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

vulkan : add dev notes

Browse source
This commit is contained in:
Georgi Gerganov 2024-05-10 15:56:25 +03:00
parent 536983b1ad
commit 397b1f8f9d
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
2 changed files with 11 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

3
ggml-vulkan.cpp
View file

@ -4285,6 +4285,9 @@ static void ggml_vk_soft_max(ggml_backend_vk_context * ctx, vk_context * subctx,
const float m0 = powf(2.0f, -(max_bias ) / n_head_log2);
const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2);
#pragma message("TODO: src2 is no longer used in soft_max - should be removed and ALiBi calculation should be updated")
#pragma message("ref: https://github.com/ggerganov/llama.cpp/pull/7192")
ggml_vk_op_f32<vk_op_soft_max_push_constants>(ctx, subctx, src0, src1, src2, dst, GGML_OP_SOFT_MAX, {
ncols,
src1 != nullptr ? nrows_y : (uint32_t)0,