wehub-resource-sync
136 lines
5.0 KiB
Python
136 lines
5.0 KiB
Python
# SPDX-License-Identifier: Apache-2.0
|
|
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
|
|
|
# Adapted from
|
|
# https://github.com/huggingface/transformers/blob/v4.28.0/src/transformers/models/qwen2/modeling_qwen2.py
|
|
# Copyright 2025 Xiaomi Corporation.
|
|
# Copyright 2024 The Qwen team.
|
|
# Copyright 2023 The vLLM team.
|
|
# Copyright 2022 EleutherAI and the HuggingFace Inc. team. All rights reserved.
|
|
#
|
|
# This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
|
|
# and OPT implementations in this library. It has been modified from its
|
|
# original forms to accommodate minor architectural differences compared
|
|
# to GPT-NeoX and OPT used by the Meta AI team that trained the model.
|
|
#
|
|
# 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.
|
|
"""Inference-only MiMo model compatible with HuggingFace weights."""
|
|
|
|
from collections.abc import Iterable
|
|
from itertools import islice
|
|
|
|
import torch
|
|
import torch.nn as nn
|
|
|
|
from vllm.compilation.decorators import support_torch_compile
|
|
from vllm.config import VllmConfig
|
|
from vllm.distributed import get_pp_group
|
|
from vllm.logger import init_logger
|
|
from vllm.model_executor.layers.logits_processor import LogitsProcessor
|
|
from vllm.model_executor.layers.vocab_parallel_embedding import ParallelLMHead
|
|
from vllm.model_executor.models.qwen2 import Qwen2ForCausalLM, Qwen2Model
|
|
from vllm.sequence import IntermediateTensors
|
|
|
|
from .utils import AutoWeightsLoader, PPMissingLayer, maybe_prefix
|
|
|
|
logger = init_logger(__name__)
|
|
|
|
|
|
@support_torch_compile(
|
|
dynamic_arg_dims={
|
|
"input_ids": 0,
|
|
"positions": -1,
|
|
"intermediate_tensors": 0,
|
|
"inputs_embeds": 0,
|
|
}
|
|
)
|
|
class MiMoModel(Qwen2Model):
|
|
def forward(
|
|
self,
|
|
input_ids: torch.Tensor | None,
|
|
positions: torch.Tensor,
|
|
intermediate_tensors: IntermediateTensors | None = None,
|
|
inputs_embeds: torch.Tensor | None = None,
|
|
) -> torch.Tensor | IntermediateTensors:
|
|
if get_pp_group().is_first_rank:
|
|
if inputs_embeds is not None:
|
|
hidden_states = inputs_embeds
|
|
else:
|
|
hidden_states = self.embed_input_ids(input_ids)
|
|
residual = None
|
|
else:
|
|
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):
|
|
hidden_states, residual = layer(
|
|
positions,
|
|
hidden_states,
|
|
residual,
|
|
)
|
|
if not get_pp_group().is_last_rank:
|
|
return IntermediateTensors(
|
|
{"hidden_states": hidden_states, "residual": residual}
|
|
)
|
|
hidden_states = hidden_states + residual
|
|
return hidden_states
|
|
|
|
|
|
class MiMoForCausalLM(Qwen2ForCausalLM, nn.Module):
|
|
def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
|
|
nn.Module.__init__(self)
|
|
config = vllm_config.model_config.hf_config
|
|
quant_config = vllm_config.quant_config
|
|
|
|
self.config = config
|
|
|
|
self.quant_config = quant_config
|
|
|
|
self.model = MiMoModel(
|
|
vllm_config=vllm_config, prefix=maybe_prefix(prefix, "model")
|
|
)
|
|
|
|
if get_pp_group().is_last_rank:
|
|
if config.tie_word_embeddings:
|
|
self.lm_head = self.model.embed_tokens
|
|
else:
|
|
self.lm_head = ParallelLMHead(
|
|
config.vocab_size,
|
|
config.hidden_size,
|
|
quant_config=quant_config,
|
|
prefix=maybe_prefix(prefix, "lm_head"),
|
|
)
|
|
else:
|
|
self.lm_head = PPMissingLayer()
|
|
|
|
self.logits_processor = LogitsProcessor(config.vocab_size)
|
|
|
|
self.make_empty_intermediate_tensors = (
|
|
self.model.make_empty_intermediate_tensors
|
|
)
|
|
|
|
def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
|
|
skip_prefixes = ["lm_head."] if self.config.tie_word_embeddings else []
|
|
# MTP layers are loaded by the draft model, not the main model.
|
|
skip_prefixes.append("model.mtp_layers.")
|
|
loader = AutoWeightsLoader(self, skip_prefixes=skip_prefixes)
|
|
return loader.load_weights(weights)
|
|
|
|
def compute_logits(
|
|
self,
|
|
hidden_states: torch.Tensor,
|
|
) -> torch.Tensor | None:
|
|
hidden_states = self.model.norm(hidden_states)
|
|
logits = self.logits_processor(self.lm_head, hidden_states)
|
|
return logits
|