Spaces:
Sleeping
Sleeping
Upload app.py
Browse files
app.py
ADDED
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
### 1. Imports
|
3 |
+
|
4 |
+
import gradio as gr
|
5 |
+
import os
|
6 |
+
import torch
|
7 |
+
|
8 |
+
from model import create_model_alexnet
|
9 |
+
from timeit import default_timer as timer
|
10 |
+
from typing import Tuple, Dict
|
11 |
+
|
12 |
+
### 2. Model and transforms preparation ###
|
13 |
+
|
14 |
+
# Create model_alexnet
|
15 |
+
model_alexnet, transforms = create_model_alexnet( num_classes=2)
|
16 |
+
|
17 |
+
# Load saved weights
|
18 |
+
model_alexnet.load_state_dict(torch.load(f="cat_dog_classifier.pth", map_location=torch.device("cpu"))) # load to CPU
|
19 |
+
|
20 |
+
### 3. Predict function ###
|
21 |
+
|
22 |
+
# Create predict function
|
23 |
+
def predict(img):
|
24 |
+
|
25 |
+
# Start the timer
|
26 |
+
start_time = timer()
|
27 |
+
|
28 |
+
model_alexnet.eval()
|
29 |
+
|
30 |
+
# Reading the image and size transformation
|
31 |
+
features = Image.open(img)
|
32 |
+
img = auto_transform(features).unsqueeze(0)
|
33 |
+
|
34 |
+
with torch.inference_mode():
|
35 |
+
output = model_alexnet(img).to(device)
|
36 |
+
_, predicted = torch.max(output, 1)
|
37 |
+
|
38 |
+
# Create a prediction label and prediction probability dictionary for each prediction class
|
39 |
+
# This is the required format for Gradio's output parameter
|
40 |
+
pred_labels_and_probs = 'dog' if predicted.item() ==1 else 'cat'
|
41 |
+
|
42 |
+
# Calculate the prediction time
|
43 |
+
pred_time = round(timer() - start_time, 5)
|
44 |
+
|
45 |
+
# Return the prediction dictionary and prediction time
|
46 |
+
return pred_labels, pred_time
|
47 |
+
|
48 |
+
|
49 |
+
### 4. Gradio app ###
|
50 |
+
|
51 |
+
import gradio as gr
|
52 |
+
|
53 |
+
# Create title, description and article strings
|
54 |
+
title = "Classification Demo"
|
55 |
+
description = "Cat/Dog classification - Transfer Learning "
|
56 |
+
|
57 |
+
# Create the Gradio demo
|
58 |
+
demo = gr.Interface(fn=predict, # mapping function from input to output
|
59 |
+
inputs=gr.Image(type='filepath'), # what are the inputs?
|
60 |
+
outputs=[gr.Label(label="Predictions"), # what are the outputs?
|
61 |
+
gr.Number(label="Prediction time (s)")], # our fn has two outputs, therefore we have two outputs
|
62 |
+
#examples=example_list,
|
63 |
+
title=title,
|
64 |
+
description=description,)
|
65 |
+
|
66 |
+
# Launch the demo!
|
67 |
+
demo.launch()
|