README.md

---
language:
- en
license: apache-2.0
tags:
- insurance
- document-classification
- modernbert
- uk-insurance
- text-classification
- bytical
library_name: transformers
pipeline_tag: text-classification
base_model: answerdotai/ModernBERT-base
datasets:
- piyushptiwari/insureos-training-data
model-index:
- name: InsureDocClassifier
  results:
  - task:
      type: text-classification
      name: Insurance Document Classification
    metrics:
    - type: f1
      value: 1.0
      name: F1 (macro)
    - type: accuracy
      value: 1.0
      name: Accuracy
---

# InsureDocClassifier — Insurance Document Classification

**Created by [Bytical AI](https://bytical.ai)** — AI agents that run insurance operations.

## Model Description

InsureDocClassifier is a 12-class insurance document classifier built on ModernBERT-base. It automatically categorizes insurance documents into their correct type, enabling automated document routing, indexing, and processing in insurance operations.

## New to These Hugging Face Repos?

INSUREOS is published as separate repos so each page stays focused:

- This page (`InsureDocClassifier`) contains model artifacts and usage.
- Full training code is in [piyushptiwari/insureos-models](https://huggingface.co/piyushptiwari/insureos-models).
- Training data is in [piyushptiwari/insureos-training-data](https://huggingface.co/datasets/piyushptiwari/insureos-training-data).
- Beginner explanation is in [LEARN.md](https://huggingface.co/piyushptiwari/insureos-models/blob/main/LEARN.md).

### Document Classes (12)

| ID | Document Type | Description |
|----|--------------|-------------|
| 0 | Policy Schedule | Policy details and coverage summary |
| 1 | Certificate of Insurance | Proof of insurance document |
| 2 | Claim Form | Insurance claim submission form |
| 3 | Loss Adjuster Report | Assessment report from loss adjuster |
| 4 | Bordereaux — Premium | Premium transaction records |
| 5 | Bordereaux — Claims | Claims transaction records |
| 6 | Endorsement | Policy amendment document |
| 7 | Renewal Notice | Policy renewal notification |
| 8 | Statement of Fact | Declaration of material facts |
| 9 | FNOL Report | First Notification of Loss report |
| 10 | Subrogation Notice | Recovery rights notification |
| 11 | Policy Wording | Full policy terms and conditions |

### Training Details

| Parameter | Value |
|-----------|-------|
| Base Model | answerdotai/ModernBERT-base |
| Training Samples | 10,000 synthetic insurance documents |
| Epochs | 5 |
| Eval Loss | 4.17e-06 |
| GPU | NVIDIA Tesla T4 16GB |

### Evaluation Results

| Metric | Score |
|--------|-------|
| **Accuracy** | **1.0** |
| **F1 (macro)** | **1.0** |
| **F1 (weighted)** | **1.0** |
| Eval Samples/sec | 32.96 |

## How to Use

```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer

model = AutoModelForSequenceClassification.from_pretrained("piyushptiwari/InsureDocClassifier")
tokenizer = AutoTokenizer.from_pretrained("piyushptiwari/InsureDocClassifier")

text = "We hereby confirm that the above-named insured holds a valid policy of insurance..."
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512)
outputs = model(**inputs)
predicted_class = outputs.logits.argmax(-1).item()

labels = {
    0: "Policy Schedule", 1: "Certificate of Insurance", 2: "Claim Form",
    3: "Loss Adjuster Report", 4: "Bordereaux — Premium", 5: "Bordereaux — Claims",
    6: "Endorsement", 7: "Renewal Notice", 8: "Statement of Fact",
    9: "FNOL Report", 10: "Subrogation Notice", 11: "Policy Wording"
}
print(f"Document type: {labels[predicted_class]}")
```

## Learn How This Model Works

New to ML or insurance AI? We wrote a plain-English guide that explains the concepts,
the data, and the **actual training code** behind every INSUREOS model:

- **Learning guide:** [LEARN.md](https://huggingface.co/piyushptiwari/insureos-models/blob/main/LEARN.md)
- **Training code:** [piyushptiwari/insureos-models](https://huggingface.co/piyushptiwari/insureos-models) — this model is trained by [`training/doc_classifier.py`](https://huggingface.co/piyushptiwari/insureos-models/blob/main/training/doc_classifier.py)

## Part of the INSUREOS Model Suite

This model is part of the **INSUREOS** — a complete AI/ML suite for insurance operations built by Bytical AI:

| Model | Task | Metric |
|-------|------|--------|
| [InsureLLM-4B](https://huggingface.co/piyushptiwari/InsureLLM-4B) | Insurance domain LLM | ROUGE-1: 0.384 |
| **InsureDocClassifier** (this model) | 12-class document classification | F1: 1.0 |
| [InsureNER](https://huggingface.co/piyushptiwari/InsureNER) | 13-entity Named Entity Recognition | F1: 1.0 |
| [InsureFraudNet](https://huggingface.co/piyushptiwari/InsureFraudNet) | Fraud detection (Motor/Property/Liability) | AUC-ROC: 1.0 |
| [InsurePricing](https://huggingface.co/piyushptiwari/InsurePricing) | Insurance pricing (GLM + EBM) | MAE: £11,132 |

## Citation

```bibtex
@misc{bytical2026insuredocclassifier,
  title={InsureDocClassifier: Insurance Document Classification with ModernBERT},
  author={Bytical AI},
  year={2026},
  url={https://huggingface.co/piyushptiwari/InsureDocClassifier}
}
```

## About Bytical AI

[Bytical](https://bytical.ai) builds AI agents that run insurance operations — claims automation, underwriting intelligence, digital sales, and core system modernization for insurers across the UK and Europe. Microsoft AI Partner | NVIDIA | Salesforce.