diff --git a/README.md b/README.md
index 9a4403d..3ee48e7 100644
--- a/README.md
+++ b/README.md
@@ -1,13 +1,38 @@
 # PyroNN layers as a tf custom op
 
+This repository contains the pyronn layers configured as a tf-custom ops for windows. 
+Use the install instructions to install one of the provided builds, or follow the step-by-step guide to build for your own system.
+
+## Install with pip
+
+- create a python env with python 3.6 (I use conda for env management here, but any python 3.6 will do)
+
+`conda create -n "pyronn" python=3.6`
+
+- use the supplied requirements.txt file to install pyronn along with all necessary dependencies
+
+`pip install -r https://raw.githubusercontent.com/maxrohleder/win-pyronn/master/requirements.txt`
+
+- the approach above guarantees, that all deps are fitting. Alternatively, you can install the major dependecies by hand:
+
+```
+- python 3.6
+- matplotlib 3.3.4
+- tensorflow 2.4.1
+- pyronn 0.1.0
+```
+Then install the v0.1.0 release with
+
+`pip install https://github.com/maxrohleder/win-pyronn/releases/download/v0.1.0/pyronn_layers-0.1.0-cp36-cp36m-win_amd64.whl`
+
+## A step-by-step guide
+
 As the [custom_ops readme](https://github.com/tensorflow/custom-op) offers very limited support for building a custom
 layer on windows, I decided to create my own and thoroughly document the build process of 
 the [pyroNN layers](https://github.com/csyben/PYRO-NN-Layers).
 
 I spent a lot of time on this issue, so here are the exact steps I used to compile the framework.
 
-### A step-by-step guide
-
 Make sure to use one of the tested combinations of build tools and dependencies listed [here](https://www.tensorflow.org/install/source_windows?hl=en#gpu).
 
 ```md
@@ -22,9 +47,9 @@ ________________________________________________________________________________
 1. download bazel 3.1.0, unzip it to some folder and add it to path (its just one executable)
 https://github.com/bazelbuild/bazel/releases/tag/3.1.0
 
-2. install cuda 10.1 and cudnn 7.6
+2. install cuda 10.1 and cudnn 7.6.5
 https://developer.nvidia.com/cuda-10.1-download-archive-base
-https://developer.nvidia.com/rdp/cudnn-archive  <-- (have to make an nvidia dev account)
+https://developer.nvidia.com/rdp/cudnn-archive  <-- (have to make an nvidia dev account [direct link](https://developer.nvidia.com/compute/machine-learning/cudnn/secure/7.6.5.32/Production/10.1_20191031/cudnn-10.1-windows10-x64-v7.6.5.32.zip))
 
 3. unpack cudnn into the cuda folder
 https://medium.com/vitrox-publication/deep-learning-frameworks-tensorflow-build-from-source-on-windows-python-c-cpu-gpu-d3aa4d0772d8