Data Science Case Study Examples Notebooks

Authors: Yodiaditya

Introduction:

I wrote data-science case study examples, start from Deep Learning, Neural Network, and many more.
Start from theory, implementation, bug, optimization memory and benchmarking.
The major difference in this repository, you will find long comments and detailed explanation on what each lines do.

I hope this will help you to understand the code and the theory behind it.
This repository is still under development, I will add more case study examples in the future.

Case Study:

Binary Classification using BERT for Spam Detection

We will using negative and positive class (num_labels=1), Simple Layer, Dropout and Adam Optimizer combined with BERT

• Binary Classification for Spam using Bert and Tensorflow : classification_spam_with_bert_tensorflow.ipynb
• Binary Classification for Spam using Bert and PyTorch classification_spam_with_bert_tensorflow.ipynb
• Binary Classification for Spam using Bert and Trainer HuggingFace classification_spam_with_bert_trainer.ipynb

Single and Multi-GPU Training Benchmark

Understanding how single and multi-gpu performs during training

• Benchmark Trainer using multi-GPU: benchmark_training_multigpu.py
• Benchmark Trainer using single-GPU: benchmark_training_singlegpu.py

Citation

    publisher = "Yodiaditya",
    url = "https://github.com/yodiaditya/datascience",
    website = "https://yodiw.com",
    linkedin = "https://www.linkedin.com/in/yodiaditya/"