Apply for community grant: Academic project
I'm a high school student competing at the Israeli young scientist competition. My project discovered a faster way to use Causal Language Models (CLMs) when generating long texts.
I want to show-case my algorithm with large language models using hugging face spaces in the competition for three days from March 12-14,.
I will need my space running with a GPU for about 24-30 hours in those days.
My project: https://github.com/yonikremer/grouped_sampling.
Very cool project @yonikremer ! Congratulations on your technique and algorithm.
We have granted you a GPU for your project with a sleep time of 15 minutes (so after 15 minutes of inactivity it goes to sleep) - but anyone visiting the space can then re-activate it. From March 12-14th I will increase that period to 72h to make sure you have availability across the competition.
Thanks!!!
@multimodalart
Seems like T4 wouldn't be enough for running medium sized language models such as llama 13 B. Can You upgrade it to a A100?
BY the way, I will need the demo for only few hours.