Running 8b-lm
#3
by
Ehsanjahanbakhsh
- opened
Is HF code for running this still in progress?
Can we finetune this model using LoRA or all-weight fine-tuning?
I never got it to work. I didn't manage to run the original t5x weights either, so it's tricky to find where the issue is without having expected layer activations to compare with.