Ich habe dieses Tensorflow Keras Model und das Script funktioniert eigentlich.
import yfinance as yf
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import tensorflow as tf
import os
from sklearn.preprocessing import MinMaxScaler
from sklearn.metrics import mean_squared_error, mean_absolute_percentage_error
from tensorflow.keras.models import Sequential, load_model
from tensorflow.keras.layers import LSTM, Dense, Dropout
# 1. Goldpreis-Daten laden
df = yf.download("GC=F", start="2015-01-01", end="2025-03-31")
df = df[['Close']].dropna()
# 2. Normalisieren
scaler = MinMaxScaler()
scaled = scaler.fit_transform(df)
# 3. Sequenzen für Seq2Seq erzeugen
def create_seq2seq_dataset(data, input_len, output_len):
X, y = [], []
for i in range(len(data) - input_len - output_len):
X.append(data[i:i + input_len])
y.append(data[i + input_len:i + input_len + output_len])
return np.array(X), np.array(y)
input_days = 365 # Anzahl der Tage für Eingabesequenzen
output_days = 30 # Anzahl der Tage für die Vorhersage
X, y = create_seq2seq_dataset(scaled, input_days, output_days)
# 4. Trainings-/Testdaten aufteilen
split = int(0.9 * len(X))
X_train, y_train = X[:split], y[:split]
X_test, y_test = X[split:], y[split:]
# 5. Modell definieren
model = Sequential([
LSTM(128, input_shape=(input_days, 1)),
Dropout(0.2),
Dense(output_days)
])
model.compile(optimizer='adam', loss='mse')
model.summary()
# 6. Modell trainieren
history = model.fit(X_train, y_train, epochs=10, batch_size=32, validation_split=0.1)
# 7. Modell speichern
model_dir = "gold_lstm_model"
os.makedirs(model_dir, exist_ok=True)
model_path = os.path.join(model_dir, "model.h5")
model.save(model_path)
# Modell optional laden
model = load_model(model_path)
# 8. Vorhersage für die nächsten 30 Tage
last_input = scaled[-input_days:].reshape(1, input_days, 1)
predicted_scaled = model.predict(last_input)[0]
# Korrekte De-Skalierung der Vorhersage
predicted_prices = scaler.inverse_transform(predicted_scaled.reshape(-1, 1))
# Zeige den letzten tatsächlichen Goldpreis und den ersten Vorhersagewert
print(f"Letzter realer Preis: {df['Close'].values[-1]}")
print(f"Vorhergesagter Preis (erste Vorhersage): {predicted_prices[0][0]}")
# 9. Visualisierung: 180 Tage Historie + 30 Tage Prognose
last_real = df['Close'].values[-input_days:]
future_index = pd.date_range(start=df.index[-1] + pd.Timedelta(days=1), periods=output_days, freq="B")
plt.figure(figsize=(14, 5))
plt.plot(df.index[-input_days:], last_real, label="Letzte 180 Tage (Real)", color='blue')
plt.plot(future_index, predicted_prices.flatten(), label="Vorhersage (30 Tage)", color='orange')
plt.axvline(x=df.index[-1], color='gray', linestyle='--', label='Heute')
plt.title("Goldpreis-Vorhersage mit Seq2Seq LSTM")
plt.xlabel("Datum")
plt.ylabel("Goldpreis in USD")
plt.legend()
plt.grid(True)
plt.tight_layout()
plt.show()
# 10. Metriken berechnen auf Testdaten
y_pred_test_scaled = model.predict(X_test)
y_pred_test = scaler.inverse_transform(y_pred_test_scaled.reshape(-1, 1))
y_test_flat = scaler.inverse_transform(y_test.reshape(-1, 1))
rmse = np.sqrt(mean_squared_error(y_test_flat, y_pred_test))
mape = mean_absolute_percentage_error(y_test_flat, y_pred_test) * 100
print(f"RMSE auf Testdaten: {rmse:.2f}")
print(f"MAPE auf Testdaten: {mape:.2f}%")
# 11. Vorhersage als CSV speichern
forecast_dates = pd.date_range(start=df.index[-1] + pd.Timedelta(days=1), periods=output_days, freq="B")
forecast_df = pd.DataFrame(predicted_prices, columns=["Forecast_Close"])
forecast_df["Date"] = forecast_dates
forecast_df = forecast_df[["Date", "Forecast_Close"]]
forecast_df.to_csv("gold_forecast.csv", index=False)
print("Vorhersage als 'gold_forecast.csv' gespeichert.")
print("Modell gespeichert unter:", model_path)Das Problem ist das ich das Model zwar speichern kann aber das Laden funktioniert dann nicht mehr. Wenn ich versuche das Model zu laden
model = load_model(model_path)
kommt dieser Fehler
/usr/local/lib/python3.11/dist-packages/keras/src/layers/rnn/rnn.py:200: UserWarning: Do not pass an input_shape/input_dim argument to a layer. When using Sequential models, prefer using an Input(shape) object as the first layer in the model instead.
super().__init__(**kwargs)
Model: "sequential_9"
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓
┃ Layer (type) ┃ Output Shape ┃ Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩
│ lstm_9 (LSTM) │ (None, 128) │ 66,560 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dropout_9 (Dropout) │ (None, 128) │ 0 │
├─────────────────────────────────┼────────────────────────┼───────────────┤
│ dense_9 (Dense) │ (None, 30) │ 3,870 │
└─────────────────────────────────┴────────────────────────┴───────────────┘
Total params: 70,430 (275.12 KB)
Trainable params: 70,430 (275.12 KB)
Non-trainable params: 0 (0.00 B)
56/56 ━━━━━━━━━━━━━━━━━━━━ 38s 470ms/step - loss: 0.0327 - val_loss: 0.0018
WARNING:absl:You are saving your model as an HDF5 file via model.save() or keras.saving.save_model(model). This file format is considered legacy. We recommend using instead the native Keras format, e.g. model.save('my_model.keras') or keras.saving.save_model(model, 'my_model.keras').
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-13-0d8e3b7a3b24> in <cell line: 0>()
58
59 # Modell optional laden
---> 60 model = load_model(model_path)
61
62 # 8. Vorhersage für die nächsten 30 Tage
7 frames
/usr/local/lib/python3.11/dist-packages/keras/src/saving/serialization_lib.py in _retrieve_class_or_fn(name, registered_name, module, obj_type, full_config, custom_objects)
801 return obj
802
--> 803 raise TypeError(
804 f"Could not locate {obj_type} '{name}'. "
805 "Make sure custom classes are decorated with "
TypeError: Could not locate function 'mse'. Make sure custom classes are decorated with @keras.saving.register_keras_serializable(). Full object config: {'module': 'keras.metrics', 'class_name': 'function', 'config': 'mse', 'registered_name': 'mse'}Ich habe leider keine Ahnung was der Fehler bedeutet und was ich machen kann um das Model speichern zu können.
