added time to model training metrics (#995)

.github/workflows/metrics.yml

@@ -39,12 +39,12 @@ jobs:
       - name: Compare performance
         run: |
           echo "## Model Benchmark" >> report.md
-          echo "<details>\n<summary>Show benchmark results</summary>\n" >> report.md
           python tests/metrics/compareMetrics.py >> report.md
       - name: Publish report
         env:
           REPO_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
+          echo "<details>\n<summary>Model training plots</summary>\n" >> report.md
           echo "## Model Training" >> report.md
           echo "### PeytonManning" >> report.md
           cml asset publish tests/metrics/PeytonManning.svg --md >> report.md

tests/test_model_performance.py

@@ -4,6 +4,7 @@
 import logging
 import os
 import pathlib
+import time
 
 import pandas as pd
 import plotly.graph_objects as go
@@ -92,9 +93,12 @@ def test_PeytonManning():
     df = pd.read_csv(PEYTON_FILE)
     m = NeuralProphet(early_stopping=True)
     df_train, df_test = m.split_df(df=df, freq="D", valid_p=0.2)
+    start = time.time()
     metrics = m.fit(df_train, validation_df=df_test, freq="D")
+    end = time.time()
 
     accuracy_metrics = metrics.to_dict("records")[-1]
+    accuracy_metrics["time"] = round(end - start, 2)
     with open(os.path.join(DIR, "tests", "metrics", "PeytonManning.json"), "w") as outfile:
         json.dump(accuracy_metrics, outfile)
 
@@ -112,9 +116,12 @@ def test_YosemiteTemps():
         early_stopping=True,
     )
     df_train, df_test = m.split_df(df=df, freq="5min", valid_p=0.2)
+    start = time.time()
     metrics = m.fit(df_train, validation_df=df_test, freq="5min")
+    end = time.time()
 
     accuracy_metrics = metrics.to_dict("records")[-1]
+    accuracy_metrics["time"] = round(end - start, 2)
     with open(os.path.join(DIR, "tests", "metrics", "YosemiteTemps.json"), "w") as outfile:
         json.dump(accuracy_metrics, outfile)
 
@@ -125,9 +132,12 @@ def test_AirPassengers():
     df = pd.read_csv(AIR_FILE)
     m = NeuralProphet(seasonality_mode="multiplicative", early_stopping=True)
     df_train, df_test = m.split_df(df=df, freq="MS", valid_p=0.2)
+    start = time.time()
     metrics = m.fit(df_train, validation_df=df_test, freq="MS")
+    end = time.time()
 
     accuracy_metrics = metrics.to_dict("records")[-1]
+    accuracy_metrics["time"] = round(end - start, 2)
     with open(os.path.join(DIR, "tests", "metrics", "AirPassengers.json"), "w") as outfile:
         json.dump(accuracy_metrics, outfile)