amd
/

Safetensors
llama
alignment-handbook
Generated from Trainer

Please use safetensors format

#1
by Qubitium - opened

pickle format for weights is prone to safety issues and hf repo does not list the modules/layers when submitted using .bin format.

AMD org

@Qubitium
Thanks so much for bringing this up! We’ve updated the weights to safetensors format, which should make things more reliable and easier to use. Hope you find it helpful!

Sign up or log in to comment