selfitcamera
commited on
Commit
·
5d333f5
1
Parent(s):
61db459
init
Browse files
.DS_Store
ADDED
Binary file (6.15 kB). View file
|
|
app.py
ADDED
@@ -0,0 +1,115 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
from mtcnn.mtcnn import MTCNN
|
3 |
+
from utils import *
|
4 |
+
|
5 |
+
|
6 |
+
face_detector = MTCNN()
|
7 |
+
|
8 |
+
# Description
|
9 |
+
title = r"""
|
10 |
+
<h1 align="center">IDM-VTON + Outfit Anyone in the Wild </h1>
|
11 |
+
"""
|
12 |
+
|
13 |
+
description = r"""
|
14 |
+
This demo combines <b>IDM-VTON </b> and <b>Outfit Anyone in the Wild </b>
|
15 |
+
1. Human pose detection and reconstruction using large human model from Outfit Anyone in the Wild.
|
16 |
+
2. Use IDM-VTON for training-free try-on.
|
17 |
+
3. Applying the refine network from Outfit Anyone in the Wild.<br>
|
18 |
+
"""
|
19 |
+
|
20 |
+
css = """
|
21 |
+
.gradio-container {width: 85% !important}
|
22 |
+
"""
|
23 |
+
|
24 |
+
|
25 |
+
def onClick(cloth_image, pose_image, category,
|
26 |
+
caption, request: gr.Request):
|
27 |
+
if pose_image is None:
|
28 |
+
yield None, f"no user image found !"
|
29 |
+
return None, "no user image found !"
|
30 |
+
elif cloth_image is None:
|
31 |
+
yield None, f"no cloth image found !"
|
32 |
+
return None, "no cloth image found !"
|
33 |
+
try:
|
34 |
+
faces = face_detector.detect_faces(pose_image[:,:,::-1])
|
35 |
+
if len(faces)==0:
|
36 |
+
print(client_ip, 'faces num is 0! ', flush=True)
|
37 |
+
yield None, "Fatal Error !!! No face detected in pose image !!! "
|
38 |
+
return None, "Fatal Error !!! No face detected in pose image !!! "
|
39 |
+
else:
|
40 |
+
x, y, w, h = faces[0]["box"]
|
41 |
+
H, W = pose_image.shape[:2]
|
42 |
+
max_face_ratio = 1/3.3
|
43 |
+
if w/W>max_face_ratio or h/H>max_face_ratio:
|
44 |
+
yield None, "Fatal Error !!! Headshot is not allowed in pose image!!!"
|
45 |
+
return None, "Fatal Error !!! Headshot is not allowed in pose image!!!"
|
46 |
+
|
47 |
+
uploads = upload_imgs(ApiUrl, UploadToken, cloth_image, pose_image)
|
48 |
+
if uploads is None:
|
49 |
+
yield None, "fail to upload"
|
50 |
+
return None, "fail to upload"
|
51 |
+
|
52 |
+
infId = publicFastSwap(ApiUrl, OpenId, ApiKey, uploads, category, caption)
|
53 |
+
if not infId:
|
54 |
+
yield None, "fail to public you task"
|
55 |
+
return None, "fail to public you task"
|
56 |
+
|
57 |
+
max_try = 30
|
58 |
+
wait_s = 3
|
59 |
+
yield None, "start to process, please wait..."
|
60 |
+
for i in range(max_try):
|
61 |
+
time.sleep(wait_s)
|
62 |
+
taskStatus = getTaskRes(ApiUrl, infId)
|
63 |
+
if taskStatus is None: continue
|
64 |
+
|
65 |
+
status = taskStatus['status']
|
66 |
+
if status in ['FAILED', 'CANCELLED', 'TIMED_OUT', ]:
|
67 |
+
yield None, f"task failed, query {i}, status {status}"
|
68 |
+
return None, f"task failed, query {i}, status {status}"
|
69 |
+
elif status in ['IN_QUEUE', 'IN_PROGRESS', 'IN_QUEUE', ]:
|
70 |
+
pass
|
71 |
+
yield None, f"task is on processing, query {i}, status {status}, please do not exit !!!"
|
72 |
+
elif status=='COMPLETED':
|
73 |
+
out = taskStatus['output']['job_results']['output1']
|
74 |
+
yield out, f"task is COMPLETED"
|
75 |
+
return out, f"{i} task COMPLETED"
|
76 |
+
yield None, "fail to query task.."
|
77 |
+
return None, "fail to query task.."
|
78 |
+
|
79 |
+
|
80 |
+
except Exception as e:
|
81 |
+
print(e)
|
82 |
+
return None, "fail to create task"
|
83 |
+
|
84 |
+
|
85 |
+
with gr.Blocks(css=css) as demo:
|
86 |
+
# description
|
87 |
+
gr.Markdown(title)
|
88 |
+
gr.Markdown(description)
|
89 |
+
|
90 |
+
with gr.Row():
|
91 |
+
with gr.Column():
|
92 |
+
with gr.Column():
|
93 |
+
cloth_image = gr.Image(value=None, type="numpy", label="cloth")
|
94 |
+
with gr.Column():
|
95 |
+
with gr.Column():
|
96 |
+
pose_image = gr.Image(value=None, type="numpy", label="user photo")
|
97 |
+
with gr.Column():
|
98 |
+
with gr.Column():
|
99 |
+
category = gr.Dropdown(value="upper_cloth", choices=["upper_cloth",
|
100 |
+
"lower_cloth", "full_body", "dresses"], interactive=True)
|
101 |
+
caption = gr.Textbox(value="", interactive=True, label='cloth caption')
|
102 |
+
|
103 |
+
info_text = gr.Textbox(value="", interactive=False, label='runtime information')
|
104 |
+
run_button = gr.Button(value="Run")
|
105 |
+
res_image = gr.Image(label="result image", value=None, type="filepath")
|
106 |
+
|
107 |
+
run_button.click(fn=onClick, inputs=[cloth_image, pose_image,
|
108 |
+
category, caption, ],
|
109 |
+
outputs=[res_image, info_text, ])
|
110 |
+
|
111 |
+
if __name__ == "__main__":
|
112 |
+
|
113 |
+
demo.queue(max_size=50)
|
114 |
+
demo.launch(server_name='0.0.0.0')
|
115 |
+
|
requirements.txt
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
opencv-python
|
2 |
+
numpy
|
3 |
+
requests
|
4 |
+
gradio==3.41.2
|
5 |
+
gradio-client==0.5.0
|
6 |
+
mtcnn
|
7 |
+
tensorflow
|
8 |
+
func_timeout
|
9 |
+
httpx==0.24.1
|
utils.py
ADDED
@@ -0,0 +1,113 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
import os
|
3 |
+
import sys
|
4 |
+
import cv2
|
5 |
+
import json
|
6 |
+
import random
|
7 |
+
import time
|
8 |
+
import requests
|
9 |
+
import func_timeout
|
10 |
+
import numpy as np
|
11 |
+
import gradio as gr
|
12 |
+
|
13 |
+
|
14 |
+
ApiUrl = os.environ['ApiUrl']
|
15 |
+
OpenId = os.environ['OpenId']
|
16 |
+
ApiKey = os.environ['ApiKey']
|
17 |
+
UploadToken = os.environ['UploadToken']
|
18 |
+
|
19 |
+
|
20 |
+
proj_dir = os.path.dirname(os.path.abspath(__file__))
|
21 |
+
data_dir = os.path.join(proj_dir, 'Datas')
|
22 |
+
# data_dir = "Datas"
|
23 |
+
tmpFolder = "tmp"
|
24 |
+
os.makedirs(tmpFolder, exist_ok=True)
|
25 |
+
|
26 |
+
|
27 |
+
|
28 |
+
def upload_imgs(apiUrl, UploadToken, cloth_image, pose_image):
|
29 |
+
folder = os.path.join(tmpFolder, str(random.randint(0, 100)))
|
30 |
+
os.makedirs(folder, exist_ok=True)
|
31 |
+
pose_path = os.path.join(folder, 'pose.jpg')
|
32 |
+
cloth_path = os.path.join(folder, 'cloth.jpg')
|
33 |
+
cv2.imwrite(pose_path, pose_image[:,:,::-1])
|
34 |
+
cv2.imwrite(cloth_path, cloth_image[:,:,::-1])
|
35 |
+
|
36 |
+
params = {'token':UploadToken,
|
37 |
+
'input1':'pose.jpg',
|
38 |
+
'input2':'cloth.jpg',
|
39 |
+
'protocol':'https'}
|
40 |
+
session = requests.session()
|
41 |
+
ret = requests.post(f"{apiUrl}/upload", data=json.dumps(params))
|
42 |
+
if ret.status_code==200:
|
43 |
+
if 'upload1' in ret.json():
|
44 |
+
data = ret.json()
|
45 |
+
with open(cloth_path, 'rb') as file:
|
46 |
+
headers = {"Content-Type": 'image/jpeg'}
|
47 |
+
response = requests.put(data['upload2'], data=file, headers=headers)
|
48 |
+
if response.status_code == 200:
|
49 |
+
print("上传成功")
|
50 |
+
else:
|
51 |
+
print(f"上传失败,状态码: {response.status_code}, 响应内容: {response.text}")
|
52 |
+
return
|
53 |
+
with open(pose_path, 'rb') as file:
|
54 |
+
response = requests.put(data['upload1'], data=file, headers=headers)
|
55 |
+
if response.status_code == 200:
|
56 |
+
print("上传成功")
|
57 |
+
else:
|
58 |
+
print(f"上传失败,状态码: {response.status_code}, 响应内容: {response.text}")
|
59 |
+
return
|
60 |
+
if os.path.exists(pose_path): os.remove(pose_path)
|
61 |
+
if os.path.exists(cloth_path): os.remove(cloth_path)
|
62 |
+
return {'pose':data['upload1'], 'cloth':data['upload2']}
|
63 |
+
|
64 |
+
def publicFastSwap(apiUrl, openId, apiKey, uploads, category, caption):
|
65 |
+
if category=="upper_cloth":
|
66 |
+
category = 1
|
67 |
+
elif category=="lower_cloth":
|
68 |
+
category = 2
|
69 |
+
elif category=="dresses":
|
70 |
+
category = 3
|
71 |
+
elif category=="full_body":
|
72 |
+
category = 4
|
73 |
+
params = {'openId':OpenId, 'apiKey':ApiKey,
|
74 |
+
'task_type':"10", 'image':str(uploads['pose']),
|
75 |
+
'mask':str(uploads['cloth']),
|
76 |
+
'param1':str(category), 'param2':str(caption),
|
77 |
+
'param3':'', 'param4':'', 'param5':'', }
|
78 |
+
session = requests.session()
|
79 |
+
ret = requests.post(f"{ApiUrl}/public", data=json.dumps(params))
|
80 |
+
if ret.status_code==200:
|
81 |
+
if 'id' in ret.json():
|
82 |
+
print('public task successfully!')
|
83 |
+
return ret.json()['id']
|
84 |
+
|
85 |
+
def getTaskRes(apiUrl, taskId):
|
86 |
+
params = {'id':taskId, 'task_type':"10"}
|
87 |
+
session = requests.session()
|
88 |
+
ret = requests.post(f"{apiUrl}/status", data=json.dumps(params))
|
89 |
+
if ret.status_code==200:
|
90 |
+
if 'status' in ret.json():
|
91 |
+
return ret.json()
|
92 |
+
else:
|
93 |
+
print(ret.json(), ret.status_code, 'call status failed')
|
94 |
+
return None
|
95 |
+
|
96 |
+
@func_timeout.func_set_timeout(10)
|
97 |
+
def check_func(ip):
|
98 |
+
session = requests.session()
|
99 |
+
ret = requests.get(f"https://webapi-pc.meitu.com/common/ip_location?ip={ip}")
|
100 |
+
for k in ret.json()['data']:
|
101 |
+
nat = ret.json()['data'][k]['nation']
|
102 |
+
if nat.lower() in Regions.lower():
|
103 |
+
print(nat, 'invalid')
|
104 |
+
return False
|
105 |
+
else:
|
106 |
+
print(nat, 'valid')
|
107 |
+
return True
|
108 |
+
def check_warp(ip):
|
109 |
+
try:
|
110 |
+
return check_func(ip)
|
111 |
+
except Exception as e:
|
112 |
+
print(e)
|
113 |
+
return True
|