From 696a43c94cb088a000ec6fea4dad6729a17956c7 Mon Sep 17 00:00:00 2001 From: sang Date: Fri, 14 Jun 2024 03:56:13 -0700 Subject: [PATCH 1/2] . --- vllm/config.py | 3 +++ 1 file changed, 3 insertions(+) diff --git a/vllm/config.py b/vllm/config.py index d9e4a619ee01..8ae23b551144 100644 --- a/vllm/config.py +++ b/vllm/config.py @@ -1092,6 +1092,9 @@ def verify_with_scheduler_config(self, scheduler_config: SchedulerConfig): "Due to limitations of the custom LoRA CUDA kernel, " "max_num_batched_tokens must be <= 65528 when " "LoRA is enabled.") + if scheduler_config.chunked_prefill_enabled: + raise ValueError( + "Lora is not supported with chunked prefill yet.") @dataclass From 7c9f3a4d1c86909af04dd8b8e280f1940a97f01f Mon Sep 17 00:00:00 2001 From: Cyrus Leung Date: Sat, 15 Jun 2024 21:13:43 +0800 Subject: [PATCH 2/2] Fix linter error and use proper capitalization --- vllm/config.py | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/vllm/config.py b/vllm/config.py index 8ae23b551144..54f36e1d6678 100644 --- a/vllm/config.py +++ b/vllm/config.py @@ -1093,8 +1093,7 @@ def verify_with_scheduler_config(self, scheduler_config: SchedulerConfig): "max_num_batched_tokens must be <= 65528 when " "LoRA is enabled.") if scheduler_config.chunked_prefill_enabled: - raise ValueError( - "Lora is not supported with chunked prefill yet.") + raise ValueError("LoRA is not supported with chunked prefill yet.") @dataclass