Upload 5 files

.gitattributes

@@ -37,3 +37,4 @@ deliveries.csv filter=lfs diff=lfs merge=lfs -text
 gru_match_simulation_plot.gif filter=lfs diff=lfs merge=lfs -text
 gru_rw_predictor.keras filter=lfs diff=lfs merge=lfs -text
 match_momentum_dashboard.gif filter=lfs diff=lfs merge=lfs -text
+match_live_predictor/gru_score_predictor_rw.keras filter=lfs diff=lfs merge=lfs -text

match_live_predictor/gru_score_predictor_rw.keras

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6b37bcf62ca560d22dc94033e02817affe1753f720bc9dbdb02086c79fd768e
+size 211036

match_live_predictor/live_predictor.py

@@ -0,0 +1,107 @@
+# live_predictor.py
+
+import streamlit as st
+import numpy as np
+import pandas as pd
+import datetime
+import os
+import matplotlib.pyplot as plt
+from tensorflow.keras.models import load_model
+import joblib
+
+# App config
+st.set_page_config(page_title="🏏 IPL Live Match Predictor", layout="wide")
+st.title("🏏 IPL Live Match Predictor")
+
+# Match Setup
+st.header("📋 Match Setup")
+col1, col2, col3 = st.columns(3)
+
+with col1:
+    team_a = st.selectbox("🏏 Team A", ["CSK", "MI", "RCB", "GT", "RR", "LSG", "KKR", "SRH"])
+with col2:
+    team_b = st.selectbox("🏏 Team B", ["CSK", "MI", "RCB", "GT", "RR", "LSG", "KKR", "SRH"])
+with col3:
+    venue = st.selectbox("📍 Venue", ["Wankhede Stadium", "Chepauk", "Eden Gardens", "Narendra Modi Stadium", "Chinnaswamy Stadium"])
+
+match_date = st.date_input("📅 Match Date", datetime.date.today())
+match_status = st.radio("📶 Match Status", ["Scheduled", "In Progress", "Completed"])
+
+# Section: Simulate Match Data
+st.header("🎲 Simulate Over-wise Runs & Wickets")
+
+if st.button("🔁 Generate Match"):
+    overs = list(range(1, 21))
+    runs_this_over = np.random.randint(0, 21, size=20)  # realistic run range
+    wickets_this_over = np.random.binomial(1, 0.3, size=20)
+
+    cumulative_runs = np.cumsum(runs_this_over)
+    cumulative_wickets = np.cumsum(wickets_this_over)
+
+    df = pd.DataFrame({
+        "Over": overs,
+        "Runs_This_Over": runs_this_over,
+        "Cumulative_Runs": cumulative_runs,
+        "Wickets_This_Over": wickets_this_over,
+        "Cumulative_Wickets": cumulative_wickets
+    })
+
+    st.session_state["live_df"] = df
+    st.success(f"📝 {team_a} vs {team_b} at {venue} on {match_date}")
+    st.dataframe(df, use_container_width=True)
+
+# GRU Prediction
+if "live_df" in st.session_state:
+    df = st.session_state["live_df"]
+    st.subheader("📈 Predicted Final Score (GRU Model)")
+
+    try:
+        # Load model and scalers
+        model_path = os.path.join("match_live_predictor", "gru_score_predictor_rw.keras")
+        input_scaler_path = os.path.join("match_live_predictor", "scaler_input_rw.save")
+        output_scaler_path = os.path.join("match_live_predictor", "scaler_output_rw.save")
+
+        model = load_model(model_path, compile=False)
+        scaler_input = joblib.load(input_scaler_path)
+        scaler_output = joblib.load(output_scaler_path)
+
+        # Prepare input
+        match_input = df[["Runs_This_Over", "Wickets_This_Over"]].values
+        padded_input = np.pad(match_input, ((0, 20 - len(match_input)), (0, 0)), mode="constant")
+
+        scaled_input = scaler_input.transform(padded_input)
+        reshaped_input = scaled_input.reshape(1, 20, 2)
+
+        # Predict final score
+        pred_scaled = model.predict(reshaped_input, verbose=0)
+        predicted_final = scaler_output.inverse_transform(pred_scaled)[0][0]
+
+        # Display predicted score
+        st.success(f"🎯 Predicted Final Score: {predicted_final:.2f} runs")
+
+        # Visualization
+        fig, ax1 = plt.subplots(figsize=(10, 4))
+
+        ax1.plot(df["Over"], df["Cumulative_Runs"], marker='o', label="Cumulative Runs", color="green")
+        ax1.set_xlabel("Overs")
+        ax1.set_ylabel("Runs", color="green")
+        ax1.tick_params(axis='y', labelcolor="green")
+        ax1.set_title("📊 Match Progression & GRU Prediction")
+
+        ax2 = ax1.twinx()
+        ax2.bar(df["Over"], df["Wickets_This_Over"], alpha=0.3, color="red", label="Wickets")
+        ax2.set_ylabel("Wickets", color="red")
+        ax2.tick_params(axis='y', labelcolor="red")
+
+        plt.axhline(y=predicted_final, color="blue", linestyle="--", label="Predicted Final")
+        fig.legend(loc="upper left", bbox_to_anchor=(0.1, 0.85))
+        st.pyplot(fig)
+
+    except Exception as e:
+        st.error(f"⚠️ Prediction module error: {e}")
+
+# Footer
+st.markdown("""
+---
+👨‍💻 Built by Dinesh Kumar | Powered by Streamlit, GRU, NumPy
+""")

match_live_predictor/train_gru_rw_model.py

@@ -0,0 +1,55 @@
+# train_gru_rw_model.py
+
+import numpy as np
+import pandas as pd
+from tensorflow.keras.models import Sequential
+from tensorflow.keras.layers import GRU, Dense, Input
+from sklearn.preprocessing import MinMaxScaler
+import joblib
+import os
+
+# Simulate realistic training data
+np.random.seed(42)
+X = []
+y = []
+
+for _ in range(1000):
+    runs = np.random.randint(1, 15, size=20)
+    wickets = np.random.binomial(1, 0.3, size=20)
+    
+    input_seq = np.column_stack((runs, wickets))
+    X.append(input_seq)
+    
+    final_score = np.sum(runs)
+    y.append(final_score)
+
+X = np.array(X)                  # Shape: (1000, 20, 2)
+y = np.array(y).reshape(-1, 1)   # Shape: (1000, 1)
+
+# Fit scalers
+scaler_input = MinMaxScaler()
+scaler_output = MinMaxScaler()
+
+X_scaled = np.array([scaler_input.fit_transform(x) for x in X])
+y_scaled = scaler_output.fit_transform(y)
+
+# Save scalers
+os.makedirs("match_live_predictor", exist_ok=True)
+joblib.dump(scaler_input, "match_live_predictor/scaler_input_rw.save")
+joblib.dump(scaler_output, "match_live_predictor/scaler_output_rw.save")
+
+# Build GRU model
+model = Sequential([
+    Input(shape=(20, 2)),
+    GRU(64, return_sequences=False),
+    Dense(32, activation='relu'),
+    Dense(1)
+])
+
+model.compile(optimizer='adam', loss='mse')
+model.fit(X_scaled, y_scaled, epochs=50, batch_size=32, verbose=1)
+
+# Save model
+model.save("match_live_predictor/gru_score_predictor_rw.keras")
+print("GRU model retrained and saved successfully!")
+