paraphrase-spanish-distilroberta

This is a sentence-transformers model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.

We follow a teacher-student transfer learning approach to train an bertin-roberta-base-spanish model using parallel EN-ES sentence pairs.

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 = ["Este es un ejemplo", "Cada oración es transformada"]

model = SentenceTransformer('hackathon-pln-es/paraphrase-spanish-distilroberta')
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
import torch.nn.functional as F

#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 = ['Este es un ejemplo", "Cada oración es transformada']

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('hackathon-pln-es/paraphrase-spanish-distilroberta')
model = AutoModel.from_pretrained('hackathon-pln-es/paraphrase-spanish-distilroberta')

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

# Normalize embeddings
sentence_embeddings = F.normalize(sentence_embeddings, p=2, dim=1)

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

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)

Evaluation Results

Similarity Evaluation on STS-2017.es-en.txt and STS-2017.es-es.txt (translated manually for evaluation purposes)

We measure the semantic textual similarity (STS) between sentence pairs in different languages:

ES-ES

ES-EN

Intended uses

Our model is intented to be used as a sentence and short paragraph encoder. Given an input text, it ouptuts a vector which captures the semantic information. The sentence vector may be used for information retrieval, clustering or sentence similarity tasks.

Background

This model is a bilingual Spanish-English model trained according to instructions in the paper Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation and the documentation accompanying its companion python package. We have used the strongest available pretrained English Bi-Encoder (paraphrase-mpnet-base-v2) as a teacher model, and the pretrained Spanish BERTIN as the student model.

We developped this model during the Hackathon 2022 NLP - Spanish, organized by hackathon-pln-es Organization.

Training data

We use the concatenation from multiple datasets with sentence pairs (EN-ES). We could check out the dataset that was used during training: parallel-sentences

Authors

• Anibal Pérez,
• Emilio Tomás Ariza,
• Lautaro Gesuelli Pinto
• Mauricio Mazuecos