Update app.py
Browse files
app.py
CHANGED
@@ -6,19 +6,20 @@ from transformers import BlipProcessor, BlipForConditionalGeneration
|
|
6 |
processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
|
7 |
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large")
|
8 |
|
9 |
-
img_url = 'https://storage.googleapis.com/sfr-vision-language-research/BLIP/demo.jpg'
|
10 |
-
raw_image = Image.open(requests.get(img_url, stream=True).raw).convert('RGB')
|
11 |
-
|
12 |
def caption(img):
|
13 |
raw_image = Image.open(img).convert('RGB')
|
14 |
|
15 |
inputs = processor(raw_image, return_tensors="pt")
|
16 |
|
17 |
-
out = model.generate(**inputs, min_length=
|
18 |
return processor.decode(out[0], skip_special_tokens=True)
|
19 |
|
20 |
def greet(img):
|
21 |
return caption(img)
|
22 |
|
23 |
-
iface = gr.Interface(fn=greet,
|
|
|
|
|
|
|
|
|
24 |
iface.launch()
|
|
|
6 |
processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
|
7 |
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large")
|
8 |
|
|
|
|
|
|
|
9 |
def caption(img):
|
10 |
raw_image = Image.open(img).convert('RGB')
|
11 |
|
12 |
inputs = processor(raw_image, return_tensors="pt")
|
13 |
|
14 |
+
out = model.generate(**inputs, min_length=40, max_length=200)
|
15 |
return processor.decode(out[0], skip_special_tokens=True)
|
16 |
|
17 |
def greet(img):
|
18 |
return caption(img)
|
19 |
|
20 |
+
iface = gr.Interface(fn=greet,
|
21 |
+
title='Blip Image Captioning Large',
|
22 |
+
inputs=gr.Image(type='filepath'),
|
23 |
+
outputs="text",
|
24 |
+
theme = gr.themes.Base(primary_hue="teal",secondary_hue="teal",neutral_hue="slate"),)
|
25 |
iface.launch()
|