diff --git a/README.md b/README.md
index 45a56fb..692da4e 100644
--- a/README.md
+++ b/README.md
@@ -10,8 +10,8 @@ to classify the papers given some label (which we specified as the paper's venue
 `DatasetReader`).  Finally, we added a script to use AllenNLP's training commands from a
 third-party repository, and an experiment configuration for running a real model on real data.
 
-To train this model, after setting up your development environment by installing `allennlp==0.4`,
-`pytorch==0.3.1`, and `spacy>=2.0`, you run:
+To train this model, after setting up your development environment by running
+`pip install -r requirements.txt`, you run:
 
 ```bash
 python -m allennlp.run train experiments/venue_classifier.json -s /tmp/your_output_dir_here --include-package my_library
diff --git a/requirements.txt b/requirements.txt
index 3eca544..1d66b46 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,3 @@
-allennlp==0.4.1
+allennlp==0.4.2
 pytest
 pytest-pythonpath