diff --git a/requirements.txt b/requirements.txt
index 32495d4b04f..9161d6009d2 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -38,9 +38,6 @@ scikit-learn
 # Install the Tensorboard application separately (part of tensorflow) to view them.
 tensorboardX>=1.2
 
-# aws commandline tools for running on Docker remotely.
-awscli>=1.11.91
-
 # Accessing files from S3 directly.
 boto3
 
diff --git a/setup.py b/setup.py
index 03df8ca0273..b49bccd7fc3 100644
--- a/setup.py
+++ b/setup.py
@@ -108,7 +108,6 @@
           'spacy>=2.0.18,<2.2',
           'numpy',
           'tensorboardX>=1.2',
-          'awscli>=1.11.91',
           'boto3',
           'flask>=1.0.2',
           'flask-cors>=3.0.7',
diff --git a/tutorials/how_to/training_transformer_elmo.md b/tutorials/how_to/training_transformer_elmo.md
index 111dcb8534e..eafe29a713b 100644
--- a/tutorials/how_to/training_transformer_elmo.md
+++ b/tutorials/how_to/training_transformer_elmo.md
@@ -11,6 +11,7 @@ This document describes how to train and use a transformer-based version of ELMo
     ```
 2. Obtain vocab.
     ```
+    pip install --user awscli
     mkdir vocabulary
     export BIDIRECTIONAL_LM_VOCAB_PATH=$PWD'/vocabulary'
     cd $BIDIRECTIONAL_LM_VOCAB_PATH