Hello community, I was working on my own model, but when I run this script to test it, the result is a green image, as I will show you bellow my script:
This is my script:
import torch
from torchvision.transforms import ToTensor
from PIL import Image
from generator_model import Generator
# Cargar el modelo del generador
generator = Generator(in_channels=3)
generator.load_state_dict(torch.load('models/generator_model.pth'))
generator.eval()
# Cargar una nueva imagen para la prueba
image_path = 'watermarks/ex1/ex1-wm.jpeg' # Ruta de la imagen que deseas probar
image = Image.open(image_path).convert('RGB')
transform = ToTensor()
input_image = transform(image).unsqueeze(0) # Agregar una dimensión adicional para el batch
# Generar una nueva imagen utilizando el modelo
with torch.no_grad():
generated_image = generator(input_image)
# Guardar la imagen generada
output_image = generated_image.squeeze(0).permute(1, 2, 0).clamp(0, 1).numpy()
output_image = (output_image * 255).astype('uint8')
output_image = Image.fromarray(output_image)
output_image.save('results/generated_image.jpg')
# Mostrar la imagen generada
output_image.show()
and this is the result:
I would appreciate it if you help me debug my script and see how to correct it. I am new to using PyTorch.