README.md

metadata

dataset_info:
  features:
    - name: context
      dtype: string
    - name: name
      dtype: string
    - name: embedding
      sequence: float32
  splits:
    - name: train
      num_bytes: 717888
      num_examples: 202
  download_size: 1005715
  dataset_size: 717888
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
task_categories:
  - sentence-similarity
language:
  - es

Modelo

• "Alibaba-NLP/gte-multilingual-base"

Puedes obtener toda la información relacionado con el modelo aquí

Busqueda

from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim
from datasets import load_dataset
import numpy as np

model_name = "Alibaba-NLP/gte-multilingual-base"
model = SentenceTransformer(model_name, trust_remote_code=True)

raw_data = load_dataset('Manyah/incrustaciones')

question = ""
question_embedding = model.encode(question)

sim = [cos_sim(raw_data['train'][i]['embedding'],question_embedding).numpy() for i in range(0,len(raw_data['train']))]

index =  sim.index(max(sim))

print(raw_data['train'][index]['context'])