Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
@@ -1,15 +1,21 @@
|
|
1 |
import gradio as gr
|
|
|
|
|
|
|
|
|
|
|
2 |
|
3 |
-
# Define the function to predict using the model
|
4 |
def predict_image(img):
|
5 |
-
|
6 |
-
|
|
|
|
|
7 |
|
8 |
# Create the interface
|
9 |
iface = gr.Interface(
|
10 |
fn=predict_image,
|
11 |
-
inputs=gr.Image(),
|
12 |
-
outputs="
|
13 |
live=True,
|
14 |
capture_session=True,
|
15 |
title="Image recognition",
|
|
|
1 |
import gradio as gr
|
2 |
+
from transformers import ViTForImageClassification, ViTProcessor
|
3 |
+
|
4 |
+
# Load the model and processor
|
5 |
+
model = ViTForImageClassification.from_pretrained("google/vit-base-patch16-224")
|
6 |
+
processor = ViTProcessor.from_pretrained("google/vit-base-patch16-224")
|
7 |
|
|
|
8 |
def predict_image(img):
|
9 |
+
inputs = processor(img, return_tensors="pt")
|
10 |
+
outputs = model(**inputs)
|
11 |
+
predictions = outputs.logits.argmax(-1)
|
12 |
+
return model.config.id2label[predictions.item()]
|
13 |
|
14 |
# Create the interface
|
15 |
iface = gr.Interface(
|
16 |
fn=predict_image,
|
17 |
+
inputs=gr.Image(shape=(224, 224)),
|
18 |
+
outputs="text",
|
19 |
live=True,
|
20 |
capture_session=True,
|
21 |
title="Image recognition",
|