Add basic Sacred support

Jenkinsfile-train

@@ -29,7 +29,7 @@ pipeline {
     stages {
         stage('Train model') {
             steps {
-                sh "python train_model.py -e ${params.EPOCHS} -b ${params.BATCHSIZE}"
+                sh "python train_model.py with 'epochs=${params.EPOCHS}' 'batch_size=${params.BATCHSIZE}'"
             }
         }
         stage('Archive model and evaluate it') {

requirements.txt

@@ -3,3 +3,4 @@ pandas==1.4.1
 torch==1.11.0
 numpy~=1.22.3
 matplotlib==3.5.2
+sacred==0.8.2

train_model.py

@@ -5,6 +5,7 @@ import pandas as pd
 import torch
 from torch import nn
 from torch.utils.data import DataLoader, Dataset
+from sacred import Experiment
 
 default_batch_size = 64
 default_epochs = 5
@@ -90,21 +91,9 @@ def test(dataloader, model, loss_fn):
     return test_loss
 
 
-def setup_args():
+def main(batch_size, epochs):
-    args_parser = argparse.ArgumentParser(prefix_chars='-')
-    args_parser.add_argument('-b', '--batchSize', type=int, default=default_batch_size)
-    args_parser.add_argument('-e', '--epochs', type=int, default=default_epochs)
-
-    return args_parser.parse_args()
-
-
-def main():
     print(f"Using {device} device")
 
-    args = setup_args()
-
-    batch_size = args.batchSize
-
     plant_test = PlantsDataset('data/Plant_1_Generation_Data.csv.test')
     plant_train = PlantsDataset('data/Plant_1_Generation_Data.csv.train')
 
@@ -121,7 +110,6 @@ def main():
 
     loss_fn = nn.MSELoss()
     optimizer = torch.optim.Adam(model.parameters(), lr=1e-3)
-    epochs = args.epochs
     for t in range(epochs):
         print(f"Epoch {t + 1}\n-------------------------------")
         train(train_dataloader, model, loss_fn, optimizer)
@@ -132,5 +120,15 @@ def main():
     print("Model saved in ./model_out file.")
 
 
-if __name__ == "__main__":
+ex = Experiment('Predict power output for a given time')
-    main()
+
+
+@ex.config
+def experiment_config():
+    batch_size = 64
+    epochs = 5
+
+
+@ex.automain
+def run(batch_size, epochs):
+    main(batch_size, epochs)