Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -6,7 +6,7 @@ import random
|
|
6 |
import os
|
7 |
from PIL import Image
|
8 |
from deep_translator import GoogleTranslator
|
9 |
-
from gradio_client import Client # Import the
|
10 |
|
11 |
# os.makedirs('assets', exist_ok=True)
|
12 |
if not os.path.exists('icon.jpg'):
|
@@ -15,108 +15,85 @@ API_URL_DEV = "https://api-inference.huggingface.co/models/black-forest-labs/FLU
|
|
15 |
API_URL = "https://api-inference.huggingface.co/models/black-forest-labs/FLUX.1-schnell"
|
16 |
timeout = 100
|
17 |
|
18 |
-
|
19 |
-
|
20 |
-
client = Client("
|
21 |
result = client.predict(
|
22 |
-
|
23 |
-
|
|
|
|
|
|
|
|
|
24 |
)
|
25 |
-
print(f"System session modified: {result}")
|
26 |
return result
|
27 |
|
28 |
-
# Function to enhance the prompt with Qwen model
|
29 |
-
def enhance_prompt_with_qwen(prompt):
|
30 |
-
client = Client("Qwen/Qwen2.5-72B-Instruct")
|
31 |
-
result = client.predict(
|
32 |
-
query=prompt,
|
33 |
-
history=[],
|
34 |
-
system="You are Qwen, an image generation prompt enhancer",
|
35 |
-
api_name="/model_chat"
|
36 |
-
)
|
37 |
-
|
38 |
-
# Extract the relevant part of the tuple, index [0], which contains the enhanced prompt.
|
39 |
-
enhanced_prompt = result[0] # This is the string we need for the image generation prompt.
|
40 |
-
|
41 |
-
print(f"Enhanced prompt: {enhanced_prompt}")
|
42 |
-
return enhanced_prompt
|
43 |
-
|
44 |
-
# Image generation query function
|
45 |
def query(prompt, is_negative=False, steps=30, cfg_scale=7, sampler="DPM++ 2M Karras", seed=-1, strength=0.7, huggingface_api_key=None, use_dev=False):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
46 |
try:
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
api_url = API_URL_DEV if use_dev else API_URL
|
55 |
-
|
56 |
-
# Check if the request is an API call by checking for the presence of the huggingface_api_key
|
57 |
-
is_api_call = huggingface_api_key is not None
|
58 |
-
|
59 |
-
if is_api_call:
|
60 |
-
# Use the environment variable for the API key in GUI mode
|
61 |
-
API_TOKEN = os.getenv("HF_READ_TOKEN")
|
62 |
-
headers = {"Authorization": f"Bearer {API_TOKEN}"}
|
63 |
-
else:
|
64 |
-
# Validate the API key if it's an API call
|
65 |
-
if huggingface_api_key == "":
|
66 |
-
raise gr.Error("API key is required for API calls.")
|
67 |
-
headers = {"Authorization": f"Bearer {huggingface_api_key}"}
|
68 |
-
|
69 |
-
if enhanced_prompt == "" or enhanced_prompt is None:
|
70 |
-
return None, None
|
71 |
-
|
72 |
-
key = random.randint(0, 999)
|
73 |
-
|
74 |
-
# Translate the enhanced prompt
|
75 |
-
enhanced_prompt = GoogleTranslator(source='ru', target='en').translate(enhanced_prompt)
|
76 |
-
print(f'\033[1mGeneration {key} translation:\033[0m {enhanced_prompt}')
|
77 |
-
|
78 |
-
enhanced_prompt = f"{enhanced_prompt} | ultra detail, ultra elaboration, ultra quality, perfect."
|
79 |
-
print(f'\033[1mGeneration {key}:\033[0m {enhanced_prompt}')
|
80 |
-
|
81 |
-
# If seed is -1, generate a random seed and use it
|
82 |
-
if seed == -1:
|
83 |
-
seed = random.randint(1, 1000000000)
|
84 |
-
|
85 |
-
payload = {
|
86 |
-
"inputs": enhanced_prompt,
|
87 |
-
"is_negative": is_negative,
|
88 |
-
"steps": steps,
|
89 |
-
"cfg_scale": cfg_scale,
|
90 |
-
"seed": seed,
|
91 |
-
"strength": strength
|
92 |
-
}
|
93 |
-
|
94 |
-
response = requests.post(api_url, headers=headers, json=payload, timeout=timeout)
|
95 |
-
if response.status_code != 200:
|
96 |
-
print(f"Error: Failed to get image. Response status: {response.status_code}")
|
97 |
-
print(f"Response content: {response.text}")
|
98 |
-
if response.status_code == 503:
|
99 |
-
raise gr.Error(f"{response.status_code} : The model is being loaded")
|
100 |
-
raise gr.Error(f"{response.status_code}")
|
101 |
|
102 |
-
|
103 |
-
|
104 |
-
|
105 |
-
|
106 |
-
print(f'\033[1mGeneration {key} completed!\033[0m ({enhanced_prompt})')
|
107 |
-
|
108 |
-
# Save the image to a file and return the file path and seed
|
109 |
-
output_path = f"./output_{key}.png"
|
110 |
-
image.save(output_path)
|
111 |
-
|
112 |
-
return output_path, seed
|
113 |
-
except Exception as e:
|
114 |
-
print(f"Error when trying to open the image: {e}")
|
115 |
-
return None, seed # If the image fails, return None for image, seed is still returned
|
116 |
-
|
117 |
-
except Exception as ex:
|
118 |
-
print(f"Error in query execution: {ex}")
|
119 |
-
return None, None # If the entire process fails, return None for both
|
120 |
|
121 |
css = """
|
122 |
#app-container {
|
|
|
6 |
import os
|
7 |
from PIL import Image
|
8 |
from deep_translator import GoogleTranslator
|
9 |
+
from gradio_client import Client # Import the Gradio Client
|
10 |
|
11 |
# os.makedirs('assets', exist_ok=True)
|
12 |
if not os.path.exists('icon.jpg'):
|
|
|
15 |
API_URL = "https://api-inference.huggingface.co/models/black-forest-labs/FLUX.1-schnell"
|
16 |
timeout = 100
|
17 |
|
18 |
+
def enhance_prompt(prompt):
|
19 |
+
"""Enhance the prompt using the Mistral Nemo prompt enhancer API."""
|
20 |
+
client = Client("K00B404/mistral-nemo-prompt-enhancer")
|
21 |
result = client.predict(
|
22 |
+
message=prompt,
|
23 |
+
system_message="You are an image generation prompt enhancer and should only respond with the enhanced version of the user input image generation prompt.",
|
24 |
+
max_tokens=512,
|
25 |
+
temperature=0.7,
|
26 |
+
top_p=0.95,
|
27 |
+
api_name="/chat"
|
28 |
)
|
|
|
29 |
return result
|
30 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
31 |
def query(prompt, is_negative=False, steps=30, cfg_scale=7, sampler="DPM++ 2M Karras", seed=-1, strength=0.7, huggingface_api_key=None, use_dev=False):
|
32 |
+
# Determine which API URL to use
|
33 |
+
api_url = API_URL_DEV if use_dev else API_URL
|
34 |
+
|
35 |
+
# Check if the request is an API call by checking for the presence of the huggingface_api_key
|
36 |
+
is_api_call = huggingface_api_key is not None
|
37 |
+
|
38 |
+
if is_api_call:
|
39 |
+
# Use the environment variable for the API key in GUI mode
|
40 |
+
API_TOKEN = os.getenv("HF_READ_TOKEN")
|
41 |
+
headers = {"Authorization": f"Bearer {API_TOKEN}"}
|
42 |
+
else:
|
43 |
+
# Validate the API key if it's an API call
|
44 |
+
if huggingface_api_key == "":
|
45 |
+
raise gr.Error("API key is required for API calls.")
|
46 |
+
headers = {"Authorization": f"Bearer {huggingface_api_key}"}
|
47 |
+
|
48 |
+
if prompt == "" or prompt is None:
|
49 |
+
return None
|
50 |
+
|
51 |
+
key = random.randint(0, 999)
|
52 |
+
|
53 |
+
prompt = GoogleTranslator(source='ru', target='en').translate(prompt)
|
54 |
+
print(f'\033[1mGeneration {key} translation:\033[0m {prompt}')
|
55 |
+
|
56 |
+
# Enhance the prompt using the API
|
57 |
+
enhanced_prompt = enhance_prompt(prompt)
|
58 |
+
print(f'\033[1mEnhanced Prompt:\033[0m {enhanced_prompt}')
|
59 |
+
|
60 |
+
prompt = f"{enhanced_prompt} | ultra detail, ultra elaboration, ultra quality, perfect."
|
61 |
+
print(f'\033[1mGeneration {key}:\033[0m {prompt}')
|
62 |
+
|
63 |
+
# If seed is -1, generate a random seed and use it
|
64 |
+
if seed == -1:
|
65 |
+
seed = random.randint(1, 1000000000)
|
66 |
+
|
67 |
+
payload = {
|
68 |
+
"inputs": prompt,
|
69 |
+
"is_negative": is_negative,
|
70 |
+
"steps": steps,
|
71 |
+
"cfg_scale": cfg_scale,
|
72 |
+
"seed": seed,
|
73 |
+
"strength": strength
|
74 |
+
}
|
75 |
+
|
76 |
+
response = requests.post(api_url, headers=headers, json=payload, timeout=timeout)
|
77 |
+
if response.status_code != 200:
|
78 |
+
print(f"Error: Failed to get image. Response status: {response.status_code}")
|
79 |
+
print(f"Response content: {response.text}")
|
80 |
+
if response.status_code == 503:
|
81 |
+
raise gr.Error(f"{response.status_code} : The model is being loaded")
|
82 |
+
raise gr.Error(f"{response.status_code}")
|
83 |
+
|
84 |
try:
|
85 |
+
image_bytes = response.content
|
86 |
+
image = Image.open(io.BytesIO(image_bytes))
|
87 |
+
print(f'\033[1mGeneration {key} completed!\033[0m ({prompt})')
|
88 |
+
|
89 |
+
# Save the image to a file and return the file path and seed
|
90 |
+
output_path = f"./output_{key}.png"
|
91 |
+
image.save(output_path)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
92 |
|
93 |
+
return output_path, seed
|
94 |
+
except Exception as e:
|
95 |
+
print(f"Error when trying to open the image: {e}")
|
96 |
+
return None, None
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
97 |
|
98 |
css = """
|
99 |
#app-container {
|