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

Guwenbert

GuwenBERT is a RoBERTa-based pretrained language model trained on Classical Chinese texts for use as a foundation for downstream literary-Chinese NLP tasks.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Guwenbert — 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 GuwenBERT when a downstream NLP project needs pretrained representations suited to Classical Chinese and can supply the task-specific modeling and data workflow.

Who it is for

  • Researchers working with Classical Chinese and historical texts.
  • Developers building or fine-tuning Classical Chinese NLP systems.

Documented uses

  • Start a Classical Chinese named-entity-recognition model.
  • Build sentence segmentation or punctuation-restoration workflows for literary Chinese.
  • Fine-tune a pretrained model for another Classical Chinese NLP task.
Before you choose

Requirements & limits

The repository provides base and large model checkpoints through Hugging Face and identifies the project as Apache-2.0 licensed.

  • Two checkpoint sizes are documented: base (12 layers, 768 hidden dimensions, 12 attention heads) and large (24 layers, 1,024 hidden dimensions, 16 attention heads).
  • The model's training corpus converted traditional characters to simplified Chinese; assess whether that text treatment fits your material.
  • The repository offers PyTorch checkpoints; TensorFlow use involves a conversion step.
  • The reported named-entity-recognition evaluation is task- and dataset-specific, rather than a general performance guarantee.

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 GuwenBERT checkpoint with Transformers

Load a pretrained GuwenBERT model and its configured tokenizer as a starting point for Classical Chinese NLP work.

Before you start

  • A Python environment with Hugging Face Transformers installed.
  • Internet access to retrieve the checkpoint, or a locally downloaded checkpoint.
  • Choose either ethanyt/guwenbert-base or ethanyt/guwenbert-large.

Steps

  1. Import AutoTokenizer and AutoModel from transformers.

  2. Call AutoTokenizer.from_pretrained with the selected checkpoint name.

  3. Call AutoModel.from_pretrained with the same checkpoint name.

Expected outcome

A tokenizer and pretrained model are loaded; the repository configuration selects BertTokenizer and RobertaModel.

Guide limitations

  • The documented checkpoints are PyTorch versions; TensorFlow use requires conversion with a Transformers conversion script.
  • Loading the model is not itself a complete sentence segmentation, punctuation restoration or named-entity-recognition pipeline.
Documentation 1 ↗
Evidence and transparency

Sources & review

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