CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-28 08:31:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml : sync ggml (add GPT-NeoX RoPE implementation)

Browse Source
This commit is contained in:
Georgi Gerganov
2023-04-20 23:32:59 +03:00
parent 9ff334f3c9
commit 12b5900dbc
3 changed files with 49 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
ggml.h
View File

@@ -630,7 +630,8 @@ struct ggml_tensor * ggml_soft_max(
// rotary position embedding
// in-place, returns view(a)
// if mode == 1, skip n_past elements
// if mode & 1 == 1, skip n_past elements
// if mode & 2 == 1, GPT-NeoX style
// TODO: avoid creating a new tensor every time
struct ggml_tensor * ggml_rope(
struct ggml_context * ctx,