mlflow

Dockerfile

@@ -9,8 +9,7 @@ RUN pip3 install matplotlib
 RUN pip3 install sklearn
 RUN pip3 install kaggle
 RUN pip3 install torch
-RUN pip3 install sacred
-RUN pip3 install pymongo
+RUN pip3 install mlflow
 
 RUN mkdir /.kaggle && chmod o+w /.kaggle
 

MLproject

@@ -0,0 +1,12 @@
+name: s444501
+
+docker_env:
+  image: zadanie
+
+entry_points:
+  main:
+    parameters:
+      epochs: {type: float, default: 100}
+    command: "python biblioteki_ml.py {epochs}"
+  eval:
+    command: "python eval.py"

biblioteki_ml.py

@@ -1,11 +1,23 @@
 import sys
+from urllib.parse import urlparse
+import numpy as np
+import mlflow
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
-from sacred.observers import FileStorageObserver, MongoObserver
 from sklearn.preprocessing import LabelEncoder
 import pandas as pd
-from sacred import Experiment
+
+# MLFlow 1
+mlflow.set_experiment("s444501")
+
+
+# Parametry z konsoli
+try:
+    epochs = int(sys.argv[1])
+except:
+    print('No epoch number passed. Defaulting to 100')
+    epochs = 100
 
 
 # Model
@@ -23,32 +35,7 @@ class Model(nn.Module):
         return x
 
 
-# Sacred
-ex = Experiment()
-ex.observers.append(FileStorageObserver('my_runs'))
-# Parametry treningu            -> my_runs/X/config.json
-# Plik z modelem jako artefakt  -> my_runs/X/model.pkl
-# Kod źródłowy                  -> my_runs/_sources/biblioteki_ml_XXXXXXXXXXX.py
-# Wyniki (ostateczny loss)      -> my_runs/X/metrics.json
-ex.observers.append(MongoObserver(url='mongodb://admin:IUM_2021@172.17.0.1:27017',
-                                  db_name='sacred'))
-
-
-@ex.config
-def my_config():
-    epochs = 100
-
-
-@ex.automain
-def train_main(epochs, _run):
-    # Parametry z konsoli
-    # try:
-    #     epochs = int(sys.argv[1])
-    # except:
-    #     print('No epoch number passed. Defaulting to 100')
-    #     epochs = 100
-
-
+def train_main(epochs, run):
     # Ładowanie danych
     train_set = pd.read_csv('d_train.csv', encoding='latin-1')
     train_set = train_set[['Rating', 'Branch', 'Reviewer_Location']]
@@ -103,7 +90,6 @@ def train_main(epochs, _run):
         optimizer.zero_grad()
         loss.backward()
         optimizer.step()
-    _run.log_scalar("training.final_loss", losses[-1].item())  # Ostateczny loss
 
 
     # Testy
@@ -115,13 +101,36 @@ def train_main(epochs, _run):
 
     df = pd.DataFrame({'Testing Y': y_test, 'Predicted Y': preds})
     df['Correct'] = [1 if corr == pred else 0 for corr, pred in zip(df['Testing Y'], df['Predicted Y'])]
-    print(f"{df['Correct'].sum() / len(df)} percent of predictions correct")
+    correct = df['Correct'].sum() / len(df)
+    print(f"{correct} percent of predictions correct")
 
 
+    # Logi
+    mlflow.log_param("epochs", epochs)
+    mlflow.log_metric("final_loss", losses[-1].item())
+    mlflow.log_metric("accuracy", correct)
+
+    signature = mlflow.models.signature.infer_signature(X_train.numpy(), np.array(preds))
+    tracking_url_type_store = urlparse(mlflow.get_tracking_uri()).scheme
+
+    if tracking_url_type_store != "file":
+        mlflow.pytorch.log_model(model,
+                                 's444501',
+                                 registered_model_name='s444501',
+                                 signature=signature,
+                                 input_example=X_test.numpy())
+    else:
+        mlflow.pytorch.log_model(model,
+                                 's444501',
+                                 signature=signature,
+                                 input_example=X_test.numpy())
+
     # Zapis do pliku
     df.to_csv('neural_network_prediction_results.csv', index=False)
     torch.save(model, "model.pkl")
 
-    # Zapis Sacred
-    ex.add_artifact("model.pkl")
-    ex.add_artifact("neural_network_prediction_results.csv")
+
+with mlflow.start_run() as run:
+    print(f"MLflow run experiment_id: {run.info.experiment_id}")
+    print(f"MLflow run artifact_uri: {run.info.artifact_uri}")
+    train_main(epochs, run)

training.Jenkinsfile

@@ -24,14 +24,14 @@ pipeline {
         stage('Train model') {
             steps {
                 withEnv(["EPOCH=${params.EPOCH}"]) {
-				    sh 'python biblioteki_ml.py with "epochs=$EPOCH"'
+				    sh 'python biblioteki_ml.py $EPOCH'
 				}
 			}
 		}
 		stage('Archive artifacts') {
             steps {
 				archiveArtifacts artifacts: 'model.pkl, neural_network_prediction_results.csv'
-				archiveArtifacts artifacts: 'my_runs/**'
+				archiveArtifacts artifacts: 'mlruns/**'
 			}
 		}
 		stage ('Model - evaluation') {