Indian Flag
Government Of India
A-
A
A+
Home/Models/NE-BERT
ownerBadal Nyalang
organizationMWire Labs
ORGANISATION

NE-BERT

NE-BERT is Northeast India's first domain-specific multilingual foundation model. Built on the ModernBERT architecture and trained on 8.3 million sentences, it supports 9 regional languages: Assamese, Khasi, Garo, Manipuri (Meitei), Mizo, Nyishi, Nagamese, Kokborok, and Pnar. It achieves State-of-the-Art performance on regional benchmarks and offers 1.6x faster inference, bridging the digital divide for low-resource languages.

About Model

NE-BERT is a state-of-the-art transformer model designed specifically for the complex, low-resource linguistic landscape of Northeast India. While generic multilingual models (like mBERT or IndicBERT) often underperform on Austroasiatic and Tibeto-Burman languages due to data scarcity, NE-BERT treats these languages as first-class citizens. Supported Languages: The model is trained on a curated corpus of ~8.3 million sentences covering: 1. Assamese [asm] (Indo-Aryan) 2. Manipuri/Meitei [mni] (Tibeto-Burman) 3. Khasi [kha] (Austroasiatic) 4. Mizo [lus] (Tibeto-Burman) 5. Garo [grt] (Tibeto-Burman) 6. Nyishi [njz] (Tibeto-Burman) 7. Nagamese Creole [nag] (Indo-Aryan) 8. Kokborok [trp] (Tibeto-Burman) 9. Pnar [pbv] (Austroasiatic) Technical Key Features: * Architecture: ModernBERT-Base (Pre-Norm, Rotary Embeddings, Flash Attention 2) * Context Window: Supports up to 8,192 tokens * Efficiency: 1.6x more efficient tokenization than mBERT * Performance: Low perplexity on Austroasiatic and Tibeto-Burman languages

NE-BERT

Metadata Metadata

Attribution 4.0 International (CC BY- 4.0)

MWirelabs

Transformers

PyTorch

Open

MWire Labs

Arts, Culture and Tourism

22/11/25 13:08:10

Badal Nyalang

0

Activity Overview Activity Overview

  • Downloads0
  • Redirect 25
  • File Size 0
  • Views 774

Tags Tags

  • Masked Language Modeling
  • low-resource-NLP
  • Assamese
  • Garo
  • Nyishi
  • Meitei
  • Nagamese
  • northeast-india
  • khasi
  • A'chik
  • modernbert
  • mwirelabs
  • northeast bert
  • token-efficiency
  • Mizo
  • kokborok
  • Pnar

License Control License Control

Attribution 4.0 International (CC BY- 4.0)

Related Models Related Models

NE-Embed
NE-Embed is a multilingual text embedding model for Northeast Indian languages, enabling semantic search, retrieval, and RAG across 10 languages including Khasi, Garo, Meitei, Bodo, Mizo, Assamese, Nyishi, Kokborok, Pnar, and Nagamese. Fine-tuned on LaBSE with 201,738 parallel pairs.
embeddings
Multilingual
rag
low-resource
retrieval
sentence-transformers
northeast-india
  • See Upvoters0
  • Downloads0
  • File Size0
  • Views63
Updated 3 month(s) ago
MWIRE LABS
View Details
share
Northeast Language Identification
NE-LID is a fast and accurate language identification model for Northeast Indian languages using character level features. It is designed for low resource and script diverse text and achieves high accuracy on short sentences.
Multilingual
language identification
low-resource
northeast-india
MWire Labs
fasttext
fastText
  • See Upvoters1
  • Downloads13
  • File Size0
  • Views1,080
Updated 8 month(s) ago
MWIRE LABS
View Details
share

Related Articles

Building Sovereign Language AI for Northeast India: The NE-Stack Story
Northeast India is home to over 220 languages spoken by 45 million people, yet remains mostly absent from modern AI systems. MWire Labs, a northeast India AI startup based in Shillong, Meghalaya, is building the NE-Stack, a foundational suite of northeast AI models covering speech recognition, machine translation, OCR, TTS, and vision-language understanding across 11+ indigenous languages. This article documents the technical and strategic approach behind northeast AI development for one of the world's most linguistically diverse and digitally underserved regions.
Building Sovereign Language AI for Northeast India: The NE-Stack Story
Badal NyalangBadal Nyalang
  • See Upvoters1
  • Read Time1 min read
View Article
share

More Models from MWire Labs More Models from MWire Labs

Northeast STT Multilingual Speech to Text Model
A multilingual Speech-to-Text (STT) model for eight Northeast Indian languages, fine-tuned from Whisper Medium using over 150,000 speech-text pairs from public and institutional datasets. The model expands speech recognition support for low-resource indigenous languages, including Khasi, Garo, Mizo, Kokborok, Nagamese, Assamese, Chakma, and Wancho.
low-resource
northeast-india
Northeast India Languages
whisper
Speech processing
Multilingual speech
Speech to Text
Multilingual
Automatic Speech Recognition
  • See Upvoters0
  • Downloads3
  • File Size0
  • Views112
Updated 2 month(s) ago
MWIRE LABS
View Details
share
NE-SpeechEmbed
NE-SpeechEmbed is a multilingual speech-text embedding model by MWire Labs for Northeast Indian languages. The model supports semantic speech search, cross-modal retrieval, and audio-text embeddings across Khasi, Garo, Mizo, Nagamese, Kokborok, Assamese, Wancho, and Chakma.
speech-text-retrieval
speech-embeddings
speech-language-model
mwire-labs
embeddings
whisper
low-resource-languages
northeast-india
retrieval
low-resource
XLM-RoBERTa
speech
multilingual
audio-text-retrieval
  • See Upvoters0
  • Downloads0
  • File Size0
  • Views56
Updated 3 month(s) ago
MWIRE LABS
View Details
share
NE-Embed
NE-Embed is a multilingual text embedding model for Northeast Indian languages, enabling semantic search, retrieval, and RAG across 10 languages including Khasi, Garo, Meitei, Bodo, Mizo, Assamese, Nyishi, Kokborok, Pnar, and Nagamese. Fine-tuned on LaBSE with 201,738 parallel pairs.
Multilingual
embeddings
northeast-india
sentence-transformers
retrieval
low-resource
rag
  • See Upvoters0
  • Downloads0
  • File Size0
  • Views63
Updated 3 month(s) ago
MWIRE LABS
View Details
share
Mizo OCR - Text Recognition for Mizo Language
OCR model for the Mizo language achieving 90.68% character accuracy on synthetic and curated printed text
Image-to-Text
trocr
Mizo
northeast-india
low-resource
OCR
  • See Upvoters0
  • Downloads4
  • File Size0
  • Views409
Updated 6 month(s) ago
MWIRE LABS
View Details
share
NE-OCR
NE-OCR is a multilingual Optical Character Recognition model developed by MWire Labs to accurately recognize printed text from documents in Northeast Indian languages. The model supports Assamese, Bodo, English, Garo, Hindi, Khasi, Kokborok, Meitei (Bengali script), Meitei (Meitei Mayek script), Mizo, Nagamese, and Nyishi. It is designed to enable reliable digitization of books, newspapers, government records, educational materials, and cultural archives from Northeast India where mainstream OCR
northeast-india
khasi
Mizo
Optical Character Recognition
OCR
doctr
vitstr
Multilingual OCR
Northeast India OCR
Printed Text Recognition
BODO
Garo
Kokborok
Nyishi
Meitei
Nagamese
  • See Upvoters0
  • Downloads9
  • File Size0
  • Views276
Updated 6 month(s) ago
MWIRE LABS
View Details
share
Nagamese Speech-to-Text
Automatic Speech Recognition (ASR) model for Nagamese speech, designed to transcribe spoken Nagamese into text for real-world usage.
Nagamese
Automatic Speech Recognition
ASR
Speech Recognition
whisper
low-resource-language
  • See Upvoters0
  • Downloads3
  • File Size0
  • Views257
Updated 6 month(s) ago
MWIRE LABS
View Details
share
Garo OCR - Text Recognition for Garo
OCR model for the Garo language achieving 93.13% character accuracy.
florence-2
northeast-india
Garo
OCR
Image-to-Text
  • See Upvoters0
  • Downloads0
  • File Size0
  • Views350
Updated 6 month(s) ago
MWIRE LABS
View Details
share
Northeast Language Identification
NE-LID is a fast and accurate language identification model for Northeast Indian languages using character level features. It is designed for low resource and script diverse text and achieves high accuracy on short sentences.
language identification
fastText
Multilingual
fasttext
MWire Labs
northeast-india
low-resource
  • See Upvoters1
  • Downloads13
  • File Size0
  • Views1,080
Updated 8 month(s) ago
MWIRE LABS
View Details
share
NortheastNER
NortheastNER is a token classification model built on XLM-RoBERTa and fine-tuned on ~25k sentences from gazetteers, news, and cultural texts across Northeast India. It detects region-specific entities, places, tribes, festivals, tourist sites, flora, fauna, and experimental local names; ideal for low-resource NER, regional search, cultural analytics, and knowledge graph applications.
Token Classification
Conservation
Northeast India
Meghalaya
northeast-india
low-resource
XLM-RoBERTa
NER
  • See Upvoters0
  • Downloads15
  • File Size0
  • Views432
Updated 10 month(s) ago
MWIRE LABS
View Details
share
Kren-M
Northeast India's first AI language model. Kren-M is a 2.6B parameter bilingual model for Khasi-English, built on Gemma-2-2B. Features Kren-NE custom tokenizer covering 7 NE languages (Khasi, Garo, Mizo, Assamese, Manipuri, Nagamese, Nyishi) with 35.7% efficiency gain. Trained on 5.43M Khasi sentences. Capabilities: bidirectional translation, natural conversation, cultural context. Designed for language preservation across Northeast India
northeast-india
Garo
low-resource
Instruction-Tuning
Indian Languages
Tokenizer
Foundational model
Northeast India Languages
Kren-M
bilingual
continued-pretraining
Northeast India
khasi
  • See Upvoters0
  • Downloads40
  • File Size0
  • Views1,151
Updated 10 month(s) ago
MWIRE LABS
View Details
share

© 2026 - Copyright AIKosh. All rights reserved.