File size: 21,511 Bytes
f618189 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 | # Copyright 2024-2025 ModelCloud.ai
# Copyright 2024-2025 qubitium@modelcloud.ai
# Contact: qubitium@modelcloud.ai, x.com/qubitium
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import json
import os.path
import re
from dataclasses import dataclass, field, fields
from enum import Enum
from os.path import join
from typing import Any, Dict, List, Optional, Tuple, Union
import torch
from packaging import version
from ..adapter.adapter import Lora, normalize_adapter
from ..utils.logger import setup_logger
log = setup_logger()
FORMAT_FIELD_CODE = "format"
FORMAT_FIELD_JSON = "checkpoint_format"
FORMAT_FIELD_COMPAT_MARLIN = "is_marlin_format"
QUANT_METHOD_FIELD = "quant_method"
PACK_DTYPE_FIELD = "pack_dtype"
QUANT_CONFIG_FILENAME = "quantize_config.json"
QUANT_CONFIG_FILENAME_COMPAT = [QUANT_CONFIG_FILENAME, "quant_config.json", "config.json"]
MIN_VERSION_WITH_V2 = "0.9.0"
META_FIELD = "meta"
# quantizer is the tool that did the quantization
META_FIELD_QUANTIZER = "quantizer"
META_QUANTIZER_GPTQMODEL = "gptqmodel"
META_FIELD_URI = "uri"
META_VALUE_URI = "https://github.com/modelcloud/gptqmodel"
META_FIELD_DAMP_PERCENT = "damp_percent"
META_FIELD_DAMP_AUTO_INCREMENT = "damp_auto_increment"
META_FIELD_STATIC_GROUPS = "static_groups"
META_FIELD_TRUE_SEQUENTIAL = "true_sequential"
META_FIELD_MSE = "mse"
META_FIELD_ACT_GROUP_AWARE = "act_group_aware"
META_FIELD_V2_ENABLED = "v2"
META_FIELD_V2_ALPHA = "v2_alpha"
META_FIELD_V2_MEMORY_DEVICE = "v2_memory_device"
ADAPTER_FIELD = "adapter"
# pkg names
PKG_AUTO_ROUND = "auto-round"
# saved formats
class FORMAT(str, Enum):
GPTQ = "gptq"
# v2 format fixed sym = False quantization
GPTQ_V2 = "gptq_v2"
MARLIN = "marlin"
BITBLAS = "bitblas"
IPEX = "ipex"
QQQ = "qqq"
# quant methods
class QUANT_METHOD(str, Enum):
GPTQ = "gptq"
AUTO_ROUND = "auto_round"
QQQ = "qqq"
QUANT_METHOD_FORMAT_MAPPING = {
QUANT_METHOD.GPTQ: {
FORMAT.GPTQ,
FORMAT.GPTQ_V2,
FORMAT.MARLIN,
FORMAT.BITBLAS,
FORMAT.IPEX,
},
QUANT_METHOD.AUTO_ROUND: {
FORMAT.GPTQ,
FORMAT.GPTQ_V2,
FORMAT.MARLIN,
FORMAT.BITBLAS,
},
QUANT_METHOD.QQQ: {
FORMAT.QQQ,
},
}
# inference only methods should go here
QUANTIZE_BLACK_LIST = {}
# compat
QUANT_CONFIG_ARG_SYNONYMS = {
"w_bit": "bits",
"q_group_size": "group_size",
# map format field (checkpoint_format) to class/code (format)
FORMAT_FIELD_JSON: FORMAT_FIELD_CODE,
}
def dict_scale_dtype_to_str(d: Dict[str, Any]) -> None:
"""
Checks whether the passed dictionary and its nested dicts have a *scale_dtype* key and if it's not None,
converts torch.dtype to a string of just the type. For example, `torch.float32` get converted into *"float32"*
string, which can then be stored in the json format.
"""
if d.get("scale_dtype", None) is not None and not isinstance(d["scale_dtype"], str):
d["scale_dtype"] = str(d["scale_dtype"]).split(".")[1]
for value in d.values():
if isinstance(value, dict):
dict_scale_dtype_to_str(value)
def dynamic_get(dynamic: Dict[str, Dict[str, Union[int, bool]]], module_name: str, key: str = None,
default: Union[int, bool] = None, sub_key: str = None) -> Union[Dict, int, bool]:
if dynamic is None:
return default
for pattern, overrides in dynamic.items():
if pattern.startswith("-:"):
if re.match(pattern.removeprefix("-:"), module_name):
return False
elif re.match(pattern.removeprefix("+:"), module_name):
if key is None:
return overrides
else:
# subkey example: Lora override format: `{ "adapter": { "rank": 512 } }`
if sub_key:
sub_value = overrides.get(key, None)
if isinstance(sub_value, Dict):
return sub_value.get(sub_key, default)
else:
log.info(f"QuantConfig: Dynamic `sub_key`: `{sub_key}` failed extraction from `sub_value`: `{sub_value}`")
else:
return overrides.get(key, default)
return default
@dataclass
class QuantizeConfig():
bits: int = field(default=4, metadata={"choices": [2, 3, 4, 8]})
# allow dynamic bitsize per layer, if None or some layer not set, use bits
dynamic: Optional[Dict[str, Dict[str, Union[int, bool]]]] = field(default=None)
# 128 offer good balance between inference speed, vram usage (bpw), and quality
# use 32 for highest quality with slower inference and higher vram usage
group_size: int = field(default=128)
# increase damp if NaN is encountered during `.quantize()` and/or increase calib dataset size
damp_percent: float = field(default=0.05)
damp_auto_increment: float = field(default=0.01)
desc_act: bool = field(default=True)
act_group_aware: bool = field(default=False)
static_groups: bool = field(default=False)
sym: bool = field(default=True)
true_sequential: bool = field(default=True)
lm_head: bool = field(default=False)
quant_method: QUANT_METHOD = field(default=QUANT_METHOD.GPTQ)
# default to gptq v1 format for maximum compat with 3rd party inference libs with minimal loss vs v2
# if you inference with gptqmodel, save to gptq_v2 format for best result
format: FORMAT = field(default=FORMAT.GPTQ)
# quantization_order: str = "activate",
# quantization_scale: str = "mse", # or absmax
# is_distributed: bool = False,
# tied_gptq_handle: Optional["GPTQ"] = None
# mean square error calculation: may reduce error loss for some models
mse: float = field(default=0.0)
# parallel packing will make ~40% speedup for many models, but may cause OOM in some large models
# if OOM, can set to False
parallel_packing: bool = field(default=True)
# properties that do not directly contributes to quantization or quant inference should be placed in meta
# i.e. quantizer tool (producer) + version, timestamp, entity who made the quant, etc
meta: Optional[Dict] = field(default=None)
# normalized to DEVICE after passing to load()
device: Optional[Union[str, torch.device]] = field(default=None)
# gptq was originally designed to pack quantized weights inside INT32 dtypes
# allowing using different dtypes used for packing quantized weights
# affects [`qweights`, `qzeros`]
pack_dtype: Optional[Union[str, torch.dtype]] = field(default=torch.int32)
# pending used field
adapter: Optional[Union[Dict[str, Any], Lora]] = field(default=None)
rotation: Optional[str] = field(default=None, metadata={"choices": ["hadamard", "random"]})
is_marlin_format: bool = False
v2: bool = False
v2_alpha: float = 0.25
v2_memory_device: str = "auto" #
# Skip all heavy computations for testing model loading
mock_quantization: bool = field(default=False, metadata={"help": "Skip heavy computations for fast model loading validation"})
def __post_init__(self):
fields_info = fields(self)
# validate/normalizes pack_dtype from string and dtype to valid dtype
if self.pack_dtype is None:
self.pack_dtype = torch.int32
else:
if isinstance(self.pack_dtype, str):
self.pack_dtype = self.pack_dtype.lower()
if self.pack_dtype not in ["int64", "int32", "int16", "int8"]:
raise ValueError(f"QuantizeConfig: Unsupported `pack_dtype`: {self.pack_dtype}")
self.pack_dtype = getattr(torch, self.pack_dtype)
elif isinstance(self.pack_dtype, torch.dtype):
if self.pack_dtype not in [torch.int64, torch.int32, torch.int16, torch.int8]:
raise ValueError(f"QuantizeConfig: Unsupported `pack_dtype`: {self.pack_dtype}")
else:
raise ValueError(f"QuantizeConfig: Unsupported `pack_dtype`: {self.pack_dtype}")
# validate quant method and format is matched
valid_formats = QUANT_METHOD_FORMAT_MAPPING.get(self.quant_method, None)
if valid_formats is None:
raise ValueError(f"QuantizeConfig: Unsupported `quant_method`: {self.quant_method}")
# TODO FIXME qqq compat which didn't have checkpoint_format before merging to gptqmodel
if self.quant_method == QUANT_METHOD.QQQ and self.format != FORMAT.QQQ:
log.info(f"QuantizeConfig: Auto fix `format` to `{FORMAT.QQQ}`")
self.format = FORMAT.QQQ
if self.format not in valid_formats:
raise ValueError(
f"QuantizeConfig: checkpoint `format` used is {self.format}, and the quantization method is {self.quant_method}. "
)
if self.bits not in fields_info[0].metadata["choices"]:
raise ValueError(f"QuantizeConfig: `bits` must be in the set of `{fields_info[0].metadata['choices']}`.")
if self.dynamic is not None:
self.dynamic = {
**{k: v for k, v in self.dynamic.items() if k.startswith('-')}, # 先添加以 "-" 开头的键
**{k: v for k, v in self.dynamic.items() if not k.startswith('-')} # 然后添加其他键
}
for layer, layer_dict in self.dynamic.items():
for key, value in layer_dict.items():
if key == "bits" and value not in fields_info[0].metadata["choices"]:
raise ValueError(f"QuantizeConfig: Layer `{layer}` only support quantization of `{fields_info[0].metadata['choices']}` bits.")
elif key == "group_size" and value != -1 and value <= 0:
raise ValueError("QuantizeConfig: `group_size` must in the value set of `[-1, 16, 32, 64, 128]`.")
if self.group_size != -1 and self.group_size <= 0:
raise ValueError("QuantizeConfig: `group_size` must in the value set of `[-1, 16, 32, 64, 128]`.")
if not (0 < self.damp_percent < 1):
raise ValueError("QuantizeConfig: `damp_percent` must between 0 and 1.")
if self.damp_auto_increment < 0:
raise ValueError("QuantizeConfig:: `damp_auto_increment` must greater than 0.")
# validate hybrid act order
if self.act_group_aware and self.desc_act:
raise ValueError("QuantizeConfig:: `act_group_aware` == `True` requires `desc_act` == `False`.")
# validate meta
if self.meta is not None:
if not isinstance(self.meta, dict):
raise ValueError("QuantizeConfig: `meta` must be a dictionary")
for key, value in self.meta.items():
if not isinstance(key, str):
raise ValueError("QuantizeConfig: `meta` keys must be strings")
else:
self.meta = {}
# adapter normalize
self.adapter = normalize_adapter(self.adapter)
#print(f"adapter: {self.adapter}")
def extension_set(self, key: str, value: Any):
if self.adapter is None:
self.adapter = {}
self.adapter[key.lower()] = value
def extension_get(self, key: str) -> Any:
return self.adapter.get(key.lower()) if self.adapter else None
def meta_set(self, key: str, value: Any):
self.meta[key] = value
def meta_get(self, key: str) -> Any:
return self.meta.get(key)
def dynamic_get(self, layer_name: str, key: str = None, default: Union[int, bool, float] = None, sub_key: str = None
) -> Union[Dict, int, bool, float]:
return dynamic_get(self.dynamic, layer_name, key, default, sub_key)
# versionable is a meta.property that pairs value with version i.e "value:1.0.0"
def meta_set_versionable(self, key: str, value: List[str]):
self.meta_set(key, value)
# versionable is a meta.property that pairs value with version i.e "value:1.0.0"
def meta_get_versionable(self, key: str) -> List[Tuple[str, str]]:
values = self.meta_get(key)
if values is None:
return []
if not isinstance(values, list):
values = [values]
result = []
for val in values:
parts = val.split(":")
if len(parts) >= 2:
result.append((parts[0].lower(), parts[1].lower()))
return result
# is quantized model quantized or packed by gptqmodel version with v2 format code
def is_quantized_by_v2(self) -> bool:
# check meta.quantizer
result = self.meta_get_versionable(META_FIELD_QUANTIZER)
if len(result) > 0:
for producer, _version in result:
if producer == META_QUANTIZER_GPTQMODEL:
return version.parse(_version) >= version.parse(MIN_VERSION_WITH_V2)
return False
def extract_adapter_rank_patterns(self) -> Optional[Dict[str, int]]:
adapter_rank_patterns = {}
# no rank can be had if there is no dynamic or adapter
if not self.dynamic or not self.adapter:
return adapter_rank_patterns
# override format: `{ "adapter": { "rank": 512 } }`
for k, v in self.dynamic.items():
adapter_override = v.get("adapter", None) # TODO use const, not str
if adapter_override and isinstance(adapter_override, Dict):
rank = adapter_override.get("rank", None)
if rank and isinstance(rank, int):
# need to strip `+:` positive prefix
adapter_rank_patterns[k.lstrip("+:")] = rank # TODO use const, not str
return adapter_rank_patterns
def save_pretrained(self, save_dir: str, **kwargs):
with open(join(save_dir, QUANT_CONFIG_FILENAME), "w", encoding="utf-8") as f:
d = self.to_dict()
json_str = json.dumps(d, indent=2)
log.info(f"Saved Quantize Config: \n{json_str}")
f.write(json_str)
@classmethod
# normalize quant config for compat and also performs validation
def from_quant_config(cls, quantize_cfg, format: str = None):
valid_formats = {FORMAT.GPTQ, FORMAT.GPTQ_V2, FORMAT.MARLIN, FORMAT.BITBLAS, FORMAT.IPEX}
format_auto_inferred = False
# compat: format can be passed in via from_quantized() if field missing from json
if format:
if format not in valid_formats:
raise ValueError(f"QuantizeConfig: Unknown quantization checkpoint format: {format}.")
if quantize_cfg.get(FORMAT_FIELD_JSON):
raise ValueError("QuantizeConfig: Conflicting quantization format passed in manually and also exists in model config.")
# compat: warn if checkpoint_format is missing
elif quantize_cfg.get(FORMAT_FIELD_JSON) is None:
format_auto_inferred = True
field_names = [field.name for field in fields(cls)]
normalized = {
QUANT_METHOD_FIELD: QUANT_METHOD.GPTQ,
# compat: default to gptq(v1) when loading models
FORMAT_FIELD_CODE: format if format else FORMAT.GPTQ,
}
for key, val in quantize_cfg.items():
key = key.lower()
# remap keys according to compat map
if key in QUANT_CONFIG_ARG_SYNONYMS and QUANT_CONFIG_ARG_SYNONYMS[key] in field_names:
key = QUANT_CONFIG_ARG_SYNONYMS[key]
if key == FORMAT_FIELD_JSON:
val = val.lower()
if val in {FORMAT.GPTQ, FORMAT.GPTQ_V2, FORMAT.MARLIN, FORMAT.BITBLAS}:
normalized[key] = val
else:
raise ValueError(f"QuantizeConfig: Unknown quantization format: `{val}`.")
elif key == QUANT_METHOD_FIELD:
val = val.lower()
# compat: some hf models use quant_method=marlin or bitblas
if val == FORMAT.MARLIN:
normalized[FORMAT_FIELD_CODE] = FORMAT.MARLIN
elif val == FORMAT.BITBLAS:
normalized[FORMAT_FIELD_CODE] = FORMAT.BITBLAS
elif val not in {QUANT_METHOD.GPTQ, QUANT_METHOD.AUTO_ROUND, QUANT_METHOD.QQQ}:
raise ValueError(f"QuantizeConfig: Unknown quantization method: `{val}`.")
else:
normalized[QUANT_METHOD_FIELD] = val
elif key in field_names:
normalized[key] = val
else:
log.info(f"QuantizeConfig: Ignoring unknown parameter in the quantization configuration: {key}.")
if format_auto_inferred:
log.info(f"QuantizeConfig: `{FORMAT_FIELD_JSON}` is missing from the quantization configuration and is automatically inferred to {normalized[FORMAT_FIELD_CODE]}")
if normalized[FORMAT_FIELD_CODE] in {FORMAT.BITBLAS}:
# AWQ and Marlin do not reorder the rows.
normalized["desc_act"] = False
if "sym" not in normalized:
log.warn(
"QuantizeConfig: config does not contain `sym` (symmetric quantization). This may result in silent errors. Defaulting to `sym=True`."
)
return cls(**normalized)
@classmethod
def from_pretrained(cls, save_dir: str, **kwargs):
format = kwargs.pop("format", None)
transformers_config = False
resolved_config_file = None
for quantize_config_filename in QUANT_CONFIG_FILENAME_COMPAT:
resolved_config_file = join(save_dir, quantize_config_filename)
if os.path.exists(resolved_config_file):
if quantize_config_filename == "config.json":
transformers_config = True
break
if resolved_config_file is None:
raise ValueError(
"QuantizeConfig: No quantize_config.json, quant_config.json or config.json file was found in the model repository."
)
with open(resolved_config_file, "r", encoding="utf-8") as f:
args_from_json = json.load(f)
if transformers_config:
args_from_json = args_from_json["quantization_config"]
return cls.from_quant_config(args_from_json, format)
def to_dict(self):
out = {
"bits": self.bits,
"dynamic": self.dynamic,
"group_size": self.group_size,
"desc_act": self.desc_act,
"sym": self.sym,
"lm_head": self.lm_head,
QUANT_METHOD_FIELD:self.quant_method,
FORMAT_FIELD_JSON: self.format,
# torch.dtype convert to string
PACK_DTYPE_FIELD: str(self.pack_dtype).split(".")[-1],
META_FIELD: self.meta,
# DO NOT EXPORT Adapter to config/json since adapter can be swapped out/in
# ADAPTER_FIELD: self.adapter.to_dict() if self.adapter else None,
}
dynamic = out["dynamic"]
if dynamic:
# dynamic adapter config is only used in the quantize phase and is deleted when saving.
for _, v in dynamic.items():
v.pop("adapter", None)
# simplify: clean keys where the value is None or empty [list, dict]
out = {k: v for k, v in out.items() if v is not None and (v not in [None, {}])}
dict_scale_dtype_to_str(out)
return out
# TODO FIX ME, g_idx int32 per infeature but infeature count is per module
def calculate_bits_per_weight(self):
if self.group_size != -1:
# naive bits is
#mlp.down_proj.g_idx: I32
#mlp.down_proj.qweight: I32
#mlp.down_proj.qzeros: I32
#mlp.down_proj.scales: F16
per_group_bits = self.group_size * self.bits # qweight: packed by group_size
per_group_bits += 16 # scales fp16: one per group
per_group_bits += self.bits # qzeros: one per group
# FIX ME: g_idx is I32, one per infeature
per_group_bits += 4 # ESTIMATE for g_idx int32: one per features/group_size item
bpw = per_group_bits / self.group_size
# normally g_idx (int32 allocated one per in_feature) is allocated in device memory
# but each module may have different infeatures we don't have enouch ctx here, use estimated `0.1` for now
bpw += 0.1
else:
# there is only one scale int32 + one qzero int32 per entire module so overall it contributes to close to 0 bpw
bpw = self.bits
log.info(f"Estimated Quantization BPW (bits per weight): {bpw} bpw, based on [bits: {self.bits}, group_size: {self.group_size}]")
# deprecated: will be removed in future update
@dataclass
class BaseQuantizeConfig(QuantizeConfig):
def __init__(self, **kwargs):
super().__init__(**kwargs)
log.warn("QuantizeConfig: BaseQuantizeConfig is re-named and pending deprecation. Please use `QuantizeConfig` instead.")
|