[CI] Remove compatibility maintenance for vllm v0.10.1 and v0.10.1.1 (#2840)
### What this PR does / why we need it?
Remove compatibility maintenance for vllm v0.10.1 and v0.10.1.1
### Does this PR introduce _any_ user-facing change?
branch main of vllm-ascend will not be compatible with vllm v0.10.1 and
v0.10.1.1
### How was this patch tested?
CI passed with existing test.
- vLLM version: v0.10.1.1
- vLLM main:
6fb2788163
---------
Signed-off-by: MengqingCao <cmq0113@163.com>
This commit is contained in:
Mengqing Cao
@@ -34,7 +34,7 @@ from vllm_ascend.ops.moe.experts_selector import select_experts
|
||||
from vllm_ascend.ops.moe.moe_comm_method import (AllGatherCommImpl,
|
||||
AlltoAllCommImpl, MC2CommImpl)
|
||||
from vllm_ascend.ops.moe.token_dispatcher import setup_token_dispatchers
|
||||
from vllm_ascend.utils import ACL_FORMAT_FRACTAL_NZ, is_310p, vllm_version_is
|
||||
from vllm_ascend.utils import ACL_FORMAT_FRACTAL_NZ, is_310p
|
||||
|
||||
original_unquantized_fused_moe_init_func = UnquantizedFusedMoEMethod.__init__
|
||||
|
||||
@@ -137,67 +137,6 @@ def unquantized_fused_moe_init_func(self, *args, **kwargs):
|
||||
self.transpose = True
|
||||
|
||||
|
||||
def forward_oot_v01011(
|
||||
self,
|
||||
layer: torch.nn.Module,
|
||||
x: torch.Tensor,
|
||||
use_grouped_topk: bool,
|
||||
top_k: int,
|
||||
router_logits: torch.Tensor,
|
||||
renormalize: bool,
|
||||
topk_group: Optional[int] = None,
|
||||
num_expert_group: Optional[int] = None,
|
||||
custom_routing_function: Optional[Callable] = None,
|
||||
scoring_func: str = "softmax",
|
||||
e_score_correction_bias: Optional[torch.Tensor] = None,
|
||||
global_num_experts: int = -1,
|
||||
expert_map: Optional[torch.Tensor] = None,
|
||||
apply_router_weight_on_input: bool = False,
|
||||
activation: str = "silu",
|
||||
enable_eplb: bool = False,
|
||||
expert_load_view: Optional[torch.Tensor] = None,
|
||||
logical_to_physical_map: Optional[torch.Tensor] = None,
|
||||
logical_replica_count: Optional[torch.Tensor] = None) -> torch.Tensor:
|
||||
|
||||
topk_weights, topk_ids, row_idx = select_experts(
|
||||
hidden_states=x,
|
||||
router_logits=router_logits,
|
||||
top_k=top_k,
|
||||
use_grouped_topk=use_grouped_topk,
|
||||
renormalize=renormalize,
|
||||
topk_group=topk_group,
|
||||
num_expert_group=num_expert_group,
|
||||
custom_routing_function=custom_routing_function,
|
||||
scoring_func=scoring_func,
|
||||
routed_scaling_factor=1.0,
|
||||
e_score_correction_bias=e_score_correction_bias,
|
||||
global_num_experts=global_num_experts)
|
||||
|
||||
if topk_ids.shape[1] < top_k or is_310p():
|
||||
assert global_num_experts is not None
|
||||
return fused_experts_moge(
|
||||
hidden_states=x,
|
||||
w1=layer.w13_weight,
|
||||
w2=layer.w2_weight,
|
||||
moe_parallel_config=self.moe.moe_parallel_config,
|
||||
topk_weights=topk_weights,
|
||||
topk_ids=topk_ids,
|
||||
top_k=top_k,
|
||||
global_num_experts=global_num_experts,
|
||||
expert_map=expert_map,
|
||||
apply_router_weight_on_input=apply_router_weight_on_input)
|
||||
|
||||
moe_comm_method = get_forward_context().moe_comm_method
|
||||
return moe_comm_method.fused_experts(hidden_states=x,
|
||||
w1=layer.w13_weight,
|
||||
w2=layer.w2_weight,
|
||||
topk_weights=topk_weights,
|
||||
topk_ids=topk_ids,
|
||||
row_idx=row_idx,
|
||||
global_num_experts=global_num_experts,
|
||||
expert_map=expert_map)
|
||||
|
||||
|
||||
def forward_oot(
|
||||
self,
|
||||
layer: torch.nn.Module,
|
||||
@@ -315,59 +254,32 @@ class AscendFusedMoE(FusedMoE):
|
||||
num_redundant_experts=0,
|
||||
has_bias=False,
|
||||
):
|
||||
if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
|
||||
super().__init__(
|
||||
num_experts,
|
||||
top_k,
|
||||
hidden_size,
|
||||
intermediate_size,
|
||||
params_dtype,
|
||||
reduce_results,
|
||||
renormalize,
|
||||
use_grouped_topk,
|
||||
num_expert_group,
|
||||
topk_group,
|
||||
quant_config,
|
||||
tp_size,
|
||||
ep_size,
|
||||
dp_size,
|
||||
prefix,
|
||||
custom_routing_function,
|
||||
scoring_func,
|
||||
e_score_correction_bias,
|
||||
apply_router_weight_on_input,
|
||||
activation,
|
||||
enable_eplb,
|
||||
num_redundant_experts,
|
||||
has_bias,
|
||||
)
|
||||
else:
|
||||
super().__init__(
|
||||
num_experts,
|
||||
top_k,
|
||||
hidden_size,
|
||||
intermediate_size,
|
||||
params_dtype,
|
||||
reduce_results,
|
||||
renormalize,
|
||||
use_grouped_topk,
|
||||
num_expert_group,
|
||||
topk_group,
|
||||
quant_config,
|
||||
tp_size,
|
||||
ep_size,
|
||||
dp_size,
|
||||
prefix,
|
||||
custom_routing_function,
|
||||
scoring_func,
|
||||
routed_scaling_fator,
|
||||
e_score_correction_bias,
|
||||
apply_router_weight_on_input,
|
||||
activation,
|
||||
enable_eplb,
|
||||
num_redundant_experts,
|
||||
has_bias,
|
||||
)
|
||||
super().__init__(
|
||||
num_experts,
|
||||
top_k,
|
||||
hidden_size,
|
||||
intermediate_size,
|
||||
params_dtype,
|
||||
reduce_results,
|
||||
renormalize,
|
||||
use_grouped_topk,
|
||||
num_expert_group,
|
||||
topk_group,
|
||||
quant_config,
|
||||
tp_size,
|
||||
ep_size,
|
||||
dp_size,
|
||||
prefix,
|
||||
custom_routing_function,
|
||||
scoring_func,
|
||||
routed_scaling_fator,
|
||||
e_score_correction_bias,
|
||||
apply_router_weight_on_input,
|
||||
activation,
|
||||
enable_eplb,
|
||||
num_redundant_experts,
|
||||
has_bias,
|
||||
)
|
||||
setup_token_dispatchers(self.moe_config.ep_size,
|
||||
top_k=self.top_k,
|
||||
num_experts=self.global_num_experts,
|
||||
@@ -529,8 +441,4 @@ class AscendSharedFusedMoE(AscendFusedMoE):
|
||||
|
||||
UnquantizedFusedMoEMethod.__init__ = unquantized_fused_moe_init_func
|
||||
UnquantizedFusedMoEMethod.process_weights_after_loading = process_weights_after_loading
|
||||
|
||||
if vllm_version_is("0.10.1.1") or vllm_version_is("0.10.1"):
|
||||
UnquantizedFusedMoEMethod.forward_oot = forward_oot_v01011
|
||||
else:
|
||||
UnquantizedFusedMoEMethod.forward_oot = forward_oot
|
||||
UnquantizedFusedMoEMethod.forward_oot = forward_oot
|
||||
|
||||
Reference in New Issue
Block a user