diff --git a/README.md b/README.md
index c34b301..8c1893b 100644
--- a/README.md
+++ b/README.md
@@ -11,6 +11,10 @@ Note: 1. The latest version of tensorflow is needed.
 A Strongly-supervised Representation for Detailed Action Understanding" International Conference on Computer Vision (ICCV). Dec 2013.):
 http://dreamdragon.github.io/PennAction/
 
+Then download pretrained alexnet model: ./models/download.sh
+
+and pretrained vgg model: https://mega.nz/#!YU1FWJrA!O1ywiCS2IiOlUCtCpI6HTJOMrneN-Qdv3ywQP5poecM (rename it vgg.npy afterwards)
+
 and then run preposs.ipynb
 
 3. Run our LSTM model and analogy network model separately.
\ No newline at end of file
diff --git a/models/download.sh b/models/download.sh
new file mode 100644
index 0000000..9fe927c
--- /dev/null
+++ b/models/download.sh
@@ -0,0 +1,3 @@
+wget http://www.cs.toronto.edu/~guerzhoy/tf_alexnet/bvlc_alexnet.npy
+mv bvlc_alexnet.npy models/alexnet.npy
+