Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

qwen35: use post-norm hidden state for MTP (#24025)

Browse Source 
* qwen35: use post-norm hidden state for MTP

* rename pre_norm to nextn

* fix step35
This commit is contained in:
Aman Gupta
2026-06-04 01:29:09 +08:00
committed by GitHub
parent c8d6a00636
commit 166fe29492
12 changed files with 132 additions and 139 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/models/step35.cpp
+6 -6
View File
@@ -294,7 +294,7 @@ llama_model_step35::graph::graph(const llama_model & model, const llm_graph_para
cb(cur, "attn_proj", il);
}
if (il == n_transformer_layers - 1 && inp_out_ids && cparams.embeddings_pre_norm_masked) {
if (il == n_transformer_layers - 1 && inp_out_ids && cparams.embeddings_nextn_masked) {
cur = ggml_get_rows(ctx0, cur, inp_out_ids);
inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
}
@@ -353,10 +353,10 @@ llama_model_step35::graph::graph(const llama_model & model, const llm_graph_para
cur = inpL;
cb(cur, "h_pre_norm", -1);
res->t_h_pre_norm = cur;
cb(cur, "h_nextn", -1);
res->t_h_nextn = cur;
if (!cparams.embeddings_pre_norm_masked && inp_out_ids) {
if (!cparams.embeddings_nextn_masked && inp_out_ids) {
cur = ggml_get_rows(ctx0, cur, inp_out_ids);
}
@@ -541,8 +541,8 @@ llama_model_step35::graph_mtp::graph_mtp(const llama_model & model, const llm_gr
cb(cur, "mtp_post_ffn", il);
// Pre-norm hidden state: used by the AR draft loop to seed the next MTP step.
cb(cur, "h_pre_norm", -1);
res->t_h_pre_norm = cur;
cb(cur, "h_nextn", -1);
res->t_h_nextn = cur;
ggml_tensor * head_norm_w = layer.nextn.shared_head_norm
? layer.nextn.shared_head_norm