Skip to content

Minnesota Case Law ModernBERT Embedder

This document provides information about the Minnesota Case Law ModernBERT embedder, which is based on the conceptofmind/teraflop-minn-caselaw model from Hugging Face.

Overview

The Minnesota Case Law ModernBERT embedder is a sentence transformer model that generates 768-dimensional embeddings for text. It is based on the conceptofmind/teraflop-minn-caselaw model, which has been fine-tuned on Minnesota case law data, making it particularly suitable for legal document analysis.

Features

  • 768-dimensional embeddings
  • Optimized for legal document analysis
  • Support for long documents
  • Supports PDF, TXT, DOCX, and HTML file types

Installation

The Minnesota Case Law ModernBERT embedder requires the sentence-transformers library:

pip install sentence-transformers>=2.2.2

Usage

Using the Embedder in Code

from opencontractserver.pipeline.embedders.minn_modern_bert_embedder import MinnModernBERTEmbedder

# Create an instance of the embedder
embedder = MinnModernBERTEmbedder()

# Generate embeddings for a text
text = "This is a sample legal text to embed."
embedding = embedder.embed_text(text)

# The embedding is a list of 768 floating-point values
print(f"Embedding dimension: {len(embedding)}")

Docker Setup

We provide a Docker setup to run the Minnesota Case Law ModernBERT embedder as a service:

  1. Build and start the service:
docker-compose -f docker-compose.minn_modernbert.yml up -d

  1. The model will be downloaded and cached in a Docker volume.

  2. The service includes a healthcheck to ensure the model is loaded correctly.

Configuration

The Minnesota Case Law ModernBERT embedder is configured in config/settings/base.py:

# Minnesota Case Law ModernBERT embedder settings
MINN_MODERNBERT_EMBEDDERS = {
    "application/pdf": "opencontractserver.pipeline.embedders.minn_modern_bert_embedder.MinnModernBERTEmbedder768",
    "text/plain": "opencontractserver.pipeline.embedders.minn_modern_bert_embedder.MinnModernBERTEmbedder768",
    "text/html": "opencontractserver.pipeline.embedders.minn_modern_bert_embedder.MinnModernBERTEmbedder768",
    "application/vnd.openxmlformats-officedocument.wordprocessingml.document": "opencontractserver.pipeline.embedders.minn_modern_bert_embedder.MinnModernBERTEmbedder768",
}

Model Details

  • Base Model: conceptofmind/teraflop-minn-caselaw
  • Output Dimensionality: 768 dimensions
  • Similarity Function: Cosine Similarity
  • Training Dataset: Minnesota case law

Performance Considerations

  • The first time the model is used, it will be downloaded from Hugging Face, which may take some time.
  • For faster startup, use the Docker setup which preloads the model.
  • Using GPU acceleration is recommended for processing large volumes of text.