From 23c99e15ca42ec8ba4a93bdbcdd6cd2c52b3801f Mon Sep 17 00:00:00 2001
From: Elad Shaham <7040645+eshaham@users.noreply.github.com>
Date: Mon, 8 Oct 2018 20:56:17 +0300
Subject: [PATCH] save final version

---
 mnist_cnn.py | 25 +++++++++++++++----------
 1 file changed, 15 insertions(+), 10 deletions(-)

diff --git a/mnist_cnn.py b/mnist_cnn.py
index 8a300c3..997e339 100644
--- a/mnist_cnn.py
+++ b/mnist_cnn.py
@@ -11,6 +11,9 @@
 
 import tensorflow as tf
 from tensorflow.examples.tutorials.mnist import input_data
+import missinglink
+
+project = missinglink.TensorFlowProject()
 
 # Input params
 NUM_CLASSES = 10  # The MNIST dataset has 10 classes, representing the digits 0 through 9.
@@ -115,18 +118,20 @@ def run_training():
         session = tf.Session()
         session.run(init)
 
-        # Start the training loop
-        for step in range(MAX_STEPS):
-            feed_dict = fill_feed_dict(data_sets.train, images_placeholder, labels_placeholder)
-
-            _, loss_value = session.run([train_op, loss], feed_dict=feed_dict)
+        with project.create_experiment() as experiment:
+          # Start the training loop
+          for step in experiment.loop(max_iterations=MAX_STEPS):
+              feed_dict = fill_feed_dict(data_sets.train, images_placeholder, labels_placeholder)
 
-            # Validate the model with the validation dataset
-            if (step + 1) % 500 == 0 or (step + 1) == MAX_STEPS:
-                print('Step %d: loss = %.2f' % (step, loss_value))
-                print('Running on validation dataset...')
-                do_eval(session, eval_correct, images_placeholder, labels_placeholder, data_sets.validation)
+              with experiment.train(monitored_metrics={'loss': loss, 'acc': eval_correct}):
+                _, loss_value = session.run([train_op, loss], feed_dict=feed_dict)
 
+              # Validate the model with the validation dataset
+              if (step + 1) % 500 == 0 or (step + 1) == MAX_STEPS:
+                  print('Step %d: loss = %.2f' % (step, loss_value))
+                  print('Running on validation dataset...')
+                  with experiment.validation(monitored_metrics={'loss': loss, 'acc': eval_correct}):
+                    do_eval(session, eval_correct, images_placeholder, labels_placeholder, data_sets.validation)
 
 if __name__ == '__main__':
     run_training()
\ No newline at end of file