Skip to content

GitLab

Commit 4f9947e6 authored by zhuwenwen's avatar zhuwenwen
Browse files

update qwen3_moe.py

parent 1e622f10
Hide whitespace changes
Inline Side-by-side
Showing with 173 additions and 240 deletions
vllm/model_executor/models/qwen3_moe.py
View file @ 4f9947e6
......@@ -22,22 +22,19 @@
# See the License for the specific language governing permissions and
# limitations under the License.
"""Inference-only Qwen3MoE model compatible with HuggingFace weights."""
import typing
from collections.abc import Callable, Iterable
from itertools import islice
from collections.abc import Iterable
from typing import Any, Optional, Union
import os
import re
import torch
from torch import nn
from transformers import PretrainedConfig
from vllm.attention import Attention
from vllm.compilation.decorators import support_torch_compile
from vllm.config import CacheConfig, VllmConfig, get_current_vllm_config
from vllm.distributed import (get_ep_group, get_pp_group,
get_tensor_model_parallel_world_size,
tensor_model_parallel_all_gather)
from vllm.config import CacheConfig, VllmConfig
from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
from vllm.logger import init_logger
from vllm.model_executor.layers.activation import SiluAndMul
from vllm.model_executor.layers.fused_moe import FusedMoE
......@@ -51,17 +48,17 @@ from vllm.model_executor.layers.quantization import QuantizationConfig
from vllm.model_executor.layers.rotary_embedding import get_rope
from vllm.model_executor.layers.vocab_parallel_embedding import (
ParallelLMHead, VocabParallelEmbedding)
from vllm.model_executor.model_loader.weight_utils import (
default_weight_loader, maybe_remap_kv_scale_name)
from vllm.model_executor.models.utils import sequence_parallel_chunk
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.sequence import IntermediateTensors
from .interfaces import MixtureOfExperts, SupportsLoRA, SupportsPP
from .utils import (AutoWeightsLoader, PPMissingLayer, extract_layer_index,
from .interfaces import SupportsPP
from .utils import (AutoWeightsLoader, extract_layer_index,
is_pp_missing_parameter,
make_empty_intermediate_tensors_factory, make_layers,
maybe_prefix)
import vllm.envs as envs
from vllm.utils import direct_register_custom_op
from vllm import _custom_ops as ops
from vllm.model_executor.utils import pad_weight, gemm_bank_conf
from vllm.utils import W8a8GetCacheJSON
......@@ -108,86 +105,49 @@ class Qwen3MoeSparseMoeBlock(nn.Module):
def __init__(
self,
vllm_config: VllmConfig,
config: PretrainedConfig,
quant_config: Optional[QuantizationConfig] = None,
prefix: str = "",
):
super().__init__()
config = vllm_config.model_config.hf_text_config
parallel_config = vllm_config.parallel_config
quant_config = vllm_config.quant_config
self.tp_size = get_tensor_model_parallel_world_size()
self.ep_group = get_ep_group().device_group
self.ep_rank = self.ep_group.rank()
self.ep_size = self.ep_group.size()
self.n_routed_experts = config.num_experts
self.is_sequence_parallel = parallel_config.use_sequence_parallel_moe
if self.tp_size > config.num_experts:
raise ValueError(
f"Tensor parallel size {self.tp_size} is greater than "
f"the number of experts {config.num_experts}.")
# Load balancing settings.
vllm_config = get_current_vllm_config()
eplb_config = vllm_config.parallel_config.eplb_config
self.enable_eplb = parallel_config.enable_eplb
self.n_logical_experts = self.n_routed_experts
self.n_redundant_experts = eplb_config.num_redundant_experts
self.n_physical_experts = (self.n_logical_experts +
self.n_redundant_experts)
self.n_local_physical_experts = self.n_physical_experts // self.ep_size
self.physical_expert_start = (self.ep_rank *
self.n_local_physical_experts)
self.physical_expert_end = (self.physical_expert_start +
self.n_local_physical_experts)
self.experts = FusedMoE(num_experts=self.n_routed_experts,
self.experts = FusedMoE(num_experts=config.num_experts,
top_k=config.num_experts_per_tok,
hidden_size=config.hidden_size,
intermediate_size=config.moe_intermediate_size,
reduce_results=True,
reduce_results=False,
renormalize=config.norm_topk_prob,
quant_config=quant_config,
prefix=f"{prefix}.experts",
enable_eplb=self.enable_eplb,
num_redundant_experts=self.n_redundant_experts,
is_sequence_parallel=self.is_sequence_parallel)
prefix=f"{prefix}.experts")
self.gate = ReplicatedLinear(config.hidden_size,
config.num_experts,
bias=False,
quant_config=quant_config,
quant_config=None,
prefix=f"{prefix}.gate")
def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
assert hidden_states.dim(
) <= 2, "Qwen3MoeSparseMoeBlock only supports 1D or 2D inputs"
is_input_1d = hidden_states.dim() == 1
num_tokens, hidden_dim = hidden_states.shape
# NOTE: hidden_states can have either 1D or 2D shape.
orig_shape = hidden_states.shape
hidden_dim = hidden_states.shape[-1]
hidden_states = hidden_states.view(-1, hidden_dim)
if self.is_sequence_parallel:
hidden_states = sequence_parallel_chunk(hidden_states)
# router_logits: (num_tokens, n_experts)
router_logits, _ = self.gate(hidden_states)
final_hidden_states = self.experts(hidden_states=hidden_states,
router_logits=router_logits)
if self.is_sequence_parallel:
final_hidden_states = tensor_model_parallel_all_gather(
final_hidden_states, 0)
final_hidden_states = final_hidden_states[:num_tokens]
if self.tp_size > 1:
final_hidden_states = self.experts.maybe_all_reduce_tensor_model_parallel( # noqa E501
final_hidden_states)
# return to 1d if input is 1d
return final_hidden_states.squeeze(0) if is_input_1d else \
final_hidden_states
return final_hidden_states.view(orig_shape)
class Qwen3MoeAttention(nn.Module):
......@@ -206,7 +166,6 @@ class Qwen3MoeAttention(nn.Module):
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
prefix: str = "",
dual_chunk_attention_config: Optional[dict[str, Any]] = None,
) -> None:
super().__init__()
self.hidden_size = hidden_size
......@@ -230,7 +189,6 @@ class Qwen3MoeAttention(nn.Module):
self.scaling = self.head_dim**-0.5
self.rope_theta = rope_theta
self.max_position_embeddings = max_position_embeddings
self.dual_chunk_attention_config = dual_chunk_attention_config
self.qkv_proj = QKVParallelLinear(hidden_size,
self.head_dim,
......@@ -252,25 +210,72 @@ class Qwen3MoeAttention(nn.Module):
max_position=max_position_embeddings,
base=rope_theta,
rope_scaling=rope_scaling,
dual_chunk_attention_config=dual_chunk_attention_config,
)
self.attn = Attention(
self.num_heads,
self.head_dim,
self.scaling,
num_kv_heads=self.num_kv_heads,
cache_config=cache_config,
quant_config=quant_config,
prefix=f"{prefix}.attn",
**{
"layer_idx": extract_layer_index(prefix),
"dual_chunk_attention_config": dual_chunk_attention_config,
} if dual_chunk_attention_config else {},
)
self.attn = Attention(self.num_heads,
self.head_dim,
self.scaling,
num_kv_heads=self.num_kv_heads,
cache_config=cache_config,
quant_config=quant_config,
prefix=f"{prefix}.attn")
self.q_norm = RMSNorm(self.head_dim, eps=rms_norm_eps)
self.k_norm = RMSNorm(self.head_dim, eps=rms_norm_eps)
def rms_rotary_embedding_fuse(
positions: torch.Tensor,
query: torch.Tensor,
key: Optional[torch.Tensor],
head_size: int,
cos_sin_cache: torch.Tensor,
is_neox_style: bool,
q_weight: torch.Tensor,
k_weight: torch.Tensor,
q_bias: Optional[torch.Tensor],
k_bias: Optional[torch.Tensor],
epsilon: float,
) -> None:
from lightop import rms_rotary_embedding_fuse as fused_kernel
fused_kernel(
positions,
query,
key,
head_size,
cos_sin_cache,
is_neox_style,
q_weight,
k_weight,
q_bias,
k_bias,
epsilon,
)
def rms_rotary_embedding_fuse_fake(
# q_out:torch.Tensor,
# k_out:torch.Tensor,
positions: torch.Tensor,
query: torch.Tensor,
key: Optional[torch.Tensor],
head_size: int,
cos_sin_cache: torch.Tensor,
is_neox_style: bool,
q_weight: torch.Tensor,
k_weight: torch.Tensor,
q_bias: Optional[torch.Tensor],
k_bias: Optional[torch.Tensor],
epsilon: float,
) -> None:
# Fake impl intentionally left as no-op for graph tracing modes.
pass
direct_register_custom_op(
op_name="rms_rotary_embedding_fuse",
op_func=rms_rotary_embedding_fuse,
mutates_args=["query", "key"],
fake_impl=rms_rotary_embedding_fuse_fake,
)
def forward(
self,
positions: torch.Tensor,
......@@ -278,23 +283,52 @@ class Qwen3MoeAttention(nn.Module):
) -> torch.Tensor:
qkv, _ = self.qkv_proj(hidden_states)
q, k, v = qkv.split([self.q_size, self.kv_size, self.kv_size], dim=-1)
# Add qk-norm
q_by_head = q.view(*q.shape[:-1], q.shape[-1] // self.head_dim,
self.head_dim)
if envs.VLLM_USE_APEX_RN:
q_by_head = self.q_norm.forward_apex(q_by_head)
else:
q_by_head = self.q_norm.forward_cuda(q_by_head)
q = q_by_head.view(q.shape)
if envs.VLLM_USE_FUSED_RMS_ROPE :
# Fused RMSNorm + RoPE path through custom op.
cos_sin_cache = self.rotary_emb.cos_sin_cache
if (cos_sin_cache.device != q.device
or cos_sin_cache.dtype != q.dtype):
cos_sin_cache = cos_sin_cache.to(q.device,
dtype=q.dtype,
non_blocking=True)
# Persist the converted cache so we don't re-copy/re-allocate
# on every forward when the original buffer starts on CPU.
self.rotary_emb.cos_sin_cache = cos_sin_cache
# # q, k 使用 continuous
q = q.contiguous()
k = k.contiguous()
torch.ops.vllm.rms_rotary_embedding_fuse(
positions,
q,
k,
self.head_dim,
cos_sin_cache,
self.rotary_emb.is_neox_style,
self.q_norm.weight,
self.k_norm.weight,
None,
None,
self.q_norm.variance_epsilon,
)
k_by_head = k.view(*k.shape[:-1], k.shape[-1] // self.head_dim,
self.head_dim)
if envs.VLLM_USE_APEX_RN:
k_by_head = self.k_norm.forward_apex(k_by_head)
else:
k_by_head = self.k_norm.forward_cuda(k_by_head)
k = k_by_head.view(k.shape)
q, k = self.rotary_emb(positions, q, k)
# Add qk-norm then RoPE (original path).
q_by_head = q.view(*q.shape[:-1], q.shape[-1] // self.head_dim,
self.head_dim)
if envs.VLLM_USE_APEX_RN:
q_by_head = self.q_norm.forward_apex(q_by_head)
else:
q_by_head = self.q_norm.forward_cuda(q_by_head)
q = q_by_head.view(q.shape)
k_by_head = k.view(*k.shape[:-1], k.shape[-1] // self.head_dim,
self.head_dim)
if envs.VLLM_USE_APEX_RN:
k_by_head = self.k_norm.forward_apex(k_by_head)
else:
k_by_head = self.k_norm.forward_cuda(k_by_head)
k = k_by_head.view(k.shape)
q, k = self.rotary_emb(positions, q, k)
attn_output = self.attn(q, k, v)
output, _ = self.o_proj(attn_output)
return output
......@@ -302,21 +336,19 @@ class Qwen3MoeAttention(nn.Module):
class Qwen3MoeDecoderLayer(nn.Module):
def __init__(self, vllm_config: VllmConfig, prefix: str = "") -> None:
def __init__(
self,
config: PretrainedConfig,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
prefix: str = "",
) -> None:
super().__init__()
config = vllm_config.model_config.hf_text_config
cache_config = vllm_config.cache_config
quant_config = vllm_config.quant_config
self.hidden_size = config.hidden_size
rope_theta = getattr(config, "rope_theta", 10000)
rope_scaling = getattr(config, "rope_scaling", None)
max_position_embeddings = getattr(config, "max_position_embeddings",
8192)
dual_chunk_attention_config = getattr(config,
"dual_chunk_attention_config",
None)
self.self_attn = Qwen3MoeAttention(
hidden_size=self.hidden_size,
num_heads=config.num_attention_heads,
......@@ -330,7 +362,6 @@ class Qwen3MoeDecoderLayer(nn.Module):
cache_config=cache_config,
quant_config=quant_config,
prefix=f"{prefix}.self_attn",
dual_chunk_attention_config=dual_chunk_attention_config,
)
# `mlp_only_layers` in the config.
......@@ -340,7 +371,8 @@ class Qwen3MoeDecoderLayer(nn.Module):
if (layer_idx not in mlp_only_layers) and (
config.num_experts > 0 and
(layer_idx + 1) % config.decoder_sparse_step == 0):
self.mlp = Qwen3MoeSparseMoeBlock(vllm_config=vllm_config,
self.mlp = Qwen3MoeSparseMoeBlock(config=config,
quant_config=quant_config,
prefix=f"{prefix}.mlp")
else:
self.mlp = Qwen3MoeMLP(hidden_size=config.hidden_size,
......@@ -384,11 +416,9 @@ class Qwen3MoeModel(nn.Module):
def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
super().__init__()
config = vllm_config.model_config.hf_text_config
config = vllm_config.model_config.hf_config
cache_config = vllm_config.cache_config
quant_config = vllm_config.quant_config
parallel_config = vllm_config.parallel_config
eplb_config = parallel_config.eplb_config
self.num_redundant_experts = eplb_config.num_redundant_experts
self.padding_idx = config.pad_token_id
self.vocab_size = config.vocab_size
......@@ -403,11 +433,12 @@ class Qwen3MoeModel(nn.Module):
self.embed_tokens = VocabParallelEmbedding(
config.vocab_size,
config.hidden_size,
quant_config=quant_config,
prefix=f"{prefix}.embed_tokens")
self.start_layer, self.end_layer, self.layers = make_layers(
config.num_hidden_layers,
lambda prefix: Qwen3MoeDecoderLayer(vllm_config=vllm_config,
lambda prefix: Qwen3MoeDecoderLayer(config=config,
cache_config=cache_config,
quant_config=quant_config,
prefix=prefix),
prefix=f"{prefix}.layers",
)
......@@ -444,7 +475,8 @@ class Qwen3MoeModel(nn.Module):
assert intermediate_tensors is not None
hidden_states = intermediate_tensors["hidden_states"]
residual = intermediate_tensors["residual"]
for layer in islice(self.layers, self.start_layer, self.end_layer):
for i in range(self.start_layer, self.end_layer):
layer = self.layers[i]
hidden_states, residual = layer(positions, hidden_states, residual)
if not get_pp_group().is_last_rank:
return IntermediateTensors({
......@@ -454,16 +486,6 @@ class Qwen3MoeModel(nn.Module):
hidden_states, _ = self.norm(hidden_states, residual)
return hidden_states
def get_expert_mapping(self) -> list[tuple[str, str, int, str]]:
# Params for weights, fp8 weight scales, fp8 activation scales
# (param_name, weight_name, expert_id, shard_id)
return FusedMoE.make_expert_params_mapping(
ckpt_gate_proj_name="gate_proj",
ckpt_down_proj_name="down_proj",
ckpt_up_proj_name="up_proj",
num_experts=self.config.num_experts,
num_redundant_experts=self.num_redundant_experts)
def load_weights(self, weights: Iterable[tuple[str,
torch.Tensor]]) -> set[str]:
stacked_params_mapping = [
......@@ -480,9 +502,16 @@ class Qwen3MoeModel(nn.Module):
".v_scale", "_v_scale", ".weight_scale",
"_weight_scale", ".input_scale", "_input_scale")
# Params for weights, fp8 weight scales, fp8 activation scales
# (param_name, weight_name, expert_id, shard_id)
expert_params_mapping = FusedMoE.make_expert_params_mapping(
ckpt_gate_proj_name="gate_proj",
ckpt_down_proj_name="down_proj",
ckpt_up_proj_name="up_proj",
num_experts=self.config.num_experts)
params_dict = dict(self.named_parameters())
loaded_params: set[str] = set()
expert_params_mapping = self.get_expert_mapping()
for name, loaded_weight in weights:
if self.use_llama_nn:
current_count = loaded_weight.current_count
......@@ -508,68 +537,35 @@ class Qwen3MoeModel(nn.Module):
# Skip layers on other devices.
if is_pp_missing_parameter(name, self):
continue
if name.endswith("scale"):
# Remapping the name of FP8 kv-scale.
name = maybe_remap_kv_scale_name(name, params_dict)
if name is None:
continue
if name not in params_dict:
continue
param = params_dict[name]
weight_loader = getattr(param, "weight_loader",
default_weight_loader)
if weight_loader == default_weight_loader:
weight_loader(param, loaded_weight)
else:
weight_loader(param, loaded_weight, shard_id)
weight_loader = param.weight_loader
weight_loader(param, loaded_weight, shard_id)
break
else:
is_expert_weight = False
for mapping in expert_params_mapping:
param_name, weight_name, expert_id, shard_id = mapping
if weight_name not in name:
continue
# Anyway, this is an expert weight and should not be
# attempted to load as other weights later
is_expert_weight = True
# Do not modify `name` since the loop may continue here
# Instead, create a new variable
name_mapped = name.replace(weight_name, param_name)
if is_pp_missing_parameter(name_mapped, self):
name = name.replace(weight_name, param_name)
# Skip layers on other devices.
if is_pp_missing_parameter(name, self):
continue
# Skip loading extra parameters for GPTQ/modelopt models.
if name_mapped.endswith(
ignore_suffixes
) and name_mapped not in params_dict:
if name.endswith(
ignore_suffixes) and name not in params_dict:
continue
param = params_dict[name_mapped]
# We should ask the weight loader to return success or not
# here since otherwise we may skip experts with other
# available replicas.
weight_loader = typing.cast(Callable[..., bool],
param.weight_loader)
success = weight_loader(param,
loaded_weight,
name_mapped,
shard_id=shard_id,
expert_id=expert_id,
return_success=True)
if success:
name = name_mapped
break
param = params_dict[name]
weight_loader = param.weight_loader
weight_loader(param,
loaded_weight,
name,
shard_id=shard_id,
expert_id=expert_id)
break
else:
if is_expert_weight:
# We've checked that this is an expert weight
# However it's not mapped locally to this rank
# So we simply skip it
continue
# Skip loading extra parameters for GPTQ/modelopt models.
if name.endswith(
ignore_suffixes) and name not in params_dict:
......@@ -639,8 +635,7 @@ class Qwen3MoeModel(nn.Module):
return loaded_params
class Qwen3MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA,
MixtureOfExperts):
class Qwen3MoeForCausalLM(nn.Module, SupportsPP):
packed_modules_mapping = {
"qkv_proj": [
"q_proj",
......@@ -657,7 +652,7 @@ class Qwen3MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA,
def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
super().__init__()
config = vllm_config.model_config.hf_text_config
config = vllm_config.model_config.hf_config
quant_config = vllm_config.quant_config
self.config = config
self.quant_config = quant_config
......@@ -665,74 +660,13 @@ class Qwen3MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA,
prefix=maybe_prefix(prefix, "model"))
self.lm_head = ParallelLMHead(config.vocab_size,
config.hidden_size,
quant_config=quant_config,
prefix=maybe_prefix(prefix, "lm_head"))
quant_config=quant_config)
if self.config.tie_word_embeddings:
self.lm_head.weight = self.model.embed_tokens.weight
self.logits_processor = LogitsProcessor(config.vocab_size)
self.make_empty_intermediate_tensors = (
self.model.make_empty_intermediate_tensors)
# Set MoE hyperparameters
self.expert_weights = []
self.moe_layers: list[FusedMoE] = []
example_layer = None
for layer in self.model.layers:
if isinstance(layer, PPMissingLayer):
continue
assert isinstance(layer, Qwen3MoeDecoderLayer)
if isinstance(layer.mlp, Qwen3MoeSparseMoeBlock):
example_layer = layer.mlp
self.moe_layers.append(layer.mlp.experts)
if example_layer is None:
raise RuntimeError("No Qwen3MoE layer found in the model.layers.")
self.num_moe_layers = len(self.moe_layers)
self.num_expert_groups = 1
self.num_shared_experts = 0
self.num_logical_experts = example_layer.n_logical_experts
self.num_physical_experts = example_layer.n_physical_experts
self.num_local_physical_experts = example_layer.n_local_physical_experts
self.num_routed_experts = example_layer.n_routed_experts
self.num_redundant_experts = example_layer.n_redundant_experts
def set_eplb_state(
self,
expert_load_view: torch.Tensor,
logical_to_physical_map: torch.Tensor,
logical_replica_count: torch.Tensor,
) -> None:
for layer_idx, layer in enumerate(self.moe_layers):
# Register the expert weights.
self.expert_weights.append(layer.get_expert_weights())
layer.set_eplb_state(
moe_layer_idx=layer_idx,
expert_load_view=expert_load_view,
logical_to_physical_map=logical_to_physical_map,
logical_replica_count=logical_replica_count,
)
def update_physical_experts_metadata(
self,
num_physical_experts: int,
num_local_physical_experts: int,
) -> None:
assert self.num_local_physical_experts == num_local_physical_experts
self.num_physical_experts = num_physical_experts
self.num_local_physical_experts = num_local_physical_experts
self.num_redundant_experts = (num_physical_experts -
self.num_logical_experts)
for layer in self.model.layers:
if isinstance(layer.mlp, Qwen3MoeSparseMoeBlock):
moe = layer.mlp
moe.n_local_physical_experts = num_local_physical_experts
moe.n_physical_experts = num_physical_experts
moe.n_redundant_experts = self.num_redundant_experts
moe.experts.update_expert_map()
def get_input_embeddings(self, input_ids: torch.Tensor) -> torch.Tensor:
return self.model.get_input_embeddings(input_ids)
......@@ -750,14 +684,13 @@ class Qwen3MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA,
def compute_logits(
self,
hidden_states: torch.Tensor,
sampling_metadata: SamplingMetadata,
) -> Optional[torch.Tensor]:
logits = self.logits_processor(self.lm_head, hidden_states)
logits = self.logits_processor(self.lm_head, hidden_states,
sampling_metadata)
return logits
def load_weights(self, weights: Iterable[tuple[str,
torch.Tensor]]) -> set[str]:
loader = AutoWeightsLoader(self)
return loader.load_weights(weights)
def get_expert_mapping(self) -> list[tuple[str, str, int, str]]:
return self.model.get_expert_mapping()
return loader.load_weights(weights)
\ No newline at end of file
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment