[Core] Support min_tokens with speculative decoding (#32642)

Signed-off-by: qianlihuang <yiliu.dong@qq.com> Co-authored-by: qianlihuang <yiliu.dong@qq.com>
2026-02-27 01:31:28 +08:00
parent 99c7892c5b
commit d940607629
7 changed files with 75 additions and 11 deletions
--- a/tests/v1/e2e/test_async_scheduling.py
+++ b/tests/v1/e2e/test_async_scheduling.py
@@ -32,8 +32,7 @@ example_prompts = [first_prompt, "In one word, the capital of France is "] + [
 default_params = dict(
    temperature=0.0,  # greedy
    max_tokens=30,
-    # spec decoding currently doesn't support min_tokens
-    # min_tokens=28,
+    min_tokens=28,
 )