The HF Text Feature Extraction deep learning package seamlessly integrates pretrained Sentence Transformers models with ArcGIS, enabling the generation of text embeddings directly within the ArcGIS environment. Using this package, input text can be converted into numerical vector representations using a selected pretrained model from the Hugging Face Hub. These embeddings can be used for downstream tasks such as semantic search, clustering, and semantic similarity analysis. The package supports Hugging Face models listed under the sentence-transformers library section for text embedding tasks.
Before running a model, ensure compliance with the licensing terms provided on the corresponding Hugging Face model page. Only use trusted models, as untrusted model weights or code may pose security risks. Since model sizes vary, ensure that sufficient CPU or GPU memory is available for inference.
Model details
This model has the following characteristics:
- Input—The input layer must be a point or polygon feature class containing the field with the text to be embedded.
- Output—A feature class with a new field named Embedding containing the embedding vector blob generated from the input text.
- Compute—This workflow can be compute-intensive depending on the selected model and input size, and a GPU with a minimum CUDA compute capability of 6.0 is recommended.
- Architecture—Uses a Hugging Face pretrained model listed under the sentence-transformers library section, identified by its model ID.
Access and download the model
Download the HF Text Feature Extraction pretrained model from ArcGIS Living Atlas of the World.
- Browse to ArcGIS Living Atlas of the World.
- Sign in with your ArcGIS Online credentials.
- Search for HF Text Feature Extraction and open the item page from the search results.
- Click the Download button to download the model.
You can use the downloaded .dlpk file directly in ArcGIS Pro, or upload and use it in ArcGIS Enterprise.
Release notes
The following are the release notes:
| Date | Description |
|---|---|
| May 2026 | First release of the HF Text Feature Extraction model |