Upload 4 files

app.py

@@ -0,0 +1,78 @@
+import os
+import uuid
+import joblib
+import json
+
+import gradio as gr
+import pandas as pd
+
+from huggingface_hub import CommitScheduler
+from pathlib import Path
+
+log_file = Path("logs/") / f"data_{uuid.uuid4()}.json"
+log_folder = log_file.parent
+
+scheduler = CommitScheduler(
+    repo_id="machine-failure-logs",
+    repo_type="dataset",
+    folder_path=log_folder,
+    path_in_repo="data",
+    every=2
+)
+
+machine_failure_predictor = joblib.load('model.joblib')
+
+air_temperature_input = gr.Number(label='Air temperature [K]')
+process_temperature_input = gr.Number(label='Process temperature [K]')
+rotational_speed_input = gr.Number(label='Rotational speed [rpm]')
+torque_input = gr.Number(label='Torque [Nm]')
+tool_wear_input = gr.Number(label='Tool wear [min]')
+type_input = gr.Dropdown(
+    ['L', 'M', 'H'],
+    label='Type'
+)
+
+model_output = gr.Label(label="Machine failure")
+
+def predict_machine_failure(air_temperature, process_temperature, rotational_speed, torque, tool_wear, type):
+    sample = {
+        'Air temperature [K]': air_temperature,
+        'Process temperature [K]': process_temperature,
+        'Rotational speed [rpm]': rotational_speed,
+        'Torque [Nm]': torque,
+        'Tool wear [min]': tool_wear,
+        'Type': type
+    }
+    data_point = pd.DataFrame([sample])
+    prediction = machine_failure_predictor.predict(data_point).tolist()
+
+    with scheduler.lock:
+        with log_file.open("a") as f:
+            f.write(json.dumps(
+                {
+                    'Air temperature [K]': air_temperature,
+                    'Process temperature [K]': process_temperature,
+                    'Rotational speed [rpm]': rotational_speed,
+                    'Torque [Nm]': torque,
+                    'Tool wear [min]': tool_wear,
+                    'Type': type,
+                    'prediction': prediction[0]
+                }
+            ))
+            f.write("\n")
+            
+    return prediction[0]
+
+demo = gr.Interface(
+    fn=predict_machine_failure,
+    inputs=[air_temperature_input, process_temperature_input, rotational_speed_input, 
+            torque_input, tool_wear_input, type_input],
+    outputs=model_output,
+    title="Machine Failure Predictor",
+    description="This API allows you to predict the machine failure status of an equipment",
+    allow_flagging="auto",
+    concurrency_limit=8
+)
+
+demo.queue()
+demo.launch(share=False)

model.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c3c382c7233f0463a9c2698c7190fa6a89f2704433ae79735d9a6a1acfb5529
+size 8439

requirements.txt

@@ -0,0 +1,2 @@
+scikit-learn==1.2.2
+numpy==1.26.4

train.py

@@ -0,0 +1,80 @@
+
+import joblib
+import pandas as pd
+
+from sklearn.preprocessing import StandardScaler, OneHotEncoder
+from sklearn.compose import make_column_transformer
+from sklearn.impute import SimpleImputer
+from sklearn.pipeline import Pipeline
+from sklearn.pipeline import make_pipeline
+
+from sklearn.model_selection import train_test_split, RandomizedSearchCV
+
+from sklearn.linear_model import LogisticRegression
+from sklearn.metrics import accuracy_score, classification_report
+
+data_df = pd.read_csv("Bank_Telemarketing.csv")
+
+target = 'subscribed'
+numerical_features = ['Age', 'Duration(Sec)', 'CC Contact Freq', 'Days Since PC','PC Contact Freq']
+categorical_features = ['Job', 'Marital Status', 'Education', 'Defaulter', 'Home Loan',
+       'Personal Loan', 'Communication Type', 'Last Contacted', 'Day of Week',
+       'PC Outcome']
+
+print("Creating data subsets")
+
+X = data_df[numerical_features + categorical_features]
+y = data_df[target]
+
+Xtrain, Xtest, ytrain, ytest = train_test_split(
+    X, y,
+    test_size=0.2,
+    random_state=42
+)
+
+numerical_pipeline = Pipeline([
+    ('imputer', SimpleImputer(strategy='median')),
+    ('scaler', StandardScaler())
+])
+
+categorical_pipeline = Pipeline([
+    ('imputer', SimpleImputer(strategy='most_frequent')),
+    ('onehot', OneHotEncoder(handle_unknown='ignore'))
+])
+
+preprocessor = make_column_transformer(
+    (numerical_pipeline, numerical_features),
+    (categorical_pipeline, categorical_features)
+)
+
+model_logistic_regression = LogisticRegression(n_jobs=-1)
+
+print("Estimating Best Model Pipeline")
+
+model_pipeline = make_pipeline(
+    preprocessor,
+    model_logistic_regression
+)
+
+param_distribution = {
+    "logisticregression__C": [0.001, 0.01, 0.1, 0.5, 1, 5, 10]
+}
+
+rand_search_cv = RandomizedSearchCV(
+    model_pipeline,
+    param_distribution,
+    n_iter=3,
+    cv=3,
+    random_state=42
+)
+
+rand_search_cv.fit(Xtrain, ytrain)
+
+print("Logging Metrics")
+print(f"Accuracy: {rand_search_cv.best_score_}")
+
+print("Serializing Model")
+
+saved_model_path = "model.joblib"
+
+joblib.dump(rand_search_cv.best_estimator_, saved_model_path)