Sync from v0.13

This commit is contained in:
2026-01-19 10:38:50 +08:00
parent b2ef04d792
commit 5aef6c175a
3714 changed files with 854317 additions and 89342 deletions

View File

@@ -1,30 +1,150 @@
from typing import Optional
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from typing import Literal
from torch import nn
from vllm.config import (DeviceConfig, LoadConfig, LoRAConfig, ModelConfig,
ParallelConfig, SchedulerConfig, VisionLanguageConfig)
from vllm.model_executor.model_loader.loader import (BaseModelLoader,
get_model_loader)
from vllm.config import ModelConfig, VllmConfig
from vllm.config.load import LoadConfig
from vllm.logger import init_logger
from vllm.model_executor.model_loader.base_loader import BaseModelLoader
from vllm.model_executor.model_loader.bitsandbytes_loader import BitsAndBytesModelLoader
from vllm.model_executor.model_loader.default_loader import DefaultModelLoader
from vllm.model_executor.model_loader.dummy_loader import DummyModelLoader
from vllm.model_executor.model_loader.gguf_loader import GGUFModelLoader
from vllm.model_executor.model_loader.runai_streamer_loader import (
RunaiModelStreamerLoader,
)
from vllm.model_executor.model_loader.sharded_state_loader import ShardedStateLoader
from vllm.model_executor.model_loader.tensorizer_loader import TensorizerLoader
from vllm.model_executor.model_loader.utils import (
get_architecture_class_name, get_model_architecture)
get_architecture_class_name,
get_model_architecture,
get_model_cls,
)
logger = init_logger(__name__)
# Reminder: Please update docstring in `LoadConfig`
# if a new load format is added here
LoadFormats = Literal[
"auto",
"hf",
"bitsandbytes",
"dummy",
"fastsafetensors",
"gguf",
"mistral",
"npcache",
"pt",
"runai_streamer",
"runai_streamer_sharded",
"safetensors",
"sharded_state",
"tensorizer",
]
_LOAD_FORMAT_TO_MODEL_LOADER: dict[str, type[BaseModelLoader]] = {
"auto": DefaultModelLoader,
"hf": DefaultModelLoader,
"bitsandbytes": BitsAndBytesModelLoader,
"dummy": DummyModelLoader,
"fastsafetensors": DefaultModelLoader,
"gguf": GGUFModelLoader,
"mistral": DefaultModelLoader,
"npcache": DefaultModelLoader,
"pt": DefaultModelLoader,
"runai_streamer": RunaiModelStreamerLoader,
"runai_streamer_sharded": ShardedStateLoader,
"safetensors": DefaultModelLoader,
"sharded_state": ShardedStateLoader,
"tensorizer": TensorizerLoader,
}
def register_model_loader(load_format: str):
"""Register a customized vllm model loader.
When a load format is not supported by vllm, you can register a customized
model loader to support it.
Args:
load_format (str): The model loader format name.
Examples:
>>> from vllm.config.load import LoadConfig
>>> from vllm.model_executor.model_loader import (
... get_model_loader,
... register_model_loader,
... )
>>> from vllm.model_executor.model_loader.base_loader import BaseModelLoader
>>>
>>> @register_model_loader("my_loader")
... class MyModelLoader(BaseModelLoader):
... def download_model(self):
... pass
...
... def load_weights(self):
... pass
>>>
>>> load_config = LoadConfig(load_format="my_loader")
>>> type(get_model_loader(load_config))
<class 'MyModelLoader'>
""" # noqa: E501
def _wrapper(model_loader_cls):
if load_format in _LOAD_FORMAT_TO_MODEL_LOADER:
logger.warning(
"Load format `%s` is already registered, and will be "
"overwritten by the new loader class `%s`.",
load_format,
model_loader_cls,
)
if not issubclass(model_loader_cls, BaseModelLoader):
raise ValueError(
"The model loader must be a subclass of `BaseModelLoader`."
)
_LOAD_FORMAT_TO_MODEL_LOADER[load_format] = model_loader_cls
logger.info(
"Registered model loader `%s` with load format `%s`",
model_loader_cls,
load_format,
)
return model_loader_cls
return _wrapper
def get_model_loader(load_config: LoadConfig) -> BaseModelLoader:
"""Get a model loader based on the load format."""
load_format = load_config.load_format
if load_format not in _LOAD_FORMAT_TO_MODEL_LOADER:
raise ValueError(f"Load format `{load_format}` is not supported")
return _LOAD_FORMAT_TO_MODEL_LOADER[load_format](load_config)
def get_model(
*, model_config: ModelConfig, load_config: LoadConfig,
device_config: DeviceConfig, parallel_config: ParallelConfig,
scheduler_config: SchedulerConfig, lora_config: Optional[LoRAConfig],
vision_language_config: Optional[VisionLanguageConfig]) -> nn.Module:
loader = get_model_loader(load_config)
return loader.load_model(model_config=model_config,
device_config=device_config,
lora_config=lora_config,
vision_language_config=vision_language_config,
parallel_config=parallel_config,
scheduler_config=scheduler_config)
*, vllm_config: VllmConfig, model_config: ModelConfig | None = None
) -> nn.Module:
loader = get_model_loader(vllm_config.load_config)
if model_config is None:
model_config = vllm_config.model_config
return loader.load_model(vllm_config=vllm_config, model_config=model_config)
__all__ = [
"get_model", "get_model_loader", "BaseModelLoader",
"get_architecture_class_name", "get_model_architecture"
"get_model",
"get_model_loader",
"get_architecture_class_name",
"get_model_architecture",
"get_model_cls",
"register_model_loader",
"BaseModelLoader",
"BitsAndBytesModelLoader",
"GGUFModelLoader",
"DefaultModelLoader",
"DummyModelLoader",
"RunaiModelStreamerLoader",
"ShardedStateLoader",
"TensorizerLoader",
]

View File

@@ -0,0 +1,57 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from abc import ABC, abstractmethod
import torch
import torch.nn as nn
from vllm.config import ModelConfig, VllmConfig
from vllm.config.load import LoadConfig
from vllm.logger import init_logger
from vllm.model_executor.model_loader.utils import (
initialize_model,
process_weights_after_loading,
)
from vllm.utils.torch_utils import set_default_torch_dtype
logger = init_logger(__name__)
class BaseModelLoader(ABC):
"""Base class for model loaders."""
def __init__(self, load_config: LoadConfig):
self.load_config = load_config
@abstractmethod
def download_model(self, model_config: ModelConfig) -> None:
"""Download a model so that it can be immediately loaded."""
raise NotImplementedError
@abstractmethod
def load_weights(self, model: nn.Module, model_config: ModelConfig) -> None:
"""Load weights into a model. This standalone API allows
inplace weights loading for an already-initialized model"""
raise NotImplementedError
def load_model(
self, vllm_config: VllmConfig, model_config: ModelConfig
) -> nn.Module:
"""Load a model with the given configurations."""
device_config = vllm_config.device_config
load_config = vllm_config.load_config
load_device = (
device_config.device if load_config.device is None else load_config.device
)
target_device = torch.device(load_device)
with set_default_torch_dtype(model_config.dtype):
with target_device:
model = initialize_model(
vllm_config=vllm_config, model_config=model_config
)
logger.debug("Loading weights on %s ...", load_device)
# Quantization does not happen in `load_weights` but after it
self.load_weights(model, model_config)
process_weights_after_loading(model, model_config, target_device)
return model.eval()

View File

@@ -0,0 +1,822 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
# ruff: noqa: SIM117
import fnmatch
import glob
import itertools
import math
import os
from collections.abc import Callable, Generator
from typing import Any
import numpy as np
import torch
from huggingface_hub import HfApi
from packaging import version
from torch import nn
from transformers.utils import SAFE_WEIGHTS_INDEX_NAME
from vllm.config import ModelConfig
from vllm.config.load import LoadConfig
from vllm.distributed import (
get_tensor_model_parallel_rank,
get_tensor_model_parallel_world_size,
)
from vllm.logger import init_logger
from vllm.model_executor.layers.fused_moe import FusedMoE
from vllm.model_executor.layers.linear import (
LinearBase,
MergedColumnParallelLinear,
QKVParallelLinear,
ReplicatedLinear,
RowParallelLinear,
)
from vllm.model_executor.model_loader.base_loader import BaseModelLoader
from vllm.model_executor.model_loader.utils import ParamMapping
from vllm.model_executor.model_loader.weight_utils import (
download_safetensors_index_file_from_hf,
download_weights_from_hf,
filter_duplicate_safetensors_files,
filter_files_not_needed_for_inference,
pt_weights_iterator,
safetensors_weights_iterator,
)
from vllm.model_executor.models import is_pooling_model
from vllm.model_executor.utils import (
get_moe_expert_mapping,
get_packed_modules_mapping,
set_weight_attrs,
)
from vllm.platforms import current_platform
from vllm.utils.torch_utils import set_default_torch_dtype
logger = init_logger(__name__)
def is_moe_model(model: torch.nn.Module) -> bool:
"""Checks if the model contains FusedMoE layers."""
return bool(any(isinstance(module, FusedMoE) for module in model.modules()))
class BitsAndBytesModelLoader(BaseModelLoader):
"""Model loader to load model weights with BitAndBytes quantization."""
possible_config_file_names = ["adapter_config.json"]
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
# Save the module names without sharding.
self.unsharded_weights_modules: list[str] = []
# Save the module names that are sharded by column.
self.column_sharded_weights_modules: list[str] = []
# Modules whose weights might have fused on disk
# we need their output_sizes to make shard in flight correctly with TP
self.maybe_fused_weights_modules: dict[str, list[int]] = {}
# Store all module names (from transformers) that support
# BNB quantization.
self.target_modules: list[str] = []
self.tp_disabled_modules: list[str] = []
# Store the mapping of expert parameters for MoE models.
self.expert_params_mapping: list[tuple[str, str, int, str]] = []
# mapping weight names from transformers to vllm.
self.weight_mapper: Callable = lambda name: name
self.pre_quant: bool = False
self.load_8bit: bool = False
self.is_pool_model: bool = False
def _get_weight_files(
self,
model_name_or_path: str,
allowed_patterns: list[str],
revision: str | None = None,
) -> tuple[str, list[str], str]:
"""Retrieve weight files. Download the files if necessary.
Return the weight files and the file pattern."""
is_local = os.path.isdir(model_name_or_path)
if is_local:
for pattern in allowed_patterns:
weight_files = glob.glob(os.path.join(model_name_or_path, pattern))
if weight_files:
return model_name_or_path, weight_files, pattern
else:
hf_api = HfApi()
repo_files = hf_api.list_repo_files(repo_id=model_name_or_path)
for pattern in allowed_patterns:
matching_files = fnmatch.filter(repo_files, pattern)
if matching_files:
hf_folder = download_weights_from_hf(
model_name_or_path,
self.load_config.download_dir,
[pattern],
revision,
ignore_patterns=self.load_config.ignore_patterns,
)
return (
hf_folder,
glob.glob(os.path.join(hf_folder, pattern)),
pattern,
)
raise RuntimeError(f"No model weights found in: `{model_name_or_path}`")
def _prepare_weights(
self, model_name_or_path: str, revision: str | None
) -> tuple[list[str], bool]:
"""Prepare weight files for the model."""
allowed_patterns = ["*.safetensors", "*.bin", "*.pt"]
hf_folder, hf_weights_files, matched_pattern = self._get_weight_files(
model_name_or_path, allowed_patterns, revision
)
use_safetensors = matched_pattern == "*.safetensors"
is_local = os.path.isdir(model_name_or_path)
index_file = SAFE_WEIGHTS_INDEX_NAME
if use_safetensors:
# For models like Mistral-7B-Instruct-v0.3
# there are both sharded safetensors files and a consolidated
# safetensors file. Using both breaks.
# Here, we download the `model.safetensors.index.json` and filter
# any files not found in the index.
if not is_local:
download_safetensors_index_file_from_hf(
model_name_or_path,
index_file,
self.load_config.download_dir,
revision,
)
hf_weights_files = filter_duplicate_safetensors_files(
hf_weights_files, hf_folder, index_file
)
else:
hf_weights_files = filter_files_not_needed_for_inference(hf_weights_files)
if len(hf_weights_files) == 0:
raise RuntimeError(
f"Cannot find any model weights with `{model_name_or_path}`"
)
return hf_weights_files, use_safetensors
def _hf_weight_iter(self, hf_weights_files, use_safetensors: bool):
def _maybe_pool_model(module_name: str):
# For pool model, we need to add the prefix `model.`
# for the weight name if possible.
if (
self.is_pool_model
and self.target_modules[0].startswith("model.")
and not module_name.startswith("model.")
):
return "model." + module_name
return module_name
if use_safetensors:
iterator = safetensors_weights_iterator(
hf_weights_files,
self.load_config.use_tqdm_on_load,
)
else:
iterator = pt_weights_iterator(
hf_weights_files,
self.load_config.use_tqdm_on_load,
self.load_config.pt_load_map_location,
)
for org_name, param in iterator:
# mapping weight names from transformers to vllm while preserving
# original names.
mapped_name = self.weight_mapper(org_name)
mapped_name = _maybe_pool_model(mapped_name)
yield org_name, mapped_name, param
def _get_quantized_weights_iterator(
self,
model_name_or_path: str,
revision: str | None,
) -> tuple[Generator[tuple[str, torch.Tensor], None, None], dict[str, Any]]:
"""Get an iterator to the model weights with bitsandbytes quantization,
as well as the quantization state dictionary."""
# only load the bitsandbytes module when needed
try:
import bitsandbytes
if version.parse(bitsandbytes.__version__) < version.parse("0.46.1"):
raise ImportError(
"bitsandbytes version is wrong. Please "
"install bitsandbytes>=0.46.1."
)
except ImportError as err:
raise ImportError(
"Please install bitsandbytes>=0.46.1 via "
"`pip install bitsandbytes>=0.46.1` to use "
"bitsandbytes quantizer."
) from err
hf_weights_files, use_safetensors = self._prepare_weights(
model_name_or_path, revision
)
quant_state_dict: dict[str, Any] = {}
if self.pre_quant:
if self.load_8bit:
return self._quantized_8bit_generator(
hf_weights_files, use_safetensors, quant_state_dict
), quant_state_dict
else:
return self._quantized_4bit_generator(
hf_weights_files, use_safetensors, quant_state_dict
), quant_state_dict
return self._unquantized_generator(
hf_weights_files, use_safetensors, quant_state_dict
), quant_state_dict
def _is_8bit_weight_name(self, weight_name: str):
quantized_suffix = {".scb", ".weight_format"}
return any(weight_name.lower().endswith(suffix) for suffix in quantized_suffix)
def _is_4bit_weight_name(self, weight_name: str):
quantized_suffix = {
"absmax",
"quant_map",
"nested_absmax",
"nested_quant_map",
"bitsandbytes",
}
suffix = weight_name.split(".")[-1]
return any(q_suffix in suffix for q_suffix in quantized_suffix)
def _quantized_8bit_generator(
self, hf_weights_files, use_safetensors, quant_state_dict
) -> Generator:
for (
org_weight_name,
mapped_weight_name,
weight_tensor,
) in self._hf_weight_iter(hf_weights_files, use_safetensors):
if not mapped_weight_name.lower().endswith(".scb"):
continue
weight_key = mapped_weight_name.lower().replace(".scb", ".weight")
quant_state_dict[weight_key] = weight_tensor
for (
org_weight_name,
mapped_weight_name,
weight_tensor,
) in self._hf_weight_iter(hf_weights_files, use_safetensors):
if self._is_8bit_weight_name(mapped_weight_name):
continue
if mapped_weight_name in quant_state_dict:
set_weight_attrs(weight_tensor, {"load_in_8bit": True})
yield org_weight_name, weight_tensor
else:
yield org_weight_name, weight_tensor
def _quantized_4bit_generator(
self, hf_weights_files, use_safetensors, quant_state_dict
) -> Generator:
from bitsandbytes.functional import QuantState
# First iterate over all quant state weights
weight_iterator = self._hf_weight_iter(hf_weights_files, use_safetensors)
temp_state_dict = {}
for (
org_weight_name,
mapped_weight_name,
weight_tensor,
) in weight_iterator:
if not self._is_4bit_weight_name(mapped_weight_name):
continue
# bitsandbytes library requires
# weight.quant_state.bitsandbytes__* in CPU
if "quant_state.bitsandbytes" in mapped_weight_name:
temp_state_dict[mapped_weight_name] = weight_tensor.cpu().data
else:
temp_state_dict[mapped_weight_name] = weight_tensor
# Closure to parse quant_state for each prequant weight
def _parse_quant_state(param_name: str, temp_state_dict: dict) -> QuantState:
quant_state = {}
for k in temp_state_dict:
if param_name + "." in k:
quant_state[k] = temp_state_dict[k]
return QuantState.from_dict(
quant_state, device=current_platform.device_type
)
# Second iterate over all prequant and normal weights
# pre quantized weights would have a quant_state
for (
org_weight_name,
mapped_weight_name,
weight_tensor,
) in self._hf_weight_iter(hf_weights_files, use_safetensors):
if self._is_4bit_weight_name(mapped_weight_name):
continue
if (
f"{mapped_weight_name}.quant_state.bitsandbytes__nf4" in temp_state_dict
) or (
f"{mapped_weight_name}.quant_state.bitsandbytes__fp4" in temp_state_dict
):
quant_state = _parse_quant_state(mapped_weight_name, temp_state_dict)
quant_state_dict[mapped_weight_name] = quant_state
yield org_weight_name, weight_tensor
else:
yield org_weight_name, weight_tensor
def _unquantized_generator(
self, hf_weights_files, use_safetensors, quant_state_dict
) -> Generator:
from bitsandbytes.functional import quantize_4bit
global_tp_size = get_tensor_model_parallel_world_size()
global_tp_rank = get_tensor_model_parallel_rank()
check_match = (
lambda weight_name, module_name: weight_name.removesuffix(".weight")
== module_name
)
for (
org_weight_name,
mapped_weight_name,
weight_tensor,
) in self._hf_weight_iter(hf_weights_files, use_safetensors):
# override tp_size and tp_rank if the module has disabled TP
if any(
tp_disabled_module in mapped_weight_name
for tp_disabled_module in self.tp_disabled_modules
):
tp_size = 1
tp_rank = 0
else:
tp_size = global_tp_size
tp_rank = global_tp_rank
if any(
target_module in mapped_weight_name
for target_module in self.target_modules
) and mapped_weight_name.endswith(".weight"):
# Without sharding
if any(
check_match(mapped_weight_name, module)
for module in self.unsharded_weights_modules
):
weight_sub_tensor = weight_tensor
# Shard by column
elif any(
check_match(mapped_weight_name, module)
for module in self.column_sharded_weights_modules
):
total_size = weight_tensor.size(-1)
start_index = total_size // tp_size * tp_rank
end_index = total_size // tp_size * (tp_rank + 1)
weight_sub_tensor = weight_tensor[..., start_index:end_index]
# Weights have fused on disk. In this case, we assume that the
# weight and module use same name.
elif any(
check_match(mapped_weight_name, module)
for module in self.maybe_fused_weights_modules
):
# special case for fused weights
# get the size of each shard weight tensor
total_shard_sizes = next(
(
sizes
for module, sizes in self.maybe_fused_weights_modules.items() # noqa: E501
if check_match(mapped_weight_name, module)
)
)
total_size = weight_tensor.size(0)
assert total_size == sum(total_shard_sizes)
# get the start/end index of each shard weight tensor
total_start_index = list(
itertools.accumulate([0] + total_shard_sizes)
)[:-1]
shard_weights_index = [
(
idx + size // tp_size * tp_rank,
idx + size // tp_size * (tp_rank + 1),
)
for idx, size in zip(total_start_index, total_shard_sizes)
]
# slice and reorder the weight tensor
weight_tensor = [
weight_tensor[start_index:end_index, ...]
for start_index, end_index in shard_weights_index
]
weight_sub_tensor = torch.cat(weight_tensor, dim=0)
# Shard by row
else:
total_size = weight_tensor.size(0)
start_index = total_size // tp_size * tp_rank
end_index = total_size // tp_size * (tp_rank + 1)
weight_sub_tensor = weight_tensor[start_index:end_index, ...]
# bitsandbytes requires data in GPU
if weight_sub_tensor.is_cuda:
loaded_weight = weight_sub_tensor
else:
loaded_weight = weight_sub_tensor.to(
device=current_platform.device_type
)
# remove the following after the issue is fixed:
# https://github.com/bitsandbytes-foundation/bitsandbytes/issues/1342
if loaded_weight.is_contiguous() is False:
loaded_weight = loaded_weight.contiguous()
with set_default_torch_dtype(torch.float32):
processed_weight, quant_state = quantize_4bit(
loaded_weight,
compress_statistics=True,
quant_type="nf4",
)
quant_state_dict[mapped_weight_name] = quant_state
else:
processed_weight = weight_tensor
yield org_weight_name, processed_weight
def _get_bnb_target_modules(self, model: nn.Module) -> None:
"""
Identify and collect all modules that support BitsAndBytes
quantization.
"""
for name, module in model.named_modules():
if isinstance(module, LinearBase) and hasattr(
module.quant_method, "quant_config"
):
if modules_info := self.modules_mapping.get_sub_modules(name):
# Map vllm's names to transformers's names.
rep_name, sub_modules = modules_info
for sub_name in sub_modules:
new_name = name.replace(rep_name, sub_name)
self.target_modules.append(new_name)
if module.disable_tp:
self.tp_disabled_modules.append(new_name)
# Add original module name even if the module has stacked map,
# in case model has a mixture of disk-merged and disk-split
# weights with same last name.
self.target_modules.append(name)
if module.disable_tp:
self.tp_disabled_modules.append(name)
elif isinstance(module, FusedMoE) and hasattr(
module.quant_method, "quant_config"
):
# TODO: support FusedMoE with prequant and 8bit.
if self.pre_quant and self.load_8bit:
raise ValueError(
"Prequant BitsAndBytes 8bit models with FusedMoE "
"is not supported yet."
)
# Get the corresponding weight name using module name and
# expert_params_mapping.
for exp in self.expert_params_mapping:
weight_name = exp[1]
rep_name = name.replace("experts", "") + weight_name.removesuffix(
"."
)
self.target_modules.append(rep_name)
assert self.target_modules, (
"vLLM currently does not support BNB quantization for"
)
f" {type(model).__name__}"
def _classify_module_sharding(self, model: nn.Module):
"""
Categorize modules based on their weight sharding requirements
for tensor parallelism.
"""
for name, module in model.named_modules():
# Some modules like `ReplicatedLinear` should not have their weights
# sharded. The reason for implementing it this way is to avoid new
# static variable in the model implementation.
if isinstance(module, (ReplicatedLinear,)):
self.unsharded_weights_modules.append(name)
# `QKVParallelLinear` and `MergedColumnParallelLinear` might have
# fused weights on disk. We need to use the output sizes of these
# modules to shard the weights correctly.
elif isinstance(module, (QKVParallelLinear, MergedColumnParallelLinear)):
self.maybe_fused_weights_modules[name] = module.output_sizes
# In TP, these weights are partitioned along the column
# dimension (dim=-1)
elif isinstance(module, (RowParallelLinear,)):
self.column_sharded_weights_modules.append(name)
elif isinstance(module, FusedMoE):
expert_mapping = self.expert_params_mapping
for exp in expert_mapping:
if exp[-1] == "w2":
weight_name = exp[1]
rep_name = name.replace(
"experts", ""
) + weight_name.removesuffix(".")
self.column_sharded_weights_modules.append(rep_name)
def _verify_model_compatibility(
self, model: nn.Module, model_config: ModelConfig
) -> None:
"""
Verify that the model is compatible with BitsAndBytes quantization.
"""
if not hasattr(model, "load_weights"):
raise AttributeError(
"The required method 'load_weights' is not defined in class"
f" {type(model).__name__}."
)
if not hasattr(model, "packed_modules_mapping"):
raise AttributeError(
f"Model {type(model).__name__} does not support BitsAndBytes "
"quantization yet. No 'packed_modules_mapping' found."
)
quant_config = getattr(model_config.hf_config, "quantization_config", None)
if quant_config and (quant_method := quant_config.get("quant_method")):
if quant_method == "bitsandbytes":
self.pre_quant = True
else:
raise ValueError(
f"BitsAndBytes loader does not support {quant_method} quantization"
)
# The quant_states in pre_quantized models cannot work with a split
# weight tensor. So TP does not work with pre_quantized bnb models.
if self.pre_quant and get_tensor_model_parallel_world_size() > 1:
raise ValueError(
"Prequant BitsAndBytes models with tensor parallelism is not "
"supported. Please try with pipeline parallelism."
)
if quant_config and self.pre_quant:
self.load_8bit = quant_config.get("load_in_8bit", False)
def _initialize_loader_state(
self, model: nn.Module, model_config: ModelConfig
) -> None:
"""
Initialize the loader's internal state based on the model and
configuration.
"""
self.is_pool_model = is_pooling_model(model)
self.modules_mapping = ParamMapping(get_packed_modules_mapping(model))
if is_moe_model(model):
self.expert_params_mapping = get_moe_expert_mapping(model)
if not self.expert_params_mapping:
raise AttributeError(
f"MoE Model {type(model).__name__} does not support "
"BitsAndBytes quantization yet. Ensure this model has "
"'get_expert_mapping' method."
)
# For some models like Molmo, we need to use hf_to_vllm_mapper
# to ensure correct loading of weights.
if hf_to_vllm_mapper := getattr(model, "hf_to_vllm_mapper", None):
self.weight_mapper = lambda name: hf_to_vllm_mapper._map_name(name)
self._get_bnb_target_modules(model)
self._classify_module_sharding(model)
def _dequantize_dq(self, quant_states: Any):
"""
When BNB employs Double Quantization, we perform the dequantization of
these constants during weight loading rather than at inference time,
thereby avoiding this computational overhead during inference. This
comes at the cost of increased memory usage.
"""
from bitsandbytes.functional import QuantState, dequantize_blockwise
def _dequantize_single_state(quant_state):
"""Helper function to dequantize a single QuantState object."""
if not (isinstance(quant_state, QuantState) and quant_state.nested):
return
# Copied from: https://github.com/bitsandbytes-foundation/bitsandbytes/blob/0.45.3/bitsandbytes/functional.py#L1352-#L1356
absmax = dequantize_blockwise(quant_state.absmax, quant_state.state2)
absmax += quant_state.offset
# Ensure float32 dtype
if absmax.dtype != torch.float32:
absmax = absmax.float()
quant_state.absmax = absmax
quant_state.nested = False
quant_state.offset = None
quant_state.state2 = None
if isinstance(quant_states, dict):
for quant_state in quant_states.values():
_dequantize_single_state(quant_state)
else:
_dequantize_single_state(quant_states)
return quant_states
def _fuse_moe_quant_states(self, model: nn.Module, quant_states_dict: dict) -> dict:
"""
This function consolidates individual expert quantization states into
fused representations for w13 and w2.
"""
from bitsandbytes.functional import QuantState
if not self.expert_params_mapping:
return dict()
expert_mapping = self.expert_params_mapping
expert_qs_dict = {}
for name, module in model.named_modules():
if not isinstance(module, FusedMoE):
continue
w1_states_lst = []
w2_states_lst = []
w3_states_lst = []
for exp in expert_mapping:
shard_id = exp[-1]
if shard_id not in ("w1", "w2", "w3"):
raise ValueError(
f"shard_id must be ['w1','w2','w3'] but got {shard_id}."
)
layer_prefix = name.split("experts")[0]
weight_qual_name = layer_prefix + exp[1] + "weight"
quant_state = self._dequantize_dq(quant_states_dict[weight_qual_name])
if shard_id == "w1":
w1_states_lst.append(quant_state)
elif shard_id == "w2":
w2_states_lst.append(quant_state)
else:
w3_states_lst.append(quant_state)
del quant_states_dict[weight_qual_name]
assert len(w1_states_lst) == len(w2_states_lst) == len(w3_states_lst)
w13_absmax_lst = []
w2_absmax_lst = []
w13_total_dim0 = 0
w2_total_dim0 = 0
for w1_qs, w2_qs, w3_qs in zip(w1_states_lst, w2_states_lst, w3_states_lst):
assert w1_qs.shape == w3_qs.shape
assert w1_qs.blocksize == w2_qs.blocksize == w3_qs.blocksize
assert w1_qs.dtype == w2_qs.dtype == w3_qs.dtype
# w1 and w3 are interleaved in storage
w13_absmax_lst.append(w1_qs.absmax)
w13_absmax_lst.append(w3_qs.absmax)
w2_absmax_lst.append(w2_qs.absmax)
w13_total_dim0 += w1_qs.shape[0] + w3_qs.shape[0]
w2_total_dim0 += w2_qs.shape[0]
w13_absmax = torch.cat(w13_absmax_lst)
w2_absmax = torch.cat(w2_absmax_lst)
# Create fused quantization state for w13.
w13_qs = QuantState(
absmax=w13_absmax,
shape=(w13_total_dim0, w1_states_lst[0].shape[1]),
code=w1_states_lst[0].code,
blocksize=w1_states_lst[0].blocksize,
quant_type="nf4",
dtype=w1_states_lst[0].dtype,
)
# Create fused quantization state for w2.
w2_qs = QuantState(
absmax=w2_absmax,
shape=(w2_total_dim0, w2_states_lst[0].shape[1]),
code=w2_states_lst[0].code,
blocksize=w2_states_lst[0].blocksize,
quant_type="nf4",
dtype=w2_states_lst[0].dtype,
)
# The weight suffixes .w13_weight and .w2_weight are consistent
# with the param in BitsAndBytesMoEMethod.
w13_weight_name = name + ".w13_weight"
w2_weight_name = name + ".w2_weight"
expert_qs_dict[w13_weight_name] = w13_qs
expert_qs_dict[w2_weight_name] = w2_qs
return expert_qs_dict
def _stack_quantization_states(
self, model: nn.Module, quant_state_dict: dict
) -> dict[str, dict[int, Any]]:
stacked_quant_state_dict: dict[str, dict[int, Any]] = {}
# TODO: Change this lazy import to normal import
# after the checks are updated to run on a new version
from vllm.model_executor.models.utils import is_pp_missing_parameter
param_dict = dict(model.named_parameters())
for quant_param_name in quant_state_dict:
if is_pp_missing_parameter(quant_param_name, model):
continue
non_stacked_param_name = quant_param_name
shard_index = 0
for shard_name, (
weight_name,
index,
) in self.modules_mapping.inverse_packed_mapping.items():
# Some models, such as MiniCPM V2.5/2.6, contain both
# module names 'kv_proj' and 'qkv_proj'. To prevent 'kv_proj'
# from being incorrectly identified as being present in
# 'vpm.encoder.layers.0.self_attn.qkv_proj.weight
shard_pos = quant_param_name.find(shard_name)
can_correct_rename = (shard_pos > 0) and (
quant_param_name[shard_pos - 1] == "."
)
# If the quant_param_name is packed, it won't occur in the
# param_dict before renaming.
new_quant_param_name = quant_param_name.replace(shard_name, weight_name)
need_rename = (quant_param_name not in param_dict) and (
new_quant_param_name in param_dict
)
if can_correct_rename and need_rename:
shard_index = index
quant_param_name = new_quant_param_name
break
# Models like Clip/Siglip may skip some layers in initialization,
# causing unused quant_param_name in state_dict.
if quant_param_name not in param_dict:
continue
if quant_param_name not in stacked_quant_state_dict:
stacked_quant_state_dict[quant_param_name] = {}
stacked_quant_state_dict[quant_param_name][shard_index] = quant_state_dict[
non_stacked_param_name
]
return stacked_quant_state_dict
def _bind_quant_states_to_params(
self, model: nn.Module, stacked_quant_state_dict: dict
) -> None:
# save quant_states and offsets as the attributes of the parameters
param_dict = dict(model.named_parameters())
for param_name, param in param_dict.items():
if param_name in stacked_quant_state_dict:
quant_states = stacked_quant_state_dict[param_name]
# Dequantize double quantized values during weight loading.
self._dequantize_dq(quant_states)
set_weight_attrs(param, {"bnb_quant_state": quant_states})
if not isinstance(quant_states, dict):
continue
pack_ratio = getattr(param, "pack_factor", -1)
if pack_ratio == -1:
raise ValueError(f"pack_factor not set for parameter {param_name}.")
num_elements = [0] * len(quant_states)
for seq, quant_state in quant_states.items():
num_elements[seq] = math.prod(quant_state.shape) // pack_ratio
offsets = np.concatenate(([0], np.cumsum(num_elements)))
# Make torch infer_schema happy
offsets = torch.tensor(offsets).cpu()
set_weight_attrs(param, {"bnb_shard_offsets": offsets})
if self.load_8bit:
set_weight_attrs(
param, {"matmul_state": [None] * len(quant_states)}
)
def load_weights(self, model: nn.Module, model_config: ModelConfig) -> None:
self._verify_model_compatibility(model, model_config)
self._initialize_loader_state(model, model_config)
logger.info(
"Loading weights with BitsAndBytes quantization. May take a while ..."
)
qweight_iterator, quant_state_dict = self._get_quantized_weights_iterator(
model_config.model,
model_config.revision,
)
weights_to_load = {name for name, _ in model.named_parameters()}
loaded_weights = model.load_weights(qweight_iterator)
# Some models may have weights loading tracker unimplemented.
if loaded_weights is not None:
weights_not_loaded = weights_to_load - loaded_weights
if weights_not_loaded:
raise ValueError(
"Following weights were not initialized from "
f"checkpoint: {weights_not_loaded}"
)
expert_quant_state_dict = self._fuse_moe_quant_states(model, quant_state_dict)
stacked_quant_state_dict = self._stack_quantization_states(
model, quant_state_dict
)
stacked_quant_state_dict = {
**expert_quant_state_dict,
**stacked_quant_state_dict,
}
self._bind_quant_states_to_params(model, stacked_quant_state_dict)
torch.cuda.empty_cache()
def download_model(self, model_config: ModelConfig) -> None:
self._prepare_weights(model_config.model, model_config.revision)

View File

@@ -0,0 +1,321 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import dataclasses
import glob
import os
import time
from collections.abc import Generator, Iterable
from typing import cast
import torch
from torch import nn
from transformers.utils import SAFE_WEIGHTS_INDEX_NAME
from vllm.config import ModelConfig
from vllm.config.load import LoadConfig
from vllm.logger import init_logger
from vllm.model_executor.layers.quantization.torchao import torchao_version_at_least
from vllm.model_executor.model_loader.base_loader import BaseModelLoader
from vllm.model_executor.model_loader.weight_utils import (
download_safetensors_index_file_from_hf,
download_weights_from_hf,
fastsafetensors_weights_iterator,
filter_duplicate_safetensors_files,
filter_files_not_needed_for_inference,
get_quant_config,
maybe_download_from_modelscope,
multi_thread_pt_weights_iterator,
multi_thread_safetensors_weights_iterator,
np_cache_weights_iterator,
pt_weights_iterator,
safetensors_weights_iterator,
)
from vllm.platforms import current_platform
from vllm.transformers_utils.repo_utils import list_filtered_repo_files
logger = init_logger(__name__)
class DefaultModelLoader(BaseModelLoader):
"""Model loader that can load different file types from disk."""
# default number of thread when enable multithread weight loading
DEFAULT_NUM_THREADS = 8
@dataclasses.dataclass
class Source:
"""A source for weights."""
model_or_path: str
"""The model ID or path."""
revision: str | None
"""The optional model revision."""
prefix: str = ""
"""A prefix to prepend to all weights."""
fall_back_to_pt: bool = True
"""Whether .pt weights can be used."""
allow_patterns_overrides: list[str] | None = None
"""If defined, weights will load exclusively using these patterns."""
counter_before_loading_weights: float = 0.0
counter_after_loading_weights: float = 0.0
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
extra_config = load_config.model_loader_extra_config
allowed_keys = {"enable_multithread_load", "num_threads"}
unexpected_keys = set(extra_config.keys()) - allowed_keys
if unexpected_keys:
raise ValueError(
f"Unexpected extra config keys for load format "
f"{load_config.load_format}: "
f"{unexpected_keys}"
)
def _prepare_weights(
self,
model_name_or_path: str,
revision: str | None,
fall_back_to_pt: bool,
allow_patterns_overrides: list[str] | None,
) -> tuple[str, list[str], bool]:
"""Prepare weights for the model.
If the model is not local, it will be downloaded."""
model_name_or_path = (
maybe_download_from_modelscope(model_name_or_path, revision)
or model_name_or_path
)
is_local = os.path.isdir(model_name_or_path)
load_format = self.load_config.load_format
use_safetensors = False
index_file = SAFE_WEIGHTS_INDEX_NAME
# First check for 'auto' format that mistral files format are present.
# This is to load mistral models with official format by default.
if load_format == "auto":
load_format = (
"mistral"
if len(
list_filtered_repo_files(
model_name_or_path=model_name_or_path,
allow_patterns=["consolidated*.safetensors"],
revision=revision,
)
)
> 0
else "hf"
)
# Some quantized models use .pt files for storing the weights.
if load_format == "hf":
allow_patterns = ["*.safetensors", "*.bin"]
elif load_format == "safetensors" or load_format == "fastsafetensors":
use_safetensors = True
allow_patterns = ["*.safetensors"]
elif load_format == "mistral":
use_safetensors = True
allow_patterns = ["consolidated*.safetensors"]
index_file = "consolidated.safetensors.index.json"
elif load_format == "pt":
allow_patterns = ["*.pt"]
elif load_format == "npcache":
allow_patterns = ["*.bin"]
else:
raise ValueError(f"Unknown load_format: {load_format}")
if fall_back_to_pt:
allow_patterns += ["*.pt"]
if allow_patterns_overrides is not None:
allow_patterns = allow_patterns_overrides
if not is_local:
hf_folder = download_weights_from_hf(
model_name_or_path,
self.load_config.download_dir,
allow_patterns,
revision,
ignore_patterns=self.load_config.ignore_patterns,
)
else:
hf_folder = model_name_or_path
hf_weights_files: list[str] = []
for pattern in allow_patterns:
hf_weights_files += glob.glob(os.path.join(hf_folder, pattern))
if len(hf_weights_files) > 0:
if pattern == "*.safetensors":
use_safetensors = True
break
if use_safetensors:
# For models like Mistral-7B-Instruct-v0.3
# there are both sharded safetensors files and a consolidated
# safetensors file. Using both breaks.
# Here, we download the `model.safetensors.index.json` and filter
# any files not found in the index.
if not is_local:
download_safetensors_index_file_from_hf(
model_name_or_path,
index_file,
self.load_config.download_dir,
revision,
)
hf_weights_files = filter_duplicate_safetensors_files(
hf_weights_files, hf_folder, index_file
)
else:
hf_weights_files = filter_files_not_needed_for_inference(hf_weights_files)
if len(hf_weights_files) == 0:
raise RuntimeError(
f"Cannot find any model weights with `{model_name_or_path}`"
)
return hf_folder, hf_weights_files, use_safetensors
def _get_weights_iterator(
self, source: "Source"
) -> Generator[tuple[str, torch.Tensor], None, None]:
"""Get an iterator for the model weights based on the load format."""
extra_config = self.load_config.model_loader_extra_config
hf_folder, hf_weights_files, use_safetensors = self._prepare_weights(
source.model_or_path,
source.revision,
source.fall_back_to_pt,
source.allow_patterns_overrides,
)
if self.load_config.load_format == "npcache":
# Currently np_cache only support *.bin checkpoints
assert use_safetensors is False
weights_iterator = np_cache_weights_iterator(
source.model_or_path,
self.load_config.download_dir,
hf_folder,
hf_weights_files,
self.load_config.use_tqdm_on_load,
)
elif use_safetensors:
if self.load_config.load_format == "fastsafetensors":
weights_iterator = fastsafetensors_weights_iterator(
hf_weights_files,
self.load_config.use_tqdm_on_load,
)
else:
if extra_config.get("enable_multithread_load"):
weights_iterator = multi_thread_safetensors_weights_iterator(
hf_weights_files,
self.load_config.use_tqdm_on_load,
max_workers=extra_config.get(
"num_threads", self.DEFAULT_NUM_THREADS
),
)
else:
weights_iterator = safetensors_weights_iterator(
hf_weights_files,
self.load_config.use_tqdm_on_load,
self.load_config.safetensors_load_strategy,
)
else:
if extra_config.get("enable_multithread_load"):
weights_iterator = multi_thread_pt_weights_iterator(
hf_weights_files,
self.load_config.use_tqdm_on_load,
self.load_config.pt_load_map_location,
max_workers=extra_config.get(
"num_threads", self.DEFAULT_NUM_THREADS
),
)
else:
weights_iterator = pt_weights_iterator(
hf_weights_files,
self.load_config.use_tqdm_on_load,
self.load_config.pt_load_map_location,
)
if current_platform.is_tpu():
from vllm.platforms.tpu import USE_TPU_INFERENCE
if not USE_TPU_INFERENCE:
# In PyTorch XLA, we should call `torch_xla.sync`
# frequently so that not too many ops are accumulated
# in the XLA program.
import torch_xla
def _xla_weights_iterator(iterator: Generator):
for weights in iterator:
yield weights
torch_xla.sync(wait=False)
weights_iterator = _xla_weights_iterator(weights_iterator)
if self.counter_before_loading_weights == 0.0:
self.counter_before_loading_weights = time.perf_counter()
# Apply the prefix.
return ((source.prefix + name, tensor) for (name, tensor) in weights_iterator)
def get_all_weights(
self,
model_config: ModelConfig,
model: nn.Module,
) -> Generator[tuple[str, torch.Tensor], None, None]:
primary_weights = DefaultModelLoader.Source(
model_config.model,
model_config.revision,
prefix="",
fall_back_to_pt=getattr(model, "fall_back_to_pt_during_load", True),
allow_patterns_overrides=getattr(model, "allow_patterns_overrides", None),
)
yield from self._get_weights_iterator(primary_weights)
secondary_weights = cast(
Iterable[DefaultModelLoader.Source],
getattr(model, "secondary_weights", ()),
)
for source in secondary_weights:
yield from self._get_weights_iterator(source)
def download_model(self, model_config: ModelConfig) -> None:
self._prepare_weights(
model_config.model,
model_config.revision,
fall_back_to_pt=True,
allow_patterns_overrides=None,
)
def load_weights(self, model: nn.Module, model_config: ModelConfig) -> None:
if model_config.quantization == "torchao":
quant_config = get_quant_config(model_config, self.load_config)
if (
hasattr(quant_config, "is_checkpoint_torchao_serialized")
and quant_config.is_checkpoint_torchao_serialized
and torchao_version_at_least("0.15.0")
):
self.load_config.safetensors_load_strategy = "torchao"
weights_to_load = {name for name, _ in model.named_parameters()}
loaded_weights = model.load_weights(self.get_all_weights(model_config, model))
self.counter_after_loading_weights = time.perf_counter()
logger.info_once(
"Loading weights took %.2f seconds",
self.counter_after_loading_weights - self.counter_before_loading_weights,
scope="local",
)
# We only enable strict check for non-quantized models
# that have loaded weights tracking currently.
if model_config.quantization is None and loaded_weights is not None:
weights_not_loaded = weights_to_load - loaded_weights
if weights_not_loaded:
raise ValueError(
"Following weights were not initialized from "
f"checkpoint: {weights_not_loaded}"
)

View File

@@ -0,0 +1,28 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import torch.nn as nn
from vllm.config import ModelConfig
from vllm.config.load import LoadConfig
from vllm.model_executor.model_loader.base_loader import BaseModelLoader
from vllm.model_executor.model_loader.weight_utils import initialize_dummy_weights
class DummyModelLoader(BaseModelLoader):
"""Model loader that will set model weights to random values."""
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
if load_config.model_loader_extra_config:
raise ValueError(
f"Model loader extra config is not supported for "
f"load format {load_config.load_format}"
)
def download_model(self, model_config: ModelConfig) -> None:
pass # Nothing to download
def load_weights(self, model: nn.Module, model_config: ModelConfig) -> None:
# NOTE(woosuk): For accurate performance evaluation, we assign
# random values to the weights.
initialize_dummy_weights(model)

View File

@@ -0,0 +1,371 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import os
from collections.abc import Generator
import gguf
import regex as re
import torch
import torch.nn as nn
from huggingface_hub import hf_hub_download
from transformers import AutoModelForCausalLM, AutoModelForImageTextToText
from vllm.config import ModelConfig, VllmConfig
from vllm.config.load import LoadConfig
from vllm.logger import init_logger
from vllm.model_executor.model_loader.base_loader import BaseModelLoader
from vllm.model_executor.model_loader.utils import (
initialize_model,
process_weights_after_loading,
)
from vllm.model_executor.model_loader.weight_utils import (
download_gguf,
get_gguf_extra_tensor_names,
get_gguf_weight_type_map,
gguf_quant_weights_iterator,
)
from vllm.transformers_utils.gguf_utils import detect_gguf_multimodal
from vllm.utils.torch_utils import set_default_torch_dtype
logger = init_logger(__name__)
class GGUFModelLoader(BaseModelLoader):
"""
Model loader that can load GGUF files. This is useful for loading models
that are quantized with GGUF and saved in the GGUF format. This loader
supports loading both full models and sharded models.
"""
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
if load_config.model_loader_extra_config:
raise ValueError(
f"Model loader extra config is not supported for "
f"load format {load_config.load_format}"
)
def _prepare_weights(self, model_config: ModelConfig):
model_name_or_path = model_config.model
if os.path.isfile(model_name_or_path):
return model_name_or_path
# for raw HTTPS link
if model_name_or_path.startswith(
("http://", "https://")
) and model_name_or_path.endswith(".gguf"):
return hf_hub_download(url=model_name_or_path)
# repo id/filename.gguf
if "/" in model_name_or_path and model_name_or_path.endswith(".gguf"):
repo_id, filename = model_name_or_path.rsplit("/", 1)
return hf_hub_download(repo_id=repo_id, filename=filename)
# repo_id:quant_type
elif "/" in model_name_or_path and ":" in model_name_or_path:
repo_id, quant_type = model_name_or_path.rsplit(":", 1)
return download_gguf(
repo_id,
quant_type,
cache_dir=self.load_config.download_dir,
revision=model_config.revision,
ignore_patterns=self.load_config.ignore_patterns,
)
raise ValueError(
f"Unrecognised GGUF reference: {model_name_or_path} "
"(expected local file, raw URL, <repo_id>/<filename>.gguf, "
"or <repo_id>:<quant_type>)"
)
def _get_gguf_weights_map(self, model_config: ModelConfig):
"""
GGUF uses this naming convention for their tensors from HF checkpoint:
`blk.N.BB.weight` and `blk.N.BB.bias`
where N signifies the block number of a layer, and BB signifies the
attention/mlp layer components.
See "Standardized tensor names" in
https://github.com/ggerganov/ggml/blob/master/docs/gguf.md for details.
"""
config = model_config.hf_config
# Get text config to handle both nested (multimodal) and flat
# (text-only) config structures. For multimodal models like
# Gemma3Config, this returns config.text_config. For text-only
# models, this returns config itself.
text_config = config.get_text_config()
model_type = config.model_type
is_multimodal = (
hasattr(config, "vision_config") and config.vision_config is not None
)
gguf_to_hf_name_map = {}
sideload_params: list[re.Pattern] = []
# hack: ggufs have a different name than transformers
if model_type == "cohere":
model_type = "command-r"
if model_type == "gemma3_text":
# Gemma3 models use "gemma3_text" in HuggingFace but
# "gemma3" in GGUF architecture naming
model_type = "gemma3"
if model_type in ("deepseek_v3", "deepseek_v2"):
model_type = "deepseek2"
# GGUF layer map assumes that we will have a merged expert weights
# so we need to map them manually
for idx in range(config.num_hidden_layers):
gguf_to_hf_name_map[f"blk.{idx}.exp_probs_b.bias"] = (
f"model.layers.{idx}.mlp.gate.e_score_correction_bias"
)
gguf_to_hf_name_map[f"blk.{idx}.ffn_down_exps.weight"] = (
f"model.layers.{idx}.mlp.experts.0.down_proj.weight"
)
gguf_to_hf_name_map[f"blk.{idx}.ffn_gate_exps.weight"] = (
f"model.layers.{idx}.mlp.experts.0.gate_proj.weight"
)
gguf_to_hf_name_map[f"blk.{idx}.ffn_up_exps.weight"] = (
f"model.layers.{idx}.mlp.experts.0.up_proj.weight"
)
sideload_params.append(
re.compile(
f"model\\.layers\\.{idx}"
r"\.mlp\.experts\.[0-9]+\.(gate|up|down)_proj\.weight"
)
)
if model_type in ("qwen2_moe", "qwen3_moe"):
model_type = model_type.replace("_", "")
# GGUF layer map assumes that we will have a merged expert weights
# so we need to map them manually
for idx in range(config.num_hidden_layers):
gguf_to_hf_name_map[f"blk.{idx}.ffn_down_exps.weight"] = (
f"model.layers.{idx}.mlp.experts.0.down_proj.weight"
)
gguf_to_hf_name_map[f"blk.{idx}.ffn_gate_exps.weight"] = (
f"model.layers.{idx}.mlp.experts.0.gate_proj.weight"
)
gguf_to_hf_name_map[f"blk.{idx}.ffn_up_exps.weight"] = (
f"model.layers.{idx}.mlp.experts.0.up_proj.weight"
)
sideload_params.append(
re.compile(
f"model\\.layers\\.{idx}"
r"\.mlp\.experts\.[0-9]+\.(gate|up|down)_proj\.weight"
)
)
arch = None
for key, value in gguf.MODEL_ARCH_NAMES.items():
if value == model_type:
arch = key
break
if arch is None:
raise RuntimeError(f"Unknown gguf model_type: {model_type}")
text_num_layers = text_config.num_hidden_layers
text_name_map = gguf.get_tensor_name_map(arch, text_num_layers)
if is_multimodal:
mm_proj_arch = gguf.MODEL_ARCH.MMPROJ
vision_num_layers = config.vision_config.num_hidden_layers
vision_name_map = gguf.get_tensor_name_map(mm_proj_arch, vision_num_layers)
else:
vision_name_map = None
# Create dummy model to extract parameter names
# For multimodal: use AutoModelForImageTextToText to get
# language + vision + projector params
# For text-only: use AutoModelForCausalLM to get language model params
auto_cls = (
AutoModelForImageTextToText if is_multimodal else AutoModelForCausalLM
)
with torch.device("meta"):
dummy_model = auto_cls.from_config(
config, trust_remote_code=model_config.trust_remote_code
)
state_dict = dummy_model.state_dict()
if hf_checkpoint_map := getattr(
dummy_model, "_checkpoint_conversion_mapping", None
):
def revert_hf_rename(name: str) -> str:
for original_name, hf_name in hf_checkpoint_map.items():
if hf_name in name:
name = name.replace(hf_name, original_name).lstrip("^")
return name
state_dict = {
revert_hf_rename(name): tensor for name, tensor in state_dict.items()
}
def find_hf_name_in_tensor_map(hf_name: str) -> str | None:
"""
Map HuggingFace parameter name to GGUF tensor name.
This function handles the mismatch between HF parameter naming
conventions and gguf-py's expected format:
1. Strips 'model.' prefix (common in multimodal models)
2. Converts '_weight' suffix to '.weight' (Gemma3 compatibility)
3. Searches vision_name_map for multimodal parameters
4. Falls back to text_name_map for language model parameters
Args:
hf_name: Full HuggingFace parameter name (e.g.,
'model.multi_modal_projector.mm_soft_emb_norm.weight')
Returns:
GGUF tensor name with suffix (e.g., 'mm.soft_emb_norm.weight')
or None if no mapping found
"""
# Strip 'language_model.' prefix for multimodal models - gguf-py
# tensor mappings expect parameter names without this prefix.
# Note: 'model.' prefix should be KEPT for text-only models as
# gguf-py expects it.
if hf_name.startswith("language_model."):
hf_name = hf_name[15:] # Remove 'language_model.'
# Parse parameter name and suffix
if hf_name.endswith((".weight", ".bias")):
base_name, suffix = hf_name.rsplit(".", 1)
else:
base_name, suffix = hf_name, ""
# Handle '_weight' suffix (Gemma3 naming: parameter ends with
# '_weight' instead of '.weight')
if base_name.endswith("_weight"):
base_name = base_name[:-7] # Remove '_weight'
suffix = "weight"
gguf_name = None
# Priority 1: Search vision/projector parameters for multimodal models
if vision_name_map is not None:
gguf_name = vision_name_map.get_name(base_name)
# Priority 2: Search text backbone parameters
if gguf_name is None:
gguf_name = text_name_map.get_name(base_name)
if gguf_name is None:
return None
return gguf_name + "." + suffix
# Build mapping and track unmapped parameters
unmapped_params = []
for hf_name in state_dict:
gguf_name_with_suffix = find_hf_name_in_tensor_map(hf_name)
# Track mapping success
if gguf_name_with_suffix is not None:
gguf_to_hf_name_map[gguf_name_with_suffix] = hf_name
logger.debug("Mapped GGUF %s → HF %s", gguf_name_with_suffix, hf_name)
elif hf_name not in gguf_to_hf_name_map.values():
# Parameter not in manual overrides either
unmapped_params.append(hf_name)
# All parameters (except those initialized by other means) must be mapped:
# both vision/projector and backbone
if unmapped_params:
unmapped_params = list(
filter(
lambda x: not any(re.fullmatch(p, x) for p in sideload_params),
unmapped_params,
)
)
if unmapped_params:
raise RuntimeError(
f"Failed to map GGUF parameters "
f"({len(unmapped_params)}): "
f"{unmapped_params}"
)
return gguf_to_hf_name_map
def _get_gguf_weight_type(
self,
model_config: ModelConfig,
model_name_or_path: str,
gguf_to_hf_name_map: dict[str, str],
) -> dict[str, str]:
weight_type_map = get_gguf_weight_type_map(
model_name_or_path, gguf_to_hf_name_map
)
is_multimodal = hasattr(model_config.hf_config, "vision_config")
if is_multimodal:
mmproj_file = detect_gguf_multimodal(model_name_or_path)
assert mmproj_file is not None, (
"Could not find mm_proj file for multimodal GGUF model"
)
logger.info("Loading extra mm_proj weights from %s...", mmproj_file)
mm_proj_weight_type_map = get_gguf_weight_type_map(
mmproj_file, gguf_to_hf_name_map
)
weight_type_map.update(mm_proj_weight_type_map)
return weight_type_map
def _get_weights_iterator(
self,
model_config: ModelConfig,
model_name_or_path: str,
gguf_to_hf_name_map: dict[str, str],
) -> Generator[tuple[str, torch.Tensor], None, None]:
"""
Iterate over GGUF model weights, loading from both main model file and
mmproj.gguf for multimodal Gemma3 models.
For Gemma3 multimodal GGUF models:
- Main file (gemma-3-*.gguf): Language model weights (model.*)
- mmproj file (mmproj*.gguf): Vision tower + projector weights (v.*, mm.*)
Yields:
Tuples of (parameter_name, tensor) for all model weights
"""
hf_config = model_config.hf_config
is_multimodal = hasattr(hf_config, "vision_config")
if is_multimodal:
# Load mm_proj (mm_encoder + projector) for multimodal weights
mmproj_file = detect_gguf_multimodal(model_name_or_path)
assert mmproj_file is not None, (
"Could not find mm_proj file for multimodal GGUF model"
)
yield from gguf_quant_weights_iterator(mmproj_file, gguf_to_hf_name_map)
yield from gguf_quant_weights_iterator(model_name_or_path, gguf_to_hf_name_map)
def download_model(self, model_config: ModelConfig) -> None:
self._prepare_weights(model_config)
def load_weights(self, model: nn.Module, model_config: ModelConfig) -> None:
local_model_path = self._prepare_weights(model_config)
gguf_weights_map = self._get_gguf_weights_map(model_config)
model.load_weights(
self._get_weights_iterator(model_config, local_model_path, gguf_weights_map)
)
def load_model(
self, vllm_config: VllmConfig, model_config: ModelConfig
) -> nn.Module:
device_config = vllm_config.device_config
local_model_path = self._prepare_weights(model_config)
gguf_weights_map = self._get_gguf_weights_map(model_config)
# we can only know if tie word embeddings after mapping weights
if "lm_head.weight" in get_gguf_extra_tensor_names(
local_model_path, gguf_weights_map
):
model_config.hf_config.update({"tie_word_embeddings": True})
weight_type_map = self._get_gguf_weight_type(
model_config, local_model_path, gguf_weights_map
)
# filter out unquantized modules to skip
unquant_names = [
name.removesuffix(".weight")
for name, weight_type in weight_type_map.items()
if weight_type in ("F32", "F16", "BF16") and name.endswith(".weight")
]
logger.debug(
"GGUF unquantized modules: %s",
unquant_names,
)
vllm_config.quant_config.unquantized_modules.extend(unquant_names)
target_device = torch.device(device_config.device)
with set_default_torch_dtype(model_config.dtype):
with target_device:
model = initialize_model(vllm_config=vllm_config)
self.load_weights(model, model_config)
process_weights_after_loading(model, model_config, target_device)
return model

View File

@@ -1,362 +0,0 @@
# ruff: noqa: SIM117
import copy
import glob
import os
from abc import ABC, abstractmethod
from typing import Any, Dict, Generator, List, Optional, Tuple, Type
import huggingface_hub
import torch
from torch import nn
from vllm.config import (DeviceConfig, LoadConfig, LoadFormat, LoRAConfig,
ModelConfig, ParallelConfig, SchedulerConfig,
VisionLanguageConfig)
from vllm.envs import VLLM_USE_MODELSCOPE
from vllm.logger import init_logger
from vllm.model_executor.layers.quantization.base_config import (
QuantizationConfig)
from vllm.model_executor.model_loader.tensorizer import (
TensorizerConfig, is_vllm_serialized_tensorizer, load_with_tensorizer,
tensorizer_weights_iterator)
from vllm.model_executor.model_loader.utils import (get_model_architecture,
set_default_torch_dtype)
from vllm.model_executor.model_loader.weight_utils import (
download_weights_from_hf, filter_files_not_needed_for_inference,
get_quant_config, initialize_dummy_weights, np_cache_weights_iterator,
pt_weights_iterator, safetensors_weights_iterator)
from vllm.model_executor.models.llava import LlavaForConditionalGeneration
_VISION_MODEL_CLASSES = [
LlavaForConditionalGeneration,
]
logger = init_logger(__name__)
def _get_quantization_config(
model_config: ModelConfig,
load_config: LoadConfig) -> Optional[QuantizationConfig]:
"""Get the quantization config."""
if model_config.quantization is not None:
quant_config = get_quant_config(model_config, load_config)
capability = torch.cuda.get_device_capability()
capability = capability[0] * 10 + capability[1]
if capability < quant_config.get_min_capability():
raise ValueError(
f"The quantization method {model_config.quantization} is not "
"supported for the current GPU. "
f"Minimum capability: {quant_config.get_min_capability()}. "
f"Current capability: {capability}.")
supported_dtypes = quant_config.get_supported_act_dtypes()
if model_config.dtype not in supported_dtypes:
raise ValueError(
f"{model_config.dtype} is not supported for quantization "
f"method {model_config.quantization}. Supported dtypes: "
f"{supported_dtypes}")
return quant_config
return None
def _get_model_initialization_kwargs(
model_class: Type[nn.Module], lora_config: Optional[LoRAConfig],
vision_language_config: Optional[VisionLanguageConfig]
) -> Dict[str, Any]:
"""Get extra kwargs for model initialization."""
extra_kwargs = {}
if hasattr(model_class, "supported_lora_modules"):
extra_kwargs["lora_config"] = lora_config
elif lora_config:
raise ValueError(
f"Model {model_class.__name__} does not support LoRA, "
"but LoRA is enabled. Support for this model may "
"be added in the future. If this is important to you, "
"please open an issue on github.")
elif model_class in _VISION_MODEL_CLASSES:
extra_kwargs["vision_language_config"] = vision_language_config
return extra_kwargs
def _initialize_model(
model_config: ModelConfig, load_config: LoadConfig,
lora_config: Optional[LoRAConfig],
vision_language_config: Optional[VisionLanguageConfig]) -> nn.Module:
"""Initialize a model with the given configurations."""
model_class = get_model_architecture(model_config)[0]
quant_config = _get_quantization_config(model_config, load_config)
return model_class(config=model_config.hf_config,
quant_config=quant_config,
**_get_model_initialization_kwargs(
model_class, lora_config, vision_language_config))
class BaseModelLoader(ABC):
"""Base class for model loaders."""
def __init__(self, load_config: LoadConfig):
self.load_config = load_config
@abstractmethod
def load_model(self, *, model_config: ModelConfig,
device_config: DeviceConfig,
lora_config: Optional[LoRAConfig],
vision_language_config: Optional[VisionLanguageConfig],
parallel_config: ParallelConfig,
scheduler_config: SchedulerConfig) -> nn.Module:
"""Load a model with the given configurations."""
...
class DefaultModelLoader(BaseModelLoader):
"""Model loader that can load different file types from disk."""
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
if load_config.model_loader_extra_config:
raise ValueError(f"Model loader extra config is not supported for "
f"load format {load_config.load_format}")
def _maybe_download_from_modelscope(
self, model: str, revision: Optional[str]) -> Optional[str]:
"""Download model from ModelScope hub if VLLM_USE_MODELSCOPE is True.
Returns the path to the downloaded model, or None if the model is not
downloaded from ModelScope."""
if VLLM_USE_MODELSCOPE:
# download model from ModelScope hub,
# lazy import so that modelscope is not required for normal use.
# pylint: disable=C.
from modelscope.hub.snapshot_download import snapshot_download
if not os.path.exists(model):
model_path = snapshot_download(
model_id=model,
cache_dir=self.load_config.download_dir,
local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE,
revision=revision,
)
else:
model_path = model
return model_path
return None
def _prepare_weights(self, model_name_or_path: str,
revision: Optional[str],
fall_back_to_pt: bool) -> Tuple[str, List[str], bool]:
"""Prepare weights for the model.
If the model is not local, it will be downloaded."""
model_name_or_path = self._maybe_download_from_modelscope(
model_name_or_path, revision) or model_name_or_path
is_local = os.path.isdir(model_name_or_path)
load_format = self.load_config.load_format
use_safetensors = False
# Some quantized models use .pt files for storing the weights.
if load_format == LoadFormat.AUTO:
allow_patterns = ["*.safetensors", "*.bin"]
elif load_format == LoadFormat.SAFETENSORS:
use_safetensors = True
allow_patterns = ["*.safetensors"]
elif load_format == LoadFormat.PT:
allow_patterns = ["*.pt"]
elif load_format == LoadFormat.NPCACHE:
allow_patterns = ["*.bin"]
else:
raise ValueError(f"Unknown load_format: {load_format}")
if fall_back_to_pt:
allow_patterns += ["*.pt"]
if not is_local:
hf_folder = download_weights_from_hf(model_name_or_path,
self.load_config.download_dir,
allow_patterns, revision)
else:
hf_folder = model_name_or_path
hf_weights_files: List[str] = []
for pattern in allow_patterns:
hf_weights_files += glob.glob(os.path.join(hf_folder, pattern))
if len(hf_weights_files) > 0:
if pattern == "*.safetensors":
use_safetensors = True
break
if not use_safetensors:
hf_weights_files = filter_files_not_needed_for_inference(
hf_weights_files)
if len(hf_weights_files) == 0:
raise RuntimeError(
f"Cannot find any model weights with `{model_name_or_path}`")
return hf_folder, hf_weights_files, use_safetensors
def _get_weights_iterator(
self, model_name_or_path: str, revision: Optional[str],
fall_back_to_pt: bool
) -> Generator[Tuple[str, torch.Tensor], None, None]:
"""Get an iterator for the model weights based on the load format."""
hf_folder, hf_weights_files, use_safetensors = self._prepare_weights(
model_name_or_path, revision, fall_back_to_pt)
if self.load_config.load_format == LoadFormat.NPCACHE:
# Currently np_cache only support *.bin checkpoints
assert use_safetensors is False
return np_cache_weights_iterator(model_name_or_path,
self.load_config.download_dir,
hf_folder, hf_weights_files)
if use_safetensors:
return safetensors_weights_iterator(hf_weights_files)
return pt_weights_iterator(hf_weights_files)
def load_model(self, *, model_config: ModelConfig,
device_config: DeviceConfig,
lora_config: Optional[LoRAConfig],
vision_language_config: Optional[VisionLanguageConfig],
parallel_config: ParallelConfig,
scheduler_config: SchedulerConfig) -> nn.Module:
with set_default_torch_dtype(model_config.dtype):
with torch.device(device_config.device):
model = _initialize_model(model_config, self.load_config,
lora_config, vision_language_config)
model.load_weights(
self._get_weights_iterator(model_config.model,
model_config.revision,
fall_back_to_pt=getattr(
model,
"fall_back_to_pt_during_load",
True)), )
for _, module in model.named_modules():
quant_method = getattr(module, "quant_method", None)
if quant_method is not None:
quant_method.process_weights_after_loading(module)
# FIXME: Remove this after Mixtral is updated
# to use quant_method.
if hasattr(module, "process_weights_after_loading"):
module.process_weights_after_loading()
return model.eval()
class DummyModelLoader(BaseModelLoader):
"""Model loader that will set model weights to random values."""
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
if load_config.model_loader_extra_config:
raise ValueError(f"Model loader extra config is not supported for "
f"load format {load_config.load_format}")
def load_model(self, *, model_config: ModelConfig,
device_config: DeviceConfig,
lora_config: Optional[LoRAConfig],
vision_language_config: Optional[VisionLanguageConfig],
parallel_config: ParallelConfig,
scheduler_config: SchedulerConfig) -> nn.Module:
with set_default_torch_dtype(model_config.dtype):
with torch.device(device_config.device):
model = _initialize_model(model_config, self.load_config,
lora_config, vision_language_config)
# NOTE(woosuk): For accurate performance evaluation, we assign
# random values to the weights.
initialize_dummy_weights(model)
return model.eval()
class TensorizerLoader(BaseModelLoader):
"""Model loader using CoreWeave's tensorizer library."""
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
if isinstance(load_config.model_loader_extra_config, TensorizerConfig):
self.tensorizer_config = load_config.model_loader_extra_config
else:
self.tensorizer_config = TensorizerConfig(
**load_config.model_loader_extra_config)
def _verify_config(self, model_config: ModelConfig,
parallel_config: ParallelConfig):
self.tensorizer_config.verify_with_model_config(model_config)
self.tensorizer_config.verify_with_parallel_config(parallel_config)
def _get_weights_iterator(
self) -> Generator[Tuple[str, torch.Tensor], None, None]:
tensorizer_args = self.tensorizer_config._construct_tensorizer_args()
return tensorizer_weights_iterator(tensorizer_args)
def _load_model_unserialized(
self, model_config: ModelConfig, device_config: DeviceConfig,
lora_config: Optional[LoRAConfig],
vision_language_config: Optional[VisionLanguageConfig]
) -> nn.Module:
"""Load an unserialized model with tensorizer.
Unserialized here means "not serialized with tensorizer". This
should still be faster than default HuggingFace loading, but will
be slower than loading a tensorizer-serialized model.
"""
with set_default_torch_dtype(model_config.dtype):
with torch.device(device_config.device):
model = _initialize_model(model_config, self.load_config,
lora_config, vision_language_config)
model.load_weights(self._get_weights_iterator())
return model.eval()
def _load_model_serialized(
self, model_config: ModelConfig, device_config: DeviceConfig,
lora_config: Optional[LoRAConfig],
vision_language_config: Optional[VisionLanguageConfig]
) -> nn.Module:
"""Load a serialized model with tensorizer.
See the examples/tensorize_vllm_model.py example "
script for serializing vLLM models."""
with set_default_torch_dtype(model_config.dtype):
with torch.device(device_config.device):
model_class = get_model_architecture(model_config)[0]
quant_config = _get_quantization_config(
model_config, self.load_config)
extra_kwargs = _get_model_initialization_kwargs(
model_class, lora_config, vision_language_config)
extra_kwargs["quant_config"] = quant_config
tensorizer_config = copy.copy(self.tensorizer_config)
tensorizer_config.model_class = model_class
tensorizer_config.hf_config = model_config.hf_config
tensorizer_config.dtype = model_config.dtype
model = load_with_tensorizer(tensorizer_config, **extra_kwargs)
return model.eval()
def load_model(self, *, model_config: ModelConfig,
device_config: DeviceConfig,
lora_config: Optional[LoRAConfig],
vision_language_config: Optional[VisionLanguageConfig],
parallel_config: ParallelConfig,
scheduler_config: SchedulerConfig) -> nn.Module:
self._verify_config(model_config, parallel_config)
if is_vllm_serialized_tensorizer(self.tensorizer_config):
return self._load_model_serialized(model_config, device_config,
lora_config,
vision_language_config)
return self._load_model_unserialized(model_config, device_config,
lora_config,
vision_language_config)
def get_model_loader(load_config: LoadConfig) -> BaseModelLoader:
"""Get a model loader based on the load format."""
if isinstance(load_config.load_format, type):
return load_config.load_format(load_config)
if load_config.load_format == LoadFormat.DUMMY:
return DummyModelLoader(load_config)
if load_config.load_format == LoadFormat.TENSORIZER:
return TensorizerLoader(load_config)
return DefaultModelLoader(load_config)

View File

@@ -1,136 +0,0 @@
"""Utilities for selecting and loading neuron models."""
import importlib
import os
from typing import Dict, Optional, Tuple
import torch
import torch.nn as nn
import transformers
from transformers import PretrainedConfig
from vllm.config import ModelConfig, ParallelConfig, SchedulerConfig
from vllm.model_executor.layers.logits_processor import LogitsProcessor
from vllm.model_executor.layers.sampler import Sampler
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.sequence import SamplerOutput
TORCH_DTYPE_TO_NEURON_AMP = {
"auto": "f32",
"half": "f16",
"float16": "f16",
"bfloat16": "bf16",
"float": "f32",
"float32": "f32",
torch.float16: "f16",
torch.bfloat16: "bf16",
torch.float32: "f32",
}
# Models supported by Neuron.
_NEURON_SUPPORTED_MODELS: Dict[str, Tuple[str, str, str]] = {
"LlamaForCausalLM": ("transformers_neuronx.llama.model",
"LlamaForSampling", "LlamaForCausalLM"),
"MistralForCausalLM": ("transformers_neuronx.mistral.model",
"MistralForSampling", "MistralForCausalLM")
}
class NeuronCasualLM(nn.Module):
def __init__(
self,
config: PretrainedConfig,
) -> None:
super().__init__()
self.config = config
self.logits_processor = LogitsProcessor(config.vocab_size,
logits_as_input=True)
self.sampler = Sampler()
# Lazy initialized
self.model: nn.Module
def forward(
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
input_block_ids: torch.Tensor,
) -> torch.Tensor:
logits = self.model(input_ids,
cache_ids=positions,
start_ids=input_block_ids)
return logits
def compute_logits(self, hidden_states: torch.Tensor,
sampling_metadata: SamplingMetadata) -> torch.Tensor:
logits = self.logits_processor(None, hidden_states, sampling_metadata)
return logits
def sample(
self,
logits: torch.Tensor,
sampling_metadata: SamplingMetadata,
) -> Optional[SamplerOutput]:
next_tokens = self.sampler(logits, sampling_metadata)
return next_tokens
def load_weights(self, model_name_or_path: str, **kwargs):
arch = _get_model_architecture(self.config)
neuronx_module_path, neuronx_model_cls_name, hf_model_cls_name = (
_NEURON_SUPPORTED_MODELS[arch])
neuronx_module = importlib.import_module(neuronx_module_path)
neuronx_model_cls = getattr(neuronx_module, neuronx_model_cls_name)
split_model_dir = f"{model_name_or_path}-split"
if os.path.isdir(os.path.join(model_name_or_path,
"pytorch_model.bin")):
split_model_dir = model_name_or_path
elif not os.path.exists(f"{model_name_or_path}-split"):
hf_model_cls = getattr(transformers, hf_model_cls_name)
from transformers_neuronx.module import save_pretrained_split
hf_model = hf_model_cls.from_pretrained(model_name_or_path,
low_cpu_mem_usage=True)
save_pretrained_split(hf_model, f"{model_name_or_path}-split")
self.model = neuronx_model_cls.from_pretrained(split_model_dir,
**kwargs)
self.model.to_neuron()
def _get_model_architecture(config: PretrainedConfig) -> str:
architectures = getattr(config, "architectures", [])
for arch in architectures:
if arch in _NEURON_SUPPORTED_MODELS:
return arch
raise ValueError(
f"Model architectures {architectures} are not supported on Neuron "
f"for now. Supported architectures: "
f"{list(_NEURON_SUPPORTED_MODELS.keys())}")
def get_neuron_model(model_config: ModelConfig,
parallel_config: ParallelConfig,
scheduler_config: SchedulerConfig) -> nn.Module:
from transformers_neuronx.config import (ContinuousBatchingConfig,
NeuronConfig)
# Create a model instance.
model = NeuronCasualLM(model_config.hf_config)
continuous_batching_config = ContinuousBatchingConfig(
batch_size_for_shared_caches=scheduler_config.max_num_seqs)
neuron_config = NeuronConfig(
continuous_batching=continuous_batching_config)
# Load the weights from the cached or downloaded files.
model.load_weights(
model_config.model,
tp_degree=parallel_config.tensor_parallel_size,
amp=TORCH_DTYPE_TO_NEURON_AMP[model_config.dtype],
neuron_config=neuron_config,
context_length_estimate=[scheduler_config.max_model_len],
n_positions=[scheduler_config.max_model_len],
batch_size=scheduler_config.max_num_seqs)
return model.eval()

View File

@@ -0,0 +1,275 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import types
from collections.abc import Iterable
import torch
from torch import nn
from vllm.config import ModelConfig
from vllm.logger import init_logger
from vllm.model_executor.model_loader.utils import process_weights_after_loading
logger = init_logger(__name__)
# Notes for Online Quantization
# In terms of state of checkpoints, quantization config and their
# correspondance to online quantization:
# | Use Case | Checkpoints | model_config.quantization |
# | no quant | high precision | None |
# | offline quant | quantized | fp8, torchao etc. |
# | online quant | high precision | torchao etc. |
#
# The process for loading non-quantized checkpoint
# 1. load non-quantized weights (load_weights)
# 2. do any additional post processing (process_weights_after_loading)
#
# The process for loading offline quantized checkpoint
# 1. load offline-quantized weights (load_weights)
# 2. do any additional post processing (process_weights_after_loading)
# The process for unquantized model reloading
# (repeated run in RL training loop)
# first run
# UI1. load_weights: load bfloat16 weights
# UI2. process_weights_after_loading: any additional post processing
# subsequent run
# UC1: load_weights: load bfloat16 weights
# (shouldn't be any issues since we didn't change any attributes
# of the weights)
# UC2: process_weights_after_loading: any additional post processing
# The process for weight reloading with online quantization
# (repeated run in RL training loop)
# first run
# I1. load_weights: load bfloat16 weights
# I2. process_weights_after_loading:
# record weight metadata and attributes for R1 and R2
# quantize weights to fp8
# subsequent run
# (beginning model weight is in fp8)
# load_weights:
# R1. restore bfloat16 model weight metadata
# R2. restore the model weight attributes
# R3. reload bfloat16 weights
# R4. quantize weights (by calling process_weights_after_loading),
# also set `process_weights_after_loading_already_called` to
# True to stop it from running again
# R5. (workaround for cudagraph), we restore the weight params to original quantized
# weights params, and use original_weight_param.copy_(updated_weight_param) so that
# the weight update work well with cudagraph
# process_weights_after_loading (if called):
# this will be skipped since it's already ran in
# load_weights
def maybe_save_metadata_and_attributes_for_weight_reloading(
model: nn.Module, model_config: ModelConfig
):
# following is to support on the fly quantization, currently only supported
# for torchao
if model_config.quantization != "torchao":
return
from vllm.model_executor.model_loader.weight_utils import get_quant_config
quant_config = get_quant_config(model_config, None)
# If checkpoint is already torchao serialized, this means it's
# pre-quantized quantization case, we'll skip saving the metadata
# Otherwise, this is Step I2 of initialization steps of
# online quantization
# This step record the weights metadata and weight attributes so we can
# restore the bfloat16 model weights during the relad step (R1 and R2)
# see Notes in online_quantization.py for more details
if not (
hasattr(quant_config, "is_checkpoint_torchao_serialized")
and not quant_config.is_checkpoint_torchao_serialized
):
return
# This is the I2 step of online quantiztion that saves
# metadata and attributes of weights so they can be used in R1 and
# R2 step, note that we only save these during initialization
# Includes two things
# 1. save floating point metadata (shape, dtype, device) for init
# 2. save weight attributes, e.g. `output_dim`, `weight_loader` for init
if getattr(model, "weight_metadata_and_attr_saved", False):
return
# save the dtype, shape and device for model parameter, used for
# restoring the model high precision parameters before
# reloading the weights
assert not hasattr(model, "original_weights_rebuild_keys")
model.original_weights_rebuild_keys = {}
for name, p in model.named_parameters():
model.original_weights_rebuild_keys[name] = {
"shape": p.shape,
"dtype": p.dtype,
"device": p.device,
}
# record the weight attributes (loader functions etc.)
# so these can be recovered later when we reload the weights
# structure: {"weight_name": {"weight_attr_key": attr}}
assert not hasattr(model, "recorded_weight_attr")
model.recorded_weight_attr = {}
for name, param in model.named_parameters():
model.recorded_weight_attr[name] = {}
for key in param.__dict__:
if hasattr(param, key):
attr = getattr(param, key)
if not callable(attr):
model.recorded_weight_attr[name][key] = attr
elif hasattr(attr, "__self__") and param is attr.__self__:
# if attr is a bonded method for an instance, and
# attr.__self__ points to the instance (param)
# we'll record the underlying function object
model.recorded_weight_attr[name][key] = attr.__func__
else:
model.recorded_weight_attr[name][key] = attr
# mark the metadata and attributes saved so we don't run it again
model._model_config = model_config
model.weight_metadata_and_attr_saved = True
def _bond_method_to_cls(func, obj):
if hasattr(func, "__self__") or not callable(func):
# If the function is already bound to an instance, return it as is
return func
else:
return types.MethodType(func, obj)
def support_quantized_model_reload_from_hp_weights(original_load_weights):
"""Decorator for `load_weights` method for AutoWeightsLoader.load_weights to support
reloading high precision (bfloat16/float16/float32) weight for an already quantized
model, this involves restoring the weights to a high precision weights and
then online quantize the weights
"""
# online quantization, right now only enabled for
# torchao
# R1, R2, R3, R4, R5 in the Notes
def patched_model_load_weights(
auto_weight_loader, weights: Iterable[tuple[str, torch.Tensor]], *, mapper=None
) -> set[str]:
model = auto_weight_loader.module
offline_quantization_or_first_run_of_online_quantization = not getattr(
model, "weight_metadata_and_attr_saved", False
)
# if we don't have `model.weight_metadata_and_attr_saved` defined and
# set to True, it means that this is either offline quantization case
# or the first run of online quantization
# see Notes in this file for more details
if offline_quantization_or_first_run_of_online_quantization:
# case 1: offline quantized checkpoint
# case 2: Step I1 first run of weight loading with
# online quantization
return original_load_weights(auto_weight_loader, weights, mapper=mapper)
model_config = model._model_config
# TODO: Add fp8 support
assert model_config.quantization == "torchao", (
"online quantization is only enabled for torchao currently"
)
# TODO: use create_weights to restore the weights to original state
# Step R1: First restore the quantized weights to original bfloat16
# weights, with original metadata (shape, dtype, device)
# and attributes, so that bfloat16 weights can be loaded properly
# TODO: maybe set remove_duplicate to True?
original_quantized_weight_dict = dict(
model.named_parameters(remove_duplicate=False)
)
named_modules = dict(model.named_modules(remove_duplicate=False))
model_device = None
for name, d in model.original_weights_rebuild_keys.items():
_shape = d["shape"]
_dtype = d["dtype"]
_device = d["device"]
if model_device is not None:
assert model_device == _device, (
"Expecting all weights "
"to be in the same device for now, got both: "
f"{model_device} and {_device}"
)
else:
model_device = _device
if name in original_quantized_weight_dict:
module_name, weight_name = name.rsplit(".", 1)
module = named_modules[module_name]
setattr(
module,
weight_name,
torch.nn.Parameter(
torch.empty(_shape, dtype=_dtype, device=_device),
requires_grad=False,
),
)
# Step R2: recover the weight attributes to the state before first loading
# recorded_weight_attr is
# {"weight_name": {"weight_attr_key": attr}}
# e.g.
# {
# {
# "layer.0.weight": {
# "weight_loader": weight_loader_function_object,
# "input_dim": 0, ...
# },
# "layer.1.weight": ...,
# }
# }
for full_weight_name, weight_attr_dict in model.recorded_weight_attr.items():
for attr_name, attr in weight_attr_dict.items():
module_name, weight_name = full_weight_name.rsplit(".", 1)
module = named_modules[module_name]
weight = getattr(module, weight_name)
if not hasattr(weight, attr_name):
setattr(weight, attr_name, _bond_method_to_cls(attr, weight))
# Step R3: reload bfloat16 / high precision weights
updated_params = original_load_weights(
auto_weight_loader, weights, mapper=mapper
)
# Step R4: online quantize the weights
# manually process weights after loading
model.process_weights_after_loading_already_called = False
if model_device is not None:
process_weights_after_loading(model, model_config, model_device)
else:
logger.warning_once(
"model_device is None, skip calling process_weights_after_loading"
)
# Step R5 (workaround for cudagraph): restore the original quantized weights
# and do a copy_ of the currents weights to the original weights
updated_quantized_weights = dict(model.named_parameters(remove_duplicate=False))
for name in model.original_weights_rebuild_keys:
if name in original_quantized_weight_dict:
original_quantized_weight = original_quantized_weight_dict[name]
updated_quantized_weight = updated_quantized_weights[name]
module_name, weight_name = name.rsplit(".", 1)
module = named_modules[module_name]
setattr(module, weight_name, original_quantized_weight)
with torch.no_grad():
original_quantized_weight.copy_(updated_quantized_weight)
del original_quantized_weight_dict
del named_modules
del updated_quantized_weight
model.process_weights_after_loading_already_called = True
return updated_params
return patched_model_load_weights

View File

@@ -0,0 +1,116 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
# ruff: noqa: SIM117
import os
from collections.abc import Generator
import torch
from torch import nn
from transformers.utils import SAFE_WEIGHTS_INDEX_NAME
from vllm.config import ModelConfig
from vllm.config.load import LoadConfig
from vllm.model_executor.model_loader.base_loader import BaseModelLoader
from vllm.model_executor.model_loader.weight_utils import (
download_safetensors_index_file_from_hf,
download_weights_from_hf,
runai_safetensors_weights_iterator,
)
from vllm.transformers_utils.runai_utils import is_runai_obj_uri, list_safetensors
class RunaiModelStreamerLoader(BaseModelLoader):
"""
Model loader that can load safetensors
files from local FS or S3 bucket.
"""
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
self._is_distributed = False
if load_config.model_loader_extra_config:
extra_config = load_config.model_loader_extra_config
if "distributed" in extra_config and isinstance(
extra_config.get("distributed"), bool
):
self._is_distributed = extra_config.get("distributed")
if "concurrency" in extra_config and isinstance(
extra_config.get("concurrency"), int
):
os.environ["RUNAI_STREAMER_CONCURRENCY"] = str(
extra_config.get("concurrency")
)
if "memory_limit" in extra_config and isinstance(
extra_config.get("memory_limit"), int
):
os.environ["RUNAI_STREAMER_MEMORY_LIMIT"] = str(
extra_config.get("memory_limit")
)
runai_streamer_s3_endpoint = os.getenv("RUNAI_STREAMER_S3_ENDPOINT")
aws_endpoint_url = os.getenv("AWS_ENDPOINT_URL")
if runai_streamer_s3_endpoint is None and aws_endpoint_url is not None:
os.environ["RUNAI_STREAMER_S3_ENDPOINT"] = aws_endpoint_url
def _prepare_weights(
self, model_name_or_path: str, revision: str | None
) -> list[str]:
"""Prepare weights for the model.
If the model is not local, it will be downloaded."""
is_object_storage_path = is_runai_obj_uri(model_name_or_path)
is_local = os.path.isdir(model_name_or_path)
safetensors_pattern = "*.safetensors"
index_file = SAFE_WEIGHTS_INDEX_NAME
hf_folder = (
model_name_or_path
if (is_local or is_object_storage_path)
else download_weights_from_hf(
model_name_or_path,
self.load_config.download_dir,
[safetensors_pattern],
revision,
ignore_patterns=self.load_config.ignore_patterns,
)
)
hf_weights_files = list_safetensors(path=hf_folder)
if not is_local and not is_object_storage_path:
download_safetensors_index_file_from_hf(
model_name_or_path, index_file, self.load_config.download_dir, revision
)
if not hf_weights_files:
raise RuntimeError(
f"Cannot find any safetensors model weights with `{model_name_or_path}`"
)
return hf_weights_files
def _get_weights_iterator(
self, model_or_path: str, revision: str
) -> Generator[tuple[str, torch.Tensor], None, None]:
"""Get an iterator for the model weights based on the load format."""
hf_weights_files = self._prepare_weights(model_or_path, revision)
return runai_safetensors_weights_iterator(
hf_weights_files, self.load_config.use_tqdm_on_load, self._is_distributed
)
def download_model(self, model_config: ModelConfig) -> None:
"""Download model if necessary"""
self._prepare_weights(model_config.model, model_config.revision)
def load_weights(self, model: nn.Module, model_config: ModelConfig) -> None:
"""Load weights into a model."""
model_weights = model_config.model
if hasattr(model_config, "model_weights"):
model_weights = model_config.model_weights
model.load_weights(
self._get_weights_iterator(model_weights, model_config.revision)
)

View File

@@ -0,0 +1,214 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import collections
import glob
import os
import time
from collections.abc import Generator
from typing import Any
import torch
from torch import nn
from vllm.config import ModelConfig
from vllm.config.load import LoadConfig
from vllm.logger import init_logger
from vllm.model_executor.model_loader.base_loader import BaseModelLoader
from vllm.model_executor.model_loader.weight_utils import (
download_weights_from_hf,
runai_safetensors_weights_iterator,
)
from vllm.transformers_utils.s3_utils import glob as s3_glob
from vllm.transformers_utils.utils import is_s3
logger = init_logger(__name__)
class ShardedStateLoader(BaseModelLoader):
"""
Model loader that directly loads each worker's model state dict, which
enables a fast load path for large tensor-parallel models where each worker
only needs to read its own shard rather than the entire checkpoint. See
`examples/offline_inference/save_sharded_state.py` for creating a sharded
checkpoint.
"""
DEFAULT_PATTERN = "model-rank-{rank}-part-{part}.safetensors"
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
extra_config = (
{}
if load_config.model_loader_extra_config is None
else load_config.model_loader_extra_config.copy()
)
self.pattern = extra_config.pop("pattern", self.DEFAULT_PATTERN)
if extra_config:
raise ValueError(
f"Unexpected extra config keys for load format "
f"{load_config.load_format}: "
f"{load_config.model_loader_extra_config.keys()}"
)
@staticmethod
def _filter_subtensors(
tensors: dict[str, torch.Tensor],
) -> dict[str, torch.Tensor]:
"""
Filter out all tensors that share the same memory or a subset of the
memory of another tensor.
"""
same_storage_groups: dict[Any, list[tuple[str, torch.Tensor]]] = (
collections.defaultdict(list)
)
for key, tensor in tensors.items():
if tensor.numel():
ptr = tensor.untyped_storage().data_ptr()
same_storage_groups[tensor.device, ptr].append((key, tensor))
def get_end_ptr(tensor: torch.Tensor) -> int:
return tensor.view(-1)[-1].data_ptr() + tensor.element_size()
result: dict[str, torch.Tensor] = {}
for group in same_storage_groups.values():
for k, t in group:
a, b = t.data_ptr(), get_end_ptr(t)
for k2, t2 in group:
if not t2.is_contiguous():
continue
a2, b2 = t2.data_ptr(), get_end_ptr(t2)
if a < a2 or b2 < b:
continue
if a2 < a or b < b2 or not t.is_contiguous():
break # t2 covers strictly more memory than t.
if k2 < k:
# Same tensors, keep the one with the smaller key.
break
else:
result[k] = t
return result
def _prepare_weights(self, model_name_or_path: str, revision: str | None):
if is_s3(model_name_or_path) or os.path.isdir(model_name_or_path):
return model_name_or_path
else:
allow_patterns = ["*.safetensors"]
return download_weights_from_hf(
model_name_or_path,
self.load_config.download_dir,
allow_patterns,
revision,
ignore_patterns=self.load_config.ignore_patterns,
)
def download_model(self, model_config: ModelConfig) -> None:
self._prepare_weights(model_config.model, model_config.revision)
def load_weights(self, model: nn.Module, model_config: ModelConfig) -> None:
from vllm.distributed import get_tensor_model_parallel_rank
model_weights = model_config.model
if hasattr(model_config, "model_weights"):
model_weights = model_config.model_weights
local_model_path = model_weights
rank = get_tensor_model_parallel_rank()
pattern = os.path.join(
local_model_path,
self.pattern.format(rank=rank, part="*"),
)
filepaths = []
if is_s3(local_model_path):
file_pattern = f"*{self.pattern.format(rank=rank, part='*')}"
filepaths = s3_glob(path=local_model_path, allow_pattern=[file_pattern])
else:
filepaths = glob.glob(pattern)
if not filepaths:
# TODO: support un-sharded checkpoints too
raise ValueError(
f"Could not find checkpoint files '{pattern}', only "
f"pre-sharded checkpoints are currently supported!"
)
state_dict = self._filter_subtensors(model.state_dict())
counter_before_loading_weights = time.perf_counter()
for key, tensor in self.iterate_over_files(filepaths):
# If loading with LoRA enabled, additional padding may
# be added to certain parameters. We only load into a
# narrowed view of the parameter data.
param_data = state_dict[key].data
param_shape = state_dict[key].shape
for dim, size in enumerate(tensor.shape):
if size < param_shape[dim]:
param_data = param_data.narrow(dim, 0, size)
if tensor.shape != param_shape:
logger.warning(
"loading tensor of shape %s into parameter '%s' of shape %s",
tensor.shape,
key,
param_shape,
)
param_data.copy_(tensor)
state_dict.pop(key)
counter_after_loading_weights = time.perf_counter()
logger.info_once(
"Loading weights took %.2f seconds",
counter_after_loading_weights - counter_before_loading_weights,
scope="local",
)
if state_dict:
raise ValueError(f"Missing keys {tuple(state_dict)} in loaded state!")
def iterate_over_files(
self, paths
) -> Generator[tuple[str, torch.Tensor], None, None]:
if self.load_config.load_format == "runai_streamer_sharded":
yield from runai_safetensors_weights_iterator(paths, True)
else:
from safetensors.torch import safe_open
for path in paths:
with safe_open(path, framework="pt") as f:
for key in f.keys(): # noqa: SIM118
tensor = f.get_tensor(key)
yield key, tensor
@staticmethod
def save_model(
model: torch.nn.Module,
path: str,
pattern: str | None = None,
max_size: int | None = None,
) -> None:
from safetensors.torch import save_file
from vllm.distributed import get_tensor_model_parallel_rank
if pattern is None:
pattern = ShardedStateLoader.DEFAULT_PATTERN
rank = get_tensor_model_parallel_rank()
part_idx = 0
total_size = 0
state_dict = ShardedStateLoader._filter_subtensors(model.state_dict())
state_dict_part: dict[str, torch.Tensor] = {}
for key, tensor in state_dict.items():
param_size = tensor.nelement() * tensor.element_size()
if max_size is not None and total_size + param_size > max_size:
filename = pattern.format(rank=rank, part=part_idx)
save_file(
state_dict_part,
os.path.join(path, filename),
)
part_idx += 1
total_size = 0
state_dict_part = {}
state_dict_part[key] = tensor
total_size += param_size
if len(state_dict_part) > 0:
filename = pattern.format(rank=rank, part=part_idx)
save_file(
state_dict_part,
os.path.join(path, filename),
)

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,151 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
# ruff: noqa: SIM117
import copy
from collections.abc import Generator
import torch
from torch import nn
from vllm.config import ModelConfig, ParallelConfig, VllmConfig
from vllm.config.load import LoadConfig
from vllm.logger import init_logger
from vllm.model_executor.model_loader.base_loader import BaseModelLoader
from vllm.model_executor.model_loader.tensorizer import (
TensorizerConfig,
deserialize_tensorizer_model,
init_tensorizer_model,
is_vllm_tensorized,
serialize_vllm_model,
tensorizer_weights_iterator,
)
from vllm.model_executor.model_loader.utils import (
get_model_architecture,
initialize_model,
)
from vllm.utils.torch_utils import set_default_torch_dtype
logger = init_logger(__name__)
BLACKLISTED_TENSORIZER_ARGS = {
"device", # vLLM decides this
"dtype", # vLLM decides this
"mode", # Not meant to be configurable by the user
}
def validate_config(config: dict):
for k, v in config.items():
if v is not None and k in BLACKLISTED_TENSORIZER_ARGS:
raise ValueError(f"{k} is not an allowed Tensorizer argument.")
class TensorizerLoader(BaseModelLoader):
"""Model loader using CoreWeave's tensorizer library."""
def __init__(self, load_config: LoadConfig):
super().__init__(load_config)
if isinstance(load_config.model_loader_extra_config, TensorizerConfig):
self.tensorizer_config = load_config.model_loader_extra_config
else:
validate_config(load_config.model_loader_extra_config)
self.tensorizer_config = TensorizerConfig(
**load_config.model_loader_extra_config["tensorizer_config"]
)
def _verify_config(
self, model_config: ModelConfig, parallel_config: ParallelConfig
):
self.tensorizer_config.verify_with_model_config(model_config)
self.tensorizer_config.verify_with_parallel_config(parallel_config)
def _get_weights_iterator(
self,
) -> Generator[tuple[str, torch.Tensor], None, None]:
tensorizer_args = self.tensorizer_config._construct_tensorizer_args()
return tensorizer_weights_iterator(tensorizer_args)
def _load_model_serialized_cpu(
self,
vllm_config: VllmConfig,
) -> nn.Module:
"""Load a serialized model with tensorizer to the CPU.
This is only necessary when the model isn't vLLM-tensorized (see
examples/others/tensorize_vllm_model.py) This should still
be faster than default HuggingFace loading, but will be slower than
loading a vLLM-tensorized model.
"""
device_config = vllm_config.device_config
model_config = vllm_config.model_config
with set_default_torch_dtype(model_config.dtype):
with torch.device(device_config.device):
model = initialize_model(vllm_config=vllm_config)
model.load_weights(self._get_weights_iterator())
return model.eval()
def download_model(self, model_config: ModelConfig) -> None:
self.tensorizer_config.verify_with_model_config(model_config)
with self.tensorizer_config.open_stream():
pass
def _patch_tensorizer_config(self, model_config: ModelConfig) -> TensorizerConfig:
model_class = get_model_architecture(model_config)[0]
tensorizer_config = copy.copy(self.tensorizer_config)
tensorizer_config.model_class = model_class
tensorizer_config.hf_config = model_config.hf_config
tensorizer_config.dtype = model_config.dtype
return tensorizer_config
def load_weights(self, model: nn.Module, model_config: ModelConfig) -> None:
"""Load serialized model weights with tensorizer.
Expects a vLLM-tensorized model. See the
examples/others/tensorize_vllm_model.py example script
for serializing vLLM models."""
if is_vllm_tensorized(self.tensorizer_config):
tensorizer_config = self._patch_tensorizer_config(model_config)
deserialize_tensorizer_model(model, tensorizer_config)
else:
model.load_weights(self._get_weights_iterator())
def load_model(
self, vllm_config: VllmConfig, model_config: ModelConfig
) -> nn.Module:
parallel_config = vllm_config.parallel_config
self._verify_config(model_config, parallel_config)
if parallel_config.tensor_parallel_size > 1:
from vllm.distributed import get_tensor_model_parallel_rank
self.tensorizer_config.tensorizer_uri = (
self.tensorizer_config.tensorizer_uri % get_tensor_model_parallel_rank()
)
if is_vllm_tensorized(self.tensorizer_config):
tensorizer_config = self._patch_tensorizer_config(model_config)
device_config = vllm_config.device_config
with set_default_torch_dtype(model_config.dtype):
with torch.device(device_config.device):
model = init_tensorizer_model(
tensorizer_config=tensorizer_config, vllm_config=vllm_config
)
self.load_weights(model, model_config)
return model
return self._load_model_serialized_cpu(vllm_config=vllm_config)
@staticmethod
def save_model(
model: torch.nn.Module,
tensorizer_config: TensorizerConfig | dict,
model_config: ModelConfig,
) -> None:
if isinstance(tensorizer_config, dict):
tensorizer_config = TensorizerConfig(**tensorizer_config)
serialize_vllm_model(
model=model,
tensorizer_config=tensorizer_config,
model_config=model_config,
)

View File

@@ -0,0 +1,118 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
import time
import torch
import torch.nn as nn
import torch_xla.core.xla_model as xm
import torch_xla.distributed.spmd as xs
from vllm.config import ModelConfig, VllmConfig
from vllm.distributed.tpu_distributed_utils import get_fqn, shard_model
from vllm.logger import init_logger
from vllm.model_executor.model_loader.default_loader import DefaultModelLoader
from vllm.model_executor.model_loader.utils import (
initialize_model,
process_weights_after_loading,
)
from vllm.utils.torch_utils import set_default_torch_dtype
logger = init_logger(__name__)
class TPUModelLoader(DefaultModelLoader):
"""
A TPU model loader for model loading under SPMD mode.
"""
def load_model(
self,
vllm_config: VllmConfig,
model_config: ModelConfig,
mesh: xs.Mesh | None = None,
) -> nn.Module:
# Initialize model and load weights on CPU. Then, during SPMD partition,
# weights are sharded and transferred to TPUs.
self.counter_before_loading_weights = time.perf_counter()
model_config = vllm_config.model_config
assert model_config.quantization is None, "Quantization not supported"
target_device = torch.device("cpu")
with set_default_torch_dtype(model_config.dtype):
with target_device:
model = initialize_model(vllm_config=vllm_config)
load_format = vllm_config.load_config.load_format
if load_format != "dummy":
weights_to_load = {name for name, _ in model.named_parameters()}
all_weights = self.get_all_weights(model_config, model)
loaded_weights = model.load_weights(all_weights)
self.counter_after_loading_weights = time.perf_counter()
logger.info(
"Loading weights took %.2f seconds",
self.counter_after_loading_weights
- self.counter_before_loading_weights,
)
# We only enable strict check for non-quantized models
# that have loaded weights tracking currently.
if model_config.quantization is None and loaded_weights is not None:
weights_not_loaded = weights_to_load - loaded_weights
if weights_not_loaded:
raise ValueError(
"Following weights were not initialized from "
f"checkpoint: {weights_not_loaded}"
)
else:
logger.info("Use dummy weight during weight loading.")
process_weights_after_loading(model, model_config, target_device)
counter_before_partition = time.perf_counter()
model = model.eval()
model = model.to("xla")
shard_model(model, mesh)
counter_after_partition = time.perf_counter()
logger.info(
"Partition model took %.2f seconds",
counter_after_partition - counter_before_partition,
)
# Ensure the model is properly loaded.
self._check_model_is_loaded(mesh, model)
# Need to torch compile after model sharding are done. Because the
# compiler hints ('xs.mark_sharding') are torch ops.
if not model_config.is_multimodal_model:
model.model = torch.compile(model.model, backend="openxla")
else:
model.language_model.model = torch.compile(
model.language_model.model, backend="openxla"
)
return model
def _check_model_is_loaded(self, mesh: xs.Mesh | None, model: nn.Module) -> None:
"""
Ensure the model is properly loaded.
1. All model parameters and buffers are on XLA device.
2. Non-SPMD friendly layers are replaced as expected.
"""
device = xm.xla_device()
device_type = str(device.type)
# Check parameters
for name, param in model.named_parameters():
assert param.device.type == device_type, (
f"Parameter {name} is on {param.device.type} instead of {device_type}"
)
# Check buffers
for name, buffer in model.named_buffers():
assert buffer.device.type == device_type, (
f"Buffer {name} is on {buffer.device.type} instead of {device_type}"
)
for module in model.modules():
if (mesh is not None) and (get_fqn(module) == "QKVParallelLinear"):
raise AssertionError(
"QKVParallelLinear should be replaced by \
XlaQKVParallelLinear under SPMD mode."
)

View File

@@ -1,41 +1,292 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
"""Utilities for selecting and loading models."""
import contextlib
from typing import Tuple, Type
import inspect
import warnings
from contextlib import contextmanager
from dataclasses import dataclass, field
import torch
from torch import nn
from typing_extensions import assert_never
from vllm.config import ModelConfig
from vllm.model_executor.models import ModelRegistry
from vllm.attention.layer import Attention, MLAAttention
from vllm.config import ModelConfig, VllmConfig, set_current_vllm_config
from vllm.logger import init_logger
from vllm.model_executor.layers.quantization.base_config import (
QuantizationConfig,
QuantizeMethodBase,
)
from vllm.model_executor.models.interfaces import SupportsQuant, supports_multimodal
from vllm.utils.platform_utils import is_pin_memory_available
logger = init_logger(__name__)
@contextlib.contextmanager
def set_default_torch_dtype(dtype: torch.dtype):
"""Sets the default torch dtype to the given dtype."""
old_dtype = torch.get_default_dtype()
torch.set_default_dtype(dtype)
yield
torch.set_default_dtype(old_dtype)
def initialize_model(
vllm_config: VllmConfig,
*,
prefix: str = "",
model_class: type[nn.Module] | None = None,
model_config: ModelConfig | None = None,
) -> nn.Module:
"""Initialize a model with the given configurations."""
if model_config is None:
model_config = vllm_config.model_config
if model_class is None:
model_class, _ = get_model_architecture(model_config)
if vllm_config.quant_config is not None:
configure_quant_config(vllm_config.quant_config, model_class)
signatures = inspect.signature(model_class.__init__)
all_params = [param.name for param in signatures.parameters.values()]
if "vllm_config" in all_params and "prefix" in all_params:
# new-style model class
with set_current_vllm_config(vllm_config, check_compile=True, prefix=prefix):
return model_class(vllm_config=vllm_config, prefix=prefix)
msg = (
"vLLM model class should accept `vllm_config` and `prefix` as "
"input arguments. Possibly you have an old-style model class"
" registered from out of tree and it is used for new vLLM version. "
"Check https://docs.vllm.ai/en/latest/design/arch_overview.html "
"for the design and update the model class accordingly."
)
warnings.warn(msg, DeprecationWarning, stacklevel=2)
logger.warning(
"Trying to guess the arguments for old-style model class %s",
model_class,
)
# try to be compatible with old-style model class
kwargs = {}
if "prefix" in all_params:
kwargs["prefix"] = prefix
if "config" in all_params:
kwargs["config"] = model_config.hf_config
if "cache_config" in all_params:
kwargs["cache_config"] = vllm_config.cache_config
if "quant_config" in all_params:
kwargs["quant_config"] = vllm_config.quant_config
if "lora_config" in all_params:
kwargs["lora_config"] = vllm_config.lora_config
if "scheduler_config" in all_params:
kwargs["scheduler_config"] = vllm_config.scheduler_config
with set_current_vllm_config(vllm_config, check_compile=True, prefix=prefix):
return model_class(**kwargs)
def get_model_architecture(
model_config: ModelConfig) -> Tuple[Type[nn.Module], str]:
def process_weights_after_loading(
model: nn.Module, model_config: ModelConfig, target_device: torch.device
) -> None:
if getattr(model, "process_weights_after_loading_already_called", False):
# In case `process_weights_after_loading` is called multiple times
# we'll skip it at later times
logger.debug_once(
"process_weights_after_loading already called for model %s", model
)
return
# to avoid circular dependency
from vllm.model_executor.model_loader.online_quantization import (
maybe_save_metadata_and_attributes_for_weight_reloading,
)
maybe_save_metadata_and_attributes_for_weight_reloading(model, model_config)
for _, module in model.named_modules():
quant_method = getattr(module, "quant_method", None)
if isinstance(quant_method, QuantizeMethodBase):
# When quant methods need to process weights after loading
# (for repacking, quantizing, etc), they expect parameters
# to be on the global target device. This scope is for the
# case where cpu offloading is used, where we will move the
# parameters onto device for processing and back off after.
with device_loading_context(module, target_device):
quant_method.process_weights_after_loading(module)
# Initialize post-load attention weights for both Attention and MLA.
# NOTE: Happens after other modules so we can easily decompress weights.
for _, module in model.named_modules():
if isinstance(module, (Attention, MLAAttention)) and hasattr(
module, "process_weights_after_loading"
):
# TODO(lucas): see if there is a way to unify the signatures
# of process_weights_after_loading
module.process_weights_after_loading(model_config.dtype)
@contextmanager
def device_loading_context(module: torch.nn.Module, target_device: torch.device):
if target_device.type == "cpu":
# If target is CPU, no need to move anything
yield module
return
original_device_states: dict[str, torch.device] = {}
# Store original device states and move parameters to GPU if they're on CPU
for name, p in module.named_parameters():
if p.device.type == "cpu":
original_device_states[name] = p.device
p.data = p.data.to(target_device)
# Parameters already on target device are not touched
try:
yield module
finally:
# Restore parameters to their original devices, ignoring new parameters
pin_memory = is_pin_memory_available()
for name, p in module.named_parameters():
if name in original_device_states:
original_device: torch.device = original_device_states[name]
if original_device.type == "cpu":
# `torch.empty_like` does not support `pin_memory` argument
cpu_data = torch.empty_strided(
size=p.data.size(),
stride=p.data.stride(),
dtype=p.data.dtype,
layout=p.data.layout,
device="cpu",
pin_memory=pin_memory,
)
cpu_data.copy_(p.data)
p.data = cpu_data
else:
p.data = p.data.to(original_device)
# New parameters or parameters already on target device are untouched
_MODEL_ARCH_BY_HASH = dict[int, tuple[type[nn.Module], str]]()
"""Caches the outputs of `_get_model_architecture`."""
def _get_model_architecture(model_config: ModelConfig) -> tuple[type[nn.Module], str]:
from vllm.model_executor.models.adapters import (
as_embedding_model,
as_seq_cls_model,
try_create_mm_pooling_model_cls,
)
architectures = getattr(model_config.hf_config, "architectures", [])
# Special handling for quantized Mixtral.
# FIXME(woosuk): This is a temporary hack.
if (model_config.quantization is not None
and model_config.quantization != "fp8"
and "MixtralForCausalLM" in architectures):
architectures = ["QuantMixtralForCausalLM"]
for arch in architectures:
model_cls = ModelRegistry.load_model_cls(arch)
if model_cls is not None:
return (model_cls, arch)
raise ValueError(
f"Model architectures {architectures} are not supported for now. "
f"Supported architectures: {ModelRegistry.get_supported_archs()}")
model_cls, arch = model_config.registry.resolve_model_cls(
architectures,
model_config=model_config,
)
if arch == model_config._get_transformers_backend_cls():
assert model_config.model_impl != "vllm"
if model_config.model_impl == "auto":
logger.warning_once(
"%s has no vLLM implementation, falling back to Transformers "
"implementation. Some features may not be supported and "
"performance may not be optimal.",
arch,
)
convert_type = model_config.convert_type
if convert_type != "none" and supports_multimodal(model_cls):
logger.debug_once("Detected conversion of Multi Modal model.")
converted = try_create_mm_pooling_model_cls(model_cls)
if converted is not None:
logger.debug_once("Creating wrapper class to forward pooler.")
return converted, arch
else:
logger.debug_once("Attempting direct conversion.")
if convert_type == "none":
pass
elif convert_type == "embed":
logger.debug_once("Converting to embedding model.")
model_cls = as_embedding_model(model_cls)
elif convert_type == "classify":
logger.debug_once("Converting to sequence classification model.")
model_cls = as_seq_cls_model(model_cls)
else:
assert_never(convert_type)
return model_cls, arch
def get_model_architecture(model_config: ModelConfig) -> tuple[type[nn.Module], str]:
key = hash(
(
model_config.model,
model_config.convert_type,
model_config.runner_type,
model_config.trust_remote_code,
model_config.model_impl,
tuple(getattr(model_config.hf_config, "architectures", [])),
)
)
if key in _MODEL_ARCH_BY_HASH:
return _MODEL_ARCH_BY_HASH[key]
model_arch = _get_model_architecture(model_config)
_MODEL_ARCH_BY_HASH[key] = model_arch
return model_arch
def get_model_cls(model_config: ModelConfig) -> type[nn.Module]:
return get_model_architecture(model_config)[0]
def get_architecture_class_name(model_config: ModelConfig) -> str:
return get_model_architecture(model_config)[1]
@dataclass
class ParamMapping:
"""
A class to handle parameter mapping for model weight loading.
It creates a bidirectional mapping between packed parameters and their
constituent parts.
"""
packed_mapping: dict[str, list[str]]
inverse_packed_mapping: dict[str, tuple[str, int]] = field(default_factory=dict)
def __post_init__(self):
for packed_name, sub_params in self.packed_mapping.items():
# Skip self-contained cases (e.g., {"W_pack": ["W_pack"]})
if len(sub_params) == 1 and sub_params[0] == packed_name:
continue
for index, param_name in enumerate(sub_params):
self.inverse_packed_mapping[param_name] = (
packed_name,
index,
)
def get_sub_modules(self, module_name: str) -> tuple[str, list[str]] | None:
for key, value in self.packed_mapping.items():
if module_name.endswith(key):
return key, value
return None
def configure_quant_config(
quant_config: QuantizationConfig, model_class: type[nn.Module]
):
"""
Pass packed_modules_mapping by reference to quant_config so that
quant_config can properly match fused modules
Note that model attributes are passed by reference to quant_config,
enabling them to be updated by model_class.__new__ (ex. chatglm, qwen)
Once the `SupportsQuant` mixin has been added to all models, this
function can be removed
"""
if not issubclass(model_class, SupportsQuant):
hf_to_vllm_mapper = getattr(model_class, "hf_to_vllm_mapper", None)
packed_mapping = getattr(model_class, "packed_modules_mapping", None)
# pass mappings by reference to quant_config
if hf_to_vllm_mapper is not None:
quant_config.apply_vllm_mapper(hf_to_vllm_mapper)
if packed_mapping is not None:
quant_config.packed_modules_mapping = packed_mapping

File diff suppressed because it is too large Load Diff