Fails to load: missing tensor blk.0.ffn_down_exps.weight
#1
by
Cheuwambel
- opened
Yes, as you see form the link you provided, you need the llama.cpp version that supports this quant. Unless this is some kind of requant-request?