Introducing + Goodies

Hi :wave: friends, course and community are super special for me as my DL journey started with 1st version of the fastai course and a lot of brave hearts from the fastai community became early adopters of when we launched our 1st version last year. is a bootstrapped startup founded by me, and my friends (4 people team), and we are based out of a small town in India. The goal of is to help people coming from different domains to focus on AI/DL and not worry about managing the infra woes.

All it takes is a few clicks and < 30 seconds for your instance to be ready and you can focus on building your favorite AI models.

The feedback from our early adopters helped us in improving the platform, and we have come a long way. I want to invite you all to try out a GPU cloud platform.

Would love to hear your experiences and any suggestions/feedback.

If you are signing into, please fill out this form, I will add a $20 credit and a 30% discount on all the GPU types.

Some important links.

$20 will provide you with close to 60 hours of reserved GPU or 105 hours of spot GPU.

You can apply for the credits and the discount till the last week of the live course.


Hey @VishnuSubramanian
Signed up and looking to use it more and more in future.
Will be happy to know more about how to get more involved and help Jarvis other than just using the service!?

On a sidenote - great name for the service


Thanks for volunteering to help.

As a Marvel I love the name too :blush:, but the name comes with a lot of challenges, particularly with SEO, there are too many companies with the same name. Hope Marvel does not ask us to change the name, as they did with :grin:

1 Like

You have any tips and tricks to prevent racking up the bill way too high!? Does the node auto shutoff?

If Marvel does object to the name, think of it as an opportunity to rebrand to something that conveys what the service is about :grinning:

You can limit the max number of hours you run an instance here.

1 Like

I am trying out Jarvis Labs, and it looks good! However, I am getting a strange OpenCV error right after starting a new instance.

I solved this by deleting all OpenCV traces from the system and re-installing it again. But the problem is that after restarting an instance, I need to redo it again, and it seems that the env is re-installed. Could you please help a bit to understand why it happens?

Thanks, @devforfu for trying Jarvislabs, I will look into this and update you soon.

We have fixed the Open CV2 issue. Now you can import it without any issue.

When you restart an instance, it starts with a fresh instance which will not have any installations you did earlier. We used to save the states earlier, but faced some limitations with that approach. So we stopped saving the states. You can use a startup script to install any new libraries you want on instance start.

1 Like

Interesting! How does one add startup scripts to that?

You can add it here

Thanks a lot! Yes, now it works.

Yeah, I was thinking to go with startup scripts in case if error persists. Currently, I just use pip install ... as the first line in a notebook to β€œrefresh” the dependencies. But the script is great, will definitely use it.

1 Like

Hi, the link seems to be broken, is it Possible to fix it? Thanks

To the startup script that is

I update the docs, a few weeks back. Thanks for letting me know. I have fixed it.

Here is the new link πŸ“œ Startup scripts |

1 Like