clips/mfaq

Instalación

pip install sentence-transformers transformers

Uso

Puedes usar MFAQ con sentence-transformers o directamente con un modelo HuggingFace.

Transformadores de Oraciones

from sentence_transformers import SentenceTransformer

question = "How many models can I host on HuggingFace?"
answer_1 = "All plans come with unlimited private models and datasets."
answer_2 = "AutoNLP is an automatic way to train and deploy state-of-the-art NLP models, seamlessly integrated with the Hugging Face ecosystem."
answer_3 = "Based on how much training data and model variants are created, we send you a compute cost and payment link - as low as $10 per job."

model = SentenceTransformer('clips/mfaq')
embeddings = model.encode([question, answer_1, answer_3, answer_3])
print(embeddings)

Transformadores HuggingFace

from transformers import AutoTokenizer, AutoModel
import torch

def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0]  # El primer elemento de model_output contiene todas las incrustaciones de tokens.
    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)

question = "How many models can I host on HuggingFace?"
answer_1 = "All plans come with unlimited private models and datasets."
answer_2 = "AutoNLP is an automatic way to train and deploy state-of-the-art NLP models, seamlessly integrated with the Hugging Face ecosystem."
answer_3 = "Based on how much training data and model variants are created, we send you a compute cost and payment link - as low as $10 per job."

tokenizer = AutoTokenizer.from_pretrained('clips/mfaq')
model = AutoModel.from_pretrained('clips/mfaq')

# Tokenizar oraciones
encoded_input = tokenizer([question, answer_1, answer_3, answer_3], padding=True, truncation=True, return_tensors='pt')

# Calcular incrustaciones de tokens
with torch.no_grad():
    model_output = model(**encoded_input)

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

Entrenamiento

Puedes encontrar el script de entrenamiento para el modelo aquí.