XLM-RoBERTa model for English and Azerbaijani

Usage (Sentence-Transformers)

pip install -U sentence-transformers

from sentence_transformers import SentenceTransformer
sentences = ['Bu nümunə cümlədir', 'Bu cümlə bir nümunədir']

model = SentenceTransformer('LocalDoc/xlm-roberta-AZ')
embeddings = model.encode(sentences)
print(embeddings)

Usage (HuggingFace Transformers)

from transformers import AutoTokenizer, AutoModel
import torch



def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0]
    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 = ['Bu nümunə cümlədir', 'Bu cümlə bir nümunədir']


tokenizer = AutoTokenizer.from_pretrained('LocalDoc/xlm-roberta-AZ')
model = AutoModel.from_pretrained('LocalDoc/xlm-roberta-AZ')


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


with torch.no_grad():
    model_output = model(**encoded_input)


sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])

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

SentenceTransformer Model Architecture

SentenceTransformer(
    (0): Transformer({
            'max_seq_length': 8192, 
            'do_lower_case': False
        }) 
        with Transformer model: XLMRobertaModel 
    (1): Pooling({
            'word_embedding_dimension': 1024,
            'pooling_mode_cls_token': False,
            'pooling_mode_mean_tokens': True,
            'pooling_mode_max_tokens': False,
            'pooling_mode_mean_sqrt_len_tokens': False,
            'pooling_mode_weightedmean_tokens': False,
            'pooling_mode_lasttoken': False,
            'include_prompt': True
        })
)