PR #4283 Support attention_bias on LLaMA architecture

Support attention_bias on LLaMA architecture #4283

ggerganov merged 3 commits into ggml-org:master from RealJosephus:llama_attention_bias

Support attention_bias on LLaMA architecture

c48679a8

slaren commented on 2023-12-01

check existence of qkvo bias while loading llama models

e192572d

Update llama.cpp

b1efaed3

RealJosephus requested a review from

ggerganov 2 years ago

ggerganov approved these changes on 2023-12-01

ggerganov merged 03562f3a into master 2 years ago

cebtenzzre commented on 2023-12-01

Reviewers

ggerganov

slaren

cebtenzzre

Assignees

No one assigned

Labels

None yet

Milestone

No milestone