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/llama-graph.h
+2 -2
View File
@@ -703,7 +703,7 @@ public:
ggml_tensor * get_logits() const { return t_logits; }
ggml_tensor * get_embd() const { return t_embd; }
ggml_tensor * get_embd_pooled() const { return t_embd_pooled; }
ggml_tensor * get_h_pre_norm() const { return t_h_pre_norm; }
ggml_tensor * get_h_nextn() const { return t_h_nextn; }
ggml_cgraph * get_gf() const { return gf; }
ggml_context * get_ctx() const { return ctx_compute.get(); }
@@ -732,7 +732,7 @@ public:
ggml_tensor * t_logits = nullptr;
ggml_tensor * t_embd = nullptr;
ggml_tensor * t_embd_pooled = nullptr;
ggml_tensor * t_h_pre_norm = nullptr; // [n_embd, n_outputs] hidden state before final output norm
ggml_tensor * t_h_nextn = nullptr; // [n_embd, n_outputs] hidden state before final output norm
std::map<llama_seq_id, ggml_tensor*> t_sampled_logits;
std::map<llama_seq_id, ggml_tensor*> t_candidates;