tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-30 12:55:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

metal : add f16 support

Browse Source
This commit is contained in:
Georgi Gerganov
2023-06-06 20:16:57 +03:00
parent d5b111f53d
commit 44f906e853
3 changed files with 31 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
llama.cpp
View File

@ -961,7 +961,6 @@ static void llama_model_load_internal(
model.hparams = ml->file_loaders.at(0)->hparams;
llama_file_version file_version = ml->file_loaders.at(0)->file_version;
auto & hparams = model.hparams;
uint32_t n_ff = ((2*(4*hparams.n_embd)/3 + hparams.n_mult - 1)/hparams.n_mult)*hparams.n_mult;
{
switch (hparams.n_layer) {
@ -975,6 +974,8 @@ static void llama_model_load_internal(
hparams.n_ctx = n_ctx;
}
const uint32_t n_ff = ((2*(4*hparams.n_embd)/3 + hparams.n_mult - 1)/hparams.n_mult)*hparams.n_mult;
{
fprintf(stderr, "%s: format = %s\n", __func__, llama_file_version_name(file_version));
fprintf(stderr, "%s: n_vocab = %u\n", __func__, hparams.n_vocab);