indic-sentence-bert-nli

IndicSBERT

This is a MuRIL model (google/muril-base-cased) trained on the NLI dataset of ten major Indian Languages. <br> The single model works for English, Hindi, Marathi, Kannada, Tamil, Telugu, Gujarati, Oriya, Punjabi, Malayalam, and Bengali. The model also has cross-lingual capabilities. <br> Released as a part of project MahaNLP: https://github.com/l3cube-pune/MarathiNLP <br>

A better sentence similarity model (fine-tuned version of this model) is shared here: https://huggingface.co/l3cube-pune/indic-sentence-similarity-sbert <br>

More details on the dataset, models, and baseline results can be found in our [paper] (https://arxiv.org/abs/2304.11434)

@article{deode2023l3cube,
  title={L3Cube-IndicSBERT: A simple approach for learning cross-lingual sentence representations using multilingual BERT},
  author={Deode, Samruddhi and Gadre, Janhavi and Kajale, Aditi and Joshi, Ananya and Joshi, Raviraj},
  journal={arXiv preprint arXiv:2304.11434},
  year={2023}
}

<a href='https://arxiv.org/abs/2211.11187'> monolingual Indic SBERT paper </a> <br> <a href='https://arxiv.org/abs/2304.11434'> multilingual Indic SBERT paper </a>

Other Monolingual Indic sentence BERT models are listed below: <br> <a href='https://huggingface.co/l3cube-pune/marathi-sentence-bert-nli'> Marathi SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/hindi-sentence-bert-nli'> Hindi SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/kannada-sentence-bert-nli'> Kannada SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/telugu-sentence-bert-nli'> Telugu SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/malayalam-sentence-bert-nli'> Malayalam SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/tamil-sentence-bert-nli'> Tamil SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/gujarati-sentence-bert-nli'> Gujarati SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/odia-sentence-bert-nli'> Oriya SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/bengali-sentence-bert-nli'> Bengali SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/punjabi-sentence-bert-nli'> Punjabi SBERT</a> <br> <a href='https://huggingface.co/l3cube-pune/indic-sentence-bert-nli'> Indic SBERT (multilingual)</a> <br>

Other Monolingual similarity models are listed below: <br> <a href='https://huggingface.co/l3cube-pune/marathi-sentence-similarity-sbert'> Marathi Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/hindi-sentence-similarity-sbert'> Hindi Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/kannada-sentence-similarity-sbert'> Kannada Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/telugu-sentence-similarity-sbert'> Telugu Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/malayalam-sentence-similarity-sbert'> Malayalam Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/tamil-sentence-similarity-sbert'> Tamil Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/gujarati-sentence-similarity-sbert'> Gujarati Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/odia-sentence-similarity-sbert'> Oriya Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/bengali-sentence-similarity-sbert'> Bengali Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/punjabi-sentence-similarity-sbert'> Punjabi Similarity </a> <br> <a href='https://huggingface.co/l3cube-pune/indic-sentence-similarity-sbert'> Indic Similarity (multilingual)</a> <br>

Usage (Sentence-Transformers)

Using this model becomes easy when you have sentence-transformers installed:

pip install -U sentence-transformers

Then you can use the model like this:

from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]

model = SentenceTransformer('{MODEL_NAME}')
embeddings = model.encode(sentences)
print(embeddings)

Usage (HuggingFace Transformers)

Without sentence-transformers, you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.

from transformers import AutoTokenizer, AutoModel
import torch


#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0] #First element of model_output contains all token embeddings
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)


# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('{MODEL_NAME}')
model = AutoModel.from_pretrained('{MODEL_NAME}')

# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')

# Compute token embeddings
with torch.no_grad():
    model_output = model(**encoded_input)

# Perform pooling. In this case, mean pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])

print("Sentence embeddings:")
print(sentence_embeddings)