mlflow

mlflow/MLproject

@@ -0,0 +1,7 @@
+name: MLflow_s464937
+conda_env: conda.yaml
+entry_points:
+  optimal_parameters:
+    parameters:
+      epochs: { type: int, default: 20 }
+    command: 'python mlflow_model.py {epochs}'

mlflow/conda.yaml

@@ -0,0 +1,12 @@
+name: MLflow_s464937
+channels:
+  - defaults
+dependencies:
+  - python=3.10
+  - pip
+  - pip:
+      - wheel
+      - mlflow
+      - tensorflow
+      - pandas
+      - scikit-learn

mlflow/mlflow_model.py

@@ -0,0 +1,55 @@
+import sys
+import pandas as pd
+from sklearn.model_selection import train_test_split
+from sklearn.preprocessing import StandardScaler, OneHotEncoder
+from sklearn.compose import ColumnTransformer
+from sklearn.pipeline import Pipeline
+from tensorflow.keras.models import Sequential
+from tensorflow.keras.layers import Dense
+import tensorflow as tf
+import mlflow
+
+
+mlflow.set_tracking_uri("http://localhost:5000")
+
+
+def main():
+    data = pd.read_csv('./data/train.csv')
+
+    data = data[['Sex', 'Age', 'BodyweightKg', 'TotalKg']].dropna()
+    data['Age'] = pd.to_numeric(data['Age'], errors='coerce')
+    data['BodyweightKg'] = pd.to_numeric(data['BodyweightKg'], errors='coerce')
+    data['TotalKg'] = pd.to_numeric(data['TotalKg'], errors='coerce')
+    features = data[['Sex', 'Age', 'BodyweightKg']]
+    target = data['TotalKg']
+
+    X_train, X_test, y_train, y_test = train_test_split(features, target, test_size=0.2, random_state=42)
+
+    preprocessor = ColumnTransformer(
+        transformers=[
+            ('num', StandardScaler(), ['Age', 'BodyweightKg']),
+            ('cat', OneHotEncoder(), ['Sex'])
+        ],
+    )
+
+    pipeline = Pipeline(steps=[
+        ('preprocessor', preprocessor),
+        ('model', Sequential([
+            Dense(64, activation='relu', input_dim=5),
+            Dense(64, activation='relu'),
+            Dense(1)
+        ]))
+    ])
+
+    pipeline['model'].compile(optimizer='adam', loss='mse', metrics=['mae'])
+
+    X_train_excluded = X_train.iloc[1:]
+    y_train_excluded = y_train.iloc[1:]
+
+    pipeline.fit(X_train_excluded, y_train_excluded, model__epochs=int(sys.argv[1]), model__validation_split=0.1)
+
+    pipeline['model'].save('powerlifting_model.h5')
+
+
+if __name__ == '__main__':
+    main()

mlruns/0/meta.yaml

@@ -0,0 +1,6 @@
+artifact_location: mlflow-artifacts:/0
+creation_time: 1716052187853
+experiment_id: '0'
+last_update_time: 1716052187853
+lifecycle_stage: active
+name: Default