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

Cino

CINO is an XLM-R-based pretrained language model project for natural-language processing involving Chinese and Chinese minority languages. It provides PyTorch checkpoints, a Transformers loading example, and the Wiki-Chinese-Minority classification dataset.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Cino — 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 CINO when an NLP project involves Chinese and one or more of its listed minority languages and needs an XLM-R-based pretrained checkpoint. For classification evaluation, the project’s WCM dataset is a relevant accompanying resource.

Who it is for

  • Researchers studying multilingual NLP for Chinese minority languages.
  • Developers building NLP workflows that can use an XLM-R-compatible pretrained model.
  • Teams evaluating language classification with the Wiki-Chinese-Minority dataset.

Documented uses

  • Load a CINO checkpoint and tokenizer for an NLP workflow.
  • Use a pretrained model component in multilingual research involving supported languages.
  • Evaluate language classification with the WCM or WCM-v2 dataset.
Before you choose

Requirements & limits

The public GitHub repository identifies an Apache-2.0 license and provides links to PyTorch checkpoints.

  • CINO-small-v2, CINO-base-v2 and CINO-large-v2 are listed at 148M, 190M and 442M parameters respectively; CINO-large is 585M parameters.
  • The v2 vocabulary is trimmed for the pretraining data and supports Chinese and the listed minority languages; CINO-large v1 additionally supports XLM-R languages.
  • The documented loading path uses XLMRobertaTokenizer and XLMRobertaModel from Transformers.
  • WCM and WCM-v2 provide classification data built from Wikipedia text and category labels.

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

Load a CINO checkpoint with Transformers

Load a listed CINO tokenizer and model using the repository’s Transformers example.

Before you start

  • Python environment with the Transformers library installed.
  • A listed model checkpoint available through its model name or in a local model directory.

Steps

  1. Choose a listed checkpoint, such as hfl/cino-base-v2, or prepare its model files in a local directory.

  2. Import XLMRobertaTokenizer and XLMRobertaModel from transformers.

  3. Call each class’s from_pretrained method with the model name or local directory path.

Expected outcome

A tokenizer and XLM-R-compatible model object are loaded for subsequent NLP work.

Guide limitations

  • The example covers loading only; it does not document task-specific fine-tuning, prediction or deployment steps.
  • Checkpoint variants differ in size and language coverage.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
iflytek/cino · Official homepage title · retained in its source language
How we review sourcesSuggest a correction