xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-05-19 10:07:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[ci] fix broken tests (#10956)

Browse Source 
Signed-off-by: youkaichao <youkaichao@gmail.com>
This commit is contained in:
youkaichao 2024-12-06 11:25:47 -08:00 committed by GitHub
parent c05cfb67da
commit dcdc3fafe5
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
vllm/worker/model_runner.py
View File

@ -1782,6 +1782,9 @@ class ModelRunner(GPUModelRunnerBase[ModelInputForGPUWithSamplingMetadata]):
kv_caches: vLLM's paged memory kv_caches: vLLM's paged memory
""" """
if self.vllm_config.kv_transfer_config is None:
return False
prefill_meta = model_input.attn_metadata.prefill_metadata prefill_meta = model_input.attn_metadata.prefill_metadata
# check if the current run is profiling # check if the current run is profiling
@ -1789,9 +1792,6 @@ class ModelRunner(GPUModelRunnerBase[ModelInputForGPUWithSamplingMetadata]):
# check if the current run is prefill # check if the current run is prefill
is_prefill_run = prefill_meta is not None is_prefill_run = prefill_meta is not None
if self.vllm_config.kv_transfer_config is None:
return False
return self.vllm_config.kv_transfer_config.is_kv_consumer and ( return self.vllm_config.kv_transfer_config.is_kv_consumer and (
not is_profile_run) and is_prefill_run not is_profile_run) and is_prefill_run
@ -1807,6 +1807,9 @@ class ModelRunner(GPUModelRunnerBase[ModelInputForGPUWithSamplingMetadata]):
kv_caches: vLLM's paged memory kv_caches: vLLM's paged memory
""" """
if self.vllm_config.kv_transfer_config is None:
return False
prefill_meta = model_input.attn_metadata.prefill_metadata prefill_meta = model_input.attn_metadata.prefill_metadata
# check if the current run is profiling # check if the current run is profiling
@ -1814,9 +1817,6 @@ class ModelRunner(GPUModelRunnerBase[ModelInputForGPUWithSamplingMetadata]):
# check if the current run is prefill # check if the current run is prefill
is_prefill_run = prefill_meta is not None is_prefill_run = prefill_meta is not None
if self.vllm_config.kv_transfer_config is None:
return False
return self.vllm_config.kv_transfer_config.is_kv_producer and ( return self.vllm_config.kv_transfer_config.is_kv_producer and (
not is_profile_run) and is_prefill_run not is_profile_run) and is_prefill_run