Update model.py
Browse files
model.py
CHANGED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import requests
|
| 2 |
+
from PIL import Image
|
| 3 |
+
from transformers import BlipProcessor, BlipForConditionalGeneration
|
| 4 |
+
|
| 5 |
+
# Initialize processor and model
|
| 6 |
+
processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
|
| 7 |
+
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large").to("cuda")
|
| 8 |
+
|
| 9 |
+
# Function to process and caption an image from a URL
|
| 10 |
+
def caption_image(image_url):
|
| 11 |
+
try:
|
| 12 |
+
# Load image from the provided URL
|
| 13 |
+
raw_image = Image.open(requests.get(image_url, stream=True).raw).convert('RGB')
|
| 14 |
+
|
| 15 |
+
# Conditional image captioning
|
| 16 |
+
text = "a photography of"
|
| 17 |
+
inputs = processor(raw_image, text, return_tensors="pt").to("cuda")
|
| 18 |
+
out = model.generate(**inputs)
|
| 19 |
+
conditional_caption = processor.decode(out[0], skip_special_tokens=True)
|
| 20 |
+
|
| 21 |
+
# Unconditional image captioning
|
| 22 |
+
inputs = processor(raw_image, return_tensors="pt").to("cuda")
|
| 23 |
+
out = model.generate(**inputs)
|
| 24 |
+
unconditional_caption = processor.decode(out[0], skip_special_tokens=True)
|
| 25 |
+
|
| 26 |
+
# Print the results
|
| 27 |
+
print("Conditional Caption:", conditional_caption)
|
| 28 |
+
print("Unconditional Caption:", unconditional_caption)
|
| 29 |
+
|
| 30 |
+
except Exception as e:
|
| 31 |
+
print(f"Error occurred: {e}")
|
| 32 |
+
|
| 33 |
+
# Get image URL from user input
|
| 34 |
+
image_url = input("Enter the image URL: ")
|
| 35 |
+
caption_image(image_url)
|