Create app.py

app.py

@@ -1,47 +1,31 @@
+from flask import Flask, request, jsonify
+import tensorflow as tf
+import numpy as np
 import os
-os.environ["CUDA_VISIBLE_DEVICES"] = "-1"  # Disable GPU and force CPU mode
 
-import gradio as gr
-import numpy as np
-import tensorflow as tf
+app = Flask(__name__)
 
+# Load the model
 model_path = "sleep_cognition_model.h5"
 
-# Check if the model file exists before loading
 if not os.path.exists(model_path):
-    raise FileNotFoundError(f"⚠️ Model file not found: {model_path}. Please upload it to Hugging Face Space.")
-
-
-# Load the trained model
-model = tf.keras.models.load_model("sleep_cognition_model.h5")  
-
-# Define prediction function
-def predict(TST, SE, WASO, REM_Sleep, Deep_Sleep, Number_of_Awakenings):
-    input_data = np.array([[TST, SE, WASO, REM_Sleep, Deep_Sleep, Number_of_Awakenings]])
-    prediction = model.predict(input_data)
-    return {
-        "Reaction Time": float(prediction[0][0]),
-        "Memory Recall Accuracy": float(prediction[0][1]),
-        "Attention Score": float(prediction[0][2]),
-        "Executive Function Score": float(prediction[0][3]),
-        "Mental Fatigue Index": float(prediction[0][4]),
-    }
-
-# Define Gradio interface
-iface = gr.Interface(
-    fn=predict,
-    inputs=[
-        gr.Number(label="Total Sleep Time (TST)"),
-        gr.Number(label="Sleep Efficiency (SE)"),
-        gr.Number(label="Wake After Sleep Onset (WASO)"),
-        gr.Number(label="REM Sleep (%)"),
-        gr.Number(label="Deep Sleep (%)"),
-        gr.Number(label="Number of Awakenings"),
-    ],
-    outputs="json",
-    title="Sleep & Cognitive Function Predictor",
-    description="Enter sleep parameters to predict cognitive function scores.",
-)
-
-# Run the app
-iface.launch()
+    raise FileNotFoundError(f"⚠️ Model file not found: {model_path}. Please check the path.")
+
+model = tf.keras.models.load_model(model_path)
+
+@app.route("/", methods=["GET"])
+def home():
+    return jsonify({"message": "Sleep Cognition Model API is running!"})
+
+@app.route("/predict", methods=["POST"])
+def predict():
+    try:
+        data = request.json["input"]
+        data = np.array(data).reshape(1, -1)  # Reshape input data if needed
+        prediction = model.predict(data)
+        return jsonify({"prediction": prediction.tolist()})
+    except Exception as e:
+        return jsonify({"error": str(e)}), 400
+
+if __name__ == "__main__":
+    app.run(host="0.0.0.0", port=5000, debug=True)