Looks like it is not.
Hello @johnnynunez , there is not a cu128 version?
No, it’s not necessary and we don’t have backup of that wheels
Looks like both the production + backup are down again
EDIT: https://pypi.jetson-ai-lab.io/ is back up as of Jul 22 8:00AM EST
I can confirm that the backup link is working. But the cu126 is missing many of the packages present in other cuda versions. How can I get the cu126 variants? It is the stock cuda version with Jetpack 6.2.1
@utku-pg, I can confirm that the 12.9 version works on a Jetson orin nx running Jetpack 6.2. I use a Docker image running CUDA 12.9, while my driver supports CUDA 12.6. Nvidia has announced general compatibility between all 12.x versions. Therefore, as long as your driver supports CUDA 12, using a different minor version shouldn’t pose any issues.
Both .dev and .io are down as of July 23, 11:00 am.
Are these hosted on someone’s laptop or something?
Any updates? It’s still not working.
Hasn’t the website been restored?
It is down again :(
`Internal Server Error
The server encountered an unexpected internal server error
(generated by waitress)`
So both https://pypi.jetson-ai-lab.dev and https://pypi.jetson-ai-lab.io are down.
https://apt.jetson-ai-lab.io, another server needed for some jetson-containers packages is also inoperable (url is up, but seems to have no content, the downloads result in 404)
this is getting quite frustrating.
+1
Still waiting on both
Yep trying to install today. Not working but it worked yesterday.
I see we are all back here..
Still not working 7/28 12:00AM EST
Still not working 7/29 15:17 HongKong
The README.md file says:
pypi.jetson-ai-lab.dev is down. We are improving security and performance.
Meanwhile please use pypi.jetson-ai-lab.io by setting the enviroment variable INDEX_HOST=jetson-ai-lab.io in .env file.
git reset --hard
echo “INDEX_HOST=jetson-ai-lab.io” > .env
Both .io and .dev don’t work as of 2025-08-01 in Central Europe. Just adding datapoints to the pile. How do you all cope with this long lasting issue? From a scale of 1 to 10, could you please tell me how involved it is to build torch and torchvision from source on device?
From what I found, your guess is close for the .dev
Looks like it was self-hosted? Feels very xkcd 2347. Wishing the best of luck to the team organizing the rehost at .io.
Hi, all
Sorry for the inconvenience.
We have changed the apt server from .dev to .io.
Please download the package from https://pypi.jetson-ai-lab.io/jp6/cu126 instead.
Thanks.