Files
xc-llm-ascend/vllm_ascend/torchair/models/torchair_deepseek_mtp.py

222 lines
8.8 KiB
Python
Raw Normal View History

#
# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
# Adapted from vllm/model_executor/models/deepseek_mtp.py
# Copyright 2023 The vLLM team.
#
# This file is a part of the vllm-ascend project.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import List, Optional
import torch
import torch.nn as nn
from transformers import PretrainedConfig
from vllm.attention.backends.abstract import AttentionMetadata
[1/N][Refactor] Refactor code to adapt with vllm main (#3612) ### What this PR does / why we need it? This is the step 1 of refactoring code to adapt with vllm main, and this pr aligned with https://github.com/vllm-project/vllm/commit/17c540a993af88204ad1b78345c8a865cf58ce44 1. refactor deepseek to the latest code arch as of https://github.com/vllm-project/vllm/commit/17c540a993af88204ad1b78345c8a865cf58ce44 2. bunches of fixes due to vllm changes - Fix `AscendScheduler` `__post_init__`, caused by https://github.com/vllm-project/vllm/pull/25075 - Fix `AscendScheduler` init got an unexpected arg `block_size`, caused by https://github.com/vllm-project/vllm/pull/26296 - Fix `KVCacheManager` `get_num_common_prefix_blocks` arg, caused by https://github.com/vllm-project/vllm/pull/23485 - Fix `MLAAttention` import,caused by https://github.com/vllm-project/vllm/pull/25103 - Fix `SharedFusedMoE` import, caused by https://github.com/vllm-project/vllm/pull/26145 - Fix `LazyLoader` improt, caused by https://github.com/vllm-project/vllm/pull/27022 - Fix `vllm.utils.swap_dict_values` improt, caused by https://github.com/vllm-project/vllm/pull/26990 - Fix `Backend` enum import, caused by https://github.com/vllm-project/vllm/pull/25893 - Fix `CompilationLevel` renaming to `CompilationMode` issue introduced by https://github.com/vllm-project/vllm/pull/26355 - Fix fused_moe ops, caused by https://github.com/vllm-project/vllm/pull/24097 - Fix bert model because of `inputs_embeds`, caused by https://github.com/vllm-project/vllm/pull/25922 - Fix MRope because of `get_input_positions_tensor` to `get_mrope_input_positions`, caused by https://github.com/vllm-project/vllm/pull/24172 - Fix `splitting_ops` changes introduced by https://github.com/vllm-project/vllm/pull/25845 - Fix multi-modality changes introduced by https://github.com/vllm-project/vllm/issues/16229 - Fix lora bias dropping issue introduced by https://github.com/vllm-project/vllm/pull/25807 - Fix structured ouput break introduced by https://github.com/vllm-project/vllm/issues/26737 ### Does this PR introduce _any_ user-facing change? ### How was this patch tested? CI passed with existing test. - vLLM version: v0.11.0rc3 - vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0 --------- Signed-off-by: MengqingCao <cmq0113@163.com> Signed-off-by: Icey <1790571317@qq.com> Co-authored-by: Icey <1790571317@qq.com>
2025-10-24 16:55:08 +08:00
from vllm.compilation.decorators import support_torch_compile
from vllm.config import CacheConfig, ModelConfig, VllmConfig
from vllm.distributed import get_tensor_model_parallel_world_size
from vllm.model_executor.layers.layernorm import RMSNorm
from vllm.model_executor.layers.logits_processor import LogitsProcessor
from vllm.model_executor.layers.quantization import QuantizationConfig
from vllm.model_executor.layers.vocab_parallel_embedding import (
ParallelLMHead, VocabParallelEmbedding)
from vllm.model_executor.models.deepseek_mtp import (
DeepSeekMTP, DeepSeekMultiTokenPredictor, DeepSeekMultiTokenPredictorLayer,
SharedHead)
from vllm.model_executor.models.utils import maybe_prefix
from vllm.sequence import IntermediateTensors
from vllm_ascend.torchair.models.torchair_deepseek_v2 import \
TorchairDeepseekV2DecoderLayer
class TorchairDeepSeekShareHead(SharedHead):
def __init__(self,
config: PretrainedConfig,
quant_config: Optional[QuantizationConfig] = None,
prefix: str = "") -> None:
nn.Module.__init__(self)
self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
self.head = ParallelLMHead(config.vocab_size,
config.hidden_size,
quant_config=quant_config,
prefix=maybe_prefix(prefix, "head"))
class TorchairDeepSeekMultiTokenPredictorLayer(DeepSeekMultiTokenPredictorLayer
):
def __init__(
self,
config: PretrainedConfig,
prefix: str,
model_config: ModelConfig,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
) -> None:
nn.Module.__init__(self)
self.tp_size = get_tensor_model_parallel_world_size()
self.enorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
self.hnorm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps)
self.eh_proj = nn.Linear(config.hidden_size * 2,
config.hidden_size,
bias=False)
self.shared_head = TorchairDeepSeekShareHead(config=config,
quant_config=quant_config,
prefix=maybe_prefix(
prefix,
"shared_head"))
self.mtp_block = TorchairDeepseekV2DecoderLayer(
config, prefix, model_config, cache_config, quant_config)
def forward(
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
kv_cache: torch.Tensor,
attn_metadata: AttentionMetadata,
previous_hidden_states: torch.Tensor,
inputs_embeds: Optional[torch.Tensor] = None,
spec_step_index: int = 0,
) -> torch.Tensor:
assert inputs_embeds is not None
# masking inputs at position 0, as not needed by MTP
inputs_embeds = torch.where((positions == 0).unsqueeze(-1),
torch.zeros_like(inputs_embeds),
inputs_embeds)
inputs_embeds = self.enorm(inputs_embeds)
previous_hidden_states = self.hnorm(previous_hidden_states)
hidden_states = self.eh_proj(
torch.cat([inputs_embeds, previous_hidden_states], dim=-1))
del inputs_embeds, previous_hidden_states
replace_allreduce = hidden_states.shape[0] % self.tp_size == 0
hidden_states, residual = self.mtp_block(
positions=positions,
hidden_states=hidden_states,
residual=None,
kv_cache=kv_cache,
attn_metadata=attn_metadata,
replace_allreduce=replace_allreduce)
hidden_states = residual + hidden_states
return hidden_states
class TorchairDeepSeekMultiTokenPredictor(DeepSeekMultiTokenPredictor):
def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
nn.Module.__init__(self)
config = vllm_config.model_config.hf_config
self.mtp_start_layer_idx = config.num_hidden_layers
self.num_mtp_layers = config.num_nextn_predict_layers
# to map the exact layer index from weights
self.layers = torch.nn.ModuleDict({
str(idx):
TorchairDeepSeekMultiTokenPredictorLayer(
config,
f"{prefix}.layers.{idx}",
model_config=vllm_config.model_config,
cache_config=vllm_config.cache_config,
quant_config=vllm_config.quant_config,
)
for idx in range(self.mtp_start_layer_idx,
self.mtp_start_layer_idx + self.num_mtp_layers)
})
self.embed_tokens = VocabParallelEmbedding(
config.vocab_size,
config.hidden_size,
)
# Note: torch._dynamo.exc.Unsupported: builtin: str
self.layers_list = [
self.layers[str(idx)]
for idx in range(self.mtp_start_layer_idx,
self.mtp_start_layer_idx + self.num_mtp_layers)
]
self.logits_processor = LogitsProcessor(config.vocab_size)
def forward(
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
kv_caches: torch.Tensor,
attn_metadata: AttentionMetadata,
previous_hidden_states: torch.Tensor,
inputs_embeds: Optional[torch.Tensor] = None,
spec_step_idx: int = 0,
) -> torch.Tensor:
if inputs_embeds is None:
inputs_embeds = self.embed_tokens(input_ids)
current_step_idx = (spec_step_idx % self.num_mtp_layers)
step_kv_cache = kv_caches[
current_step_idx] if kv_caches is not None else None
return self.layers_list[current_step_idx](
input_ids,
positions,
step_kv_cache,
attn_metadata,
previous_hidden_states,
inputs_embeds,
current_step_idx,
)
def compute_logits(
self,
hidden_states: torch.Tensor,
spec_step_idx: int = 0,
) -> torch.Tensor:
current_step_idx = (spec_step_idx % self.num_mtp_layers)
mtp_layer = self.layers_list[current_step_idx]
logits = self.logits_processor(mtp_layer.shared_head.head,
mtp_layer.shared_head(hidden_states))
return logits
[1/N][Refactor] Refactor code to adapt with vllm main (#3612) ### What this PR does / why we need it? This is the step 1 of refactoring code to adapt with vllm main, and this pr aligned with https://github.com/vllm-project/vllm/commit/17c540a993af88204ad1b78345c8a865cf58ce44 1. refactor deepseek to the latest code arch as of https://github.com/vllm-project/vllm/commit/17c540a993af88204ad1b78345c8a865cf58ce44 2. bunches of fixes due to vllm changes - Fix `AscendScheduler` `__post_init__`, caused by https://github.com/vllm-project/vllm/pull/25075 - Fix `AscendScheduler` init got an unexpected arg `block_size`, caused by https://github.com/vllm-project/vllm/pull/26296 - Fix `KVCacheManager` `get_num_common_prefix_blocks` arg, caused by https://github.com/vllm-project/vllm/pull/23485 - Fix `MLAAttention` import,caused by https://github.com/vllm-project/vllm/pull/25103 - Fix `SharedFusedMoE` import, caused by https://github.com/vllm-project/vllm/pull/26145 - Fix `LazyLoader` improt, caused by https://github.com/vllm-project/vllm/pull/27022 - Fix `vllm.utils.swap_dict_values` improt, caused by https://github.com/vllm-project/vllm/pull/26990 - Fix `Backend` enum import, caused by https://github.com/vllm-project/vllm/pull/25893 - Fix `CompilationLevel` renaming to `CompilationMode` issue introduced by https://github.com/vllm-project/vllm/pull/26355 - Fix fused_moe ops, caused by https://github.com/vllm-project/vllm/pull/24097 - Fix bert model because of `inputs_embeds`, caused by https://github.com/vllm-project/vllm/pull/25922 - Fix MRope because of `get_input_positions_tensor` to `get_mrope_input_positions`, caused by https://github.com/vllm-project/vllm/pull/24172 - Fix `splitting_ops` changes introduced by https://github.com/vllm-project/vllm/pull/25845 - Fix multi-modality changes introduced by https://github.com/vllm-project/vllm/issues/16229 - Fix lora bias dropping issue introduced by https://github.com/vllm-project/vllm/pull/25807 - Fix structured ouput break introduced by https://github.com/vllm-project/vllm/issues/26737 ### Does this PR introduce _any_ user-facing change? ### How was this patch tested? CI passed with existing test. - vLLM version: v0.11.0rc3 - vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0 --------- Signed-off-by: MengqingCao <cmq0113@163.com> Signed-off-by: Icey <1790571317@qq.com> Co-authored-by: Icey <1790571317@qq.com>
2025-10-24 16:55:08 +08:00
@support_torch_compile
class TorchairDeepSeekMTP(DeepSeekMTP):
# NOTE 1.The quantized MTP layer of deepseek on the NPU is not quantized;
# NOTE 2.The description file generated by the current msmodelslim tool does not have
# MTP layer info. Please manually add it and set the value to FLOAT.
packed_modules_mapping = {
"gate_up_proj": ["gate_proj", "up_proj"],
"experts":
["experts.0.gate_proj", "experts.0.up_proj", "experts.0.down_proj"]
}
def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
nn.Module.__init__(self)
self.config = vllm_config.model_config.hf_config
self.model = TorchairDeepSeekMultiTokenPredictor(
vllm_config=vllm_config, prefix=maybe_prefix(prefix, "model"))
def forward(
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
kv_caches: Optional[List[torch.Tensor]] = None,
attn_metadata: Optional[AttentionMetadata] = None,
hidden_states: Optional[torch.Tensor] = None,
intermediate_tensors: Optional[IntermediateTensors] = None,
inputs_embeds: Optional[torch.Tensor] = None,
spec_step_idx: int = 0,
) -> torch.Tensor:
hidden_states = self.model(input_ids, positions, kv_caches,
attn_metadata, hidden_states, inputs_embeds,
spec_step_idx)
return hidden_states