That isn’t an error … its how git works.
Do a …
git stash
then a
git pull
and you should be golden.
You can’t run the fastai DL stuff or pytorch under windows or its Linux subsystem
I thought it was possible, I’m sorry.
Looking at the link https://docs.google.com/document/d/1qApf-M3xxHQYeyUTG5SsuC7Dbc0zsB5fUeMRevqYh2I/edit and following the steps
I supposed it was possible to use my GPU directly from the conda environment under linux subsystem.
Big misunderstanding. I have to install ubuntu on another partition or use Crestle/Paperspace.
Thank you for the clarification,
W.
Sorry for not making that clear!
Hello @Waltz,
I put together a gist on GitHub with a group of scripts for installing the fastai repo with working environment on Ubuntu: I thought you might find it convenient in case you try to do a local installation.
I used the scripts also for installing on AWS, but this is not so relevant since we are going to use the official AWS AMI released by @jeremy.
1 Like
Hi Roberto,
thank you!! Your help is greatly appreciated. The script is really useful. Sunday I installed Ubuntu and yesterday night followed the second lesson. As soon as I have some free time I’ll try to install what is missing to me, the cuda stuf.
I was able to run the lesson1 notebook using the CPU.
It worked perfectly the first time. BTW not clear to me why some hours later, the same notebook gave me an error regarding the missing module cv2. Mah!
When in the notebook check that the selected kernel is the ‘fast’ one and see if the error keeps showing