[V0 Deprecation] Remove V0 executors (#27142)
Signed-off-by: Nick Hill <nhill@redhat.com>
This commit is contained in:
@@ -39,7 +39,7 @@ from vllm.v1.engine.exceptions import EngineDeadError, EngineGenerateError
|
||||
from vllm.v1.engine.output_processor import OutputProcessor, RequestOutputCollector
|
||||
from vllm.v1.engine.parallel_sampling import ParentRequest
|
||||
from vllm.v1.engine.processor import Processor
|
||||
from vllm.v1.executor.abstract import Executor
|
||||
from vllm.v1.executor import Executor
|
||||
from vllm.v1.metrics.loggers import (
|
||||
StatLoggerFactory,
|
||||
StatLoggerManager,
|
||||
|
||||
+2
-11
@@ -60,7 +60,7 @@ from vllm.v1.engine.utils import (
|
||||
EngineZmqAddresses,
|
||||
get_device_indices,
|
||||
)
|
||||
from vllm.v1.executor.abstract import Executor
|
||||
from vllm.v1.executor import Executor
|
||||
from vllm.v1.kv_cache_interface import KVCacheConfig
|
||||
from vllm.v1.metrics.stats import SchedulerStats
|
||||
from vllm.v1.outputs import ModelRunnerOutput
|
||||
@@ -322,7 +322,6 @@ class EngineCore:
|
||||
with self.log_error_detail(scheduler_output):
|
||||
model_output = self.model_executor.execute_model(scheduler_output)
|
||||
|
||||
assert isinstance(model_output, ModelRunnerOutput)
|
||||
engine_core_outputs = self.scheduler.update_from_output(
|
||||
scheduler_output, model_output
|
||||
)
|
||||
@@ -364,7 +363,7 @@ class EngineCore:
|
||||
if self.scheduler.has_requests():
|
||||
scheduler_output = self.scheduler.schedule()
|
||||
future = self.model_executor.execute_model(scheduler_output, non_block=True)
|
||||
batch_queue.appendleft((future, scheduler_output)) # type: ignore[arg-type]
|
||||
batch_queue.appendleft((future, scheduler_output))
|
||||
|
||||
model_executed = scheduler_output.total_num_scheduled_tokens > 0
|
||||
if (
|
||||
@@ -463,14 +462,6 @@ class EngineCore:
|
||||
) -> list[_R]:
|
||||
return self.model_executor.collective_rpc(method, timeout, args, kwargs)
|
||||
|
||||
def save_tensorized_model(
|
||||
self,
|
||||
tensorizer_config,
|
||||
) -> None:
|
||||
self.model_executor.save_tensorized_model(
|
||||
tensorizer_config=tensorizer_config,
|
||||
)
|
||||
|
||||
def preprocess_add_request(self, request: EngineCoreRequest) -> tuple[Request, int]:
|
||||
"""Preprocess the request.
|
||||
|
||||
|
||||
@@ -46,7 +46,7 @@ from vllm.v1.engine.utils import (
|
||||
CoreEngineProcManager,
|
||||
launch_core_engines,
|
||||
)
|
||||
from vllm.v1.executor.abstract import Executor
|
||||
from vllm.v1.executor import Executor
|
||||
from vllm.v1.serial_utils import MsgpackDecoder, MsgpackEncoder, bytestr
|
||||
|
||||
logger = init_logger(__name__)
|
||||
|
||||
@@ -32,7 +32,7 @@ from vllm.v1.engine.core_client import EngineCoreClient
|
||||
from vllm.v1.engine.output_processor import OutputProcessor
|
||||
from vllm.v1.engine.parallel_sampling import ParentRequest
|
||||
from vllm.v1.engine.processor import Processor
|
||||
from vllm.v1.executor.abstract import Executor
|
||||
from vllm.v1.executor import Executor
|
||||
from vllm.v1.metrics.loggers import StatLoggerFactory, StatLoggerManager
|
||||
from vllm.v1.metrics.reader import Metric, get_metrics_snapshot
|
||||
from vllm.v1.metrics.stats import IterationStats
|
||||
|
||||
@@ -23,7 +23,7 @@ from vllm.ray.ray_env import get_env_vars_to_copy
|
||||
from vllm.utils import get_mp_context
|
||||
from vllm.utils.network_utils import get_open_zmq_ipc_path, zmq_socket_ctx
|
||||
from vllm.v1.engine.coordinator import DPCoordinator
|
||||
from vllm.v1.executor.abstract import Executor
|
||||
from vllm.v1.executor import Executor
|
||||
from vllm.v1.utils import get_engine_client_zmq_addr, shutdown
|
||||
|
||||
if TYPE_CHECKING:
|
||||
|
||||
Reference in New Issue
Block a user