Day 23 #31DaysofML by @pvergadia
π‘Tips to get started with #MachineLearningπ‘
Pratham Prasoon is a 16 year old with passion for making #ML simple. We sat down for a virtual chat where he shares some of his favorite resources. google.smh.re/byM
π So excited to share this with you on day 30 #31DaysofML
Built a sound classification model for these sounds
πΈGuitar
π¦ Bird
π Clap
Also deployed the model on Cloud Run!
Want to try it yourself? code π goo.gle/3kF9yPS#MachineLearning#NoCode#GoogleCloud
Day 16 of #31DaysofML
Here are some use cases of computer vision
β Industrial inspection
π Product search
π Document classification
πΌ Image search
Checkout the architectures to implement using Vision API with
#NoCode#MachineLearninggoogle.smh.re/3LN
Day 29 #31DaysofML
π€ What is Hyperparameter tuning?
The process of selecting the right set of hyperparameters for your #ML app
π€ What are hyperparameters?
Variables that govern the training process & the topology of an ML model.
A π§΅ π 1/3
Day 26 #31DaysofML
π€ What is MLOps?
#MachineLearning Model Operationalization Management
π§ But what is it?
End-to-end #ML development process to design, build & manage reproducible, testable, & evolvable ML-powered systems
π€ Here's how MLOps looks like on #GoogleCloud
1/2
Day 22 #31DaysofML
What is Scikit-learn?
It is #MachineLearning library that offers a rich suite of tools for doing things such as:
πΉ Dataset loading & manipulation
πΉ Preprocessing pipelines and metrics
πΉ Comes with large number of ML algorithms
google.smh.re/2k7
Day 16 of #31DaysofML
Here are some use cases of computer vision
β Industrial inspection
π Product search
π Document classification
πΌ Image search
Checkout the architectures to implement using Vision API with
#NoCode#MachineLearninggoogle.smh.re/2gc
π‘Tips to get started with #MachineLearningπ‘
A lot of us are fascinated by @PrasoonPratham who is on a mission to make #ML simple. So, for today's #31DaysofML
I invited him for a quick virtual chat. Watch & follow πΉ π goo.gle/3bzQbDP
Day 23 #31DaysofML
Today I created a tweet sentiment analysis model with absolutely #nocode using #GoogleCloud Natural Language AutoML which supports π types datasets:
πΉ Single & Multi Label classification
πΉ Entity extraction
πΉ Sentiment Analysis
See how I did π
π Here is how you train the model
from sklearn import svm
clf = svm.SVC() --- Support vector classifier model
clf.fit(X_train, y_train) -- train the model using fit function
#31DaysofML
π Here is how you split the data using scikit-learn
from sklearn.model_selection import train_test_split
all_X, all_y = preprocess(data)
X_train, X_test, y_train, y_test = train_test_split(all_X, all_y)
#31DaysofML
Scikit-learn is also a great way to learn what different types of models do and gain some intuition around how the various parameters for a model perform.
#31DaysofML
Day 16 of #31DaysofML
Here are some use cases of computer vision
β Industrial inspection
π Product search
π Document classification
πΌ Image search
Checkout the architectures to implement using Vision API with
#NoCode#MachineLearninggoogle.smh.re/1vD