thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 4

common : refactor common_sampler + grammar logic changes (#17937)

Browse source 
* common : refactor common_sampler + grammar logic changes

* tests : increase max_tokens to get needed response

* batched : fix uninitialized samplers
This commit is contained in:
Georgi Gerganov 2025-12-14 10:11:13 +02:00 • committed by GitHub
parent 3238b1400c
commit 254098a279
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
27 changed files with 372 additions and 293 deletions
Download patch file Download diff file Expand all files Collapse all files

30
examples/batched/batched.cpp
View file

@ -2,6 +2,7 @@
#include "common.h"
#include "log.h"
#include "llama.h"
#include "sampling.h"
#include <algorithm>
#include <cstdio>
@ -64,17 +65,23 @@ int main(int argc, char ** argv) {
ctx_params.n_ctx = n_kv_req;
ctx_params.n_batch = std::max(n_predict, n_parallel);
llama_context * ctx = llama_init_from_model(model, ctx_params);
auto sparams = llama_sampler_chain_default_params();
sparams.no_perf = false;
llama_sampler * smpl = llama_sampler_chain_init(sparams);
std::vector<llama_sampler *> samplers;
llama_sampler_chain_add(smpl, llama_sampler_init_top_k(params.sampling.top_k));
llama_sampler_chain_add(smpl, llama_sampler_init_top_p(params.sampling.top_p, params.sampling.min_keep));
llama_sampler_chain_add(smpl, llama_sampler_init_temp (params.sampling.temp));
llama_sampler_chain_add(smpl, llama_sampler_init_dist (params.sampling.seed));
for (int32_t i = 0; i < n_parallel; ++i) {
llama_sampler * smpl = llama_sampler_chain_init(sparams);
llama_sampler_chain_add(smpl, llama_sampler_init_top_k(params.sampling.top_k));
llama_sampler_chain_add(smpl, llama_sampler_init_top_p(params.sampling.top_p, params.sampling.min_keep));
llama_sampler_chain_add(smpl, llama_sampler_init_temp (params.sampling.temp));
llama_sampler_chain_add(smpl, llama_sampler_init_dist (params.sampling.seed));
samplers.push_back(smpl);
}
llama_context * ctx = llama_init_from_model(model, ctx_params);
if (ctx == NULL) {
LOG_ERR("%s: error: failed to create the llama_context\n" , __func__);
@ -173,7 +180,7 @@ int main(int argc, char ** argv) {
continue;
}
const llama_token new_token_id = llama_sampler_sample(smpl, ctx, i_batch[i]);
const llama_token new_token_id = llama_sampler_sample(samplers[i], ctx, i_batch[i]);
// is it an end of generation? -> mark the stream as finished
if (llama_vocab_is_eog(vocab, new_token_id) || n_cur == n_predict) {
@ -229,14 +236,17 @@ int main(int argc, char ** argv) {
__func__, n_decode, (t_main_end - t_main_start) / 1000000.0f, n_decode / ((t_main_end - t_main_start) / 1000000.0f));
LOG("\n");
llama_perf_sampler_print(smpl);
llama_perf_sampler_print(samplers[0]);
llama_perf_context_print(ctx);
fprintf(stderr, "\n");
llama_batch_free(batch);
llama_sampler_free(smpl);
for (auto & sampler_config : samplers) {
llama_sampler_free(sampler_config);
}
llama_free(ctx);
llama_model_free(model);