Unverified Commit 93d2b9fe authored by Daniël de Kok's avatar Daniël de Kok Committed by GitHub
Browse files

Split up `layers.marlin` into several files (#2292)

The marlin.py file was getting large, split it up.
parent 86422506
...@@ -8,7 +8,10 @@ from text_generation_server.utils.weights import ( ...@@ -8,7 +8,10 @@ from text_generation_server.utils.weights import (
) )
from text_generation_server.layers.gptq import GPTQWeight, GPTQWeightsLoader from text_generation_server.layers.gptq import GPTQWeight, GPTQWeightsLoader
from text_generation_server.layers.exl2 import Exl2Weight, Exl2WeightsLoader from text_generation_server.layers.exl2 import Exl2Weight, Exl2WeightsLoader
from text_generation_server.layers.marlin import MarlinWeight, MarlinWeightsLoader from text_generation_server.layers.marlin.marlin import (
MarlinWeight,
MarlinWeightsLoader,
)
from types import SimpleNamespace from types import SimpleNamespace
from typing import List, Optional, Dict, Union from typing import List, Optional, Dict, Union
from pathlib import Path from pathlib import Path
......
from typing import List, Tuple
import torch
from text_generation_server.layers.marlin.fp8 import GPTQMarlinFP8Linear
from text_generation_server.layers.marlin.gptq import (
GPTQMarlinLinear,
GPTQMarlinWeight,
can_use_gptq_marlin,
repack_gptq_for_marlin,
)
from text_generation_server.layers.marlin.marlin import MarlinWeightsLoader
__all__ = [
"GPTQMarlinFP8Linear",
"GPTQMarlinLinear",
"GPTQMarlinWeight",
"MarlinWeightsLoader",
"can_use_gptq_marlin",
"repack_gptq_for_marlin",
]
from typing import Optional
import torch
import torch.nn as nn
from loguru import logger
from text_generation_server.layers.fp8 import fp8_quantize
from text_generation_server.layers.marlin.gptq import _check_valid_shape
from text_generation_server.layers.marlin.util import (
_check_marlin_kernels,
permute_scales,
)
from text_generation_server.utils.log import log_once
try:
import marlin_kernels
except ImportError:
marlin_kernels = None
MARLIN_TILE_SIZE = 16
class GPTQMarlinFP8Linear(nn.Module):
"""
FP8 GPTQ-Marlin linear layer.
"""
def __init__(
self,
qweight: torch.Tensor,
scales: torch.Tensor,
bias: Optional[torch.Tensor],
) -> None:
super().__init__()
_check_marlin_kernels()
assert marlin_kernels is not None
log_once(logger.info, "GPU does not support FP8, using Marlin FP8 kernel")
scales = scales.unsqueeze(0)
if scales.shape[1] == 1:
out_features, in_features = qweight.shape
scales = scales.repeat(1, out_features)
qweight, scales = repack_fp8_for_marlin(qweight, scales)
in_features = qweight.shape[0] * MARLIN_TILE_SIZE
out_features = scales.shape[1]
_check_valid_shape(in_features=in_features, out_features=out_features)
self.qweight = qweight
self.scales = scales
self.bias = bias if bias is not None else None
self.workspace = torch.zeros(
out_features // 64 * 16, dtype=torch.int, device=qweight.device
)
@classmethod
def from_unquant(cls, weight, bias, dtype):
qweight, scales = fp8_quantize(weight)
return cls(qweight=qweight, scales=scales.to(dtype), bias=bias)
@classmethod
def from_fp8(cls, weight, scale, _input_scale, bias, dtype):
return cls(qweight=weight, scales=scale.to(dtype), bias=bias)
def forward(self, A: torch.Tensor) -> torch.Tensor:
assert marlin_kernels is not None
A_flat = A.view(-1, A.shape[-1])
C = marlin_kernels.fp8_marlin_gemm(
A_flat,
self.qweight,
self.scales,
self.workspace,
8,
A_flat.shape[0],
self.scales.shape[1],
A_flat.shape[1],
)
C = C.reshape(A.shape[:-1] + (self.scales.shape[1],))
if self.bias is not None:
C += self.bias
return C
def pack_fp8_as_int32(fp8_tensor: torch.Tensor) -> torch.Tensor:
"""
Repack FP8 weights to gptq format (packed int32 elements).
"""
assert fp8_tensor.dtype == torch.float8_e4m3fn
if fp8_tensor.shape[0] % 4 != 0:
raise ValueError(
f"Leading tensor dimension is not divisable by 4: {fp8_tensor.shape[0]}"
)
# Reshape to prepare for packing
reshaped = fp8_tensor.reshape(-1, 4, *fp8_tensor.shape[1:])
# Convert fp8 to uint8 (byte) representation
byte_tensor = reshaped.view(torch.uint8)
# Pack 4 uint8 values into one int32
packed = torch.zeros(
fp8_tensor.shape[0] // 4,
fp8_tensor.shape[1],
dtype=torch.int32,
device=fp8_tensor.device,
)
for i in range(4):
packed.bitwise_or_(byte_tensor[:, i].to(torch.int32) << i * 8)
return packed
def repack_fp8_for_marlin(weight: torch.Tensor, scales: torch.Tensor):
"""
Repack FP8 tensor for GPTQ-Marlin.
"""
out_features, in_features = weight.shape
# Torch linear layers weights with shape [out_features, in_features],
# GPTQ-quantized weights use [in_feateres/pack_factor, in_features],
# so transpose before packing.
qweight = pack_fp8_as_int32(weight.t())
perm = torch.empty(0, dtype=torch.int, device=qweight.device)
repacked = marlin_kernels.gptq_marlin_repack(
qweight, perm, in_features, out_features, 8
)
scales = permute_scales(scales)
return repacked, scales
from dataclasses import dataclass
from typing import Optional
import numpy
import torch
import torch.nn as nn
from loguru import logger
from text_generation_server.layers.marlin.util import (
_check_marlin_kernels,
marlin_zero_points,
permute_scales,
unpack_cols,
)
from text_generation_server.utils.import_utils import SYSTEM
from text_generation_server.utils.log import log_once
from text_generation_server.utils.weights import Weight
try:
import marlin_kernels
except ImportError:
marlin_kernels = None
try:
major, _minor = torch.cuda.get_device_capability()
has_sm_8_0 = major >= 8
except Exception:
has_sm_8_0 = False
GPTQ_MARLIN_BITS = [4, 8]
GPTQ_MARLIN_GROUP_SIZES = [-1, 32, 64, 128]
MARLIN_TILE_SIZE = 16
def can_use_gptq_marlin(
*, bits: int, groupsize: int, quant_method: str, quantize: str, sym: bool
) -> bool:
return (
SYSTEM == "cuda"
and marlin_kernels is not None
and has_sm_8_0
and quantize in {"awq", "gptq"}
and quant_method in {"awq", "gptq"}
and bits in GPTQ_MARLIN_BITS
and groupsize in GPTQ_MARLIN_GROUP_SIZES
# We only suppord asymmetric quantization for AWQ.
and (sym or quant_method == "awq")
)
@dataclass
class GPTQMarlinWeight(Weight):
"""
Repacked GPTQ Marlin weights.
"""
qweight: torch.Tensor
qzeros: torch.Tensor
scales: torch.Tensor
g_idx: torch.Tensor
perm: torch.Tensor
bits: int
is_full_k: bool
def __post_init__(self):
assert self.qweight.dtype == torch.int32
assert self.scales.dtype == torch.float16
assert self.g_idx.dtype == torch.int32
assert self.perm.dtype == torch.int32
def get_linear(self, bias: torch.Tensor):
return GPTQMarlinLinear(
weight=self,
bias=bias,
)
def repack_gptq_for_marlin(
*,
qweight: torch.Tensor,
qzeros: Optional[torch.Tensor],
scales: torch.Tensor,
g_idx: Optional[torch.Tensor],
bits: int,
desc_act: bool,
groupsize: int,
quant_method: str,
sym: bool,
sharded_infeatures: bool,
) -> GPTQMarlinWeight:
"""Convert GPTQ weights to a layout that's compatible with GPTQ-Marlin kernels."""
_check_marlin_kernels()
assert marlin_kernels is not None
if bits not in GPTQ_MARLIN_BITS:
supported_bits = ", ".join(str(b) for b in GPTQ_MARLIN_BITS)
raise RuntimeError(
f"Repacking {bits}-bit GPTQ weights as Marlin is not supported, must be one of: {supported_bits}"
)
if groupsize not in GPTQ_MARLIN_GROUP_SIZES:
supported_sizes = ", ".join(str(b) for b in GPTQ_MARLIN_GROUP_SIZES)
raise RuntimeError(
f"Repacking GPTQ weights with group size {groupsize} as Marlin is not supported, must be one of: {supported_sizes}"
)
if not (sym or quant_method == "awq"):
raise RuntimeError(
"Repacking GPTQ weights with asymmetric quantization as Marlin is not supported."
)
log_once(logger.info, f"Converting {quant_method} model to Marlin packing format.")
weights_per_int = 32 // bits
in_features = qweight.shape[0]
out_features = qweight.shape[1]
# AWQ uses column packing, GPTQ uses row packing
if quant_method == "awq":
out_features *= weights_per_int
else:
in_features *= weights_per_int
if in_features % groupsize != 0:
raise ValueError(
f"Number of input features ({in_features}) not divisible by group size ({groupsize})"
)
if g_idx is not None and desc_act and groupsize != -1:
perm = torch.argsort(g_idx).to(torch.int)
g_idx = g_idx[perm]
else:
perm = torch.empty(0, dtype=torch.int, device=qweight.device)
g_idx = torch.empty(0, dtype=torch.int, device=qweight.device)
if quant_method == "awq":
repacked = marlin_kernels.awq_marlin_repack(
qweight, in_features, out_features, bits
)
if qzeros is not None:
qzeros = awq_to_marlin_zero_points(
qzeros,
in_features // groupsize,
out_features,
bits,
)
else:
repacked = marlin_kernels.gptq_marlin_repack(
qweight, perm, in_features, out_features, bits
)
if qzeros is None:
qzeros = torch.empty(0, dtype=torch.int, device=qweight.device)
scales = permute_scales(scales)
is_full_k = not (desc_act and sharded_infeatures)
return GPTQMarlinWeight(
qweight=repacked,
qzeros=qzeros,
scales=scales,
g_idx=g_idx,
perm=perm,
bits=bits,
is_full_k=is_full_k,
)
class GPTQMarlinLinear(nn.Module):
"""
Linear layer for GPTQ weights that were converted for the GPTQ-Marlin
kernels.
"""
def __init__(
self,
*,
weight: GPTQMarlinWeight,
bias: Optional[torch.Tensor],
):
super().__init__()
_check_marlin_kernels()
assert marlin_kernels is not None
in_features = weight.qweight.shape[0] * MARLIN_TILE_SIZE
out_features = weight.scales.shape[1]
_check_valid_shape(in_features=in_features, out_features=out_features)
self.bits = weight.bits
self.is_full_k = weight.is_full_k
self.qweight = weight.qweight
self.qzeros = weight.qzeros
self.scales = weight.scales
self.g_idx = weight.g_idx
self.perm = weight.perm
if bias is not None:
self.bias = bias
else:
self.bias = None
self.workspace = torch.zeros(
out_features // 64 * 16, dtype=torch.int, device=weight.qweight.device
)
def forward(self, A: torch.Tensor) -> torch.Tensor:
assert marlin_kernels is not None
A_flat = A.view(-1, A.shape[-1])
C = marlin_kernels.gptq_marlin_gemm(
A_flat,
self.qweight,
self.scales,
self.qzeros,
self.g_idx,
self.perm,
self.workspace,
self.bits,
A_flat.shape[0],
self.scales.shape[1],
A_flat.shape[1],
self.is_full_k,
self.qzeros.numel() > 0,
)
C = C.reshape(A.shape[:-1] + (self.scales.shape[1],))
if self.bias is not None:
C += self.bias
return C
def awq_to_marlin_zero_points(
q_zp_packed: torch.Tensor, size_k: int, size_n: int, num_bits: int
) -> torch.Tensor:
# AWQ zero-points are quantized and packed on the column dim.
# In addition, the values are permuted based on dequantizer.
# Here we undo both of these, and then apply marlin permutation
# and pack it back.
q_zp = unpack_cols(q_zp_packed, num_bits, size_k, size_n)
# Undo interleaving (use argsort(..) to get inverse perm)
if num_bits == 4:
undo_interleave = numpy.argsort(numpy.array([0, 2, 4, 6, 1, 3, 5, 7]))
elif num_bits == 8:
undo_interleave = numpy.argsort(numpy.array([0, 2, 1, 3]))
else:
raise Exception("num_bits must be 4 or 8, got {}".format(num_bits))
q_zp = q_zp.reshape((-1, len(undo_interleave)))[:, undo_interleave].ravel()
q_zp = q_zp.reshape((-1, size_n)).contiguous()
marlin_zp = marlin_zero_points(q_zp, size_k, size_n, num_bits)
return marlin_zp
def _check_valid_shape(in_features: int, out_features: int):
if (in_features % 128 != 0 or out_features % 64 != 0) and (
in_features % 64 != 0 or out_features % 128 != 0
):
raise ValueError(
f"The GPTQ Marlin kernel does not have a valid thread configuration for weight matrix with shape ({out_features}, {in_features})."
" The shape elements must be divisible by (128, 64) or (64, 128)."
)
from dataclasses import dataclass from dataclasses import dataclass
from typing import List, Optional, Tuple, Union from typing import List, Optional, Tuple, Union
import numpy
import torch import torch
import torch.nn as nn import torch.nn as nn
from loguru import logger from text_generation_server.layers.marlin.util import _check_marlin_kernels
from text_generation_server.layers.fp8 import fp8_quantize
from text_generation_server.utils.import_utils import SYSTEM
from text_generation_server.utils.log import log_once
from text_generation_server.utils.weights import Weight, Weights, WeightsLoader from text_generation_server.utils.weights import Weight, Weights, WeightsLoader
try: try:
...@@ -15,17 +11,6 @@ try: ...@@ -15,17 +11,6 @@ try:
except ImportError: except ImportError:
marlin_kernels = None marlin_kernels = None
try:
major, _minor = torch.cuda.get_device_capability()
has_sm_8_0 = major >= 8
except Exception:
has_sm_8_0 = False
GPTQ_MARLIN_BITS = [4, 8]
GPTQ_MARLIN_GROUP_SIZES = [-1, 32, 64, 128]
MARLIN_TILE_SIZE = 16
class MarlinWeightsLoader(WeightsLoader): class MarlinWeightsLoader(WeightsLoader):
"""Loader for Marlin-quantized weights.""" """Loader for Marlin-quantized weights."""
...@@ -168,244 +153,73 @@ class MarlinWeightsLoader(WeightsLoader): ...@@ -168,244 +153,73 @@ class MarlinWeightsLoader(WeightsLoader):
return weight return weight
def can_use_gptq_marlin(
*, bits: int, groupsize: int, quant_method: str, quantize: str, sym: bool
) -> bool:
return (
SYSTEM == "cuda"
and marlin_kernels is not None
and has_sm_8_0
and quantize in {"awq", "gptq"}
and quant_method in {"awq", "gptq"}
and bits in GPTQ_MARLIN_BITS
and groupsize in GPTQ_MARLIN_GROUP_SIZES
# We only suppord asymmetric quantization for AWQ.
and (sym or quant_method == "awq")
)
def _check_marlin_kernels():
if not (SYSTEM == "cuda" and has_sm_8_0):
raise NotImplementedError(
"Using quantized Marlin models requires a GPU with CUDA capability 8.0 or later."
)
if marlin_kernels is None:
raise NotImplementedError(
"marlin is not installed, install it with: pip install server/marlin"
)
def _check_valid_shape(in_features: int, out_features: int):
if (in_features % 128 != 0 or out_features % 64 != 0) and (
in_features % 64 != 0 or out_features % 128 != 0
):
raise ValueError(
f"The GPTQ Marlin kernel does not have a valid thread configuration for weight matrix with shape ({out_features}, {in_features})."
" The shape elements must be divisible by (128, 64) or (64, 128)."
)
# https://github.com/IST-DASLab/marlin/blob/2f6d7c10e124b3c5fa29ff8d77d568bd7af3274c/marlin/__init__.py#L40C1-L68C54
def _get_perms() -> Tuple[List[int], List[int]]:
scale_perm = []
for i in range(8):
scale_perm.extend([i + 8 * j for j in range(8)])
scale_perm_single = []
for i in range(4):
scale_perm_single.extend([2 * i + j for j in [0, 1, 8, 9, 16, 17, 24, 25]])
return scale_perm, scale_perm_single
_scale_perm, _scale_perm_single = _get_perms()
def permute_scales(scales: torch.Tensor):
out_features = scales.shape[1]
if scales.shape[0] == 1:
scales = scales.reshape((-1, len(_scale_perm_single)))[:, _scale_perm_single]
else:
scales = scales.reshape((-1, len(_scale_perm)))[:, _scale_perm]
return scales.reshape((-1, out_features)).contiguous()
@dataclass @dataclass
class GPTQMarlinWeight(Weight): class MarlinWeight(Weight):
""" """
Repacked GPTQ Marlin weights. Marlin weights.
Attributes:
B (torch.Tensor): int4-quantized weights packed into int32.
s (torch.Tensor): bfloat16/float16 scales.
""" """
qweight: torch.Tensor B: torch.Tensor
qzeros: torch.Tensor s: torch.Tensor
scales: torch.Tensor
g_idx: torch.Tensor
perm: torch.Tensor
bits: int
is_full_k: bool
def __post_init__(self): def __post_init__(self):
assert self.qweight.dtype == torch.int32 assert self.B.dtype == torch.int32
assert self.scales.dtype == torch.float16 assert self.s.dtype in [torch.float16, torch.bfloat16]
assert self.g_idx.dtype == torch.int32
assert self.perm.dtype == torch.int32
def get_linear(self, bias: torch.Tensor): def get_linear(self, bias: torch.Tensor):
return GPTQMarlinLinear( return MarlinLinear(weight=self, bias=bias)
weight=self,
bias=bias,
)
def repack_gptq_for_marlin(
*,
qweight: torch.Tensor,
qzeros: Optional[torch.Tensor],
scales: torch.Tensor,
g_idx: Optional[torch.Tensor],
bits: int,
desc_act: bool,
groupsize: int,
quant_method: str,
sym: bool,
sharded_infeatures: bool,
) -> GPTQMarlinWeight:
"""Convert GPTQ weights to a layout that's compatible with GPTQ-Marlin kernels."""
_check_marlin_kernels()
assert marlin_kernels is not None
if bits not in GPTQ_MARLIN_BITS:
supported_bits = ", ".join(str(b) for b in GPTQ_MARLIN_BITS)
raise RuntimeError(
f"Repacking {bits}-bit GPTQ weights as Marlin is not supported, must be one of: {supported_bits}"
)
if groupsize not in GPTQ_MARLIN_GROUP_SIZES:
supported_sizes = ", ".join(str(b) for b in GPTQ_MARLIN_GROUP_SIZES)
raise RuntimeError(
f"Repacking GPTQ weights with group size {groupsize} as Marlin is not supported, must be one of: {supported_sizes}"
)
if not (sym or quant_method == "awq"):
raise RuntimeError(
"Repacking GPTQ weights with asymmetric quantization as Marlin is not supported."
)
log_once(logger.info, f"Converting {quant_method} model to Marlin packing format.")
weights_per_int = 32 // bits
in_features = qweight.shape[0]
out_features = qweight.shape[1]
# AWQ uses column packing, GPTQ uses row packing
if quant_method == "awq":
out_features *= weights_per_int
else:
in_features *= weights_per_int
if in_features % groupsize != 0:
raise ValueError(
f"Number of input features ({in_features}) not divisible by group size ({groupsize})"
)
if g_idx is not None and desc_act and groupsize != -1:
perm = torch.argsort(g_idx).to(torch.int)
g_idx = g_idx[perm]
else:
perm = torch.empty(0, dtype=torch.int, device=qweight.device)
g_idx = torch.empty(0, dtype=torch.int, device=qweight.device)
if quant_method == "awq":
repacked = marlin_kernels.awq_marlin_repack(
qweight, in_features, out_features, bits
)
if qzeros is not None:
qzeros = awq_to_marlin_zero_points(
qzeros,
in_features // groupsize,
out_features,
bits,
)
else:
repacked = marlin_kernels.gptq_marlin_repack(
qweight, perm, in_features, out_features, bits
)
if qzeros is None:
qzeros = torch.empty(0, dtype=torch.int, device=qweight.device)
scales = permute_scales(scales)
is_full_k = not (desc_act and sharded_infeatures)
return GPTQMarlinWeight(
qweight=repacked,
qzeros=qzeros,
scales=scales,
g_idx=g_idx,
perm=perm,
bits=bits,
is_full_k=is_full_k,
)
class GPTQMarlinLinear(nn.Module):
"""
Linear layer for GPTQ weights that were converted for the GPTQ-Marlin
kernels.
"""
def __init__( class MarlinLinear(nn.Module):
self, def __init__(self, *, weight: MarlinWeight, bias: Optional[torch.Tensor]):
*,
weight: GPTQMarlinWeight,
bias: Optional[torch.Tensor],
):
super().__init__() super().__init__()
_check_marlin_kernels() _check_marlin_kernels()
assert marlin_kernels is not None assert marlin_kernels is not None
in_features = weight.qweight.shape[0] * MARLIN_TILE_SIZE in_features = weight.B.shape[0] * MARLIN_TILE_SIZE
out_features = weight.scales.shape[1] out_features = weight.s.shape[1]
_check_valid_shape(in_features=in_features, out_features=out_features) assert (
in_features % 128 == 0
), f"Number of input features ({in_features}) not divisable by 128"
assert (
out_features % 256 == 0
), f"Number of output features ({out_features}) not divisable by 256"
self.bits = weight.bits groupsize = -1 if weight.s.shape[0] == 1 else in_features // weight.s.shape[0]
self.is_full_k = weight.is_full_k assert groupsize in {
-1,
128,
}, f"Group size must be -1 or 128, was {groupsize}"
self.qweight = weight.qweight self.B = weight.B
self.qzeros = weight.qzeros self.s = weight.s
self.scales = weight.scales
self.g_idx = weight.g_idx
self.perm = weight.perm
if bias is not None: if bias is not None:
self.bias = bias self.bias = bias
else: else:
self.bias = None self.bias = None
self.workspace = torch.zeros( self.workspace = torch.zeros(
out_features // 64 * 16, dtype=torch.int, device=weight.qweight.device out_features // 64 * 16, dtype=torch.int, device=weight.B.device
) )
def forward(self, A: torch.Tensor) -> torch.Tensor: def forward(self, A: torch.Tensor) -> torch.Tensor:
assert marlin_kernels is not None assert marlin_kernels is not None
A_flat = A.view(-1, A.shape[-1]) C = marlin_kernels.marlin_gemm(
C = marlin_kernels.gptq_marlin_gemm( A.view(-1, A.shape[-1]),
A_flat, self.B,
self.qweight, self.s,
self.scales,
self.qzeros,
self.g_idx,
self.perm,
self.workspace, self.workspace,
self.bits, A.shape[0],
A_flat.shape[0], self.s.shape[1],
self.scales.shape[1], A.shape[1],
A_flat.shape[1],
self.is_full_k,
self.qzeros.numel() > 0,
) )
C = C.reshape(A.shape[:-1] + (self.scales.shape[1],)) C = C.reshape(A.shape[:-1] + (self.s.shape[1],))
if self.bias is not None: if self.bias is not None:
C += self.bias C += self.bias
...@@ -418,6 +232,7 @@ GPTQ_MARLIN_24_MIN_THREAD_K = 128 ...@@ -418,6 +232,7 @@ GPTQ_MARLIN_24_MIN_THREAD_K = 128
GPTQ_MARLIN_24_MAX_PARALLEL = 64 GPTQ_MARLIN_24_MAX_PARALLEL = 64
GPTQ_MARLIN_24_SUPPORTED_NUM_BITS = [4, 8] GPTQ_MARLIN_24_SUPPORTED_NUM_BITS = [4, 8]
GPTQ_MARLIN_24_SUPPORTED_GROUP_SIZES = [-1, 128] GPTQ_MARLIN_24_SUPPORTED_GROUP_SIZES = [-1, 128]
MARLIN_TILE_SIZE = 16
@dataclass @dataclass
...@@ -456,8 +271,10 @@ class GPTQMarlin24Linear(nn.Module): ...@@ -456,8 +271,10 @@ class GPTQMarlin24Linear(nn.Module):
_check_marlin_kernels() _check_marlin_kernels()
assert marlin_kernels is not None assert marlin_kernels is not None
if weight.bits not in GPTQ_MARLIN_BITS: if weight.bits not in GPTQ_MARLIN_24_SUPPORTED_NUM_BITS:
supported_bits = ", ".join(str(b) for b in GPTQ_MARLIN_BITS) supported_bits = ", ".join(
str(b) for b in GPTQ_MARLIN_24_SUPPORTED_NUM_BITS
)
raise RuntimeError( raise RuntimeError(
f"{weight.bits}-bit GPTQ Sparse 2:4 Marlin is not supported, must be one of: {supported_bits}" f"{weight.bits}-bit GPTQ Sparse 2:4 Marlin is not supported, must be one of: {supported_bits}"
) )
...@@ -527,310 +344,3 @@ class GPTQMarlin24Linear(nn.Module): ...@@ -527,310 +344,3 @@ class GPTQMarlin24Linear(nn.Module):
C += self.bias C += self.bias
return C return C
class GPTQMarlinFP8Linear(nn.Module):
"""
FP8 GPTQ-Marlin linear layer.
"""
def __init__(
self,
qweight: torch.Tensor,
scales: torch.Tensor,
bias: Optional[torch.Tensor],
) -> None:
super().__init__()
_check_marlin_kernels()
assert marlin_kernels is not None
log_once(logger.info, "GPU does not support FP8, using Marlin FP8 kernel")
scales = scales.unsqueeze(0)
if scales.shape[1] == 1:
out_features, in_features = qweight.shape
scales = scales.repeat(1, out_features)
qweight, scales = repack_fp8_for_marlin(qweight, scales)
in_features = qweight.shape[0] * MARLIN_TILE_SIZE
out_features = scales.shape[1]
_check_valid_shape(in_features=in_features, out_features=out_features)
self.qweight = qweight
self.scales = scales
self.bias = bias if bias is not None else None
self.workspace = torch.zeros(
out_features // 64 * 16, dtype=torch.int, device=qweight.device
)
@classmethod
def from_unquant(cls, weight, bias, dtype):
qweight, scales = fp8_quantize(weight)
return cls(qweight=qweight, scales=scales.to(dtype), bias=bias)
@classmethod
def from_fp8(cls, weight, scale, _input_scale, bias, dtype):
return cls(qweight=weight, scales=scale.to(dtype), bias=bias)
def forward(self, A: torch.Tensor) -> torch.Tensor:
assert marlin_kernels is not None
A_flat = A.view(-1, A.shape[-1])
C = marlin_kernels.fp8_marlin_gemm(
A_flat,
self.qweight,
self.scales,
self.workspace,
8,
A_flat.shape[0],
self.scales.shape[1],
A_flat.shape[1],
)
C = C.reshape(A.shape[:-1] + (self.scales.shape[1],))
if self.bias is not None:
C += self.bias
return C
def pack_fp8_as_int32(fp8_tensor: torch.Tensor) -> torch.Tensor:
"""
Repack FP8 weights to gptq format (packed int32 elements).
"""
assert fp8_tensor.dtype == torch.float8_e4m3fn
if fp8_tensor.shape[0] % 4 != 0:
raise ValueError(
f"Leading tensor dimension is not divisable by 4: {fp8_tensor.shape[0]}"
)
# Reshape to prepare for packing
reshaped = fp8_tensor.reshape(-1, 4, *fp8_tensor.shape[1:])
# Convert fp8 to uint8 (byte) representation
byte_tensor = reshaped.view(torch.uint8)
# Pack 4 uint8 values into one int32
packed = torch.zeros(
fp8_tensor.shape[0] // 4,
fp8_tensor.shape[1],
dtype=torch.int32,
device=fp8_tensor.device,
)
for i in range(4):
packed.bitwise_or_(byte_tensor[:, i].to(torch.int32) << i * 8)
return packed
def repack_fp8_for_marlin(weight: torch.Tensor, scales: torch.Tensor):
"""
Repack FP8 tensor for GPTQ-Marlin.
"""
out_features, in_features = weight.shape
# Torch linear layers weights with shape [out_features, in_features],
# GPTQ-quantized weights use [in_feateres/pack_factor, in_features],
# so transpose before packing.
qweight = pack_fp8_as_int32(weight.t())
perm = torch.empty(0, dtype=torch.int, device=qweight.device)
repacked = marlin_kernels.gptq_marlin_repack(
qweight, perm, in_features, out_features, 8
)
scales = permute_scales(scales)
return repacked, scales
@dataclass
class MarlinWeight(Weight):
"""
Marlin weights.
Attributes:
B (torch.Tensor): int4-quantized weights packed into int32.
s (torch.Tensor): bfloat16/float16 scales.
"""
B: torch.Tensor
s: torch.Tensor
def __post_init__(self):
assert self.B.dtype == torch.int32
assert self.s.dtype in [torch.float16, torch.bfloat16]
def get_linear(self, bias: torch.Tensor):
return MarlinLinear(weight=self, bias=bias)
class MarlinLinear(nn.Module):
def __init__(self, *, weight: MarlinWeight, bias: Optional[torch.Tensor]):
super().__init__()
_check_marlin_kernels()
assert marlin_kernels is not None
in_features = weight.B.shape[0] * MARLIN_TILE_SIZE
out_features = weight.s.shape[1]
assert (
in_features % 128 == 0
), f"Number of input features ({in_features}) not divisable by 128"
assert (
out_features % 256 == 0
), f"Number of output features ({out_features}) not divisable by 256"
groupsize = -1 if weight.s.shape[0] == 1 else in_features // weight.s.shape[0]
assert groupsize in {
-1,
128,
}, f"Group size must be -1 or 128, was {groupsize}"
self.B = weight.B
self.s = weight.s
if bias is not None:
self.bias = bias
else:
self.bias = None
self.workspace = torch.zeros(
out_features // 64 * 16, dtype=torch.int, device=weight.B.device
)
def forward(self, A: torch.Tensor) -> torch.Tensor:
assert marlin_kernels is not None
C = marlin_kernels.marlin_gemm(
A.view(-1, A.shape[-1]),
self.B,
self.s,
self.workspace,
A.shape[0],
self.s.shape[1],
A.shape[1],
)
C = C.reshape(A.shape[:-1] + (self.s.shape[1],))
if self.bias is not None:
C += self.bias
return C
# Functions below are from vLLM
def get_pack_factor(bits: int) -> int:
if 32 % bits != 0:
raise ValueError(f"Cannot {bits} bit values into uint32")
return 32 // bits
def pack_cols(
q_w: torch.Tensor,
num_bits: int,
size_k: int,
size_n: int,
):
assert q_w.shape == (size_k, size_n)
pack_factor = get_pack_factor(num_bits)
assert size_n % pack_factor == 0
orig_device = q_w.device
q_w = q_w.cpu().numpy().astype(numpy.uint32)
q_res = numpy.zeros((size_k, size_n // pack_factor), dtype=numpy.uint32)
for i in range(pack_factor):
q_res |= q_w[:, i::pack_factor] << num_bits * i
q_res = torch.from_numpy(q_res.astype(numpy.int32)).to(orig_device)
q_res = q_res.contiguous()
return q_res
def unpack_cols(
packed_q_w: torch.Tensor,
num_bits: int,
size_k: int,
size_n: int,
):
pack_factor = get_pack_factor(num_bits)
assert size_n % pack_factor == 0
assert packed_q_w.shape == (
size_k,
size_n // pack_factor,
), "packed_q_w.shape = {} size_k = {}, size_n = {} pack_Factor = {}".format(
packed_q_w.shape, size_k, size_n, pack_factor
)
orig_device = packed_q_w.device
packed_q_w_cpu = packed_q_w.cpu().numpy().astype(numpy.uint32)
q_res = numpy.zeros((size_k, size_n), dtype=numpy.uint32)
mask = (1 << num_bits) - 1
for i in range(pack_factor):
vals = packed_q_w_cpu & mask
packed_q_w_cpu >>= num_bits
q_res[:, i::pack_factor] = vals
q_res = torch.from_numpy(q_res.astype(numpy.int32)).to(orig_device)
q_res = q_res.contiguous()
return q_res
def marlin_zero_points(
zp: torch.Tensor, size_k: int, size_n: int, num_bits: int
) -> torch.Tensor:
# Permute zero-points in a similar way to scales, but do not use the
# "single" permutation, since zero-points are applied on every MMA
zp = zp.reshape((-1, len(_scale_perm)))[:, _scale_perm]
# Interleave column dim (for the dequantize code) and pack it to int32
if num_bits == 4:
interleave = numpy.array([0, 2, 4, 6, 1, 3, 5, 7])
elif num_bits == 8:
interleave = numpy.array([0, 2, 1, 3])
else:
raise Exception("num_bits must be 4 or 8, got {}".format(num_bits))
zp = zp.reshape((-1, len(interleave)))[:, interleave].ravel()
zp = zp.reshape((-1, size_n)).contiguous()
zp = pack_cols(zp, num_bits, size_k, size_n)
return zp
def awq_to_marlin_zero_points(
q_zp_packed: torch.Tensor, size_k: int, size_n: int, num_bits: int
) -> torch.Tensor:
# AWQ zero-points are quantized and packed on the column dim.
# In addition, the values are permuted based on dequantizer.
# Here we undo both of these, and then apply marlin permutation
# and pack it back.
q_zp = unpack_cols(q_zp_packed, num_bits, size_k, size_n)
# Undo interleaving (use argsort(..) to get inverse perm)
if num_bits == 4:
undo_interleave = numpy.argsort(numpy.array([0, 2, 4, 6, 1, 3, 5, 7]))
elif num_bits == 8:
undo_interleave = numpy.argsort(numpy.array([0, 2, 1, 3]))
else:
raise Exception("num_bits must be 4 or 8, got {}".format(num_bits))
q_zp = q_zp.reshape((-1, len(undo_interleave)))[:, undo_interleave].ravel()
q_zp = q_zp.reshape((-1, size_n)).contiguous()
marlin_zp = marlin_zero_points(q_zp, size_k, size_n, num_bits)
return marlin_zp
import functools
from typing import List, Tuple
import numpy
import torch
from text_generation_server.utils.import_utils import SYSTEM
try:
import marlin_kernels
except ImportError:
marlin_kernels = None
try:
major, _minor = torch.cuda.get_device_capability()
has_sm_8_0 = major >= 8
except Exception:
has_sm_8_0 = False
def _check_marlin_kernels():
if not (SYSTEM == "cuda" and has_sm_8_0):
raise NotImplementedError(
"Using quantized Marlin models requires a GPU with CUDA capability 8.0 or later."
)
if marlin_kernels is None:
raise NotImplementedError(
"marlin is not installed, install it with: pip install server/marlin"
)
# https://github.com/IST-DASLab/marlin/blob/2f6d7c10e124b3c5fa29ff8d77d568bd7af3274c/marlin/__init__.py#L40C1-L68C54
@functools.cache
def get_perms() -> Tuple[List[int], List[int]]:
scale_perm = []
for i in range(8):
scale_perm.extend([i + 8 * j for j in range(8)])
scale_perm_single = []
for i in range(4):
scale_perm_single.extend([2 * i + j for j in [0, 1, 8, 9, 16, 17, 24, 25]])
return scale_perm, scale_perm_single
def permute_scales(scales: torch.Tensor):
scale_perm, scale_perm_single = get_perms()
out_features = scales.shape[1]
if scales.shape[0] == 1:
scales = scales.reshape((-1, len(scale_perm_single)))[:, scale_perm_single]
else:
scales = scales.reshape((-1, len(scale_perm)))[:, scale_perm]
return scales.reshape((-1, out_features)).contiguous()
# Functions below are from vLLM
def get_pack_factor(bits: int) -> int:
if 32 % bits != 0:
raise ValueError(f"Cannot {bits} bit values into uint32")
return 32 // bits
def pack_cols(
q_w: torch.Tensor,
num_bits: int,
size_k: int,
size_n: int,
):
assert q_w.shape == (size_k, size_n)
pack_factor = get_pack_factor(num_bits)
assert size_n % pack_factor == 0
orig_device = q_w.device
q_w = q_w.cpu().numpy().astype(numpy.uint32)
q_res = numpy.zeros((size_k, size_n // pack_factor), dtype=numpy.uint32)
for i in range(pack_factor):
q_res |= q_w[:, i::pack_factor] << num_bits * i
q_res = torch.from_numpy(q_res.astype(numpy.int32)).to(orig_device)
q_res = q_res.contiguous()
return q_res
def unpack_cols(
packed_q_w: torch.Tensor,
num_bits: int,
size_k: int,
size_n: int,
):
pack_factor = get_pack_factor(num_bits)
assert size_n % pack_factor == 0
assert packed_q_w.shape == (
size_k,
size_n // pack_factor,
), "packed_q_w.shape = {} size_k = {}, size_n = {} pack_Factor = {}".format(
packed_q_w.shape, size_k, size_n, pack_factor
)
orig_device = packed_q_w.device
packed_q_w_cpu = packed_q_w.cpu().numpy().astype(numpy.uint32)
q_res = numpy.zeros((size_k, size_n), dtype=numpy.uint32)
mask = (1 << num_bits) - 1
for i in range(pack_factor):
vals = packed_q_w_cpu & mask
packed_q_w_cpu >>= num_bits
q_res[:, i::pack_factor] = vals
q_res = torch.from_numpy(q_res.astype(numpy.int32)).to(orig_device)
q_res = q_res.contiguous()
return q_res
def marlin_zero_points(
zp: torch.Tensor, size_k: int, size_n: int, num_bits: int
) -> torch.Tensor:
scale_perm, _ = get_perms()
# Permute zero-points in a similar way to scales, but do not use the
# "single" permutation, since zero-points are applied on every MMA
zp = zp.reshape((-1, len(scale_perm)))[:, scale_perm]
# Interleave column dim (for the dequantize code) and pack it to int32
if num_bits == 4:
interleave = numpy.array([0, 2, 4, 6, 1, 3, 5, 7])
elif num_bits == 8:
interleave = numpy.array([0, 2, 1, 3])
else:
raise Exception("num_bits must be 4 or 8, got {}".format(num_bits))
zp = zp.reshape((-1, len(interleave)))[:, interleave].ravel()
zp = zp.reshape((-1, size_n)).contiguous()
zp = pack_cols(zp, num_bits, size_k, size_n)
return zp
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment