# Problem with usin google colab GPU

**URL:** <https://discuss.ray.io/t/problem-with-usin-google-colab-gpu/4562>\
**Category:** RLlib\
**Created:** [January 3, 2022, 12:54pm UTC](https://discuss.ray.io/t/problem-with-usin-google-colab-gpu/4562 "2022-01-03T12:54:51Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![dezhi](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/dezhi/32/2189_2.png) [@dezhi](https://discuss.ray.io/u/dezhi)\
**Post date:** [January 3, 2022, 12:54pm UTC](https://discuss.ray.io/t/problem-with-usin-google-colab-gpu/4562/1 "2022-01-03T12:54:51Z")

</div>

Hi everyone. I am using google colab. my problem is that Ray finds the GPU but does not use it. I have attached 4 photos named a1,a2,a3 and a4 to explain more.  
a1 photo is when I use only CPU and it is normal. a2 and a3 are using 0.5 and 1 GPU respectively.  
Can anyone please explain me the dashboard status? am I correct that the Ray finds the GPU?  
my code is on a4, which is very simple. I appreciate any guide to use the colab GPU.  
a1 = [a1 — ImgBB](https://ibb.co/fvn93Pp)  
a4 = [a4 — ImgBB](https://ibb.co/LQGYSyf)

since I am allowed to share only one photo and two links I put a3, which is my main problem here:

 ![a3](https://us1.discourse-cdn.com/flex020/uploads/ray/original/2X/d/d35f8a254dbb727d9b8a1a3a1505123852bc0041.jpeg)

---

<div class="post-metadata">

**Author:** ![avnishn](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/avnishn/32/1544_2.png) [@avnishn](https://discuss.ray.io/u/avnishn)\
**Post date:** [January 5, 2022, 8:43pm UTC](https://discuss.ray.io/t/problem-with-usin-google-colab-gpu/4562/2 "2022-01-05T20:43:18Z")

</div>

Hi @dezhi, it would be helpful if you could provide your script from google colab, as opposed to only screenshots, so that we can reproduce your issue. Thanks!

---

<div class="post-metadata">

**Author:** ![dezhi](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/dezhi/32/2189_2.png) [@dezhi](https://discuss.ray.io/u/dezhi)\
**Post date:** [January 6, 2022, 8:33am UTC](https://discuss.ray.io/t/problem-with-usin-google-colab-gpu/4562/3 "2022-01-06T08:33:29Z")

</div>

dear avnish, Since I can not share my data it is pointless to give the whole script. I just want to know that with such GPU usage on ray dashboard I can be sure that I am using GPU?  
I dug more into my own problem and I realized that I have a simple model so I am using more CPU power for sampling and in fact the neural network is not much time and computation consuming. however I am still confused with ray dashboard stats. I know that simple NNs are better to be trained on CPU (it is faster) and the fact that I am getting faster results with 2CPU cores compare to 1CPU+1GPU is a support for my knowledge but I am not sure and need confirmation.