DeCLUTR-sci-base

Por favor, consulte nuestro repositorio para obtener detalles completos. Un ejemplo simple se muestra a continuación.

Con SentenceTransformers:

from scipy.spatial.distance import cosine
from sentence_transformers import SentenceTransformer

# Cargar el modelo
model = SentenceTransformer("johngiorgi/declutr-sci-base")

# Preparar texto para incrustar
text = [
"Oncogenic KRAS mutations are common in cancer.",
"Notably, c-Raf has recently been found essential for development of K-Ras-driven NSCLCs.",
]

# Incrustar el texto
embeddings = model.encode(texts)

# Calcular una similitud semántica a través de la distancia coseno
semantic_sim = 1 - cosine(embeddings[0], embeddings[1])

Con 🤗 Transformers:

import torch
from scipy.spatial.distance import cosine
from transformers import AutoModel, AutoTokenizer

# Cargar el modelo
tokenizer = AutoTokenizer.from_pretrained("johngiorgi/declutr-sci-base")
model = AutoModel.from_pretrained("johngiorgi/declutr-sci-base")

# Preparar texto para incrustar
text = [
"Oncogenic KRAS mutations are common in cancer.",
"Notably, c-Raf has recently been found essential for development of K-Ras-driven NSCLCs.",
]
inputs = tokenizer(text, padding=True, truncation=True, return_tensors="pt")

# Incrustar el texto
with torch.no_grad():
sequence_output = model(**inputs)[0]

# Agrupar medio las incrustaciones a nivel de token para obtener incrustaciones a nivel de oración
embeddings = torch.sum(
sequence_output * inputs["attention_mask"].unsqueeze(-1), dim=1
) / torch.clamp(torch.sum(inputs["attention_mask"], dim=1, keepdims=True), min=1e-9)

# Calcular una similitud semántica a través de la distancia coseno
semantic_sim = 1 - cosine(embeddings[0], embeddings[1])