U M L S Bert_ E N G

Introduction

UMLSBERT_ENG is a model developed for biomedical applications. It focuses on feature extraction using Transformers, implemented in PyTorch, and supports the English language. The model is tagged for its relevance to BERT architecture and its applicability in the biomedical domain. It is designed for use in inference endpoints and is licensed under Apache 2.0.

Architecture

The model utilizes the BERT architecture, which is well-suited for tasks in natural language processing and understanding. It infuses knowledge into cross-lingual medical term embedding, facilitating term normalization within biomedical contexts.

Training

UMLSBERT_ENG, previously known as CODER, was developed with a focus on embedding medical terms across languages. The model employs techniques such as knowledge graph embedding and contrastive learning to enhance its performance in medical term normalization and representation.

Guide: Running Locally

To run UMLSBERT_ENG locally, follow these steps:

1. Clone the Repository:
   Clone the repository from GitHub using the link provided in the model card.

2. Set Up the Environment:
   Install the necessary libraries, primarily PyTorch and Transformers, to ensure compatibility.

3. Download Model Files:
   Acquire the model files from the repository or use the Hugging Face model hub.

4. Run Inference:
   Utilize the model for feature extraction tasks in biomedical text processing.

For optimal performance, consider using cloud GPUs such as those provided by AWS, Google Cloud, or Azure.

License

The UMLSBERT_ENG model is licensed under the Apache 2.0 License, permitting wide usage and distribution with compliance to the license terms.