convert : add Llama4ForCausalLM (#16042)

* convert : add Llama4ForCausalLM

* handle swa

* half working version

* fix use_kq_norm

* fix use_kq_norm
This commit is contained in:
Xuan-Son Nguyen
2025-09-18 00:18:21 +07:00
committed by GitHub
parent c959b676be
commit 8f8f2274ee
4 changed files with 50 additions and 12 deletions

View File

@@ -149,7 +149,7 @@ struct llama_hparams {
bool causal_attn = true;
bool use_alibi = false;
bool attn_soft_cap = false;
bool use_kq_norm = true;
bool use_kq_norm = false;
// for Classifiers
uint32_t n_cls_out = 1;