Skip to content

Commit 3e6ef8e

Browse files
authored
Probable typo (LostRuins#1287)
1 parent 263d49d commit 3e6ef8e

File tree

1 file changed

+2
-2
lines changed

1 file changed

+2
-2
lines changed

gpttype_adapter.cpp

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -629,14 +629,14 @@ struct kcpp_embd_batch { //duplcated from llava_embd_batch
629629
};
630630

631631
//loads a model for speculative decoding.
632-
static void speculative_decoding_setup(std::string spec_model_filename, const llama_model_params & base_model_params, const llama_context_params & base_ctx_params, int base_n_vocab, const float * draft_gpusplit, int draftgpulayers)
632+
static void speculative_decoding_setup(std::string spec_model_filename, const llama_model_params & base_model_params, const llama_context_params & base_ctx_params, int base_n_vocab, const float * draft_gpusplit, int draft_gpulayers)
633633
{
634634
llama_model_params draft_model_params = llama_model_default_params();
635635
llama_context_params draft_ctx_params = llama_context_default_params();
636636

637637
draft_model_params.use_mmap = base_model_params.use_mmap;
638638
draft_model_params.use_mlock = base_model_params.use_mlock;
639-
draft_model_params.n_gpu_layers = draftgpulayers; //layers offload the speculative model.
639+
draft_model_params.n_gpu_layers = draft_gpulayers; //layers offload the speculative model.
640640
draft_ctx_params.n_ctx = base_ctx_params.n_ctx;
641641
draft_ctx_params.logits_all = false;
642642
draft_ctx_params.offload_kqv = base_ctx_params.offload_kqv;

0 commit comments

Comments
 (0)