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

Uie Pytorch

Uie Pytorch is a Python implementation that ports PaddleNLP’s Universal Information Extraction model to PyTorch and includes prediction, annotation conversion, fine-tuning, evaluation and ONNX export tools.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Uie Pytorch — 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 want a PyTorch-based UIE implementation for schema-driven prediction or a workflow that extends into data preparation, fine-tuning, evaluation and model export.

Who it is for

  • Developers building text-extraction workflows with PyTorch.
  • Teams adapting information extraction to labeled domain data.
  • Practitioners evaluating or exporting UIE inference models.

Documented uses

  • Extract entities, relations, events or review opinions from text using custom natural-language schemas.
  • Convert annotated data for training and fine-tune a model for a specific extraction task.
  • Evaluate a model and export an inference model to ONNX.
Before you choose

Requirements & limits

The repository identifies an Apache-2.0 license. Using FasterTokenizer with the ErnieM tokenizer requires installing the faster_tokenizer package.

  • The implementation is for PyTorch, while its source UIE model is from PaddleNLP.
  • The predictor documentation includes Chinese and English model examples.
  • The documentation says the ernie-3.0-base-zh conversion requires fine-tuning before prediction.
  • ONNX export is documented, but specific deployment runtimes are not named.

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 a PyTorch UIE model for extraction

Convert a supported UIE model to the repository’s PyTorch format and use the predictor for schema-based text extraction.

Before you start

  • The HUSTAI/uie_pytorch repository and its Python scripts.
  • A model name supported for automatic download, such as uie-base, or a local model directory.
  • For the default conversion validation, PaddleNLP; conversion can instead be run with validation disabled.

Steps

  1. Run the command below to download and convert the default uie-base model into uie_base_pytorch.

    Documented command
    python convert.py

  2. If PaddleNLP is not installed, run the command below; this skips importing PaddleNLP and validating the conversion result.

    Documented command
    python convert.py --no_validate_output

  3. Create a UIEPredictor with a model and extraction schema, then call it with the text to extract. The README example uses UIEPredictor(model='uie-base', schema=['时间', '选手', '赛事名称']).

  4. To change extraction targets on an existing predictor, call set_schema with the new schema before calling the predictor.

Expected outcome

A converted PyTorch model is available, and the predictor returns extracted information for the requested schema when called with text.

Guide limitations

  • The --no_validate_output option disables conversion-result validation.
  • The README lists supported automatic-download model names; uie-tiny is deprecated and replaced by uie-medium.
  • The README’s predictor example uses a model name and does not document a complete custom-model prediction setup here.
  • The ernie-3.0-base-zh model is not validated during conversion and requires fine-tuning before prediction.
Documentation 1 ↗
Source-based practical guide

Install the FasterTokenizer dependency for UIE-M tokenization

Install the optional tokenizer dependency identified for using the ErnieM tokenizer with UIE-M models.

Before you start

  • Python and the repository’s UIE-M/ErnieM tokenizer support.

Steps

  1. Install the FasterTokenizer library with the command below.

    Documented command
    pip install faster_tokenizer

  2. Use the ErnieM tokenizer for text preprocessing as part of UIE-M support.

Expected outcome

The FasterTokenizer dependency is installed for the ErnieM tokenizer’s text-preprocessing path.

Guide limitations

  • The README identifies this dependency for the ErnieM tokenizer but does not document further tokenizer configuration or a complete UIE-M inference workflow.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Uie PytorchExtract entities, relations, events or review opinions from text using custom natural-language schemas.CLI, Linux, macOS, Self-hostedThe repository identifies an Apache-2.0 license. Using FasterTokenizer with the ErnieM tokenizer requires installing the faster_tokenizer package.
Pulse Of MotionEstimate a video's physical frame rate from its visual dynamics.LinuxThe repository is public. Its README describes a released checkpoint that downloads from Hugging Face on first inference, with a manual download option. The strongest internal model is not fully released; the README directs interested users to contact the project about access. No license information is provided in the repository evidence.
Neural Image AssessmentTrain the model using AVA images and rating annotations.Cloud, Linux, macOS, Windows
Evidence and transparency

Sources & review

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