Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -1,36 +1,52 @@
|
|
1 |
import gradio as gr
|
2 |
import tensorflow as tf
|
3 |
import numpy as np
|
|
|
4 |
|
5 |
-
#
|
6 |
-
model_path =
|
7 |
-
|
|
|
8 |
|
9 |
-
#
|
10 |
-
|
11 |
-
|
|
|
|
|
12 |
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
# Example placeholder for preprocessing:
|
17 |
-
question_input = preprocess_question(question) # Replace with actual preprocessing
|
18 |
-
|
19 |
-
# Get model prediction
|
20 |
-
prediction = model.predict(question_input)
|
21 |
-
|
22 |
-
# Postprocess the prediction into readable text
|
23 |
-
answer = postprocess_prediction(prediction) # Replace with actual postprocessing
|
24 |
-
|
25 |
-
return answer
|
26 |
|
27 |
-
|
28 |
-
|
29 |
-
inputs="text",
|
30 |
-
outputs="text",
|
31 |
-
title="Doctor AI - Offline",
|
32 |
-
description="Ask Doctor AI your medical questions. **This chatbot works completely offline**.",
|
33 |
-
article="This model is trained on a custom dataset and designed to work without an internet connection.")
|
34 |
|
35 |
-
#
|
36 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
import tensorflow as tf
|
3 |
import numpy as np
|
4 |
+
import pickle
|
5 |
|
6 |
+
# Paths to model and tokenizer
|
7 |
+
model_path = "doctor_ai_model.h5"
|
8 |
+
tokenizer_path = "tokenizer.pkl"
|
9 |
+
label_encoder_path = "label_encoder.pkl"
|
10 |
|
11 |
+
# Load the trained model
|
12 |
+
try:
|
13 |
+
model = tf.keras.models.load_model(model_path)
|
14 |
+
except Exception as e:
|
15 |
+
print(f"Error loading model: {e}")
|
16 |
|
17 |
+
# Load the tokenizer and label encoder
|
18 |
+
with open(tokenizer_path, 'rb') as handle:
|
19 |
+
tokenizer = pickle.load(handle)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
|
21 |
+
with open(label_encoder_path, 'rb') as handle:
|
22 |
+
label_encoder = pickle.load(handle)
|
|
|
|
|
|
|
|
|
|
|
23 |
|
24 |
+
# Define the prediction function
|
25 |
+
def predict_answer(question):
|
26 |
+
try:
|
27 |
+
# Tokenize the input question
|
28 |
+
seq = tokenizer.texts_to_sequences([question])
|
29 |
+
padded_seq = tf.keras.preprocessing.sequence.pad_sequences(seq, maxlen=27) # Adjust maxlen to match your model
|
30 |
+
|
31 |
+
# Make prediction
|
32 |
+
prediction = model.predict(padded_seq)
|
33 |
+
|
34 |
+
# Convert prediction to label
|
35 |
+
predicted_label = label_encoder.inverse_transform(np.argmax(prediction, axis=1))
|
36 |
+
|
37 |
+
return predicted_label[0]
|
38 |
+
except Exception as e:
|
39 |
+
return f"Error during prediction: {e}"
|
40 |
+
|
41 |
+
# Define the Gradio interface
|
42 |
+
iface = gr.Interface(
|
43 |
+
fn=predict_answer,
|
44 |
+
inputs=gr.inputs.Textbox(lines=2, placeholder="Enter your question..."),
|
45 |
+
outputs="text",
|
46 |
+
title="Doctor AI Chatbot",
|
47 |
+
description="This AI chatbot provides answers based on your medical-related questions. Works completely offline."
|
48 |
+
)
|
49 |
+
|
50 |
+
# Launch the Gradio app
|
51 |
+
if __name__ == "__main__":
|
52 |
+
iface.launch()
|