diff --git a/cv/classification/resnet50/pytorch/README.md b/cv/classification/resnet50/pytorch/README.md
index 5a4f9318ca88b8ad6fb6135e5a595322c691cc2d..c388ce51cec2e50c43c1404bb6d5c5ede3cbe435 100644
--- a/cv/classification/resnet50/pytorch/README.md
+++ b/cv/classification/resnet50/pytorch/README.md
@@ -6,6 +6,7 @@ Residual Networks, or ResNets, learn residual functions with reference to the la
 ```bash
 pip3 install torch torchvision
 ```
+Sign up and login in [imagenet official website](https://www.image-net.org/index.php), then choose 'Download' to download the whole imagenet dataset. Specify `/path/to/imagenet` to your imagenet path in later training process.
 :beers: Done!
 
 ## Step 2: Training
diff --git a/cv/classification/shufflenetv2/pytorch/README.md b/cv/classification/shufflenetv2/pytorch/README.md
index 6a4542cda1fe20c093ef5f2cbb6a489b07dcbfde..6c9725a500616e6621e2e71c3bf8b265fe4a9d12 100644
--- a/cv/classification/shufflenetv2/pytorch/README.md
+++ b/cv/classification/shufflenetv2/pytorch/README.md
@@ -5,6 +5,7 @@ ShuffleNet v2 is a convolutional neural network optimized for a direct metric (s
 ```bash
 pip3 install -r requirements.txt
 ```
+Sign up and login in [imagenet official website](https://www.image-net.org/index.php), then choose 'Download' to download the whole imagenet dataset. Specify `/path/to/imagenet` to your imagenet path in later training process.
 :beers: Done!
 
 ## Step 2: Training
diff --git a/cv/classification/vgg/pytorch/README.md b/cv/classification/vgg/pytorch/README.md
index f5176b322e11291f61be098411f3459814b5a018..56618b647f42fd3f8175572d8b7a9d5e519d846c 100644
--- a/cv/classification/vgg/pytorch/README.md
+++ b/cv/classification/vgg/pytorch/README.md
@@ -7,6 +7,7 @@ VGG is a classical convolutional neural network architecture. It was based on an
 ```bash
 pip3 install -r requirements.txt
 ```
+Sign up and login in [imagenet official website](https://www.image-net.org/index.php), then choose 'Download' to download the whole imagenet dataset. Specify `/path/to/imagenet` to your imagenet path in later training process.
 :beers: Done!
 
 ## Step 2: Training
diff --git a/cv/semantic_segmentation/unet3d/pytorch/README.md b/cv/semantic_segmentation/unet3d/pytorch/README.md
index da531c3f2a7fa326e305c475bce0a6626bbc384c..5ecbccff27eda4510648dee6e0337d7359c79d3a 100644
--- a/cv/semantic_segmentation/unet3d/pytorch/README.md
+++ b/cv/semantic_segmentation/unet3d/pytorch/README.md
@@ -31,7 +31,7 @@ else:
 
 ```shell
 
-bash prepare.sh /url/to/kits19/
+bash prepare.sh
 
 ```
 
diff --git a/cv/semantic_segmentation/unet3d/pytorch/prepare.sh b/cv/semantic_segmentation/unet3d/pytorch/prepare.sh
index 844fc8465efba828f527504dcc3cf75bb841e558..70c8207966ead21f6c9587d4993e2a22983772c0 100644
--- a/cv/semantic_segmentation/unet3d/pytorch/prepare.sh
+++ b/cv/semantic_segmentation/unet3d/pytorch/prepare.sh
@@ -1,18 +1,16 @@
 #!/bin/bash
 
 echo "prepare data..."
-DATA_URL=$1
 
-mkdir -p data/kits19/train
-
-data_file="data/kits19/train/kits19.tar.gz"
+git clone https://github.com/neheller/kits19
+cd kits19
+pip3 install -r requirements.txt
+python3 -m starter_code.get_imaging
+cd ..
 
-if [ ! -e ${data_file} ]; then
-    echo "ERROR: Invalid data file ${data_file}"
-    wget -P data/kits19/train ${DATA_URL}
-fi
+mkdir -p data/kits19/train
 
-echo "Uncompressing the kits19!"
-tar -xf ${data_file} -C data/kits19/train
+python3 preprocess_dataset.py --data_dir kits19/data --results_dir data/kits19/train
 
+echo "data done!"