diff --git a/call_for_contributions.md b/call_for_contributions.md
index dc17271b91..eabb81affc 100644
--- a/call_for_contributions.md
+++ b/call_for_contributions.md
@@ -62,6 +62,7 @@ The dataset should have at least 50k samples and there should be at least a doze
 ## DETR: End-to-End Object Detection with Transformers
 
 [Reference implementation](https://github.com/facebookresearch/detr)
+[TF/Keras implementation](https://github.com/auvisusAI/detr-tensorflow)
 
 
 ## 3D image classification
@@ -72,4 +73,3 @@ The model should use `Conv3D` layers.
 
 
 
-
diff --git a/templates/getting_started/intro_to_keras_for_engineers.md b/templates/getting_started/intro_to_keras_for_engineers.md
index 1aa8b9f5e6..de0fe48bcc 100644
--- a/templates/getting_started/intro_to_keras_for_engineers.md
+++ b/templates/getting_started/intro_to_keras_for_engineers.md
@@ -27,20 +27,20 @@ Are you a machine learning engineer looking to use Keras
 to ship deep-learning powered features in real products? This guide will serve
 as your first introduction to core Keras API concepts.
 
-In this guide, you will learn about:
+In this guide, you will learn how to:
 
-- How to prepare your data before training a model (by turning it into either NumPy
+- Prepare your data before training a model (by turning it into either NumPy
  arrays or `tf.data.Dataset` objects).
-- How to do data preprocessing, for instance feature normalization or vocabulary
+- Do data preprocessing, for instance feature normalization or vocabulary
  indexing.
-- How to build a model that turns your data into useful predictions,
+- Build a model that turns your data into useful predictions,
 using the Keras Functional API.
-- How to train your model with the built-in Keras `fit()` method, while being
+- Train your model with the built-in Keras `fit()` method, while being
 mindful of checkpointing, metrics monitoring, and fault tolerance.
-- How to evaluate your model on a test data and how to use it for inference on new data.
-- How to customize what `fit()` does, for instance to build a GAN.
-- How to speed up training by leveraging multiple GPUs.
-- How to refine your model through hyperparameter tuning.
+- Evaluate your model on a test data and how to use it for inference on new data.
+- Customize what `fit()` does, for instance to build a GAN.
+- Speed up training by leveraging multiple GPUs.
+- Refine your model through hyperparameter tuning.
 
 At the end of this guide, you will get pointers to end-to-end examples to solidify
  these concepts:
@@ -562,9 +562,9 @@ Trainable params: 118,282
 Non-trainable params: 0
 _________________________________________________________________
 Fit on NumPy data
-938/938 [==============================] - 1s 1ms/step - loss: 0.2671
+938/938 [==============================] - 1s 804us/step - loss: 0.2696
 Fit on Dataset
-938/938 [==============================] - 1s 1ms/step - loss: 0.1156
+938/938 [==============================] - 1s 964us/step - loss: 0.1162
 
 ```
 </div>
@@ -580,7 +580,7 @@ print(history.history)
 
 <div class="k-default-codeblock">
 ```
-{'loss': [0.11557048559188843]}
+{'loss': [0.11615095287561417]}
 
 ```
 </div>
@@ -611,7 +611,7 @@ history = model.fit(dataset, epochs=1)
 
 <div class="k-default-codeblock">
 ```
-938/938 [==============================] - 1s 1ms/step - loss: 0.0801 - acc: 0.9760
+938/938 [==============================] - 1s 928us/step - loss: 0.0809 - acc: 0.9753
 
 ```
 </div>
@@ -628,7 +628,7 @@ history = model.fit(dataset, epochs=1, validation_data=val_dataset)
 
 <div class="k-default-codeblock">
 ```
-938/938 [==============================] - 1s 1ms/step - loss: 0.0547 - acc: 0.9835 - val_loss: 0.1121 - val_acc: 0.9671
+938/938 [==============================] - 1s 1ms/step - loss: 0.0556 - acc: 0.9829 - val_loss: 0.1163 - val_acc: 0.9670
 
 ```
 </div>
@@ -713,8 +713,8 @@ print("acc: %.2f" % acc)
 
 <div class="k-default-codeblock">
 ```
-157/157 [==============================] - 0s 667us/step - loss: 0.1121 - acc: 0.9671
-loss: 0.11
+157/157 [==============================] - 0s 907us/step - loss: 0.1163 - acc: 0.9670
+loss: 0.12
 acc: 0.97
 
 ```
@@ -899,9 +899,9 @@ model.fit(dataset)
 
 <div class="k-default-codeblock">
 ```
-1/1 [==============================] - 0s 867us/step - loss: 0.5251
+1/1 [==============================] - 0s 1ms/step - loss: 0.4867
 
-<tensorflow.python.keras.callbacks.History at 0x183298190>
+<tensorflow.python.keras.callbacks.History at 0x17ef6d290>
 
 ```
 </div>
@@ -925,9 +925,9 @@ model.fit(dataset)
 
 <div class="k-default-codeblock">
 ```
-1/1 [==============================] - 0s 894us/step - loss: 0.5090
+1/1 [==============================] - 0s 967us/step - loss: 0.4994
 
-<tensorflow.python.keras.callbacks.History at 0x1832e8890>
+<tensorflow.python.keras.callbacks.History at 0x17efe1190>
 
 ```
 </div>