Superminaren/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-09-27 21:46:57 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge remote-tracking branch 'upstream/master' into backend-sampling

Browse Source
This commit is contained in:
Daniel Bevenius
2025-11-21 16:38:32 +01:00
parent 9b2439347f 23bc779a6e
commit 79b8cf2a75
63 changed files with 1519 additions and 1118 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-sampling.cpp
+3 -6
View File
@@ -534,9 +534,6 @@ static void llama_sampler_chain_reset(struct llama_sampler * smpl) {
for (auto * smpl : chain->samplers) {
llama_sampler_reset(smpl);
}
chain->t_sample_us = 0;
chain->n_sample = 0;
}
static struct llama_sampler * llama_sampler_chain_clone(const struct llama_sampler * smpl) {
@@ -2854,8 +2851,7 @@ struct llama_perf_sampler_data llama_perf_sampler(const struct llama_sampler * c
void llama_perf_sampler_print(const struct llama_sampler * chain) {
const auto data = llama_perf_sampler(chain);
LLAMA_LOG_INFO("%s: sampling time = %10.2f ms / %5d runs (%8.2f ms per token, %8.2f tokens per second)\n",
__func__, data.t_sample_ms, data.n_sample, data.t_sample_ms / data.n_sample, 1e3 / data.t_sample_ms * data.n_sample);
LLAMA_LOG_INFO("%s: samplers time = %10.2f ms / %5d runs\n", __func__, data.t_sample_ms, data.n_sample);
}
void llama_perf_sampler_reset(struct llama_sampler * chain) {
@@ -2865,5 +2861,6 @@ void llama_perf_sampler_reset(struct llama_sampler * chain) {
auto * ctx = (struct llama_sampler_chain *) chain->ctx;
ctx->t_sample_us = ctx->n_sample = 0;
ctx->t_sample_us = 0;
ctx->n_sample = 0;
}