File size: 2,500 Bytes
b117d08 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 |
import os
os.system("pip install gradio==2.9b23")
import random
import gradio as gr
from PIL import Image
import torch
from random import randint
import sys
from subprocess import call
import psutil
torch.hub.download_url_to_file('http://people.csail.mit.edu/billf/project%20pages/sresCode/Markov%20Random%20Fields%20for%20Super-Resolution_files/100075_lowres.jpg', 'bear.jpg')
def run_cmd(command):
try:
print(command)
call(command, shell=True)
except KeyboardInterrupt:
print("Process interrupted")
sys.exit(1)
run_cmd("wget https://github.com/xinntao/Real-ESRGAN/releases/download/v0.1.0/RealESRGAN_x4plus.pth -P .")
run_cmd("pip install basicsr")
run_cmd("pip freeze")
os.system("wget https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.2.4/RealESRGAN_x4plus_anime_6B.pth -P .")
def inference(img,mode):
_id = randint(1, 10000)
INPUT_DIR = "/tmp/input_image" + str(_id) + "/"
OUTPUT_DIR = "/tmp/output_image" + str(_id) + "/"
run_cmd("rm -rf " + INPUT_DIR)
run_cmd("rm -rf " + OUTPUT_DIR)
run_cmd("mkdir " + INPUT_DIR)
run_cmd("mkdir " + OUTPUT_DIR)
basewidth = 256
wpercent = (basewidth/float(img.size[0]))
hsize = int((float(img.size[1])*float(wpercent)))
img = img.resize((basewidth,hsize), Image.ANTIALIAS)
img.save(INPUT_DIR + "1.jpg", "JPEG")
if mode == "base":
run_cmd("python inference_realesrgan.py -n RealESRGAN_x4plus -i "+ INPUT_DIR + " -o " + OUTPUT_DIR)
else:
os.system("python inference_realesrgan.py -n RealESRGAN_x4plus_anime_6B -i "+ INPUT_DIR + " -o " + OUTPUT_DIR)
return os.path.join(OUTPUT_DIR, "1_out.jpg")
title = "Real-ESRGAN"
description = "Gradio demo for Real-ESRGAN. To use it, simply upload your image, or click one of the examples to load them. Read more at the links below. Please click submit only once"
article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2107.10833'>Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data</a> | <a href='https://github.com/xinntao/Real-ESRGAN'>Github Repo</a></p>"
gr.Interface(
inference,
[gr.inputs.Image(type="pil", label="Input"),gr.inputs.Radio(["base","anime"], type="value", default="base", label="model type")],
gr.outputs.Image(type="file", label="Output"),
title=title,
description=description,
article=article,
examples=[
['bear.jpg','base'],
['anime.png','anime']
]).launch() |