Kaggle has started banning colabcode juypterlab
kuro7766 opened this issue · comments
It seems that kaggle also starts to ban colabcode juypterlab. Notebook kernel will shutdown when installing jupyter related pip dependencies.
Damn. My research has come to a COMPLETE stop because of this. I was really hoping this wasn't the case. I am facing the same issue from the last 3 days. This is very sad. My whole research depends on this library !!!
Using colabcode was a huge pivotal moment in my AI research. Using kaggle cells sucks!! I cant create production level code using kaggle cells.
Any other workaround?
I think there may be an alternative method using VS Code containers?
here: https://towardsdatascience.com/how-to-kaggle-the-engineer-way-act-1-vs-code-containers-b3279970c029
here: https://code.visualstudio.com/docs/remote/containers
I solved it with code-server and pyngrok
- Install Code-server
!curl -fsSL https://code-server.dev/install.sh | sh
- Install pyngrok & get the tunnel URL (you can't run background process in kaggle)
!pip install pyngrok
- Input ngrok auth token and get tunnel http 80.
from pyngrok import conf, ngrok
conf.get_default().auth_token = "<YOUR TOKEN>"
http_tunnel = ngrok.connect()
ssh_tunnel = ngrok.connect(80, "http")
- Get the tunnel URL
from pyngrok import ngrok
# [<NgrokTunnel: "http://<public_sub>.ngrok.io" -> "http://localhost:80">]
tunnels = ngrok.get_tunnels()
tunnels
- Get code-server default password
!cat ~/.config/code-server/config.yaml
- run code-server
!code-server --port 80
Hopefully this workaround will work for a while.
Does this work still?
Does this work still?
this still work , thanks/🥰
@hongvin you are a hero, thanks