Raj Sangani – Medium

Raj Sangani

Pinned

Raj Sangani
in
Towards AI

Advanced Recipes for Contrastive Learning

How to maximize batch size and choose in-batch negatives smartly to get the most out of your retrieval and search systems

Sep 6

Advanced Recipes for Contrastive Learning

Sep 6

Pinned

Raj Sangani
in
Towards Data Science

Adding Custom Layers on Top of a Hugging Face Model

Learn how to extract the hidden states from a Hugging Face model body, modify/add task-specific layers on top of it and train the whole…

Jan 26, 2022

Adding Custom Layers on Top of a Hugging Face Model

Jan 26, 2022

Pinned

Raj Sangani
in
Towards Data Science

Powerful Text Augmentation using NLPAUG !

Dealing with class imbalances in NLP classification problems through Text Augmentation techniques using the amazing NLPAug!

Jun 12, 2021

Powerful Text Augmentation using NLPAUG !

Jun 12, 2021

Pinned

Raj Sangani
in
Towards Data Science

Regex essential for NLP

Understanding various regular expressions and applying them to frequently encountered situations in Natural Language Processing

Jul 14, 2021

Regex essential for NLP

Jul 14, 2021

Raj Sangani
in
Towards Data Science

A Comprehensive Guide on Model Calibration: What, When, and How

Part 1: Learn about calibrating machine learning models to obtain sensible and interpretable probabilities as outputs

Sep 12, 2022

A Comprehensive Guide on Model Calibration: What, When, and How

Sep 12, 2022

Raj Sangani
in
Towards Data Science

Interpreting an LSTM through LIME

Learn how to interpret a Keras LSTM through LIME and dive into the internal working of the LIME library for text classifiers

Nov 11, 2021

Interpreting an LSTM through LIME

Nov 11, 2021

Raj Sangani
in
Towards Data Science

Locality Sensitive Hashing in NLP

A hands-on tutorial on how to speed up document retrieval by reducing the search space through Locality Sensitive Hashing (LSH)

Oct 1, 2021

Locality Sensitive Hashing in NLP

Oct 1, 2021

Raj Sangani
in
Towards Data Science

Overfitting is not the only problem Regularisation can help with

Understanding mathematically how Ridge Regression helps in cases where the number of features exceeds data points

Sep 8, 2021

Overfitting is not the only problem Regularisation can help with

Sep 8, 2021

Raj Sangani
in
Geek Culture

How do transformers fare on a grammar test?

Find out how well BERT and RoBERTa deal with questions from grammar quizzes

Aug 11, 2021

How do transformers fare on a grammar test?

Aug 11, 2021

Raj Sangani
in
Towards Data Science

Dealing with features that have high cardinality

A simple utility I use to address categorical features with many unique values

Aug 4, 2021

Dealing with features that have high cardinality

Aug 4, 2021

Raj Sangani

Raj Sangani

Machine Learning R&D at Keysight Technologies | UCI NLP Group , LinkedIn @https://www.linkedin.com/in/raj-sangani-85305a196/

Following

Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams