public_git/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/whisper.cpp.git synced 2026-10-18 13:17:59 +02:00
Code Issues Packages Projects Releases Wiki Activity

talk-llama : sync llama.cpp

Browse Source
This commit is contained in:
Georgi Gerganov
2026-06-08 14:36:36 +03:00
parent 4df9a57df2
commit 84bd03a438
151 changed files with 3439 additions and 871 deletions
Download Patch File Download Diff File Expand all files Collapse all files
examples/talk-llama/models/modern-bert.cpp
+11 -2
View File
@@ -14,7 +14,15 @@ void llama_model_modern_bert::load_arch_hparams(llama_model_loader & ml) {
ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
switch (hparams.n_layer) {
// Some ModernBert derivatives (e.g. IBM Granite Embedding 97m R2) use
// SiLU/SwiGLU in the FFN instead of the default GELU/GeGLU.
hparams.llm_ffn_op = LLM_FFN_GEGLU;
std::string hidden_act;
if (ml.get_key(LLM_KV_HIDDEN_ACT, hidden_act, false)) {
hparams.llm_ffn_op = llm_ffn_op_type_from_string(hidden_act, LLM_FFN_GEGLU);
}
switch (hparams.n_layer()) {
case 12:
type = LLM_TYPE_47M; break; // granite-embedding-small
case 22:
@@ -144,7 +152,8 @@ llama_model_modern_bert::graph::graph(const llama_model & model, const llm_graph
NULL, NULL, NULL,
model.layers[il].ffn_down, NULL, NULL,
NULL,
LLM_FFN_GEGLU, LLM_FFN_SEQ, il);
hparams.llm_ffn_op,
LLM_FFN_SEQ, il);
// attentions bypass the intermediate layer
cur = ggml_add(ctx0, cur, ffn_inp);