---
title: GneissWeb.Sci_classifier
canonical_url: "https://www.modelscope.cn/models/AI-ModelScope/GneissWeb.Sci_classifier"
md_url: "https://www.modelscope.cn/models/AI-ModelScope/GneissWeb.Sci_classifier.md"
repository: AI-ModelScope/GneissWeb.Sci_classifier
last_updated: 2025-02-22
license: apache-2.0
library_name:
  - pytorch
frameworks:
  - pytorch
language:
  - en
downloads: 726
stars: 0
---

# GneissWeb.Sci_classifier

> GneissWeb.Sci_classifier - AI-ModelScope 在 ModelScope 开源的模型。In order to be able to reproduce GneissWeb, we provide here GneissWeb.Sciclassifier - a science category fastText classifier. This fastText model is used as part of the ensemble filter in GneissWeb…

AI-ModelScope/GneissWeb.Sci_classifier 是 ModelScope 魔搭社区上的机器学习模型，采用 apache-2.0 许可。

- **Repository**: AI-ModelScope/GneissWeb.Sci_classifier
- **License**: apache-2.0
- **Downloads**: 726
- **Stars**: 0
- **Last updated**: 2025-02-22

Source: https://www.modelscope.cn/models/AI-ModelScope/GneissWeb.Sci_classifier

---

**Model Summary**

In order to be able to reproduce GneissWeb, we provide here GneissWeb.Sci_classifier - a science category fastText classifier. This fastText model is used as part of the ensemble filter in GneissWeb to detect documents with science content.

Please refer to the [GneissWeb](https://huggingface.co/datasets/ibm-granite/GneissWeb) for more details.

     **Developers**: IBM Research

     **Release Date**: Feb 21st, 2025

     **License**: Apache 2.0.


**Training Data**

The model is trained on 800k documents, labeled using the [WatsonNLP hierachical categorization](https://www.ibm.com/docs/en/watsonx/saas?topic=catalog-hierarchical-categorization). Please refer to [fastText text classification tutorial](https://fasttext.cc/docs/en/python-module.html) for details. 
Training data is selected as follows:

- *Positive documents*: 400k documents randomly sampled from the documents labeled with science category with a confidence score 0.95 and above.
- *Negative documents*: 400k documents randomly sampled from the documents labeled with any category other than science, education, medical, and technology categories with a confidence score of 0.95 and above.