Use a single workspace for flashinfer (#1077)

2024-08-14 19:25:37 -07:00
parent 6767e2229f
commit 326df4bab2
5 changed files with 16 additions and 18 deletions
--- a/benchmark/gsm8k/bench_sglang.py
+++ b/benchmark/gsm8k/bench_sglang.py
@@ -64,7 +64,7 @@ def main(args):
    @sgl.function
    def few_shot_gsm8k(s, question):
        s += few_shot_examples + question
-        s += sgl.gen("answer", max_tokens=512, stop="Question")
+        s += sgl.gen("answer", max_tokens=512, stop=["Question", "Assistant:"])

    #####################################
    ########## SGL Program End ##########