[feat] group gemm
This commit is contained in:
@@ -15,7 +15,7 @@ command:
|
||||
- -tp
|
||||
- '4'
|
||||
- --max-num-seqs
|
||||
- '1'
|
||||
- '2'
|
||||
- --disable-log-requests
|
||||
- --disable-frontend-multiprocessing
|
||||
- --max-num-batched-tokens
|
||||
@@ -46,4 +46,4 @@ env:
|
||||
- name: BI100_GDN_RESTORE_MODE
|
||||
value: hybrid64
|
||||
- name: BI100_MOE_COREX_TOPK_SOFTMAX
|
||||
value: '1'
|
||||
value: '1'
|
||||
Reference in New Issue
Block a user