parametrized

Jenkins/Jenkinsfile.training

@@ -7,9 +7,9 @@ node {
                 pipelineTriggers([upstream(threshold: hudson.model.Result.SUCCESS, upstreamProjects: "s444452-create-dataset")]),
                 parameters([
                 string(
-                 defaultValue: ".",
-                 description: 'Arguments for model training: arg1,arg2,arg3',
-                 name: 'TRAIN_ARGS'
+                 defaultValue: ".,14000,1,50,100",
+                 description: 'Train params: data_path,num_words,epochs,batch_size,pad_length',
+                 name: 'TRAIN_PARAMS'
                 )
                 ])
             ])
@@ -20,8 +20,8 @@ node {
             copyArtifacts filter: 'dev_data.csv', fingerprintArtifacts: true, projectName: 's444452-create-dataset'
         }
         stage('Run script') {
-            withEnv(["TRAIN_ARGS=${params.TRAIN_ARGS}"]) {
-                sh "python3 Scripts/train_neural_network.py $TRAIN_ARGS"
+            withEnv(["TRAIN_ARGS=${params.TRAIN_PARAMS}"]) {
+                sh "python3 Scripts/train_neural_network.py $TRAIN_PARAMS"
             }
         }
         stage('Archive artifacts') {
@@ -39,16 +39,8 @@ def notifyBuild(String buildStatus = 'STARTED') {
     buildStatus = buildStatus ?: 'SUCCESS'
 
     def subject = "Job: ${env.JOB_NAME}"
-    def details = "Build nr: ${env.BUILD_NUMBER}, status: ${buildStatus} \n url: ${env.BUILD_URL}"
+    def details = "Build nr: ${env.BUILD_NUMBER}, status: ${buildStatus} \n url: ${env.BUILD_URL} \n build params: ${params.TRAIN_PARAMS}"
 
-    // Override default values based on build status
-    if (buildStatus == 'SUCCESS') {
-        color = 'GREEN'
-        colorCode = '#00FF00'
-    } else {
-        color = 'RED'
-        colorCode = '#FF0000'
-    }
     emailext (
         subject: subject,
         body: details,

Scripts/train_neural_network.py

@@ -13,16 +13,23 @@ import logging
 
 logging.getLogger("tensorflow").setLevel(logging.ERROR)
 
+data_path = ''
+num_words = 0
+epochs = 0
+batch_size = 0
+pad_length = 0
 
-def tokenize(x, x_train, x_test, max_len):
-    tokenizer = Tokenizer(num_words=14000)
+
+def tokenize(x, x_train, x_test):
+    global pad_length, num_words
+    tokenizer = Tokenizer(num_words=num_words)
     tokenizer.fit_on_texts(x)
     train_x = tokenizer.texts_to_sequences(x_train)
     test_x = tokenizer.texts_to_sequences(x_test)
     vocabulary_length = len(tokenizer.word_index) + 1
 
-    train_x = pad_sequences(train_x, padding='post', maxlen=max_len)
-    test_x = pad_sequences(test_x, padding='post', maxlen=max_len)
+    train_x = pad_sequences(train_x, padding='post', maxlen=pad_length)
+    test_x = pad_sequences(test_x, padding='post', maxlen=pad_length)
     return train_x, test_x, vocabulary_length
 
 
@@ -47,14 +54,16 @@ def save_model(model):
 
 
 def train_model(model, x_train, y_train):
-    model.fit(x_train, y_train, epochs=1, verbose=False, batch_size=50)
+    global epochs, batch_size
+    model.fit(x_train, y_train, epochs=epochs, verbose=False, batch_size=batch_size)
 
 
-def get_model(output_dim, vocabulary_length):
+def get_model(vocabulary_length):
+    global pad_length, batch_size
     model = Sequential()
     model.add(layers.Embedding(input_dim=vocabulary_length,
-                               output_dim=output_dim,
-                               input_length=100))
+                               output_dim=batch_size,
+                               input_length=pad_length))
     model.add(layers.Flatten())
     model.add(layers.Dense(10, activation='relu'))
     model.add(layers.Dense(1, activation='sigmoid'))
@@ -74,15 +83,21 @@ def load_data(data_path, filename) -> pd.DataFrame:
     return pd.read_csv(os.path.join(data_path, filename))
 
 
+def read_params():
+    global data_path, num_words, epochs, batch_size, pad_length
+    data_path, num_words, epochs, batch_size, pad_length = sys.argv[1].split(',')
+
+
 def main():
-    data_path = sys.argv[1]
+    read_params()
+    global data_path
     abs_data_path = os.path.abspath(data_path)
     train_data = load_data(abs_data_path, 'train_data.csv')
     test_data = load_data(abs_data_path, 'test_data.csv')
     x_train, y_train = split_data(train_data)
     x_test, y_test = split_data(test_data)
-    x_train, x_test, vocab_size = tokenize(pd.concat([x_train, x_test]), x_train, x_test, 100)
-    model = get_model(50, vocab_size)
+    x_train, x_test, vocab_size = tokenize(pd.concat([x_train, x_test]), x_train, x_test)
+    model = get_model(vocab_size)
     train_model(model, x_train, y_train)
     save_model(model)
     evaluate_and_save(model, x_test, y_test, abs_data_path)