Spaces:
Runtime error
Runtime error
Commit
·
7ce51a2
1
Parent(s):
322db57
Fix some issues (#1)
Browse files- Fix some issues (4ef6d61e1037e33d2d04cb68ea77232b8dc35624)
Co-authored-by: Apolinário from multimodal AI art <multimodalart@users.noreply.huggingface.co>
app.py
CHANGED
|
@@ -5,6 +5,9 @@ from PIL import Image
|
|
| 5 |
import json
|
| 6 |
import os
|
| 7 |
import logging
|
|
|
|
|
|
|
|
|
|
| 8 |
|
| 9 |
logging.basicConfig(level=logging.DEBUG)
|
| 10 |
|
|
@@ -31,27 +34,39 @@ def run_lora(prompt, selected_state, progress=gr.Progress(track_tqdm=True)):
|
|
| 31 |
selected_lora = loras[selected_lora_index]
|
| 32 |
api_url = f"https://api-inference.huggingface.co/models/{selected_lora['repo']}"
|
| 33 |
trigger_word = selected_lora["trigger_word"]
|
| 34 |
-
token = os.getenv("API_TOKEN")
|
| 35 |
payload = {"inputs": f"{prompt} {trigger_word}"}
|
| 36 |
|
| 37 |
-
headers = {"Authorization": f"Bearer {token}"}
|
| 38 |
|
| 39 |
# Add a print statement to display the API request
|
| 40 |
print(f"API Request: {api_url}")
|
| 41 |
-
print(f"API Headers: {headers}")
|
| 42 |
print(f"API Payload: {payload}")
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 50 |
|
| 51 |
|
| 52 |
|
| 53 |
with gr.Blocks(css="custom.css") as app:
|
| 54 |
-
title = gr.
|
| 55 |
selected_state = gr.State()
|
| 56 |
with gr.Row():
|
| 57 |
gallery = gr.Gallery(
|
|
@@ -71,11 +86,16 @@ with gr.Blocks(css="custom.css") as app:
|
|
| 71 |
update_selection,
|
| 72 |
outputs=[prompt, selected_state]
|
| 73 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
button.click(
|
| 75 |
fn=run_lora,
|
| 76 |
inputs=[prompt, selected_state],
|
| 77 |
outputs=[result]
|
| 78 |
)
|
| 79 |
|
| 80 |
-
app.queue(max_size=20)
|
| 81 |
-
app.launch()
|
|
|
|
| 5 |
import json
|
| 6 |
import os
|
| 7 |
import logging
|
| 8 |
+
import math
|
| 9 |
+
from tqdm import tqdm
|
| 10 |
+
import time
|
| 11 |
|
| 12 |
logging.basicConfig(level=logging.DEBUG)
|
| 13 |
|
|
|
|
| 34 |
selected_lora = loras[selected_lora_index]
|
| 35 |
api_url = f"https://api-inference.huggingface.co/models/{selected_lora['repo']}"
|
| 36 |
trigger_word = selected_lora["trigger_word"]
|
| 37 |
+
#token = os.getenv("API_TOKEN")
|
| 38 |
payload = {"inputs": f"{prompt} {trigger_word}"}
|
| 39 |
|
| 40 |
+
#headers = {"Authorization": f"Bearer {token}"}
|
| 41 |
|
| 42 |
# Add a print statement to display the API request
|
| 43 |
print(f"API Request: {api_url}")
|
| 44 |
+
#print(f"API Headers: {headers}")
|
| 45 |
print(f"API Payload: {payload}")
|
| 46 |
+
|
| 47 |
+
error_count = 0
|
| 48 |
+
pbar = tqdm(total=None, desc="Loading model")
|
| 49 |
+
while(True):
|
| 50 |
+
response = requests.post(api_url, json=payload)
|
| 51 |
+
if response.status_code == 200:
|
| 52 |
+
return Image.open(io.BytesIO(response.content))
|
| 53 |
+
elif response.status_code == 503:
|
| 54 |
+
#503 is triggered when the model is doing cold boot. It also gives you a time estimate from when the model is loaded but it is not super precise
|
| 55 |
+
time.sleep(1)
|
| 56 |
+
pbar.update(1)
|
| 57 |
+
elif response.status_code == 500 and error_count < 5:
|
| 58 |
+
print(response.content)
|
| 59 |
+
time.sleep(1)
|
| 60 |
+
error_count += 1
|
| 61 |
+
continue
|
| 62 |
+
else:
|
| 63 |
+
logging.error(f"API Error: {response.status_code}")
|
| 64 |
+
raise gr.Error("API Error: Unable to fetch the image.") # Raise a Gradio error here
|
| 65 |
|
| 66 |
|
| 67 |
|
| 68 |
with gr.Blocks(css="custom.css") as app:
|
| 69 |
+
title = gr.Markdown("# artificialguybr LoRA portfolio")
|
| 70 |
selected_state = gr.State()
|
| 71 |
with gr.Row():
|
| 72 |
gallery = gr.Gallery(
|
|
|
|
| 86 |
update_selection,
|
| 87 |
outputs=[prompt, selected_state]
|
| 88 |
)
|
| 89 |
+
prompt.submit(
|
| 90 |
+
fn=run_lora,
|
| 91 |
+
inputs=[prompt, selected_state],
|
| 92 |
+
outputs=[result]
|
| 93 |
+
)
|
| 94 |
button.click(
|
| 95 |
fn=run_lora,
|
| 96 |
inputs=[prompt, selected_state],
|
| 97 |
outputs=[result]
|
| 98 |
)
|
| 99 |
|
| 100 |
+
app.queue(max_size=20, concurrency_count=5)
|
| 101 |
+
app.launch()
|