Fix log message in scheduler (#652)

20044cab · Lily Liu · GitHub · 64f23c29 · 20044cab
Unverified Commit 20044cab authored Aug 02, 2023 by Lily Liu Committed by GitHub Aug 02, 2023
Hide whitespace changes
Inline Side-by-side

Showing with 5 additions and 5 deletions

vllm/core/scheduler.py vllm/core/scheduler.py +5 -5

No files found.
--- a/vllm/core/scheduler.py
+++ b/vllm/core/scheduler.py
@@ -190,13 +190,13 @@ class Scheduler:
                    break
                num_prompt_tokens = seq_group.get_seqs()[0].get_len()
-                if num_prompt_tokens > min(
+                prompt_limit = min(
-                        self.scheduler_config.max_model_len,
+                    self.scheduler_config.max_model_len,
-                        self.scheduler_config.max_num_batched_tokens):
+                    self.scheduler_config.max_num_batched_tokens)
+                if num_prompt_tokens > prompt_limit:
                    logger.warning(
                        f"Input prompt ({num_prompt_tokens} tokens) is too long"
-                        " and exceeds limit of "
+                        f" and exceeds limit of {prompt_limit}")
-                        f"{self.scheduler_config.max_model_len}")
                    for seq in seq_group.get_seqs():
                        seq.status = SequenceStatus.FINISHED_IGNORED
                    ignored_seq_groups.append(seq_group)