OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Model Zoo

Model Zoo is a repository of pretrained Russian-language NLP models, with README examples for using ruBERT, ruRoBERTa and ruT5 through Hugging Face Transformers. It covers masked-token prediction, text-to-text generation, and attention visualization with BertViz.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Model Zoo — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you need a documented starting point for selecting and loading a listed Russian NLP checkpoint, especially for fill-mask inference or a ruT5 text-to-text experiment.

Who it is for

  • Developers experimenting with Russian-language NLP inference.
  • Researchers or practitioners comparing listed masked-language and text-to-text model variants.
  • Colab users who want to explore the models through the linked notebook.

Documented uses

  • Fill a masked token in a Russian sentence using ruBERT or ruRoBERTa.
  • Explore Russian text-to-text generation or transformation with ruT5 checkpoints.
  • Inspect model attention using the documented BertViz example.
Before you choose

Requirements & limits

The repository is public and specifies the Apache-2.0 license. The README points to a Colab notebook and describes running examples locally.

  • The listed ruBERT and ruRoBERTa checkpoints are for mask filling; ruT5 checkpoints are for text-to-text generation.
  • The README provides an executable-style Transformers pipeline example for ruRoBERTa fill-mask, while its ruT5 section identifies the task and links checkpoints without showing a complete generation snippet.
  • The README reports parameter counts ranging from 178 million to 737 million across listed models, a relevant consideration when selecting a variant.
  • The README lists BertViz attention visualization for ruRoBERTa, in addition to inference examples.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up ruRoBERTa fill-mask inference

Load the documented ruRoBERTa-large checkpoint through a Hugging Face Transformers pipeline and pass it a sentence with a masked token.

Before you start

  • Python environment for the repository requirements
  • Install dependencies with pip install -r requirements.txt
  • Hugging Face Transformers

Steps

  1. Install the repository requirements.

  2. Create a Transformers pipeline with task fill-mask and model sberbank-ai/ruRoberta-large.

  3. Pass a sentence containing the ` token to the pipeline and set top_k` if you want to limit returned predictions.

Expected outcome

The pipeline returns a fill-mask prediction for the supplied sentence.

Guide limitations

  • The README's pipeline example demonstrates ruRoBERTa-large for fill-mask inference; it does not give performance or hardware requirements.
  • Model loading and inference are described, not independently tested here.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
18 October 2026
Original website title
ai-forever/model-zoo · Official homepage title · retained in its source language
How we review sourcesSuggest a correction