Spaces:
Running
Hey from Hugging Face 🤗
Hey @CISCai - I'm VB, I work on the Open Source team at Hugging Face. We'd be more than happy to help you with CPU + Storage grants.
In addition to this, it'd be great to add some documentation to the space on how to use it and what kind of use-cases does it cover.
Feel free to let me know how you'd like us to proceed.
Wow, that sounds great, thank you! :)
As far as I can tell the free plan is limited to 50GB ephemeral storage. Right now this space is barely using the CPU (I doubt that will change much) and no storage (all data is streamed), however if I were to add repository cloning, a single GGUF file can be up to 50GB in itself, so to be able to perform this the space would temporarily eat up many times that.
What would you think is reasonable?
Regarding the documentation my aim is to make usage as intuitive as possible, but the actual use-cases might not be very obvious to people unless you already have some intimate knowledge of the GGUF metadata and how/if it is used during inferencing.
My plan is to partially improve that by adding some common examples in an About tab, and also add some auto-suggested actions for typical issues when detected. Additionally I'm thinking about adding a short help text for every known metadata explaining what it's for and why you might want to modify/add/remove it (or why you should not if that could potentially break the GGUF).