7B, 33B and 65B versions?
#2
by
flashvenom
- opened
Are you planning on training adapters for different versions? Also if you don't mind, do you have the code / dataset you used to train these? Would love to experiment more with these SuperHOT LoRA's
I am not releasing the dataset yet, since I am still experimenting with it.
I am currently training 30B 8K, I will upload it shortly. I encourage anyone to make their own LoRA/finetune with this method.
Fair enough, thanks for the information!
qq: are you going to release a 7B version, curious to see how it works