Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,11 +1,13 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
import tensorflow as tf
|
| 3 |
import numpy as np
|
| 4 |
-
import cv2
|
| 5 |
from huggingface_hub import hf_hub_download
|
|
|
|
| 6 |
|
| 7 |
# Descargar modelo desde Hugging Face
|
| 8 |
model_path = hf_hub_download(repo_id="Bmo411/DenoisingAutoencoder", filename="autoencoder_complete_model_Fourier.keras")
|
|
|
|
| 9 |
|
| 10 |
from tensorflow.keras.saving import register_keras_serializable
|
| 11 |
|
|
@@ -65,15 +67,28 @@ def Denoiser(imagen):
|
|
| 65 |
# Expandir dimensiones para que tenga el formato correcto para el modelo (batch_size, h, w, c)
|
| 66 |
noisy_image_input = np.expand_dims(noisy_image, axis=0)
|
| 67 |
|
|
|
|
|
|
|
|
|
|
| 68 |
# Predecir con el autoencoder
|
| 69 |
reconstructed = model.predict(noisy_image_input)[0] # Quitar batch_size
|
| 70 |
|
|
|
|
|
|
|
|
|
|
| 71 |
# Asegurarse de que la imagen restaurada esté en el rango [0, 255] para la visualización
|
| 72 |
noisy_image = np.uint8(noisy_image * 255)
|
| 73 |
reconstructed = np.uint8(reconstructed * 255)
|
| 74 |
|
| 75 |
return noisy_image, reconstructed
|
| 76 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 77 |
# Crear interfaz en Gradio con dos salidas de imagen
|
| 78 |
demo = gr.Interface(
|
| 79 |
fn=Denoiser,
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
import tensorflow as tf
|
| 3 |
import numpy as np
|
| 4 |
+
import cv2
|
| 5 |
from huggingface_hub import hf_hub_download
|
| 6 |
+
import time
|
| 7 |
|
| 8 |
# Descargar modelo desde Hugging Face
|
| 9 |
model_path = hf_hub_download(repo_id="Bmo411/DenoisingAutoencoder", filename="autoencoder_complete_model_Fourier.keras")
|
| 10 |
+
model = tf.keras.models.load_model(model_path)
|
| 11 |
|
| 12 |
from tensorflow.keras.saving import register_keras_serializable
|
| 13 |
|
|
|
|
| 67 |
# Expandir dimensiones para que tenga el formato correcto para el modelo (batch_size, h, w, c)
|
| 68 |
noisy_image_input = np.expand_dims(noisy_image, axis=0)
|
| 69 |
|
| 70 |
+
# Medir el tiempo de la predicción
|
| 71 |
+
start_time = time.time()
|
| 72 |
+
|
| 73 |
# Predecir con el autoencoder
|
| 74 |
reconstructed = model.predict(noisy_image_input)[0] # Quitar batch_size
|
| 75 |
|
| 76 |
+
prediction_time = time.time() - start_time
|
| 77 |
+
print(f"Tiempo de predicción: {prediction_time} segundos")
|
| 78 |
+
|
| 79 |
# Asegurarse de que la imagen restaurada esté en el rango [0, 255] para la visualización
|
| 80 |
noisy_image = np.uint8(noisy_image * 255)
|
| 81 |
reconstructed = np.uint8(reconstructed * 255)
|
| 82 |
|
| 83 |
return noisy_image, reconstructed
|
| 84 |
|
| 85 |
+
# Verificar si TensorFlow está utilizando la GPU
|
| 86 |
+
physical_devices = tf.config.list_physical_devices('GPU')
|
| 87 |
+
if physical_devices:
|
| 88 |
+
print("TensorFlow está utilizando la GPU")
|
| 89 |
+
else:
|
| 90 |
+
print("TensorFlow no está utilizando la GPU")
|
| 91 |
+
|
| 92 |
# Crear interfaz en Gradio con dos salidas de imagen
|
| 93 |
demo = gr.Interface(
|
| 94 |
fn=Denoiser,
|