Spaces:
Runtime error
Runtime error
app.py
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from transformers import AutoFeatureExtractor, ResNetForImageClassification
|
2 |
+
import torch
|
3 |
+
import gradio as gr
|
4 |
+
|
5 |
+
# load model
|
6 |
+
feature_extractor = AutoFeatureExtractor.from_pretrained("microsoft/resnet-50")
|
7 |
+
model = ResNetForImageClassification.from_pretrained("microsoft/resnet-50")
|
8 |
+
|
9 |
+
def predict(image):
|
10 |
+
|
11 |
+
inputs = feature_extractor(image, return_tensors="pt")
|
12 |
+
with torch.no_grad():
|
13 |
+
logits = model(**inputs).logits
|
14 |
+
|
15 |
+
# model predicts one of the 1000 ImageNet classes
|
16 |
+
predicted_label = logits.argmax(-1).item()
|
17 |
+
print(model.config.id2label[predicted_label])
|
18 |
+
|
19 |
+
# setup Gradio interface
|
20 |
+
title = "Image classifier"
|
21 |
+
description = "Image classification with pretrained resnet50 model"
|
22 |
+
#examples = ['elephant.jpg']
|
23 |
+
interpretation='default'
|
24 |
+
enable_queue=True
|
25 |
+
|
26 |
+
gr.Interface(
|
27 |
+
fn=predict,
|
28 |
+
inputs=gr.inputs.Image(),
|
29 |
+
outputs=gr.outputs.Label(num_top_classes=1),
|
30 |
+
title=title,
|
31 |
+
description=description,
|
32 |
+
#examples=examples,
|
33 |
+
interpretation=interpretation,
|
34 |
+
enable_queue=enable_queue
|
35 |
+
).launch()
|