Create main.py
Browse files
main.py
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from tensorflow import keras
|
3 |
+
import numpy as np
|
4 |
+
from huggingface_hub import HfApi
|
5 |
+
import h5py
|
6 |
+
from io import BytesIO
|
7 |
+
|
8 |
+
# Authenticate and read the custom model from Hugging Face Spaces
|
9 |
+
hf_api = HfApi()
|
10 |
+
model_url = hf_api.presigned_url('dhhd255', 'idk_test', filename='best_model.h5', token='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')
|
11 |
+
r = requests.get(model_url)
|
12 |
+
model_file = h5py.File(BytesIO(r.content), 'r')
|
13 |
+
|
14 |
+
# Load your custom model
|
15 |
+
model = keras.models.load_model(model_file)
|
16 |
+
|
17 |
+
def image_classifier(inp):
|
18 |
+
# Preprocess the input image
|
19 |
+
inp = np.array(inp)
|
20 |
+
inp = inp / 255.0
|
21 |
+
inp = np.expand_dims(inp, axis=0)
|
22 |
+
|
23 |
+
# Use your custom model for inference
|
24 |
+
predictions = model.predict(inp)
|
25 |
+
|
26 |
+
# Process the predictions and return the result
|
27 |
+
result = {}
|
28 |
+
for i, prediction in enumerate(predictions[0]):
|
29 |
+
label = f'Label {i+1}'
|
30 |
+
result[label] = prediction
|
31 |
+
|
32 |
+
return result
|
33 |
+
|
34 |
+
demo = gr.Interface(fn=image_classifier, inputs='image', outputs='label')
|
35 |
+
demo.launch()
|