UBC-NLP

/

AraT5v2-base-1024

like 20

Hi there , first i congratulate you all on the amazing work you are doing within this org, i'am a huge fan since 2021 and always waiting for new releases from your Lab.
i'am working this year on my undergrad thesis and we want the outcome of this project to be an LLM for Arabic using fine-tuning instead of training from scratch. I believe you have been doing pretty much same sometimes and i'm interested to know the cost of such fine-tuning processes, of course i'am aware that the cost is related to many factors (model size, dataset size, hardware etc) but would love to hear an estimation for a specific case and then i believe i can adapt it to my configuration in order to have an estimated cost.
Hope to hear from you soon