diff --git a/Exercises/Exercise 2 - Handwriting Recognition/Exercise2-Answer.ipynb b/Exercises/Exercise 2 - Handwriting Recognition/Exercise2-Answer.ipynb
index 4fac34a1..5c6f39cb 100644
--- a/Exercises/Exercise 2 - Handwriting Recognition/Exercise2-Answer.ipynb	
+++ b/Exercises/Exercise 2 - Handwriting Recognition/Exercise2-Answer.ipynb	
@@ -37,7 +37,7 @@
         "\n",
         "class myCallback(tf.keras.callbacks.Callback):\n",
         "  def on_epoch_end(self, epoch, logs={}):\n",
-        "    if(logs.get('accuracy')>0.99):\n",
+        "    if((logs.get('acc') is not None) and (logs.get('acc')>0.99)):\n",
         "      print(\"\\nReached 99% accuracy so cancelling training!\")\n",
         "      self.model.stop_training = True\n",
         "\n",
diff --git a/README.md b/README.md
new file mode 100644
index 00000000..a7b2a1a1
--- /dev/null
+++ b/README.md
@@ -0,0 +1,6 @@
+# dlaicourse
+# 'acc' or 'accuracy'
+Notebooks for learning deep learning
+# I see a lot of confusion relation to using 'acc' or 'accuracy' here. 
+# Apparently tf 2.5.0 supports 'acc' while tf 2.6.0 supports 'accuracy'
+# Since the jupyter notebooks used in the course run on 2.5.0, dated 19-09-2021, clearly 'acc' will help you run the right code there!