Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -44,14 +44,14 @@ def feifeichat(image):
|
|
| 44 |
yield "Error processing image"
|
| 45 |
return
|
| 46 |
|
| 47 |
-
#
|
| 48 |
messages = [
|
| 49 |
{
|
| 50 |
"role": "user",
|
| 51 |
"content": [
|
| 52 |
{
|
| 53 |
"type": "text",
|
| 54 |
-
"text": "Analyze the image and return strictly valid JSON format with keys 'title', 'description', 'meta_title' (max 60 chars), 'meta_description' (max 155 chars), 'tags' (array of 5-8 strings), 'style_breakdown' (object with keys: aesthetic, top, bottom, legwear, accessories), and 'faqs' (array of two objects with 'q' and 'a'),
|
| 55 |
},
|
| 56 |
{
|
| 57 |
"type": "image_url",
|
|
@@ -63,7 +63,6 @@ def feifeichat(image):
|
|
| 63 |
}
|
| 64 |
]
|
| 65 |
|
| 66 |
-
# response_format forcing JSON mode guarantees error-free data structural rendering
|
| 67 |
response = client.chat.completions.create(
|
| 68 |
model="meta-llama/llama-4-scout-17b-16e-instruct",
|
| 69 |
messages=messages,
|
|
|
|
| 44 |
yield "Error processing image"
|
| 45 |
return
|
| 46 |
|
| 47 |
+
# Expanded prompt explicitly pushing for narrative depth, key phrasing, and aesthetic richness
|
| 48 |
messages = [
|
| 49 |
{
|
| 50 |
"role": "user",
|
| 51 |
"content": [
|
| 52 |
{
|
| 53 |
"type": "text",
|
| 54 |
+
"text": "Analyze the image and return strictly valid JSON format with keys 'title', 'description', 'meta_title' (max 60 chars), 'meta_description' (max 155 chars), 'tags' (array of 5-8 strings), 'style_breakdown' (object with keys: aesthetic, top, bottom, legwear, accessories), and 'faqs' (array of two objects with 'q' and 'a'). The language must be creative, highly descriptive, and stylish. Focus heavily on body details (slender, petite, curvy), specific age tones (teen, young adult), and descriptive vocabulary (sexy, cute, alluring, trendy, aesthetic) woven naturally into fluent paragraphs. Ensure titles are catchy, click-worthy lookbook headlines."
|
| 55 |
},
|
| 56 |
{
|
| 57 |
"type": "image_url",
|
|
|
|
| 63 |
}
|
| 64 |
]
|
| 65 |
|
|
|
|
| 66 |
response = client.chat.completions.create(
|
| 67 |
model="meta-llama/llama-4-scout-17b-16e-instruct",
|
| 68 |
messages=messages,
|