vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama : adapt to F16 KQ_pos

Browse source
This commit is contained in:
Georgi Gerganov 2024-02-19 13:10:24 +02:00
parent 31109ca00a
commit f249c997a8
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
4 changed files with 13 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml.c
View file

@ -5192,7 +5192,7 @@ static struct ggml_tensor * ggml_soft_max_impl(
GGML_ASSERT(mask->type == GGML_TYPE_F16);
GGML_ASSERT(ggml_is_contiguous(mask));
GGML_ASSERT(ggml_is_matrix(mask));
GGML_ASSERT(ggml_can_repeat_rows(mask, a));
GGML_ASSERT(mask->ne[1] >= a->ne[1]);
}
if (pos) {