happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

llama: fix early stop in params_fit if ctx is set (#18070)

This commit is contained in:
Johannes Gäßler 2025-12-16 14:24:00 +01:00 committed by GitHub
parent 59977eba7b
commit ec98e20021
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 7 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
src/llama.cpp
View File

@ -241,6 +241,13 @@ static void llama_params_fit_impl(
global_surplus += memory_reduction;
LLAMA_LOG_INFO("%s: context size reduced from %" PRIu32 " to %" PRIu32 " -> need %" PRId64 " MiB less memory in total\n",
__func__, hp_nct, cparams->n_ctx, memory_reduction/MiB);
if (global_surplus >= 0) {
if (nd == 1) {
LLAMA_LOG_INFO("%s: entire model can be fit by reducing context\n", __func__);
return;
}
LLAMA_LOG_INFO("%s: entire model should be fit across devices by reducing context\n", __func__);
}
} else {
LLAMA_LOG_INFO("%s: default model context size is %" PRIu32 " which is <= the min. context size of %" PRIu32 " -> no change\n",
__func__, hp_nct, n_ctx_min);
@ -249,10 +256,6 @@ static void llama_params_fit_impl(
LLAMA_LOG_INFO("%s: context size set by user to %" PRIu32 " -> no change\n", __func__, cparams->n_ctx);
}
}
if (global_surplus >= 0) {
LLAMA_LOG_INFO("%s: entire model can be fit across devices by reducing context\n", __func__);
return;
}
}
if (mparams->n_gpu_layers != default_mparams.n_gpu_layers) {