[AsyncScheduling] Don't schedule past request max_tokens (#27922)

Signed-off-by: Nick Hill <nhill@redhat.com>
2025-11-04 09:06:28 -08:00
parent c9f66da8fd
commit 938a81692e
3 changed files with 14 additions and 4 deletions
--- a/tests/v1/e2e/test_spec_decode.py
+++ b/tests/v1/e2e/test_spec_decode.py
@@ -155,7 +155,6 @@ def test_suffix_decoding_acceptance(
    )

    # Run several times and check that the accepted tokens increase.
-    spec_llm.chat(test_prompts, sampling_config)
    num_draft = []
    num_accept = []
    for i in range(10):  # Run multiple times to warm up the cache.