FASTAI2 HWL -- hardware compatibility

Is there any HWL or platform compatibility list for Fastai2 and different courses. I just wonder how worthy of doing that. How worthy to write code to share the platform/HW info from the notebook? If something is not working, how do we figure it out is hardware compatibility, platform issues, or coding problem. It would be like if Fastai NLP can help with that. How difficult to classified the error messages to resolve the issues.

Generally this is why it is recommended to use the pre-built images and cloud-based resources available. Dealing with hardware, drivers, and such is complicated problem, best left to after going through the course generally.

Hi Molly,
Thank you for your reply. There is a trade off, to have standardized image, but we still need to have the hardware match with it. Especially, if there are many learner are running the program at the same time, with the same setting. I just wonder whether there will be the overloading the service provider. If there is a button or a cell to share the platform or config with a date stamp about what’s working and what’s not, and send those info back somewhere. Then the new users may find the best option for them. As Jeremy mentioned that learning AI, do not require expensive machine. Even there are free GPU machine providers, e.g. like Collab. They may change their policy and their product significantly. It would be a good, if there can be a deal between fastai and the service provider, like google, especially free service or subscription by fastai, and share by the fastai student at a lower cost.

Okay, the current supported platforms are here: FAQ, resources, and official course updates ✅

If there are any issues with platforms then they would be removed.

1 Like