Are there any plans to publish a version of the model with only pruning and no distillation?
#11
by
kurogane
- opened
Do you plan to publish a version of the model with pruning only and no distillation?
I want to reproduce and test this pruning and distillation procedure, so I would like a version of the model without distillation.
Or do you plan to publish information on which parameters of llama-3.1-8b-base you actually removed?
I am interested in what happens when the exact same parameter deletion is done for another llama-3.1-8b-based model.
Thank you for taking the time to read this.
Translated by DeepL so sorry it was not proper English.
I would be happy to be considered.γ