vrr/prima.cpp
2
0
Fork 0
mirror of https://github.com/Lizonghang/prima.cpp.git synced 2025-09-09 16:54:34 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Removed some unnecessary synchronization logic and added n_chunks communication content

Browse source
This commit is contained in:
leeetao 2025-06-27 07:04:10 +00:00
parent a3becb586a
commit 48b7f53abb
4 changed files with 97 additions and 218 deletions
Download patch file Download diff file Expand all files Collapse all files

3
common/common.cpp
View file

@ -1879,7 +1879,7 @@ struct llama_init_result llama_init_from_gpt_params(gpt_params & params) {
params.sparams.ignore_eos = false;
}
if (params.warmup) {
if (0) {
LOG_WRN("%s: warming up the model with an empty run - please wait ...\n", __func__);
const uint32_t my_rank = cparams.rank;
@ -2006,7 +2006,6 @@ struct llama_context_params llama_context_params_from_gpt_params(const gpt_param
}
cparams.next_node_ip = new char[params.next_node_ip.length() + 1];
std::strcpy(cparams.next_node_ip, params.next_node_ip.c_str());
cparams.n_ctx = params.n_ctx;
cparams.n_predict = params.n_predict;
cparams.n_seq_max = params.n_parallel;