Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -5,12 +5,53 @@ import time
|
|
5 |
import os
|
6 |
import spaces
|
7 |
import torch
|
|
|
|
|
8 |
|
9 |
zero = torch.Tensor([0]).cuda()
|
10 |
print(zero.device) # <-- 'cpu' π€
|
11 |
|
12 |
names = ['prompt', 'negative_prompt', 'subject', 'number_of_outputs', 'number_of_images_per_pose', 'randomise_poses', 'output_format', 'output_quality', 'seed']
|
13 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
@spaces.GPU
|
15 |
def predict(request: gr.Request, *args, progress=gr.Progress(track_tqdm=True)):
|
16 |
print(zero.device) # <-- 'cuda:0' π€
|
@@ -18,7 +59,6 @@ def predict(request: gr.Request, *args, progress=gr.Progress(track_tqdm=True)):
|
|
18 |
|
19 |
payload = {"input": {}}
|
20 |
|
21 |
-
|
22 |
base_url = "http://0.0.0.0:7860"
|
23 |
for i, key in enumerate(names):
|
24 |
value = args[i]
|
@@ -29,7 +69,6 @@ def predict(request: gr.Request, *args, progress=gr.Progress(track_tqdm=True)):
|
|
29 |
|
30 |
response = requests.post("http://0.0.0.0:5000/predictions", headers=headers, json=payload)
|
31 |
|
32 |
-
|
33 |
if response.status_code == 201:
|
34 |
follow_up_url = response.json()["urls"]["get"]
|
35 |
response = requests.get(follow_up_url, headers=headers)
|
@@ -132,4 +171,7 @@ with gr.Blocks(css=css) as app:
|
|
132 |
show_api = False
|
133 |
)
|
134 |
|
|
|
|
|
|
|
135 |
app.queue(max_size=12, api_open=False).launch(share=False, show_api=False)
|
|
|
5 |
import os
|
6 |
import spaces
|
7 |
import torch
|
8 |
+
import subprocess
|
9 |
+
import signal
|
10 |
|
11 |
zero = torch.Tensor([0]).cuda()
|
12 |
print(zero.device) # <-- 'cpu' π€
|
13 |
|
14 |
names = ['prompt', 'negative_prompt', 'subject', 'number_of_outputs', 'number_of_images_per_pose', 'randomise_poses', 'output_format', 'output_quality', 'seed']
|
15 |
|
16 |
+
def check_cog_server():
|
17 |
+
try:
|
18 |
+
# Start the Cog server in the background
|
19 |
+
cog_process = subprocess.Popen(["python3", "-m", "cog.server.http", "--threads=10"], cwd="/src")
|
20 |
+
|
21 |
+
# Wait for the Cog server to start on port 5000
|
22 |
+
counter1 = 0
|
23 |
+
while True:
|
24 |
+
try:
|
25 |
+
requests.get("http://localhost:5000")
|
26 |
+
print("Cog server is running on port 5000.")
|
27 |
+
break
|
28 |
+
except requests.exceptions.ConnectionError:
|
29 |
+
print("Waiting for Cog server to start on port 5000...")
|
30 |
+
time.sleep(5)
|
31 |
+
counter1 += 1
|
32 |
+
if counter1 >= 250:
|
33 |
+
raise Exception("Error: Cog server did not start on port 5000 after 250 attempts.")
|
34 |
+
|
35 |
+
# Wait for the Cog server to be fully ready
|
36 |
+
counter2 = 0
|
37 |
+
while True:
|
38 |
+
response = requests.get("http://localhost:5000/health-check")
|
39 |
+
status = response.json().get("status")
|
40 |
+
if status == "READY":
|
41 |
+
print("Cog server is fully ready.")
|
42 |
+
break
|
43 |
+
else:
|
44 |
+
print("Waiting for Cog server (models loading) on port 5000...")
|
45 |
+
time.sleep(5)
|
46 |
+
counter2 += 1
|
47 |
+
if counter2 >= 250:
|
48 |
+
raise Exception("Error: Cog server did not become fully ready after 250 attempts.")
|
49 |
+
|
50 |
+
except Exception as e:
|
51 |
+
print(f"Error: {str(e)}")
|
52 |
+
cog_process.send_signal(signal.SIGINT) # Send interrupt signal to the Cog process
|
53 |
+
raise e
|
54 |
+
|
55 |
@spaces.GPU
|
56 |
def predict(request: gr.Request, *args, progress=gr.Progress(track_tqdm=True)):
|
57 |
print(zero.device) # <-- 'cuda:0' π€
|
|
|
59 |
|
60 |
payload = {"input": {}}
|
61 |
|
|
|
62 |
base_url = "http://0.0.0.0:7860"
|
63 |
for i, key in enumerate(names):
|
64 |
value = args[i]
|
|
|
69 |
|
70 |
response = requests.post("http://0.0.0.0:5000/predictions", headers=headers, json=payload)
|
71 |
|
|
|
72 |
if response.status_code == 201:
|
73 |
follow_up_url = response.json()["urls"]["get"]
|
74 |
response = requests.get(follow_up_url, headers=headers)
|
|
|
171 |
show_api = False
|
172 |
)
|
173 |
|
174 |
+
# Check the Cog server's readiness before launching the Gradio app
|
175 |
+
check_cog_server()
|
176 |
+
|
177 |
app.queue(max_size=12, api_open=False).launch(share=False, show_api=False)
|