Commit ·
dc508e1
1
Parent(s): f8abcf5
Update README.md (#4)
Browse files- Update README.md (774f12081c7df00467e8d0233e14a49f8d913eca)
Co-authored-by: Yao Zhao <zheyishine@users.noreply.huggingface.co>
README.md
CHANGED
|
@@ -121,7 +121,7 @@ python -m sglang.launch_server \
|
|
| 121 |
--dp-size 1 \
|
| 122 |
--mamba-scheduler-strategy extra_buffer \
|
| 123 |
--speculative-algorithm NEXTN \
|
| 124 |
-
--speculative-num-steps
|
| 125 |
--speculative-eagle-topk 1 \
|
| 126 |
--speculative-num-draft-tokens 4 \
|
| 127 |
--mem-fraction-static 0.75 \
|
|
|
|
| 121 |
--dp-size 1 \
|
| 122 |
--mamba-scheduler-strategy extra_buffer \
|
| 123 |
--speculative-algorithm NEXTN \
|
| 124 |
+
--speculative-num-steps 3 \
|
| 125 |
--speculative-eagle-topk 1 \
|
| 126 |
--speculative-num-draft-tokens 4 \
|
| 127 |
--mem-fraction-static 0.75 \
|