Hi, I am Muhammad Fakhar, a Computer Science student, mainly focused in Data Science and Machine Learning. As a Machine Learning practitioner, I have been writing posts and blogs on Data science and ML, here too. Thank you for visiting my blog page. I hope you enjoy my articles! Thank you.
End-to-End Projects
Machine Price Prediction
In this app, I implemented Random Forest model to generate the price of Machinery based on previous auction data.
Students Performance Predictor
In this app, I implemented the model to predict the maths score for students based on their profiles.
Journey with fastai Library
Computer Vision
Images Regression
Practice 1: Centre of Head
In this model, I implemented images regression, to find the coordinates of the centre of the head. It uses MSELoss.
Sinlge Label Classification
Practice 1: Dog vs Cat Classifier
I made a computer vision model hosted in gradio, on huggingfaces, in Fastai, which classifies between dogs and cats. Have a look in my repo! Click the title!
Practice 2: Player Classifier
I made a computer vision model, hosted in gradio on huggingfaces, which differs between two categories of people. Ronaldo and Messi. Funny!
Practice 3: Bear Detector
In this model, I classified between 3 types of bears, black, teddy and grizzly. With 100% accuracy. This app prototype is also hosted on huggingfaces spaces.
Practice 4: MNIST Digits Classification
In this model, I just took2 digits, as provided as Samples in the dataset. I implemented the Neural Network from scratch to classify the single label. This uses softmax activation function and CrossEntropyLoss.
Practice 5: Pet Breed Classification
In this model, I implemented classifying between 37 different cats and dog breeds. This uses softmax activation function, and CrossEntropyLoss.
Multi-Label Classification
Practice 1: Predicting Multipple classes
In this model, I predicted multiple classes present in the picture. This uses sigmoid activation function and BCELossWithLogits loss.
Tabular Data
Practice 1: Movies Prediction/Collaborative Filtering
In this collaborative friltering, I worked on movies dataset, to predict the movie for a user based on his reviews.
Practice 2: Price Prediction from previous auctions
In this model. based on previous prices of machines on auctions, I predicted the price for new machinery. I used the data from kaggle.(bluebook for bulldozers)
Natural Language Processing
Practice 1: Generating text for movie reviews/ Classifying the reviews
In this Language Model, I predicted the text to write the reviews for movies. Also, I classified them as positive review or negative review.
Posts
Unfreezing and Transfer Learning in Deep Learning
The Dillema of Bias in Data
Deploying ML models?
Image Recognizers for Non-Image Tasks?
subscribe via RSS