are some coding tutorials on large language models (LLMs): 🧵1/N 2/N

Here's a thread on how ChatGPT works: ChatGPT is a large language model trained by OpenAI to generate text based on a given prompt.

Do you want to learn time-series analysis for free? Check out this thread 🧵 1/12 Konrad Banchewicz has been making time-series tutorials on my YouTube channel. The first episode was: Curve fitting is (almost) all you need.
Video:
Notebook: kaggle.com/code/konradb/t… 2/12

Here is a simple t-h-r-e-a-d to show you how easy and fun it is to fine-tune almost any transformer model for sentiment classification on imdb dataset (or any other binary classification dataset) using the new version of Tez ⬇️ 1/N First, import all the cool stuff you need 2/N

We are starting the next community competition THIS WEEK (within 1-3 days)! It will be a computer vision problem :wink:

**Top-3 will get a brand new NVIDIA RTX 3080Ti GPU each!**

How to join? See this t-h-r-e-a-d ;) 1/4 To be eligible for the prize, follow these steps:
1: Register for GTC using this link: nvidia.com/gtc/?ncid=ref-…
2: wait for the competition to launch
3: attend GTC sessions.
*Prizes will be awarded only to those who register using the link above and attend some sessions.* 2/4

🧵If you want to learn time series analysis, take a look at these great resources created by Konrad Banachewicz ⬇️ 1/N Konrad is creating content for time series analysis and is taking time out of his busy schedule to deliver them on my youtube channel. Here is the first video where he discusses curve fitting: 2/N

Understanding self-attention. A Twitter thread 🧵
To understand what self-attention is and how it works, you need to know only the following terms:
- dot product
- matrix multiplication
- softmax
Note: it's not rocket science. let's keep it simple. 1/ - in nlp, each sentence is represented by a bunch of tokens.
- each token maps to some number
- each number is represented by a vector (embedding)
- embedding carries the meaning of the token 2/

"Attention is all you need" implementation from scratch in PyTorch. A Twitter thread:
1/ There are two parts: encoder and decoder. Encoder takes source embeddings and source mask as inputs and decoder takes target embeddings and target mask. Decoder inputs are shifted right. What does shifted right mean? Keep reading the thread. 2/

🧵For Kaggle's 30 days of machine learning competition, I made several videos and notebooks on getting started. If you are planning to get started with Kaggle competitions, maybe you can give this thread a try 1/10 How to create cross-validation folds before starting a competition:

Video:
Notebook: kaggle.com/abhishek/30-da…

2/10

As promised, I have removed my book from Amazon India. It's sad that @amazon @amazonIN supports & allows sellers who sell printout copies of lower quality but don't support the original authors and allow these fake copies to be on their platform @AmazonHelp has been of no use 1/ Due to amazon's algorithm, my ebook price is also reduced automatically when these fake sellers start selling the paperback at a price lower than the ebook. Yet again, @AmazonKDP and @AmazonHelp have been of no use. I have all the conversations since last year 2/

Kaggle's #30daysofml till now (day 1 to 12): 🧵

1/ 🧵 Getting started with data science and machine learning.

- the first step is to know what data science and machine learning mean and is this field worth getting into?
- if you ask me, I would say yes. the world is data-centric. 2/ in every industry, data is useful and will be for a long long time.
- if you are a developer, it will be easier for you to get into data science
- but it also means you have to work twice as hard since you are already working

💥 Did you know that there are problems other than MNIST and iris that you can solve (or try to solve) to learn deep learning and computer vision? Here is a list of my favourite Kaggle competitions to learn deep learning and computer vision from ⬇️ 1/13 kaggle.com/c/dogs-vs-cats 2/13

👉 Want to learn Natural Language Processing by solving problems? Here is a list of my favourite NLP competitions on @kaggle to learn from ⬇️ 1/15 kaggle.com/c/predict-clos… 2/15

🚀 If you are starting with machine learning / deep learning and get a new dataset to work on, either on kaggle or in real-world or just for fun. There are a few things you must always take care of to squeeze the most out of your model and make it awesome: ⬇️⬇️⬇️
1/6 🔹 Look at the data carefully. Do EDA.
🔹 Look at the targets. See how they are distributed and what kind of problem this is.
🔹 Choose the right metric to evaluate your models
2/6

Not surprised that none of the nay-sayers were not able to respond. That's what happens when you start accusing and abusing someone without understanding the context.

https://twitter.com/abhi1thakur/status/1358674182121418752

Here are some solutions in this thread 🔽🔽🔽 1/7 Here is a solution using pandas. Time taken: 191.77s
2/7

So, people who called me names here is a test for you. You need to use python.

- You have 100k CSVs in a folder.
- Read all files in the folder
- Combine them in a single CSV
- Save the combined file for feature engineering using pandas
- All files share the same header
1/4 where do I find 100k CSVs in a folder? Well, in many scenarios and real-life situations. I have made it easy for you: github.com/abhishekkrthak…

Those who called me names must use pandas.
Those who are willing to learn, scroll below.

2/4

Stacking in machine learning 🔽 1/4 - Divide the training data into folds.
- Train a bunch of models: M1, M2.....Mn.
- Create full training predictions (using out of fold training) and test predictions using all these models. 2/4

Have you had troubles or having troubles arranging your machine learning projects? This thread should give you some idea on how to arrange machine learning / deep learning projects. See the folder structure: 1/6 🔽 input/: This folder consists of all the input files and data for your machine learning project. If you are working on NLP projects, you can keep your embeddings here. If you are working on image projects, all images go to a subfolder inside this folder. 2/6

In this thread, I will tell you how to learn python for data science in 1 hour 👇 1/N It's not possible 2/N