Spaces:
Sleeping
Sleeping
Commit
·
e092e8d
1
Parent(s):
712c19c
Added Xai
Browse files
app.py
CHANGED
@@ -1,7 +1,6 @@
|
|
1 |
# %%
|
2 |
import gradio as gr
|
3 |
import tensorflow as tf
|
4 |
-
import numpy as np
|
5 |
import cv2
|
6 |
import os
|
7 |
|
@@ -18,12 +17,25 @@ if not os.path.exists(destination):
|
|
18 |
print('Repository cloned successfully.')
|
19 |
except subprocess.CalledProcessError as e:
|
20 |
print(f'Error cloning repository: {e.output.decode()}')
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
# %%
|
22 |
with open(f'{model_folder}/labels.txt', 'r') as f:
|
23 |
labels = f.read().split('\n')
|
24 |
|
25 |
# model = tf.saved_model.load(f'{model_folder}/last_layer.hdf5')
|
26 |
-
model = tf.keras.models.load_model(f'{model_folder}/last_layer.hdf5')
|
|
|
27 |
# %%
|
28 |
def classify_image(inp):
|
29 |
inp = cv2.resize(inp, (224,224,))
|
@@ -34,7 +46,30 @@ def classify_image(inp):
|
|
34 |
confidences = {labels[i]: float(prediction[i]) for i in range(len(labels))}
|
35 |
return confidences
|
36 |
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
# %%
|
2 |
import gradio as gr
|
3 |
import tensorflow as tf
|
|
|
4 |
import cv2
|
5 |
import os
|
6 |
|
|
|
17 |
print('Repository cloned successfully.')
|
18 |
except subprocess.CalledProcessError as e:
|
19 |
print(f'Error cloning repository: {e.output.decode()}')
|
20 |
+
|
21 |
+
if not os.path.exists(destination):
|
22 |
+
import subprocess
|
23 |
+
repo_url = os.getenv("GIT_CORE")
|
24 |
+
command = f'git clone {repo_url}'
|
25 |
+
try:
|
26 |
+
subprocess.check_output(command, stderr=subprocess.STDOUT, shell=True)#, env=env)
|
27 |
+
print('Repository cloned successfully.')
|
28 |
+
except subprocess.CalledProcessError as e:
|
29 |
+
print(f'Error cloning repository: {e.output.decode()}')
|
30 |
+
|
31 |
+
from explainer_tf_mobilenetv2.explainer import explainer
|
32 |
# %%
|
33 |
with open(f'{model_folder}/labels.txt', 'r') as f:
|
34 |
labels = f.read().split('\n')
|
35 |
|
36 |
# model = tf.saved_model.load(f'{model_folder}/last_layer.hdf5')
|
37 |
+
# model = tf.keras.models.load_model(f'{model_folder}/last_layer.hdf5')
|
38 |
+
model = tf.keras.models.load_model(f'{model_folder}/MobileNetV2_last_layer.hdf5')
|
39 |
# %%
|
40 |
def classify_image(inp):
|
41 |
inp = cv2.resize(inp, (224,224,))
|
|
|
46 |
confidences = {labels[i]: float(prediction[i]) for i in range(len(labels))}
|
47 |
return confidences
|
48 |
|
49 |
+
def explainer_wrapper(inp):
|
50 |
+
return explainer(inp, model)
|
51 |
+
|
52 |
+
with gr.Blocks() as demo:
|
53 |
+
with gr.Column():
|
54 |
+
with gr.Row():
|
55 |
+
with gr.Column():
|
56 |
+
image = gr.inputs.Image(shape=(224, 224))
|
57 |
+
with gr.Row():
|
58 |
+
classify = gr.Button("Classify")
|
59 |
+
interpret = gr.Button("Interpret")
|
60 |
+
|
61 |
+
label = gr.outputs.Label(num_top_classes=3)
|
62 |
+
interpretation = gr.Plot(label="Interpretation")
|
63 |
+
# interpretation = gr.outputs.Image(type="numpy", label="Interpretation")
|
64 |
+
gr.Examples(["TomatoHealthy2.jpg", "TomatoYellowCurlVirus3.jpg", "AppleCedarRust3.jpg"],
|
65 |
+
inputs=[image],)
|
66 |
+
classify.click(classify_image, image, label, queue=True)
|
67 |
+
interpret.click(explainer_wrapper, image, interpretation, queue=True)
|
68 |
+
|
69 |
+
|
70 |
+
demo.queue(concurrency_count=3).launch()
|
71 |
+
#%%
|
72 |
+
# gr.Interface(fn=classify_image,
|
73 |
+
# inputs=gr.Image(shape=(224, 224)),
|
74 |
+
# outputs=gr.Label(num_top_classes=3),
|
75 |
+
# examples=["TomatoHealthy2.jpg", "TomatoYellowCurlVirus3.jpg", "AppleCedarRust3.jpg"]).launch()
|