6464c97ff9
vLLM 0.11.0 didn't bring PR (https://github.com/vllm-project/vllm/pull/25805) thus missing the prefix of mtp's SharedHead. This PR fixes this bug with a patch to vllm's deepseek_mtp. --------- Signed-off-by: whx-sjtu <2952154980@qq.com>
56 lines
2.0 KiB
Python
56 lines
2.0 KiB
Python
import torch
|
|
import torch.nn as nn
|
|
from transformers import PretrainedConfig
|
|
from vllm.config import VllmConfig
|
|
from vllm.model_executor.layers.layernorm import RMSNorm
|
|
from vllm.model_executor.layers.quantization import QuantizationConfig
|
|
from vllm.model_executor.layers.vocab_parallel_embedding import ParallelLMHead
|
|
from vllm.model_executor.models.deepseek_mtp import \
|
|
DeepSeekMultiTokenPredictorLayer
|
|
from vllm.model_executor.models.deepseek_v2 import DeepseekV2DecoderLayer
|
|
from vllm.model_executor.models.utils import maybe_prefix
|
|
|
|
|
|
class SharedHead(nn.Module):
|
|
|
|
def __init__(
|
|
self,
|
|
config: PretrainedConfig,
|
|
prefix: str,
|
|
quant_config: QuantizationConfig = None,
|
|
) -> None:
|
|
super().__init__()
|
|
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"),
|
|
)
|
|
|
|
def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
|
|
return self.norm(hidden_states)
|
|
|
|
|
|
def predictor_init(self, vllm_config: VllmConfig, prefix: str) -> None:
|
|
nn.Module.__init__(self)
|
|
config = vllm_config.model_config.hf_config
|
|
quant_config = vllm_config.quant_config
|
|
|
|
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)
|
|
|
|
# We don't need topk_indices_buffer in Ascend
|
|
topk_indices_buffer = None
|
|
self.shared_head = SharedHead(config=config,
|
|
prefix=prefix,
|
|
quant_config=quant_config)
|
|
self.mtp_block = DeepseekV2DecoderLayer(vllm_config, prefix,
|
|
topk_indices_buffer)
|
|
|
|
|
|
DeepSeekMultiTokenPredictorLayer.__init__ = predictor_init
|