---
language:
- en
- es
- fr
- de
- it
- pt
- sv
- nl
- pl
- ru
- bn
- hi
- fa
- zh
- uk
license: mit
tags:
- ner
- named-entity-recognition
- token-classification
- universal-ner
- unified-ner
- mdeberta-v3
- mdeberta
- deberta
- lora
- multilingual
- fine-grained-ner
- cross-lingual
- information-extraction
- nlp
- sequence-labeling
- bio-tagging
- entity-extraction
- text-mining
- deep-learning
- transformer
- wikiann
- multinerd
- multiconer
- multiconer-2023
base_model: microsoft/mdeberta-v3-base
library_name: transformers
datasets:
- unimelb-nlp/wikiann
- MultiCoNER/multiconer_v2
- Babelscape/multinerd
pipeline_tag: token-classification
widget:
- text: "Barack Obama visited the United Nations headquarters in New York City."
  example_title: "English Example"
- text: "Albert Einstein entwickelte die Relativitätstheorie in Bern und Zürich."
  example_title: "German Example"
- text: "Emmanuel Macron est le président de la République française à Paris."
  example_title: "French Example"
- text: "El delantero Lionel Messi marcó tres goles en el estadio Camp Nou de Barcelona."
  example_title: "Spanish Example"
- text: "Юрий Гагарин совершил первый космический полёт на корабле Восток-1."
  example_title: "Russian Example"
- text: "जी-20 शिखर सम्मेलन का आयोजन नई दिल्ली के भारत मंडपम में किया गया।"
  example_title: "Hindi Example"
---

# Unified Multilingual NER (mDeBERTa-v3-base)

This model is a fine-grained **Multilingual Named Entity Recognition (NER)** model fine-tuned on **1.53+ million sentences** across three major benchmark suites:
* [**WikiANN**](https://huggingface.co/datasets/unimelb-nlp/wikiann)
* [**MultiCoNER 2023 (v2)**](https://huggingface.co/datasets/MultiCoNER/multiconer_v2)
* [**MultiNERD**](https://huggingface.co/datasets/Babelscape/multinerd)

Built on top of [`microsoft/mdeberta-v3-base`](https://huggingface.co/microsoft/mdeberta-v3-base), it supports **47 fine-grained entity categories** (95 BIO logit tags) across **12+ languages**.

Developed by [**Rishabh Kumar**](https://www.linkedin.com/in/rishabh-kumar-922072195/).

---

## 1. Model Description & Architecture

- **Base Model**: [`microsoft/mdeberta-v3-base`](https://huggingface.co/microsoft/mdeberta-v3-base) (278M parameters, 12 Transformer layers, 768 hidden size, 250k vocabulary)
- **Fine-Tuning Method**: LoRA (Low-Rank Adaptation, r=64, α=64, target modules: `query_proj`, `key_proj`, `value_proj`, `intermediate.dense`, `output.dense`) merged into standalone base weights.
- **Precision**: FP16 (`torch.float16`, ~530 MB footprint)
- **Task**: Multilingual Token Classification / Named Entity Recognition
- **Languages Supported (15)**: English (`en`), Spanish (`es`), French (`fr`), German (`de`), Italian (`it`), Portuguese (`pt`), Swedish (`sv`), Dutch (`nl`), Polish (`pl`), Russian (`ru`), Farsi (`fa`), Bengali (`bn`), Hindi (`hi`), Ukrainian (`uk`), Chinese (`zh`), and multilingual mixed text.

---

## 2. Dataset Links & Entity Taxonomy

The model combines the taxonomy of **three major Hugging Face NER datasets** into a single unified classification head (47 × 2 + 1 = **95 BIO output logits**):

### A. [WikiANN](https://huggingface.co/datasets/unimelb-nlp/wikiann) (3 Core Broad Categories)
* **Languages Included (7)**: German (`de`), English (`en`), Spanish (`es`), French (`fr`), Italian (`it`), Portuguese (`pt`), Swedish (`sv`)
* **Entity Classes**: `PER` (Person), `LOC` (Location), `ORG` (Organization)

### B. [MultiNERD](https://huggingface.co/datasets/Babelscape/multinerd) (15 Multi-Domain Categories)
* **Languages Included (10)**: German (`de`), English (`en`), Spanish (`es`), French (`fr`), Italian (`it`), Dutch (`nl`), Polish (`pl`), Portuguese (`pt`), Russian (`ru`), Chinese (`zh`)
* **Standard**: `PER`, `LOC`, `ORG`, `TIME`
* **Specialized Domains**:
  * `ANIM` (Animal species, breeds, fauna)
  * `PLANT` (Plants, trees, botanical species, fungi)
  * `FOOD` (Food, dishes, beverages, culinary items)
  * `DIS` (Diseases, disorders, medical conditions)
  * `CEL` (Celestial bodies, stars, planets, galaxies)
  * `EVE` (Historical, sports, cultural events)
  * `INST` (Public institutions, international bodies)
  * `MEDIA` (Newspapers, networks, press outlets)
  * `MYTH` (Mythological entities, deities, folklore)
  * `VEHI` (Vehicles, ships, aircraft, spacecraft)
  * `BIO` (Biological compounds, genes, proteins)

### C. [MultiCoNER 2023 (v2)](https://huggingface.co/datasets/MultiCoNER/multiconer_v2) (33 Fine-Grained Categories)
* **Languages Included (13)**: Bangla (`BN`), German (`DE`), English (`EN`), Spanish (`ES`), Farsi (`FA`), French (`FR`), Hindi (`HI`), Italian (`IT`), Portuguese (`PT`), Swedish (`SV`), Ukrainian (`UK`), Chinese (`ZH`), and Multilingual Mixed (`MULTI`)
* **Person Subtypes**: `Artist`, `Athlete`, `Cleric`, `Politician`, `Scientist`, `SportsManager`, `OtherPER`
* **Organization Subtypes**: `AerospaceManufacturer`, `CarManufacturer`, `MusicalGRP`, `PrivateCorp`, `PublicCorp`, `SportsGRP`
* **Location Subtypes**: `Facility`, `HumanSettlement`, `Station`, `OtherLOC`
* **Products & Creative Works**: `ArtWork`, `MusicalWork`, `VisualWork`, `WrittenWork`, `Software`, `Clothing`, `Drink`, `Food`, `OtherPROD`
* **Medical & Healthcare**: `AnatomicalStructure`, `Disease`, `MedicalProcedure`, `Medication/Vaccine`, `Symptom`

### D. Complete Unified 47 Entity Category List
```text
 1. ANIM                   13. Disease               25. Medication/Vaccine    37. Scientist
 2. AerospaceManufacturer  14. Drink                 26. MusicalGRP            38. Software
 3. AnatomicalStructure    15. EVE                   27. MusicalWork           39. SportsGRP
 4. ArtWork                16. FOOD                  28. ORG                   40. SportsManager
 5. Artist                 17. Facility              29. OtherLOC              41. Station
 6. Athlete                18. Food                  30. OtherPER              42. Symptom
 7. BIO                    19. HumanSettlement       31. OtherPROD             43. TIME
 8. CEL                    20. INST                  32. PER                   44. VEHI
 9. CarManufacturer        21. Cleric                33. PLANT                 45. Vehicle
10. Cleric                 22. MEDIA                 34. Politician            46. VisualWork
11. Clothing               23. MYTH                  35. PrivateCorp           47. WrittenWork
12. DIS                    24. MedicalProcedure      36. PublicCorp
```

---

## 3. Training Results & Metrics

The model was fine-tuned for **5 epochs** (~28.3 hours across 1.53M training sentences) using AdamW (`lr=2e-5`, linear scheduler, effective batch size 16).

### Final Validation Evaluation (Epoch 5 / 5)

| Metric | Score |
| :--- | :---: |
| **Validation Entity F1** | **0.7693** |
| **Validation Precision** | **0.8449** |
| **Validation Recall** | **0.7061** |
| **Validation Accuracy** | **0.9587** |
| **Validation Loss** | **0.0276** |

### Per-Epoch Progression
* **Epoch 1**: F1 = 0.7183 | Precision = 0.8016 | Recall = 0.6506 | Loss = 0.0372
* **Epoch 2**: F1 = 0.7322 | Precision = 0.8151 | Recall = 0.6645 | Loss = 0.0317
* **Epoch 3**: F1 = 0.7544 | Precision = 0.8346 | Recall = 0.6883 | Loss = 0.0293
* **Epoch 4**: F1 = 0.7669 | Precision = 0.8460 | Recall = 0.7013 | Loss = 0.0281
* **Epoch 5**: F1 = 0.7693 | Precision = 0.8449 | Recall = 0.7061 | Loss = 0.0276

---

### Overall Language Benchmarks (15 Languages, Sample-Weighted Aggregate)

To evaluate cross-lingual generalization across diverse datasets, the table below aggregates performance per language weighted by sentence support across all test splits:

| Language | Total Test Sentences | Precision | Recall | **Entity-Span F1** | Token Accuracy | Source Benchmarks Included |
| :--- | :---: | :---: | :---: | :---: | :---: | :--- |
| 🇷🇺 **Russian (RU)** | 8,338 | 82.11% | 71.70% | **`76.56%`** | 96.31% | MultiNERD |
| 🇳🇱 **Dutch (NL)** | 17,181 | 83.13% | 66.41% | **`73.84%`** | 96.47% | MultiNERD |
| 🇵🇱 **Polish (PL)** | 19,555 | 80.46% | 63.69% | **`71.10%`** | 95.10% | MultiNERD |
| 🇩🇪 **German (DE)** | 45,907 | 70.64% | 67.17% | **`68.72%`** | 93.13% | WikiANN + MultiNERD + MultiCoNER 2023 |
| 🇮🇹 **Italian (IT)** | 276,098 | 66.20% | 65.93% | **`66.00%`** | 92.20% | WikiANN + MultiNERD + MultiCoNER 2023 |
| 🇸🇪 **Swedish (SV)** | 241,190 | 64.97% | 63.46% | **`64.21%`** | 91.68% | WikiANN + MultiCoNER 2023 |
| 🇫🇷 **French (FR)** | 277,481 | 61.47% | 60.23% | **`60.83%`** | 91.01% | WikiANN + MultiNERD + MultiCoNER 2023 |
| 🇺🇦 **Ukrainian (UK)** | 238,296 | 61.23% | 59.94% | **`60.58%`** | 92.32% | MultiCoNER 2023 |
| 🇪🇸 **Spanish (ES)** | 274,299 | 60.23% | 60.98% | **`60.55%`** | 91.86% | WikiANN + MultiNERD + MultiCoNER 2023 |
| 🇵🇹 **Portuguese (PT)** | 257,305 | 60.23% | 59.93% | **`60.02%`** | 91.59% | WikiANN + MultiNERD + MultiCoNER 2023 |
| 🇬🇧 **English (EN)** | 276,434 | 58.13% | 59.35% | **`58.72%`** | 90.70% | WikiANN + MultiNERD + MultiCoNER 2023 |
| 🇮🇳 **Hindi (HI)** | 18,399 | 53.08% | 55.75% | **`54.38%`** | 93.12% | MultiCoNER 2023 |
| 🇧🇩 **Bangla (BN)** | 19,859 | 50.29% | 52.75% | **`51.49%`** | 90.79% | MultiCoNER 2023 |
| 🇮🇷 **Farsi (FA)** | 219,168 | 52.86% | 46.06% | **`49.23%`** | 90.72% | MultiCoNER 2023 |
| 🇨🇳 **Chinese (ZH)** | 39,842 | 36.69% | 36.51% | **`36.11%`** | 84.05% | MultiNERD + MultiCoNER 2023 |

> **Note on Evaluated Splits & Benchmarks:**
> * **3-Benchmark Languages (DE, EN, ES, FR, IT, PT)** appear in all three datasets (WikiANN + MultiNERD + MultiCoNER 2023). Their scores represent joint generalizability across coarse Wikipedia text, domain-specific corpora, and noisy query contexts.
> * **2-Benchmark Languages (SV, ZH)** appear across two datasets (SV: WikiANN + MultiCoNER; ZH: MultiNERD + MultiCoNER).
> * **1-Benchmark Languages (RU, NL, PL, UK, HI, BN, FA)** are evaluated on their respective source benchmark splits (RU, NL, PL from MultiNERD; UK, HI, BN, FA from MultiCoNER 2023).

---

### Dataset-Specific Test Benchmarks (32-Suite)
Evaluation across **WikiANN**, **MultiNERD** (10 languages + full), and **MultiCoNER 2023** (13 languages + full):

| Dataset | Subset / Lang | Sentences | Precision | Recall | **F1 Score** | Accuracy |
| :--- | :--- | :---: | :---: | :---: | :---: | :---: |
| **`combined`** | **`combined`** | **536,661** | **0.6639** | **0.6298** | **`0.6464`** | **0.9313** |
| **`multinerd`** | 🔹 **`multinerd_full`** | **167,993** | **0.8906** | **0.7298** | **`0.8022`** | **0.9646** |
| `multinerd` | 🇬🇧 `EN-English` | 16,454 | 0.9097 | 0.8103 | **`0.8571`** | 0.9771 |
| `multinerd` | 🇩🇪 `DE-German` | 15,762 | 0.9025 | 0.7874 | **`0.8410`** | 0.9761 |
| `multinerd` | 🇷🇺 `RU-Russian` | 8,338 | 0.8211 | 0.7170 | **`0.7656`** | 0.9631 |
| `multinerd` | 🇳🇱 `NL-Dutch` | 17,181 | 0.8313 | 0.6641 | **`0.7384`** | 0.9647 |
| `multinerd` | 🇫🇷 `FR-French` | 17,695 | 0.7738 | 0.6725 | **`0.7196`** | 0.9668 |
| `multinerd` | 🇵🇱 `PL-Polish` | 19,555 | 0.8046 | 0.6369 | **`0.7110`** | 0.9510 |
| `multinerd` | 🇪🇸 `ES-Spanish` | 17,399 | 0.7893 | 0.6374 | **`0.7053`** | 0.9653 |
| `multinerd` | 🇵🇹 `PT-Portuguese` | 17,815 | 0.7471 | 0.6083 | **`0.6706`** | 0.9585 |
| `multinerd` | 🇮🇹 `IT-Italian` | 18,217 | 0.7424 | 0.5844 | **`0.6540`** | 0.9619 |
| `multinerd` | 🇨🇳 `ZH-Chinese` | 19,577 | 0.3193 | 0.2353 | **`0.2709`** | 0.8184 |
| **`wikiann`** | 🔹 **`wikiann_full`** | **70,000** | **0.6214** | **0.6199** | **`0.6202`** | **0.8099** |
| `wikiann` | `sv` (Swedish) | 10,000 | 0.7174 | 0.6662 | **`0.6909`** | 0.8843 |
| `wikiann` | `de` (German) | 10,000 | 0.6874 | 0.6352 | **`0.6603`** | 0.8938 |
| `wikiann` | `it` (Italian) | 10,000 | 0.6506 | 0.6459 | **`0.6482`** | 0.8362 |
| `wikiann` | `es` (Spanish) | 10,000 | 0.6076 | 0.6560 | **`0.6309`** | 0.7726 |
| `wikiann` | `fr` (French) | 10,000 | 0.6017 | 0.6080 | **`0.6049`** | 0.7581 |
| `wikiann` | `pt` (Portuguese) | 10,000 | 0.5613 | 0.5976 | **`0.5789`** | 0.7517 |
| `wikiann` | `en` (English) | 10,000 | 0.5238 | 0.5304 | **`0.5271`** | 0.7729 |
| **`multiconer2023`** | 🔹 **`multiconer2023_full`** | **1,991,359** | **0.5957** | **0.5923** | **`0.5937`** | **0.9160** |
| `multiconer2023` | `IT-Italian` | 247,881 | 0.6566 | 0.6654 | **`0.6609`** | 0.9225 |
| `multiconer2023` | `SV-Swedish` | 231,190 | 0.6468 | 0.6332 | **`0.6400`** | 0.9182 |
| `multiconer2023` | `UK-Ukrainian` | 238,296 | 0.6123 | 0.5994 | **`0.6058`** | 0.9232 |
| `multiconer2023` | `FR-French` | 249,786 | 0.6039 | 0.5971 | **`0.6005`** | 0.9122 |
| `multiconer2023` | `ES-Spanish` | 246,900 | 0.5889 | 0.6060 | **`0.5974`** | 0.9212 |
| `multiconer2023` | `PT-Portuguese` | 229,490 | 0.5928 | 0.5987 | **`0.5957`** | 0.9197 |
| `multiconer2023` | `MULTI-Multilingual` | 358,668 | 0.5787 | 0.5840 | **`0.5813`** | 0.9118 |
| `multiconer2023` | `DE-German` | 20,145 | 0.5625 | 0.5993 | **`0.5803`** | 0.9148 |
| `multiconer2023` | `EN-English` | 249,980 | 0.5620 | 0.5818 | **`0.5718`** | 0.9078 |
| `multiconer2023` | `HI-Hindi` | 18,399 | 0.5308 | 0.5575 | **`0.5438`** | 0.9312 |
| `multiconer2023` | `BN-Bangla` | 19,859 | 0.5029 | 0.5275 | **`0.5149`** | 0.9079 |
| `multiconer2023` | `FA-Farsi` | 219,168 | 0.5286 | 0.4606 | **`0.4923`** | 0.9072 |
| `multiconer2023` | `ZH-Chinese` | 20,265 | 0.4129 | 0.4904 | **`0.4483`** | 0.8619 |

---

## 4. Inference Mechanism & Softmax Logit Decoding (95 Logit Classes)

During inference, for any input text, our fine-tuned model (`microsoft/mdeberta-v3-base` with merged LoRA parameters) passes subword token representations through its 12 Transformer layers to output a **95-dimensional logit vector** per token. PyTorch applies a **softmax function** across all 95 output tags to compute normalized probability confidence scores:

```
P(tag_k | token_i) = exp(z_k) / Σ exp(z_j)   for k ∈ {1, 2, ..., 95}
```

- **Hard Argmax (Default Token Classification)**: Selects the single highest-probability tag (argmax).
- **Soft Confidence Distribution**: For ambiguous entity tokens (e.g. *"Harvard"*), the model distributes probability across related candidate classes (`B-INST`: **72.4%**, `B-ORG`: **21.1%**, `B-Facility`: **5.3%**), capturing semantic cross-taxonomy relationships.

---

## 5. Intended Use & Out-of-Scope Use

### Intended Use
- **Information extraction** from multilingual text (news articles, Wikipedia, social media, web content)
- **Downstream NLP pipelines** requiring structured entity extraction across multiple languages
- **Research** in cross-lingual transfer learning, fine-grained NER, and multilingual NLP
- **Knowledge graph construction** from unstructured multilingual text

### Out-of-Scope Use
- **Domain-specific NER** (e.g., legal, financial, or clinical NER) — may require additional fine-tuning on domain data
- **Languages not listed** in the supported 15-language set — the model was not trained on those languages

---

## 6. How to Use

### Quick Inference via Hugging Face Pipeline

```python
from transformers import pipeline

ner_pipeline = pipeline(
    "token-classification",
    model="Rishabh157/unified-multilingual-ner-mdeberta",
    aggregation_strategy="simple"
)

text = "Barack Obama visited the United Nations headquarters in New York City."
entities = ner_pipeline(text)

for entity in entities:
    print(f"{entity['word']:<25} | {entity['entity_group']:<15} | Score: {entity['score']:.4f}")
```

### PyTorch & Transformers Direct Usage

```python
import torch
from transformers import AutoTokenizer, AutoModelForTokenClassification

model_name = "Rishabh157/unified-multilingual-ner-mdeberta"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForTokenClassification.from_pretrained(model_name)

text = "Emmanuel Macron est le président de la République française à Paris."
inputs = tokenizer(text, return_tensors="pt")

with torch.no_grad():
    outputs = model(**inputs)
    predictions = torch.argmax(outputs.logits, dim=2)

tokens = tokenizer.convert_ids_to_tokens(inputs["input_ids"][0])
labels = [model.config.id2label[p.item()] for p in predictions[0]]

for token, label in zip(tokens, labels):
    if label != "O":
        print(f"{token:<20} -> {label}")
```

---

## 7. Known Considerations & Future Improvements

- **European Languages Lead Performance**: The model achieves highest F1 scores on European languages (English, German, Dutch, Russian — F1 `0.76`–`0.86` on MultiNERD) due to richer training data availability.
- **East Asian & South Asian Languages**: Chinese (`ZH`), Farsi (`FA`), Bangla (`BN`), and Hindi (`HI`) show room for improvement and can benefit from additional fine-tuning on language-specific corpora.
- **Unified Taxonomy Design**: The model unifies 47 entity types from three separate datasets into a single classification head. Some categories across datasets are semantically related (e.g., `Food` from MultiCoNER and `FOOD` from MultiNERD) and are treated as distinct classes to preserve dataset fidelity.
- **Continual Improvement**: Performance on lower-resource languages can be improved by fine-tuning further on domain-specific or language-specific NER datasets.

---

## 8. Citation

If you use this model, dataset pipeline, or findings in your research, please cite:

```bibtex
@misc{kumar2026multilingualner,
  author = {Rishabh Kumar},
  title = {Unified Fine-Grained Multilingual NER with mDeBERTa-v3},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/Rishabh157/unified-multilingual-ner-mdeberta}}
}
```