diff --git a/vllm/distributed/device_communicators/shm_object_storage.py b/vllm/distributed/device_communicators/shm_object_storage.py index 352e7525d4c8..0310fc14da25 100644 --- a/vllm/distributed/device_communicators/shm_object_storage.py +++ b/vllm/distributed/device_communicators/shm_object_storage.py @@ -253,7 +253,7 @@ class SingleWriterShmRingBuffer: Args: nbytes (int, optional): The size of the buffer to free. If None, - frees the maximum size of the ring buffer. + frees the maximum size of the ring buffer. ''' assert self.is_writer, "Only the writer can free buffers." diff --git a/vllm/entrypoints/openai/serving_engine.py b/vllm/entrypoints/openai/serving_engine.py index d391cc50ad23..4eb1f8b89d64 100644 --- a/vllm/entrypoints/openai/serving_engine.py +++ b/vllm/entrypoints/openai/serving_engine.py @@ -697,9 +697,7 @@ class OpenAIServing: add_special_tokens: bool = True, ) -> TextTokensPrompt: """ - A simpler implementation of - [`_tokenize_prompt_input_or_inputs`][vllm.entrypoints.openai.serving_engine.OpenAIServing._tokenize_prompt_input_or_inputs] - that assumes single input. + A simpler implementation that tokenizes a single prompt input. """ async for result in self._tokenize_prompt_inputs_async( request, @@ -718,9 +716,7 @@ class OpenAIServing: add_special_tokens: bool = True, ) -> AsyncGenerator[TextTokensPrompt, None]: """ - A simpler implementation of - [`_tokenize_prompt_input_or_inputs`][vllm.entrypoints.openai.serving_engine.OpenAIServing._tokenize_prompt_input_or_inputs] - that assumes multiple inputs. + A simpler implementation that tokenizes multiple prompt inputs. """ for prompt in prompt_inputs: if isinstance(prompt, str):