CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-04 09:32:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : fix worst case graph build

Browse Source
This commit is contained in:
Georgi Gerganov
2023-09-19 11:05:08 +03:00
parent fa0e677820
commit daf4c6d360
3 changed files with 100 additions and 81 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
common/common.cpp
View File

@@ -781,7 +781,7 @@ std::tuple<struct llama_model *, struct llama_context *> llama_init_from_gpt_par
std::vector<llama_token> tmp = { llama_token_bos(lctx), llama_token_eos(lctx), };
llama_decode(lctx, llama_batch_get_one(tmp.data(), std::min(tmp.size(), (size_t) params.n_batch), 0, 0), params.n_threads);
llama_kv_cache_keep_seq(lctx, -1);
llama_kv_cache_rm_tokens(lctx, -1, -1);
llama_reset_timings(lctx);
}