root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-11-15 15:29:53 +00:00
Code Issues Actions 13 Packages Projects Releases Wiki Activity

metal : fix wrong number of tokens per sequence in SSM_SCAN
Some checks failed
flake8 Lint / Lint (push) Has been cancelled
Details

Browse Source
This commit is contained in:
Francis Couture-Harpin 2024-10-02 21:35:50 -04:00
parent 5b8ec2b978
commit 62b09b343c
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-metal.m
View File

@ -1690,7 +1690,7 @@ static void ggml_metal_encode_node(
const int64_t d_inner = ne01; const int64_t d_inner = ne01;
const int64_t n_head = ne02; const int64_t n_head = ne02;
const int64_t n_group = ne41; const int64_t n_group = ne41;
const int64_t n_seq_tokens = ne11; const int64_t n_seq_tokens = ne12;
const int64_t n_seqs = ne13; const int64_t n_seqs = ne13;
id<MTLComputePipelineState> pipeline = nil; id<MTLComputePipelineState> pipeline = nil;