Fix chunked prefill condition (#1594)

This commit is contained in:
Ke Bao
2024-10-07 14:34:14 +08:00
committed by GitHub
parent 6a5b352aaf
commit 68f8b60d22
2 changed files with 15 additions and 1 deletions

View File

@@ -249,7 +249,10 @@ class PrefillAdder:
return AddReqResult.NO_TOKEN
tokens_freed += tokens_occupied
if req.extend_input_len <= self.rem_chunk_tokens:
if (
self.rem_chunk_tokens is None
or req.extend_input_len <= self.rem_chunk_tokens
):
self.can_run_list.append(req)
self._prefill_one_req(
0,