Refactor app.py: enhance prediction function, update title and description, and improve image handling
Browse files
app.py
CHANGED
@@ -1,47 +1,72 @@
|
|
1 |
-
|
2 |
import gradio as gr
|
3 |
import tensorflow as tf
|
4 |
import cv2
|
5 |
|
6 |
-
#
|
7 |
-
|
8 |
|
9 |
-
#
|
10 |
-
|
11 |
-
"<center>"
|
12 |
-
"<img src='./mnist-classes.png' width=400>"
|
13 |
-
"The robot was trained to classify numbers (from 0 to 9). To test it, write your number in the space provided."
|
14 |
-
"</center>"
|
15 |
-
)
|
16 |
|
17 |
-
|
18 |
-
|
|
|
|
|
19 |
|
20 |
-
#
|
21 |
-
|
|
|
22 |
|
23 |
-
#
|
24 |
-
|
|
|
25 |
|
26 |
-
#
|
27 |
-
|
|
|
|
|
|
|
28 |
|
29 |
-
#
|
30 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
31 |
|
32 |
-
|
33 |
-
|
34 |
-
|
|
|
35 |
|
36 |
-
|
37 |
-
|
|
|
38 |
|
39 |
-
|
40 |
-
|
41 |
|
42 |
-
#
|
43 |
-
|
44 |
|
45 |
-
|
46 |
-
|
47 |
-
interface.launch(share=True)
|
|
|
1 |
+
import numpy as np
|
2 |
import gradio as gr
|
3 |
import tensorflow as tf
|
4 |
import cv2
|
5 |
|
6 |
+
# Load the trained MNIST model
|
7 |
+
model = tf.keras.models.load_model("./sketch_recognition_numbers_model.h5")
|
8 |
|
9 |
+
# Class names (0 to 9)
|
10 |
+
labels = ["zero", "one", "two", "three", "four", "five", "six", "seven", "eight", "nine"]
|
|
|
|
|
|
|
|
|
|
|
11 |
|
12 |
+
def predict(data):
|
13 |
+
# Extract the 'composite' key from the input dictionary
|
14 |
+
img = data["composite"]
|
15 |
+
img = np.array(img)
|
16 |
|
17 |
+
# Convert RGBA to RGB if needed
|
18 |
+
if img.shape[-1] == 4: # RGBA
|
19 |
+
img = cv2.cvtColor(img, cv2.COLOR_RGBA2RGB)
|
20 |
|
21 |
+
# Convert RGB to Grayscale
|
22 |
+
if img.shape[-1] == 3: # RGB
|
23 |
+
img = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
|
24 |
|
25 |
+
# Resize image to 28x28
|
26 |
+
img = cv2.resize(img, (28, 28))
|
27 |
+
|
28 |
+
# Normalize pixel values to [0, 1]
|
29 |
+
img = img / 255.0
|
30 |
|
31 |
+
# Reshape to match model input (1, 28, 28, 1)
|
32 |
+
img = img.reshape(1, 28, 28, 1)
|
33 |
+
|
34 |
+
# Model predictions
|
35 |
+
preds = model.predict(img)[0]
|
36 |
+
|
37 |
+
# Get top 3 classes
|
38 |
+
top_3_classes = np.argsort(preds)[-3:][::-1]
|
39 |
+
top_3_probs = preds[top_3_classes]
|
40 |
+
class_names = [labels[i] for i in top_3_classes]
|
41 |
+
|
42 |
+
# Return top 3 predictions as a dictionary
|
43 |
+
return {class_names[i]: float(top_3_probs[i]) for i in range(3)}
|
44 |
+
|
45 |
+
# Title and description
|
46 |
+
title = "Welcome to your first sketch recognition app!"
|
47 |
+
head = (
|
48 |
+
"<center>"
|
49 |
+
"<img src='./mnist-classes.png' width=400>"
|
50 |
+
"<p>The model is trained to classify numbers (from 0 to 9). "
|
51 |
+
"To test it, draw your number in the space provided (use the editing tools in the image editor).</p>"
|
52 |
+
"</center>"
|
53 |
+
)
|
54 |
+
ref = "Find the complete code [here](https://github.com/ovh/ai-training-examples/tree/main/apps/gradio/sketch-recognition)."
|
55 |
|
56 |
+
with gr.Blocks(title=title) as demo:
|
57 |
+
# Display title and description
|
58 |
+
gr.Markdown(head)
|
59 |
+
gr.Markdown(ref)
|
60 |
|
61 |
+
with gr.Row():
|
62 |
+
# Using ImageEditor with type='numpy'
|
63 |
+
im = gr.ImageEditor(type="numpy", label="Draw your digit here (use brush and eraser)")
|
64 |
|
65 |
+
# Output label (top 3 predictions)
|
66 |
+
label = gr.Label(num_top_classes=3, label="Predictions")
|
67 |
|
68 |
+
# Trigger prediction whenever the image changes
|
69 |
+
im.change(predict, inputs=im, outputs=label, show_progress="hidden")
|
70 |
|
71 |
+
if __name__ == "__main__":
|
72 |
+
demo.launch(share=True)
|
|