diff --git a/Dockerfile b/Dockerfile
index 103fe81..ef1f9e6 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,5 +1,5 @@
 FROM ubuntu:22.04
 
 RUN apt update && apt install -y vim make python3 python3-pip python-is-python3 gcc g++ golang wget unzip git
-RUN pip install pandas matplotlib scikit-learn tensorflow
+RUN pip install pandas matplotlib scikit-learn tensorflow sacred
 CMD "bash"
diff --git a/src/tf_train.py b/src/tf_train.py
index caeb2d1..2d6bb24 100755
--- a/src/tf_train.py
+++ b/src/tf_train.py
@@ -1,15 +1,15 @@
 #!/usr/bin/env python3
+from sklearn.preprocessing import LabelEncoder
 import pandas as pd
 import tensorflow as tf
-from sklearn.preprocessing import LabelEncoder
 
 pd.set_option('display.float_format', lambda x: '%.5f' % x)
 
 le = LabelEncoder()
 le.fit([x.strip() for x in pd.read_csv('./stop_times.categories.tsv', sep='\t')['stop_headsign'].to_numpy()])
 
-def load_data(path: str, le: LabelEncoder):
-    train = pd.read_csv(path, sep='\t', dtype={ 'departure_time': float, 'stop_id': str, 'stop_headsign': str })
+def load_data(path: str, le: LabelEncoder, open_file):
+    train = pd.read_csv(open_file(path), sep='\t', dtype={ 'departure_time': float, 'stop_id': str, 'stop_headsign': str })
 
     departure_time = train['departure_time'].to_numpy()
     stop_id        = train['stop_id'].to_numpy()
@@ -23,7 +23,7 @@ def load_data(path: str, le: LabelEncoder):
 num_classes = len(le.classes_)
 
 
-def train(epochs: int):
+def train(epochs: int, open_file, add_artifact):
     global le
 
     model = tf.keras.Sequential([
@@ -38,18 +38,19 @@ def train(epochs: int):
                   loss=tf.keras.losses.SparseCategoricalCrossentropy(),
                   metrics=['accuracy'])
 
-    train_x, train_y, _ = load_data('./stop_times.train.tsv', le)
+    train_x, train_y, _ = load_data('./stop_times.train.tsv', le, open_file)
     train_x = tf.convert_to_tensor(train_x, dtype=tf.float32)
     train_y = tf.convert_to_tensor(train_y)
 
-    valid_x, valid_y, _ = load_data('./stop_times.valid.tsv', le)
+    valid_x, valid_y, _ = load_data('./stop_times.valid.tsv', le, open_file)
     valid_x = tf.convert_to_tensor(valid_x, dtype=tf.float32)
     valid_y = tf.convert_to_tensor(valid_y)
 
-    model.fit(train_x, train_y, validation_data=(valid_x, valid_y), epochs=epochs, batch_size=1024)
+    history = model.fit(train_x, train_y, validation_data=(valid_x, valid_y), epochs=epochs, batch_size=1024)
     model.save('model.keras')
+    add_artifact('model.keras', history.history)
 
 if __name__ == "__main__":
     import sys
     epochs = int('2' if len(sys.argv) != 2 else sys.argv[1])
-    train(epochs)
+    train(epochs, lambda x: x, lambda _1, _2: ())
diff --git a/src/tf_train_sacred.py b/src/tf_train_sacred.py
new file mode 100644
index 0000000..5da441e
--- /dev/null
+++ b/src/tf_train_sacred.py
@@ -0,0 +1,21 @@
+from tf_train import *
+from sacred import Experiment
+from sacred.observers import FileStorageObserver
+
+ex = Experiment('s452639')
+ex.observers.append(FileStorageObserver('experiments'))
+
+@ex.config
+def params():
+    epochs = 2
+
+@ex.automain
+def train_sacred(epochs: int):
+    def postprocess(artifact_path, history):
+        ex.add_artifact(artifact_path)
+
+        for name, values in history.items():
+            for value in values:
+                ex.log_scalar(name, value)
+
+    train(epochs, lambda path: ex.open_resource(path), postprocess)
diff --git a/training.jenkinsfile b/training.jenkinsfile
index 7eb3f0d..fe78894 100644
--- a/training.jenkinsfile
+++ b/training.jenkinsfile
@@ -32,8 +32,8 @@ node {
 				flatten: true,
 				target: 'src/'
 
-			sh 'cd src; python tf_train.py $EPOCHS'
-			archiveArtifacts artifacts: 'src/model.keras', followSymlinks: false
+			sh 'cd src; python tf_train_sacred.py with epochs=$EPOCHS'
+			archiveArtifacts artifacts: 'src/model.keras,src/experiments/**/*', followSymlinks: false
 		}
 	}