Spaces:
Runtime error
Runtime error
| import gradio as gr | |
| import tensorflow as tf | |
| import requests | |
| from PIL import Image | |
| import numpy as np | |
| # Cargando el modelo | |
| inception_net = tf.keras.applications.MobileNetV2() | |
| # Obteniendo las etiquetas | |
| respuesta = requests.get("https://git.io/JJkYN") | |
| etiquetas = respuesta.text.split("\n") | |
| def redimensionar_imagen(img_array, target_size=(224, 224)): | |
| img = Image.fromarray(img_array) | |
| img = img.resize(target_size) | |
| return np.array(img) | |
| def clasifica_imagen(inp): | |
| # Redimensionar la imagen | |
| inp = redimensionar_imagen(inp) | |
| # Verificar la forma actual de la imagen | |
| if inp.shape != (224, 224, 3): | |
| raise ValueError(f"Expected input shape (224, 224, 3), but got {inp.shape}") | |
| # Hacer prediccion | |
| inp = tf.keras.applications.mobilenet_v2.preprocess_input(inp) | |
| prediction = inception_net.predict(inp.reshape((-1, 224, 224, 3))).flatten() | |
| confidences = {etiquetas[i]: float(prediction[i]) for i in range(1000)} | |
| return confidences | |
| demo = gr.Interface(fn=clasifica_imagen, | |
| inputs=gr.Image(), | |
| outputs=gr.Label(num_top_classes=3), | |
| ) | |
| demo.launch(debug=True) | |