talk-llama : sync llama.cpp

This commit is contained in:
Georgi Gerganov
2026-07-10 13:06:42 +03:00
parent afa6bffd2b
commit 289ecb05c9
32 changed files with 4809 additions and 313 deletions
+2
View File
@@ -121,6 +121,8 @@ llama_model_qwen3next::graph::graph(const llama_model & model, const llm_graph_p
ggml_tensor * inp_out_ids = build_inp_out_ids();
for (int il = 0; il < n_layer; ++il) {
res->t_layer_inp[il] = inpL;
ggml_tensor * inpSA = inpL;
cur = build_norm(inpL, model.layers[il].attn_norm, nullptr, LLM_NORM_RMS, il);