tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-27 12:05:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : move random seed generation to the samplers (#9398)

Browse Source 
* llama_sampler_penalties : clamp penalty_last_n to zero
This commit is contained in:
slaren
2024-09-10 18:04:25 +02:00
committed by GitHub
parent 00ba2ff781
commit 49006c67b4
10 changed files with 92 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
examples/infill/infill.cpp
View File

@ -159,8 +159,6 @@ int main(int argc, char ** argv) {
print_build_info();
LOG_TEE("%s: seed = %u\n", __func__, params.sparams.seed);
LOG("%s: llama backend init\n", __func__);
llama_backend_init();
llama_numa_init(params.numa);
@ -301,6 +299,9 @@ int main(int argc, char ** argv) {
LOG_TEE("Input suffix: '%s'\n", params.input_suffix.c_str());
}
}
smpl = gpt_sampler_init(model, sparams);
LOG_TEE("sampling seed: %u\n", gpt_sampler_get_seed(smpl));
LOG_TEE("sampling: \n%s\n", sparams.print().c_str());
LOG_TEE("generate: n_ctx = %d, n_batch = %d, n_predict = %d, n_keep = %d\n", n_ctx, params.n_batch, params.n_predict, params.n_keep);
LOG_TEE("\n\n");
@ -340,8 +341,6 @@ int main(int argc, char ** argv) {
std::vector<llama_token> embd;
smpl = gpt_sampler_init(model, sparams);
while (n_remain != 0 || params.interactive) {
// predict
if (!embd.empty()) {