From 1dc151f0f03d273223a82b60f39163e80644a437 Mon Sep 17 00:00:00 2001 From: Chenyaaang Date: Mon, 23 Jun 2025 18:43:17 +0000 Subject: [PATCH] fix tpu model runner test Signed-off-by: Chenyaaang --- tests/v1/tpu/worker/test_tpu_model_runner.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tests/v1/tpu/worker/test_tpu_model_runner.py b/tests/v1/tpu/worker/test_tpu_model_runner.py index 0e7d305fef9e..d22ddf5c7e58 100644 --- a/tests/v1/tpu/worker/test_tpu_model_runner.py +++ b/tests/v1/tpu/worker/test_tpu_model_runner.py @@ -6,6 +6,7 @@ from vllm.attention.layer import Attention from vllm.config import (CacheConfig, ModelConfig, SchedulerConfig, VllmConfig, set_current_vllm_config) +from vllm.pooling_params import PoolingParams from vllm.sampling_params import SamplingParams from vllm.utils import GiB_bytes from vllm.v1.core.kv_cache_utils import (estimate_max_model_len, @@ -71,6 +72,7 @@ def _schedule_new_request(*req_ids: str) -> SchedulerOutput: mm_hashes=[], mm_positions=[], sampling_params=SamplingParams(), + pooling_params=PoolingParams(), block_ids=([0], ), # block_ids should be tuple[list[int]] num_computed_tokens=0, lora_request=None,