CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-07 09:57:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

context : minor naming fix

Browse Source
This commit is contained in:
Georgi Gerganov
2025-02-12 12:41:36 +02:00
parent 8da7f612b7
commit d146a14f77
3 changed files with 13 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/llama-context.cpp
View File

@@ -189,7 +189,7 @@ bool llama_context::apply_adapter_cvec(
return cvec.apply(model, data, len, n_embd, il_start, il_end); return cvec.apply(model, data, len, n_embd, il_start, il_end);
} }
llama_perf_context_data llama_context::get_perf() const { llama_perf_context_data llama_context::perf_get_data() const {
llama_perf_context_data data = {}; llama_perf_context_data data = {};
data.t_start_ms = 1e-3 * t_start_us; data.t_start_ms = 1e-3 * t_start_us;

12
src/llama-context.h
View File

@@ -94,7 +94,6 @@ struct llama_context {
// //
virtual int decode(llama_batch & inp_batch) = 0; virtual int decode(llama_batch & inp_batch) = 0;
// encode a batch of tokens by evaluating the encoder part of the transformer // encode a batch of tokens by evaluating the encoder part of the transformer
// //
// - lctx: llama context // - lctx: llama context
@@ -296,7 +295,7 @@ struct llama_context {
// perf // perf
virtual llama_perf_context_data get_perf() const; virtual llama_perf_context_data perf_get_data() const;
virtual void perf_reset(); virtual void perf_reset();
// members // members
@@ -326,20 +325,21 @@ protected:
bool has_evaluated_once = false; bool has_evaluated_once = false;
mutable int64_t t_start_us; mutable int64_t t_start_us = 0;
mutable int64_t t_load_us; mutable int64_t t_load_us = 0;
mutable int64_t t_p_eval_us = 0; mutable int64_t t_p_eval_us = 0;
mutable int64_t t_eval_us = 0; mutable int64_t t_eval_us = 0;
mutable int64_t t_compute_start_us = 0; mutable int64_t t_compute_start_us = 0;
mutable int64_t n_queued_tokens = 0; mutable int64_t n_queued_tokens = 0;
mutable int32_t n_p_eval = 0; // number of tokens in eval calls for the prompt (with batch size > 1) mutable int32_t n_p_eval = 0; // number of tokens in eval calls for the prompt (with batch size > 1)
mutable int32_t n_eval = 0; // number of eval calls mutable int32_t n_eval = 0; // number of eval calls
}; };
// TODO: make implementation details private // TODO: make implementation details private
struct llama_context_unified : public llama_context { class llama_context_unified : public llama_context {
public:
struct batch_manager; struct batch_manager;
// TODO: tmp until llama-model starts implementing the graph build function // TODO: tmp until llama-model starts implementing the graph build function

12
src/llama.cpp
View File

@@ -54,11 +54,11 @@ enum llm_norm_type {
}; };
struct llm_build_context { struct llm_build_context {
llama_context & lctx; llama_context & lctx;
const llama_model & model; const llama_model & model;
const llama_hparams & hparams; const llama_hparams & hparams;
const llama_cparams & cparams; const llama_cparams & cparams;
const llama_ubatch & ubatch; const llama_ubatch & ubatch;
const int64_t n_embd; const int64_t n_embd;
const int64_t n_layer; const int64_t n_layer;
@@ -7854,7 +7854,7 @@ struct llama_perf_context_data llama_perf_context(const struct llama_context * c
return data; return data;
} }
data = ctx->get_perf(); data = ctx->perf_get_data();
return data; return data;
} }