Spaces:
Sleeping
Sleeping
File size: 4,443 Bytes
dccac59 5ae4080 dccac59 5ae4080 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 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 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 | import spaces
import json
import gradio as gr
from inference import (
cnn_traffic,
lstm_energy,
gnn_fraud,
transformer_reports,
autoencoder_leaks,
)
@spaces.GPU(duration=15)
def cnn_fn(image):
if image is None:
return {"error": "No image provided"}
result = cnn_traffic.predict(image)
return {item["label"]: item["confidence"] for item in result["top_k"]}
# ---------- Module 02: LSTM ----------
def lstm_fn(history_text):
try:
values = [float(x.strip()) for x in history_text.replace("\n", ",").split(",") if x.strip()]
result = lstm_energy.forecast(values)
return f"Predicted next-day consumption: {result['forecast_kwh']:.2f} kWh"
except Exception as e:
return f"❌ Error: {e}"
# ---------- Module 03: GNN ----------
def gnn_fn(features_text, edges_text):
try:
features = json.loads(features_text)
edges = json.loads(edges_text)
result = gnn_fraud.predict_nodes(features, edges)
lines = [f"Node {p['node']}: {p['label']} (fraud prob: {p['fraud_probability']:.2%})"
for p in result["predictions"]]
return "\n".join(lines)
except Exception as e:
return f"❌ Error: {e}"
# ---------- Module 04: Transformer ----------
def transformer_fn(text):
result = transformer_reports.route(text)
return result["department"], result["all_scores"]
# ---------- Module 05: Autoencoder ----------
def leak_fn(readings_text):
try:
readings = [float(x.strip()) for x in readings_text.split(",") if x.strip()]
result = autoencoder_leaks.detect(readings)
emoji = {"normal": "🟢", "moderate": "🟡", "high": "🔴"}[result["severity"]]
return (
f"{emoji} Severity: {result['severity'].upper()}\n"
f"Anomaly score: {result['anomaly_score']:.6f}\n"
f"Threshold: {result['threshold']:.6f}\n"
f"Is leak: {result['is_leak']}"
)
except Exception as e:
return f"❌ Error: {e}"
# ---------- Build UI ----------
with gr.Blocks(title="Smart City Monitor") as demo:
gr.Markdown("# 🏙️ Smart City Monitor")
gr.Markdown("Five neural network architectures in one demo.")
with gr.Tab("🚦 Traffic Sign (CNN)"):
with gr.Row():
img_in = gr.Image(type="filepath", label="Upload a traffic sign")
lbl_out = gr.Label(num_top_classes=3, label="Prediction")
img_in.change(cnn_fn, inputs=img_in, outputs=lbl_out)
with gr.Tab("⚡ Energy Forecast (LSTM)"):
hist_in = gr.Textbox(
label="Last N daily kWh values (comma-separated)",
value="1200,1150,1300,1250,1400,1350,1280,1220,1190,1310,1420,1380,1260,1240,1210,1330,1450,1400,1290,1270,1230,1180,1300,1440,1390,1270,1250,1220,1200,1340",
lines=3,
)
hist_out = gr.Textbox(label="Forecast")
gr.Button("Predict").click(lstm_fn, inputs=hist_in, outputs=hist_out)
with gr.Tab("🕸️ Fraud Detection (GNN)"):
with gr.Row():
feat_in = gr.Code(
label="Node features (JSON)",
value='[[1.0,0.5],[0.5,1.0],[0.1,0.1],[5.0,5.0]]',
language="json",
)
edge_in = gr.Code(
label="Edge index (JSON)",
value='[[0,1,2,3],[1,2,3,0]]',
language="json",
)
gnn_out = gr.Textbox(label="Predictions", lines=6)
gr.Button("Detect").click(gnn_fn, inputs=[feat_in, edge_in], outputs=gnn_out)
with gr.Tab("📨 Citizen Reports (Transformer)"):
text_in = gr.Textbox(
label="Citizen report",
value="There is a broken pipe leaking water on Elm Street.",
lines=2,
)
dept_out = gr.Textbox(label="Routed to")
scores_out = gr.Label(label="All departments")
gr.Button("Route").click(transformer_fn, inputs=text_in, outputs=[dept_out, scores_out])
with gr.Tab("💧 Leak Detection (Autoencoder)"):
read_in = gr.Textbox(
label="Sensor readings (comma-separated)",
value="5.0,5.0,0.1,0.1,5.0,5.0,0.1,0.1,5.0,5.0",
lines=2,
)
leak_out = gr.Textbox(label="Result", lines=4)
gr.Button("Detect").click(leak_fn, inputs=read_in, outputs=leak_out)
if __name__ == "__main__":
demo.launch(server_name="0.0.0.0", server_port=7860)
|