[MoE Refactor] Rename FusedMoE to FusedMoEFactory (#44941)
Signed-off-by: Bill Nell <bnell@redhat.com> Signed-off-by: bnellnm <49004751+bnellnm@users.noreply.github.com> Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
@@ -1,7 +1,7 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
# Test that the interaction between EPLB and FusedMoE Layer is okay
|
||||
# Test that the interaction between EPLB and FusedMoEFactory Layer is okay
|
||||
|
||||
from dataclasses import dataclass
|
||||
|
||||
@@ -16,7 +16,7 @@ from vllm.distributed.parallel_state import (
|
||||
get_eplb_group,
|
||||
get_tp_group,
|
||||
)
|
||||
from vllm.model_executor.layers.fused_moe.layer import FusedMoE
|
||||
from vllm.model_executor.layers.fused_moe import FusedMoEFactory, MoERunner
|
||||
|
||||
from .eplb_utils import distributed_run, set_env_vars_and_device
|
||||
|
||||
@@ -69,8 +69,8 @@ def make_fused_moe_layer(
|
||||
rank: int,
|
||||
layer_idx: int,
|
||||
test_config: TestConfig,
|
||||
) -> FusedMoE:
|
||||
fml = FusedMoE(
|
||||
) -> MoERunner:
|
||||
fml = FusedMoEFactory(
|
||||
num_experts=test_config.num_experts,
|
||||
top_k=test_config.num_topk,
|
||||
hidden_size=test_config.hidden_size,
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
# Test that the interaction between EPLB and FusedMoE Layer is okay for DP w/ NVFP4
|
||||
# Test that the interaction between EPLB and MoERunner Layer is okay for DP w/ NVFP4
|
||||
|
||||
from dataclasses import dataclass
|
||||
|
||||
@@ -19,7 +19,7 @@ from vllm.distributed.parallel_state import (
|
||||
get_eplb_group,
|
||||
)
|
||||
from vllm.forward_context import set_forward_context
|
||||
from vllm.model_executor.layers.fused_moe.layer import FusedMoE
|
||||
from vllm.model_executor.layers.fused_moe.layer import FusedMoEFactory, MoERunner
|
||||
from vllm.model_executor.layers.quantization.modelopt import (
|
||||
ModelOptNvFp4Config,
|
||||
ModelOptNvFp4FusedMoE,
|
||||
@@ -44,7 +44,7 @@ def make_fused_moe_layer(
|
||||
rank: int,
|
||||
layer_idx: int,
|
||||
test_config: TestConfig,
|
||||
) -> FusedMoE:
|
||||
) -> MoERunner:
|
||||
quant_config = None
|
||||
|
||||
device = torch.device(f"cuda:{rank}")
|
||||
@@ -55,7 +55,7 @@ def make_fused_moe_layer(
|
||||
exclude_modules=[],
|
||||
)
|
||||
|
||||
fml = FusedMoE(
|
||||
fml = FusedMoEFactory(
|
||||
num_experts=test_config.num_experts,
|
||||
top_k=test_config.num_topk,
|
||||
hidden_size=test_config.hidden_size,
|
||||
|
||||
Reference in New Issue
Block a user