feat(grpc): add KV event source discovery (#50033)
Signed-off-by: Connor Carpenter <connorc@nvidia.com> Co-authored-by: Nick Hill <nickhill123@gmail.com>
This commit is contained in:
co-authored by
Nick Hill
parent
2ecd8645d8
commit
48fc2e2797
@@ -11,6 +11,7 @@ import msgspec
|
||||
import numpy as np
|
||||
import torch
|
||||
|
||||
from vllm.config.kv_events import KVEventsConfig
|
||||
from vllm.lora.request import LoRARequest
|
||||
from vllm.multimodal.inputs import MultiModalFeatureSpec
|
||||
from vllm.pooling_params import PoolingParams
|
||||
@@ -90,6 +91,7 @@ class EngineCoreReadyResponse:
|
||||
# KV cache capacity (None for encoder-only/attention-free models).
|
||||
kv_cache_size_tokens: int | None = None
|
||||
kv_cache_max_concurrency: float | None = None
|
||||
kv_events_config: KVEventsConfig | None = None
|
||||
|
||||
|
||||
class EngineCoreRequest(
|
||||
|
||||
@@ -1639,6 +1639,7 @@ class EngineCoreProc(EngineCore):
|
||||
max_num_seqs=scheduler_config.max_num_seqs,
|
||||
max_num_batched_tokens=scheduler_config.max_num_batched_tokens,
|
||||
instance_id=self.vllm_config.instance_id,
|
||||
kv_events_config=self.scheduler.get_kv_event_publisher_config(),
|
||||
)
|
||||
|
||||
def process_input_sockets(
|
||||
|
||||
Reference in New Issue
Block a user