Persian Embeddings

🚀 heydariAI/persian-embeddings

This model is a fine - tuned version of xlm - roberta - base, trained on Persian data to generate high - quality contextual embeddings for Persian text. It excels in Persian semantic search, clustering, and similarity tasks and also supports English - Persian multilingual tasks.

My Github: @heydaari

My Linkedin: Mohammad Hassan Heydari

This model is a fine - tuned version of xlm - roberta - base. It has been specifically trained on a large corpus of Persian data to generate high - quality contextual embeddings for Persian sentences and paragraphs. It is well - suited for tasks such as semantic search, clustering, and measuring contextual similarity for Persian text. Moreover, it also supports multilingual tasks involving English and Persian.

The fine - tuning process aimed to adapt the pre - trained multilingual XLM - RoBERTa model to better capture the linguistic nuances of Persian, making it highly effective for tasks that require embeddings tailored to the Persian language.

Note that persian - embeddings is a clone of persian - sentence - transformers - news - wiki - pairs - v4.

🚀 Quick Start

✨ Features

• Fine - tuned from xlm - roberta - base on a large Persian corpus.
• Generates high - quality contextual embeddings for Persian text.
• Supports multilingual tasks in English and Persian.
• Effective for semantic search, clustering, and contextual similarity tasks.

📦 Installation

To use this model, you need to install sentence - transformers if you want to use the simpler approach:

pip install -U sentence-transformers

💻 Usage Examples

Basic Usage (Sentence - Transformers)

If you have sentence - transformers installed, using this model is straightforward:

from sentence_transformers import SentenceTransformer
sentences = ['What are Large Language Models?','مدل های زبانی بزرگ چه هستند؟']

model = SentenceTransformer('heydariAI/persian-embeddings')
embeddings = model.encode(sentences)
print(embeddings)

Advanced Usage (HuggingFace Transformers)

Without sentence - transformers, you can use the model as follows: First, pass your input through the transformer model, and then apply the appropriate 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 = ['what are Large Language Models?', 'مدل های زبانی بزرگ چه هستند؟']

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('heydariAI/persian-embeddings')
model = AutoModel.from_pretrained('heydariAI/persian-embeddings')

# 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)

📄 License

This project is licensed under the Apache - 2.0 license.