ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-02-27 16:44:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

add dry sampler (#513)

Browse Source 
* add dry sampler

* use vocab instead of model in dry_init function

* fix compile error for build test

---------

Co-authored-by: firecoperana <firecoperana>
This commit is contained in:
firecoperana
2025-06-19 02:24:53 -05:00
committed by GitHub
parent c5368148cf
commit 3f111ad7bb
21 changed files with 743 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
examples/speculative/speculative.cpp
View File

@@ -179,7 +179,7 @@ int main(int argc, char ** argv) {
bool has_eos = false;
// target model sampling context
struct llama_sampling_context * ctx_sampling = llama_sampling_init(params.sparams);
struct llama_sampling_context * ctx_sampling = llama_sampling_init(llama_get_model_vocab(model_tgt), params.sparams);
// draft sequence data
std::vector<seq_draft> drafts(n_seq_dft);
@@ -190,7 +190,7 @@ int main(int argc, char ** argv) {
}
for (int s = 0; s < n_seq_dft; ++s) {
drafts[s].ctx_sampling = llama_sampling_init(params.sparams);
drafts[s].ctx_sampling = llama_sampling_init(llama_get_model_vocab(model_dft), params.sparams);
}
llama_batch batch_dft = llama_batch_init(params.n_ctx, 0, 1);