---
language:
- en
license: mit
tags:
- token-classification
- entity-recognition
- foundation-model
- feature-extraction
- RoBERTa
- generic
datasets:
- numind/NuNER
pipeline_tag: token-classification
inference: false
---

# SOTA Entity Recognition English Foundation Model by NuMind 🔥

This model provides embeddings for the Entity Recognition task in English. It is an improved version of the model from our [**paper**](https://arxiv.org/abs/2402.15343).

**Checkout other models by NuMind:**
* SOTA Zero-shot NER Model [NuNER Zero](https://huggingface.co/numind/NuNER_Zero)
* SOTA Multilingual Entity Recognition Foundation Model: [link](https://huggingface.co/numind/entity-recognition-multilingual-general-sota-v1)
* SOTA Sentiment Analysis Foundation Model: [English](https://huggingface.co/numind/generic-sentiment-v1), [Multilingual](https://huggingface.co/numind/generic-sentiment-multi-v1)

## About

[Roberta-base](https://huggingface.co/roberta-base) fine-tuned on the expanded version of [NuNER data](https://huggingface.co/datasets/numind/NuNER) using contrastive learning from [**NuNER: Entity Recognition Encoder Pre-training via LLM-Annotated Data**](https://arxiv.org/abs/2402.15343).

**Metrics:**

Read more about evaluation protocol & datasets in our [NuNER data](https://huggingface.co/datasets/numind/NuNER) using contrastive learning from [**paper**](https://arxiv.org/abs/2402.15343).

Here is the aggregated performance of the models over several datasets:

k=X means that as training data, we took only X examples for each class, trained the model, and evaluated it on the full test set.

| Model | k=1 | k=4 | k=16 | k=64 |
|----------|----------|----------|----------|----------|
|   RoBERTa-base  |  24.5   |   44.7 | 58.1 | 65.4
|   RoBERTa-base + NER-BERT pre-training | 32.3 | 50.9 | 61.9 | 67.6 |
|   NuNER v0.1  |   34.3 | 54.6 | 64.0 | 68.7 |
|   NuNER v1.0  |   39.4 | 59.6 | 67.8 | 71.5 |
|   **NuNER v2.0**  |   **43.6** | **61.0** | **68.2** | **72.0** |

NuNER v1.0 has similar performance to 7B LLMs (70 times bigger than NuNER v1.0) created specifically for the NER task. Thus NuNER v2.0 should be even better than the 7b LLM.

| Model | k=8~16| k=64~128 |
|----------|----------|----------|
| UniversalNER (7B) | 57.89 ± 4.34 | 71.02 ± 1.53 |
|   NuNER v1.0 (100M) | 58.75 ± 0.93 | 70.30 ± 0.35 |

## Usage

Embeddings can be used out of the box or fine-tuned on specific datasets. 

Get embeddings:


```python
import torch
import transformers


model = transformers.AutoModel.from_pretrained(
    'numind/NuNER-v2.0'
)
tokenizer = transformers.AutoTokenizer.from_pretrained(
    'numind/NuNER-v2.0'
)

text = [
    "NuMind is an AI company based in Paris and USA.",
    "See other models from us on https://huggingface.co/numind"
]
encoded_input = tokenizer(
    text,
    return_tensors='pt',
    padding=True,
    truncation=True
)
output = model(**encoded_input)

emb = output.last_hidden_state
```

## Citation
```
@misc{bogdanov2024nuner,
      title={NuNER: Entity Recognition Encoder Pre-training via LLM-Annotated Data}, 
      author={Sergei Bogdanov and Alexandre Constantin and Timothée Bernard and Benoit Crabbé and Etienne Bernard},
      year={2024},
      eprint={2402.15343},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
```