test.py

#!/usr/bin/python

import torch
import os

import numpy as np
import argparse
from PIL import Image
import torchvision.transforms as transforms
import torchvision.utils as vutils
from network.Transformer import Transformer

# import sys
# from torch.autograd import Variable

parser = argparse.ArgumentParser()
parser.add_argument("--input_file", default="input/test.jpg")
parser.add_argument("--load_size", default=500)
parser.add_argument("--model_path", default="./pretrained_model")
parser.add_argument("--style", default="Paprika")
parser.add_argument("--output_dir", default="output")
parser.add_argument("--gpu", type=int, default=-1)

opt = parser.parse_args()

valid_ext = [".jpg", ".png"]

if not os.path.exists(opt.output_dir):
    os.mkdir(opt.output_dir)

# load pretrained model
model = Transformer()
model.load_state_dict(
    torch.load(os.path.join(opt.model_path, opt.style + "_net_G_float.pth"))
)
model.eval()

if opt.gpu > -1:
    print("GPU mode")
    model.cuda()
else:
    print("CPU mode")
    model.float()

file = opt.input_file
basename = os.path.basename(file)

dir_path = os.path.dirname(os.path.realpath(__file__))

ext = os.path.splitext(file)[1]
if ext not in valid_ext:
    print("Bad ext!")

# load image
input_image = Image.open(os.path.join(dir_path, file)).convert("RGB")
# resize image, keep aspect ratio

h = input_image.size[0]
w = input_image.size[1]
ratio = h * 1.0 / w
if ratio > 1:
    h = int(opt.load_size)
    w = int(h * 1.0 / ratio)
else:
    w = int(opt.load_size)
    h = int(w * ratio)

input_image = input_image.resize((h, w), Image.BICUBIC)
input_image = np.asarray(input_image)

# RGB -> BGR
input_image = input_image[:, :, [2, 1, 0]]
input_image = transforms.ToTensor()(input_image).unsqueeze(0)
# preprocess, (-1, 1)
input_image = -1 + 2 * input_image

with torch.no_grad():
    if opt.gpu > -1:
        input_image = input_image.cuda()
    else:
        input_image = input_image.float()

# forward
output_image = model(input_image)
output_image = output_image[0]

# BGR -> RGB
output_image = output_image[[2, 1, 0], :, :]
# deprocess, (0, 1)
output_image = output_image.data.cpu().float() * 0.5 + 0.5

dest = os.path.join(opt.output_dir, basename[:-4] + "_" + opt.style + ".png")

# save
vutils.save_image(output_image, dest)

# print("Done!")

print(dest)

# sys.stdout.write(dest + "\n")
# sys.stdout.flush()
# sys.exit(0)