Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -1,3 +1,30 @@
|
|
1 |
import gradio as gr
|
|
|
|
|
|
|
2 |
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
+
from transformers import CLIPProcessor, CLIPModel
|
3 |
+
from PIL import Image
|
4 |
+
import requests
|
5 |
|
6 |
+
# Load the model and processor
|
7 |
+
model = CLIPModel.from_pretrained("geolocal/StreetCLIP")
|
8 |
+
processor = CLIPProcessor.from_pretrained("geolocal/StreetCLIP")
|
9 |
+
|
10 |
+
def classify_image(image):
|
11 |
+
# Preprocess the image
|
12 |
+
inputs = processor(images=image, return_tensors="pt")
|
13 |
+
# Perform the inference
|
14 |
+
outputs = model(**inputs)
|
15 |
+
# Postprocess the outputs
|
16 |
+
logits_per_image = outputs.logits_per_image # this is the image-text similarity score
|
17 |
+
probs = logits_per_image.softmax(dim=1) # we can use softmax to get probabilities
|
18 |
+
return probs
|
19 |
+
|
20 |
+
# Define Gradio interface
|
21 |
+
iface = gr.Interface(
|
22 |
+
fn=classify_image,
|
23 |
+
inputs=gr.inputs.Image(type="pil"),
|
24 |
+
outputs="text",
|
25 |
+
title="Geolocal StreetCLIP Classification",
|
26 |
+
description="Upload an image to classify using Geolocal StreetCLIP"
|
27 |
+
)
|
28 |
+
|
29 |
+
# Launch the interface
|
30 |
+
iface.launch()
|