AI Developer Advocate
@google
- Python🐍
- Machine Learning 🤖🧠
- Google AI ⚙️🧠
- DevRel 🥑🗣️
I've shared a new baseline notebook for the AMP®-Parkinson's Disease Progression Prediction competition on Kaggle
https://www.kaggle.com/code/gusthema/parkinson-s-disease-progression-prediction-w-tfdf
It will teach you:
🧑🔬 Feature pre-processing and engineering
📏 How to implement and use the SMAPE metric
🌳 TFDF
Leave +1 and comment!
Guess the #GoogleIO lineup!
1. 🤔🆕🟧🌊
2. 👌⚙️👩🏫📊📂
3. 🦺🛟🔨🤖💬
Go on the website to find out and get a preview of the upcoming ML/AI content ⬇️
Put this on your calendar!!! NOW!!!👀
Thank me later!
https://io.google/2023/program/e695ebdd-b968-4b85-98d9-0a722892e842/
Do you wanna learn some ML and apply to a real problem without all the hype?
Try this notebook here:
https://kaggle.com/code/gusthema/student-performance-w-tensorflow-decision-forests
You'll get a good baseline and you'll learn:
- How to load BIG tabular datasets👀
- ensemble models!🤯
- fast training with CPU only HW!⚡️
If you are here on Mastodon, there's a good chance you came from Twitter.
And you'll probably enjoy this article too!
"It almost doesn’t matter if this is deliberate sabotage by Musk or the blundering stupidity of a clueless idiot. The upshot is the same: Twitter is dying"
🎉🤩🎊🎇
TensorFlow 2.12 is live and with some cool stuff!
to name a few:
- new Keras save format
- saved model fingerprinting
- improvements to tf.function
- improvements to tf.data
What feature would you like to see on TensorFlow?
https://blog.tensorflow.org/2023/03/whats-new-in-tensorflow-212.html
Check it out this post I helped create
TensorFlow Decision Forests is great and you should definitely try it if you want to do Machine Learning on Tabular data!
among the many features, there are:
- Training models in Google Sheets
- Easier hyper-parameter tuning
- Hyper-parameters templates
- Easy to Serve models on Google Cloud
- Distributed training on billions of examples
Do you miss a feature?
please let me know and we can try to help you!!
🏠📈
I'm having a blast with TensorFlow Decision Forests!🌳🤖
Check out this Kaggle notebook where I predict housing prices using #TFDF 🏡💰
https://www.kaggle.com/gusthema/house-prices-prediction-using-tfdf
It's straightforward and easy to replicate so anyone can copy and give it a try
Let me know what you think!
4⃣Argparse
Parsing command line parameters is an easy task, right?
is it?
https://docs.python.org/3/library/argparse.html
👇🏾
#Python 🐍, beyond a great language for ML, it is also very powerful for non-ML tasks!
Knowing the right library to help you will make you write less code and be more productive overall
If you like this thread, follow me so you don't miss the next ones and RT
👍
2⃣Itertools
Bring many building blocks to work with iterables (eg: collections) for fast, memory efficient and useful by themselves or in combination
This is one of the few documentations that I liked reading top to bottom! (more than once)
https://docs.python.org/3/library/itertools.html
👇🏾
3⃣Pillow
Python Imaging Library for image processing
• extensive file format support
• efficient internal representation
• powerful image processing capabilities
https://pillow.readthedocs.io/en/stable/
👇🏾
