Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# EJEMPLO DE USO
|
2 |
+
|
3 |
+
|
4 |
+
|
5 |
+
## Cargar librerías
|
6 |
+
|
7 |
+
import torch
|
8 |
+
from transformers import XLMRobertaForSequenceClassification, XLMRobertaTokenizer, AutoTokenizer
|
9 |
+
|
10 |
+
## Cargar el modelo y el tokenizador
|
11 |
+
model_path = "nmarinnn/bert-bregman"
|
12 |
+
model = XLMRobertaForSequenceClassification.from_pretrained(model_path)
|
13 |
+
tokenizer = XLMRobertaTokenizer.from_pretrained(model_path)
|
14 |
+
loaded_tokenizer = AutoTokenizer.from_pretrained(model_path)
|
15 |
+
|
16 |
+
## Función para predecir etiqueta
|
17 |
+
|
18 |
+
def predict(text):
|
19 |
+
inputs = loaded_tokenizer(text, return_tensors="pt", truncation=True, padding=True, max_length=512)
|
20 |
+
with torch.no_grad():
|
21 |
+
outputs = model(**inputs)
|
22 |
+
|
23 |
+
probabilities = torch.nn.functional.softmax(outputs.logits, dim=-1)
|
24 |
+
predicted_class = torch.argmax(probabilities, dim=-1).item()
|
25 |
+
|
26 |
+
class_labels = {0: "negativo", 1: "neutro", 2: "positivo"}
|
27 |
+
predicted_label = class_labels[predicted_class]
|
28 |
+
predicted_probability = probabilities[0][predicted_class].item()
|
29 |
+
|
30 |
+
return predicted_label, predicted_probability, probabilities[0].tolist()
|
31 |
+
|
32 |
+
# Ejemplo de uso
|
33 |
+
text_to_classify = "vamos rusa"
|
34 |
+
predicted_label, predicted_prob, class_probabilities = predict(text_to_classify)
|
35 |
+
|
36 |
+
print(f"Clase predicha: {predicted_label} (probabilidad = {predicted_prob:.2f})")
|
37 |
+
print(f"Probabilidades de todas las clases: Negativo: {class_probabilities[0]:.2f}, Neutro: {class_probabilities[1]:.2f}, Positivo: {class_probabilities[2]:.2f}")
|