using your fancy GPU with tensorflow

 welcome to my first post . since this blog was mostly focused on cyber security  i think it's  time to change it i will be focusing on problems that arised  with me . 

so we all have  Graphical Processing Unit(GPU)  which we use to play games with , ever wondered why that setup of your gets so hot  while doing it .

Answer , simple  your GPU has 1000  of  simple cores compared to few complex cores of CPU

solution: 

1. install anaconda  (trust me )  setup path automatically from the installer (easy that way)

https://www.anaconda.com/products/individual

2. open powershell 

if you are on windows 10 

press windows +x  and press a

type    conda create --name tfgpu tensorflow-gpu

press y if prompted wait for few minutes for installation to complete

3. type 

conda activate tfgpu

this is activate your virtual env named tfgpu

4. type 

pip install jupyter notebook

pip install ipykernel

(you need this for it to  work with jupyter notebook)

5 now final step to create kernel

ipython kernel install --user --name=tfgpu

now run jupyter notebook from powershell

and u will be seeing tfgpu in place of your usual python3 kernel

Note: this method has been tested in windows only