vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-11 01:24:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

lite bugfixes, buffer size changes, fixed a topk bug.

Browse source
This commit is contained in:
Concedo 2023-06-06 22:38:25 +08:00
parent 2e5edc80e0
commit c046db5197
5 changed files with 12 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

2
otherarch/gpt2_v3.cpp
View file

@ -370,7 +370,7 @@ bool gpt2_eval(
static void * buf = malloc(buf_size);
if (mem_per_token > 0 && (mem_per_token*N*2 + 64u*1024*1024) > buf_size) {
const size_t buf_size_new = 320u*1024*1024 + 1.7*(mem_per_token*N); // add 10% to account for ggml object overhead
const size_t buf_size_new = 320u*1024*1024 + 1.6*(mem_per_token*N); // add 10% to account for ggml object overhead
//printf("\n%s: reallocating buffer from %zu to %zu bytes\n", __func__, buf_size, buf_size_new);
// reallocate