Open in app

Sign In

Write

Sign In

David Dale
David Dale

112 Followers

Home

About

Published in Towards Data Science

·Dec 14, 2021

Compressing unsupervised fastText models

How to reduce word embeddings models by 300 times, with almost the same performance on downstream NLP tasks — FastText is a method for encoding words as numeric vectors, developed in 2016 by Facebook. Pretrained fastText embeddings help in solving problems such as text classification or named entity recognition and are much faster and easier to maintain than deep neural networks such as BERT. However, typical fastText models are…

NLP

6 min read

Compressing unsupervised fastText models
Compressing unsupervised fastText models
NLP

6 min read


Published in Towards Data Science

·May 4, 2021

How to adapt a multilingual T5 model for a single language

Load embeddings only for the tokens of your language to reduce the model size — T5 is an encoder-decoder transformer from Google that once was SOTA on several NLU and NLG problems and is still very useful as a base for seq2seq tasks such as text summarization. The first T5 model was for English only, and then the massively multilingual version followed. …

NLP

4 min read

How to adapt a multilingual T5 model for a single language
How to adapt a multilingual T5 model for a single language
NLP

4 min read


Feb 13, 2018

Do you have to try to love math?

If you don’t like and don’t understand math, does it mean you are stupid? Do you need to love math to achieve at finance/engineering/science? No. Where you possibly are People say that math is the queen of all sciences, that without mathematical thinking it’s impossible to survive and to strive, that everyone should be…

Mathematics

6 min read

Do you have to try to love math?
Do you have to try to love math?
Mathematics

6 min read


Published in The Startup

·Jan 14, 2018

A machine learning model to understand fancy abbreviations, trained on Tolkien

Recently I bumped into a question on Stackoverflow, how to recover phrases from abbreviations, e.g. turn “wtrbtl” into “water bottle”, and “bsktball” into “basketball”. The question had an additional complication: lack of comprehensive list of words. That means, we need an algorithm able to invent new likely words. I was…

Machine Learning

9 min read

A machine learning model to understand fancy abbreviations, trained on Tolkien
A machine learning model to understand fancy abbreviations, trained on Tolkien
Machine Learning

9 min read

David Dale

David Dale

112 Followers

NLP researcher, chatbot developer, teacher of applied math. See daviddale.ru/en

Following
  • Nikolai Liubimov

    Nikolai Liubimov

  • Olga Vorona

    Olga Vorona

  • Aira Mongush

    Aira Mongush

  • [Y|J]unoTravel

    [Y|J]unoTravel

  • Alexandr Usov

    Alexandr Usov

Help

Status

Writers

Blog

Careers

Privacy

Terms

About

Text to speech