# BPE tokenization

A Transformer encoder model based on BERT architecture, specifically designed for generating RNA sequence embeddings

GENA-LM is an open-source foundational model family for long DNA sequences, based on a Transformer masked language model trained on human DNA sequences.

A bidirectional encoder model pre-trained for Russian language, trained on large-scale text corpora using standard masked language modeling objectives

This is a GPT-2 model based on the Transformer decoder, adjusted specifically for single-turn Spanish dialogue tasks, with 345 million parameters.

This is a small Greek language model based on the RoBERTa architecture, with a parameter scale approximately half of the base model. It is suitable for the masked filling task of Greek text.

A RoBERTa model based on Turkish, pre-trained with masked language modeling (MLM) objective, case-insensitive.

A RoBERTa model pre-trained on Turkish with masked language modeling (MLM) objective, case-insensitive, medium-sized architecture.

Pre-trained Russian base BERT model by SberDevices team, with 178 million parameters, trained on 30GB Russian text

This is an automatic speech recognition (ASR) pretrained model based on the ESPnet2 framework, trained on the mini-an4 dataset and supports English speech recognition.

A Russian large language model pre-trained by SberDevices team, based on Transformer architecture with 427 million parameters and 30GB training data