Skip to content

Commit 76c4ded

Browse files
committed
up
1 parent 2cc19a3 commit 76c4ded

1 file changed

Lines changed: 5 additions & 2 deletions

File tree

.github/configs/nvidia-master.yaml

Lines changed: 5 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -798,7 +798,7 @@ dsr1-fp8-gb200-dynamo-sglang:
798798
additional-settings:
799799
- "PREFILL_NODES=4"
800800
- "N_ADDITIONAL_FRONTENDS=9"
801-
- "SCRIPT_MODE=max-tpt"
801+
- "SCRIPT_MODE=1k1k-max-tpt"
802802
decode:
803803
num-worker: 1
804804
tp: 1
@@ -820,7 +820,7 @@ dsr1-fp8-gb200-dynamo-sglang:
820820
additional-settings:
821821
- "PREFILL_NODES=1"
822822
- "N_ADDITIONAL_FRONTENDS=9"
823-
- "SCRIPT_MODE=1p_4d"
823+
- "SCRIPT_MODE=1k1k-low-latency"
824824
decode:
825825
num-worker: 4
826826
tp: 1
@@ -842,6 +842,7 @@ dsr1-fp8-gb200-dynamo-sglang:
842842
additional-settings:
843843
- "PREFILL_NODES=6"
844844
- "N_ADDITIONAL_FRONTENDS=9"
845+
- "SCRIPT_MODE=1k1k-max-tpt"
845846
decode:
846847
num-worker: 1
847848
tp: 1
@@ -864,6 +865,7 @@ dsr1-fp8-gb200-dynamo-sglang:
864865
additional-settings:
865866
- "PREFILL_NODES=1"
866867
- "N_ADDITIONAL_FRONTENDS=8"
868+
- "SCRIPT_MODE=8k1k-low-latency"
867869
decode:
868870
num-worker: 1
869871
tp: 1
@@ -883,6 +885,7 @@ dsr1-fp8-gb200-dynamo-sglang:
883885
additional-settings:
884886
- "PREFILL_NODES=2"
885887
- "N_ADDITIONAL_FRONTENDS=8"
888+
- "SCRIPT_MODE=8k1k-max-tpt"
886889
decode:
887890
num-worker: 1
888891
tp: 1

0 commit comments

Comments
 (0)