Spaces:
Running
on
Zero
Running
on
Zero
Delete traffic_density.py
Browse files- traffic_density.py +0 -43
traffic_density.py
DELETED
@@ -1,43 +0,0 @@
|
|
1 |
-
import gradio as gr
|
2 |
-
import spaces
|
3 |
-
from transformers import AutoImageProcessor
|
4 |
-
from transformers import SiglipForImageClassification
|
5 |
-
from transformers.image_utils import load_image
|
6 |
-
from PIL import Image
|
7 |
-
import torch
|
8 |
-
|
9 |
-
# Load model and processor
|
10 |
-
model_name = "prithivMLmods/Traffic-Density-Classification"
|
11 |
-
model = SiglipForImageClassification.from_pretrained(model_name)
|
12 |
-
processor = AutoImageProcessor.from_pretrained(model_name)
|
13 |
-
|
14 |
-
@spaces.GPU
|
15 |
-
def traffic_density_classification(image):
|
16 |
-
"""Predicts traffic density category for an image."""
|
17 |
-
image = Image.fromarray(image).convert("RGB")
|
18 |
-
inputs = processor(images=image, return_tensors="pt")
|
19 |
-
|
20 |
-
with torch.no_grad():
|
21 |
-
outputs = model(**inputs)
|
22 |
-
logits = outputs.logits
|
23 |
-
probs = torch.nn.functional.softmax(logits, dim=1).squeeze().tolist()
|
24 |
-
|
25 |
-
labels = {
|
26 |
-
"0": "high-traffic", "1": "low-traffic", "2": "medium-traffic", "3": "no-traffic"
|
27 |
-
}
|
28 |
-
predictions = {labels[str(i)]: round(probs[i], 3) for i in range(len(probs))}
|
29 |
-
|
30 |
-
return predictions
|
31 |
-
|
32 |
-
# Create Gradio interface
|
33 |
-
iface = gr.Interface(
|
34 |
-
fn=traffic_density_classification,
|
35 |
-
inputs=gr.Image(type="numpy"),
|
36 |
-
outputs=gr.Label(label="Prediction Scores"),
|
37 |
-
title="Traffic Density Classification",
|
38 |
-
description="Upload an image to classify it into one of the 4 traffic density categories."
|
39 |
-
)
|
40 |
-
|
41 |
-
# Launch the app
|
42 |
-
if __name__ == "__main__":
|
43 |
-
iface.launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|