xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 03:35:17 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Kernel] Add punica dimension for Qwen1.5-32B LoRA (#4850)

Browse Source 
Co-authored-by: Silencio <silencio@adsl-99-6-187-6.dsl.irvnca.sbcglobal.net>
This commit is contained in:
Silencio 2024-05-17 02:16:09 +08:00 committed by GitHub
parent 10fa9eea21
commit 8435b207af
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 3 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
csrc/punica/bgmv/bgmv_config.h
View File

@ -53,6 +53,7 @@ void bgmv_kernel(out_T *__restrict__ Y, const in_T *__restrict__ X,
f(in_T, out_T, W_T, narrow, 22016) \
f(in_T, out_T, W_T, narrow, 24576) \
f(in_T, out_T, W_T, narrow, 27392) \
f(in_T, out_T, W_T, narrow, 27648) \
f(in_T, out_T, W_T, narrow, 28672) \
f(in_T, out_T, W_T, narrow, 32000) \
f(in_T, out_T, W_T, narrow, 32256) \
@ -121,6 +122,7 @@ void bgmv_kernel(out_T *__restrict__ Y, const in_T *__restrict__ X,
f(in_T, out_T, W_T, 22016, narrow) \
f(in_T, out_T, W_T, 24576, narrow) \
f(in_T, out_T, W_T, 27392, narrow) \
f(in_T, out_T, W_T, 27648, narrow) \
f(in_T, out_T, W_T, 28672, narrow) \
f(in_T, out_T, W_T, 32000, narrow) \
f(in_T, out_T, W_T, 32256, narrow) \

1
tests/lora/test_punica.py
View File

@ -79,6 +79,7 @@ H1 = H2 = [
22016,
24576,
27392,
27648,
32000,
32256,
32512,