diff --git a/.gitignore b/.gitignore
index beb1c04..62bfceb 100644
--- a/.gitignore
+++ b/.gitignore
@@ -2,4 +2,6 @@
 
 .ipynb_checkpoints
 __pycache__
-.env/*
\ No newline at end of file
+.env/*
+
+torchvggish.egg-info/
diff --git a/README.md b/README.md
index 159f625..c1c716a 100644
--- a/README.md
+++ b/README.md
@@ -20,6 +20,13 @@ except: urllib.request.urlretrieve(url, filename)
 model.forward(filename)
 ```
 
+If you want to modify hyperparameters in `torchvggish.vggish_params`,
+add the following `install_requires` to your `setup.py`:
+```
+"torchvggish @ git+https://github.com/neuralaudio/torchvggish.git@setup.py#egg=torchvggish",
+```
+and then `import torchvggish.vggish_params` in your code.
+
 <hr>
 [1]  S. Hershey et al., ‘CNN Architectures for Large-Scale Audio Classification’,\
     in International Conference on Acoustics, Speech and Signal Processing (ICASSP),2017\
diff --git a/setup.py b/setup.py
new file mode 100644
index 0000000..af04a6e
--- /dev/null
+++ b/setup.py
@@ -0,0 +1,22 @@
+#!/usr/bin/env python3
+from setuptools import find_packages, setup
+
+long_description = open("README.md", "r", encoding="utf-8").read()
+
+setup(
+    name="torchvggish",
+    version="v0.1.1",
+    description="Pytorch port of Google Research's VGGish model used for extracting audio features.",
+    author="Harri Taylor",
+    author_email="harritaylor@protonmail.com",
+    url="https://github.com/harritaylor/torchvggish",
+    license="Apache-2.0",
+    long_description=long_description,
+    long_description_content_type="text/markdown",
+    packages=find_packages(),
+    python_requires=">=3.6",
+    install_requires=['torch', 'numpy', 'resampy', 'soundfile'],
+    extras_require={
+    },
+)
+
diff --git a/torchvggish/__init__.py b/torchvggish/__init__.py
new file mode 100644
index 0000000..e69de29