upstream-archive/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-10-02 02:47:26 -05:00
Code Issues Packages Projects Releases Wiki Activity

mimo2 : expose t_layer_inp for the DFlash drafter

Browse Source 
Assisted-by: pi:llama.cpp/Qwen3.8-27B
This commit is contained in:
Georgi Gerganov
2026-09-28 18:05:48 +03:00
parent 1ea03b5701
commit 52cfe0f4aa
1 changed files with 12 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/models/mimo2.cpp
+12 -1
View File
@@ -102,9 +102,12 @@ llama_model_mimo2::graph::graph(const llama_model & model, const llm_graph_param
const float v_scale = hparams.f_attn_value_scale;
const bool emit_h_nextn = cparams.embeddings_nextn;
const bool crop_last_layer = inp_out_ids && (!emit_h_nextn || cparams.embeddings_nextn_masked);
const bool extract_final_inp = (size_t) n_layer < cparams.embeddings_layer_inp.size() && cparams.embeddings_layer_inp[n_layer];
const bool crop_last_layer = inp_out_ids && (!emit_h_nextn || cparams.embeddings_nextn_masked) && !extract_final_inp;
for (int il = 0; il < n_layer; ++il) {
res->t_layer_inp[il] = inpL;
ggml_tensor * inpSA = inpL;
uint32_t n_head_l = hparams.n_head(il);
@@ -232,6 +235,14 @@ llama_model_mimo2::graph::graph(const llama_model & model, const llm_graph_param
cur = inpL;
if (extract_final_inp) {
res->t_layer_inp[n_layer] = cur;
if (inp_out_ids && (!emit_h_nextn || cparams.embeddings_nextn_masked)) {
cur = ggml_get_rows(ctx0, cur, inp_out_ids);
}
}
if (emit_h_nextn) {
cb(cur, "h_nextn", -1);
res->t_h_nextn = cur;