Name: Leonid Boytsov
Type: User
Company: This is my personal account
Bio: Sr. Research Scientist @ Amazon, pH-D from CMU, working on unnatural language processing, speaking πtorch & C++. I remember dependency parsing & boosted trees.
Twitter: srchvrs
Location: Pittsburgh
Blog: http://searchivarius.org/about
Leonid Boytsov's Projects
🚀 A simple way to train and use PyTorch models with multi-GPU, TPU, mixed-precision
Improving the effectiveness Lucene's BM25 (and testing it using Yahoo! Answers and Stack Overflow collections)
Benchmarks of approximate nearest neighbor libraries in Python
Structured information retrieval using SOLR (archival version)
A Lucene toolkit for replicable information retrieval research
Code used in Leonid Boytsov's blog: http://searchivarius.org
A patched clearnlp 2.0.2
Pytorch implementations of various Deep NLP models in cs-224n(Stanford Univ)
A collection of OpenEphyra components necessary for question analysis
Fast implementations of the scancount algorithm: C++ header-only library
Gradient boosted models
Indexing TREC corpora and Wikipedia using Lucene
Scripts to reproduce InPars light paper
pubmed-query-with-entities
Metric learning algorithms in Python
A word alignment tool based on famous GIZA++, extended to support multi-threading, resume training and incremental training.
Submission archive for the MS MARCO document ranking leaderboard
Submission archive for the MS MARCO passage ranking leaderboard
TOROS N2 - lightweight approximate Nearest Neighbor library which runs faster even with large datasets
Automatically exported from code.google.com/p/nndes
ntcir-qalab-cmu-baseline
Open Source Neural Machine Translation in PyTorch
Permutation algorithms to test statistical significance of experimental results.
Python bindings for the fast integer compression library FastPFor.
Simple structured learning framework for python
A slightly modified version of the older version of the transformer library pytorch-pretrained-BERT
A C++ library to compress and intersect sorted lists of integers using SIMD instructions
A nearly SVMLight (but without the class label) Python writer