Vadim Markovtsev
vmarkovtsev
ML engineer, backend engineer, system programmer. Public speaker.
Languages
Repos
232
Stars
440
Forks
116
Top Language
Python
Loading contributions...
Top Repositories
SIGSEGV handler which prints the debug stack trace for C/C++ projects on Linux. It is configurable and rather robust.
Two-way deep RNN for sentiment classification.
Fast Longest Common Substring in Go
libudt4 Python wrapper written with Cython
Python script to fetch GitHub repos metadata.
API Blueprint Python parser
Repositories
232Google Groups raw email crawler and parser
A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on Hopper, Ada and Blackwell GPUs, to provide better performance with lower memory utilization in both training and inference.
SIGSEGV handler which prints the debug stack trace for C/C++ projects on Linux. It is configurable and rather robust.
Home Assistant integration of Sistena Onix Firebase API for smart HVAC
The home of the default HACS repositories.
🎨 Brands for Home Assistant
Fast Longest Common Substring in Go
Home Assistant integration for spanish energy data providers (e.g., datadis)
Python script to fetch GitHub repos metadata.
Fast Fourier Transform Frontend
API Blueprint Python parser
source{d} neural code completion research - temporary public fork
Recurrent neural network to split code snippets from text.
A high-throughput and memory-efficient inference and serving engine for LLMs
codec2 speech codec with some performance optimizations
TensorRT-LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and build TensorRT engines that contain state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT-LLM also contains components to create Python and C++ runtimes that execute those TensorRT engines.
libudt4 Python wrapper written with Cython
Interesting links & research papers related to Machine Learning applied to source code
Stripped Jasper (JPEG2000 library) with fixed 64-bit portability issues and various security patches applied, compiled with cmake
Utils for streaming large files (S3, HDFS, gzip, bz2...)
The official implementation of “Sophia: A Scalable Stochastic Second-order Optimizer for Language Model Pre-training”
Template reverser for Go
Unsupervised text tokenizer for Neural Network-based text generation.
Hackable and optimized Transformers building blocks, supporting a composable construction.
Fast and memory-efficient exact attention
gopy generates a CPython extension module from a go package.
Two-way deep RNN for sentiment classification.
Fun with malloc() and free(): PoC implementation, tracing, training deep neural networks
Log Web History to Firebase
Samsung Extended Autotools Project Template