mirror of
https://github.com/comfyanonymous/ComfyUI.git
synced 2026-07-18 20:38:15 +08:00
fix: register_buffer for int4 tensors, kron-based LoKr path with ceiling-div padding, remove __doc__ overrides
This commit is contained in:
JWLHS
parent
9c576caa7f
commit
2990416508
@ -12,240 +12,240 @@ import torch
|
||||
import torch.nn as nn
|
||||
import torch.nn.functional as F
|
||||
from torchao.quantization.quantize_.workflows.int4.int4_plain_int32_tensor import (
|
||||
Int4PlainInt32Tensor,
|
||||
Int4PlainInt32Tensor,
|
||||
)
|
||||
|
||||
|
||||
class TINT4Linear(nn.Module):
|
||||
"""
|
||||
INT4 weight-only linear layer.
|
||||
"""
|
||||
INT4 weight-only linear layer.
|
||||
|
||||
Stores raw int32 qdata + fp16 scales + int8 zero-points.
|
||||
Int4PlainInt32Tensor is lazily constructed and rebuilt on device change.
|
||||
Stores raw int32 qdata + fp16 scales + int8 zero-points.
|
||||
Int4PlainInt32Tensor is lazily constructed and rebuilt on device change.
|
||||
|
||||
Slots (set externally by loaders):
|
||||
_use_quarot: bool — enable activation Hadamard rotation
|
||||
_group_size: int — QuaRot group size
|
||||
_hadamard_H: Tensor|None — normalized Hadamard matrix
|
||||
_tint4_lora_entries: dict|None — {lora_name: [(type, ...), ...]}
|
||||
_lora_prerotated: bool — True if caller already rotated A/w2
|
||||
into QuaRot basis (default False).
|
||||
When False, rotation is done in forward().
|
||||
"""
|
||||
Slots (set externally by loaders):
|
||||
_use_quarot: bool — enable activation Hadamard rotation
|
||||
_group_size: int — QuaRot group size
|
||||
_hadamard_H: Tensor|None — normalized Hadamard matrix
|
||||
_tint4_lora_entries: dict|None — {lora_name: [(type, ...), ...]}
|
||||
_lora_prerotated: bool — True if caller already rotated A/w2
|
||||
into QuaRot basis (default False).
|
||||
When False, rotation is done in forward().
|
||||
"""
|
||||
|
||||
def __init__(self, in_features: int, out_features: int,
|
||||
qdata: torch.Tensor, scale: torch.Tensor,
|
||||
zp: torch.Tensor, block_size: tuple,
|
||||
bias: torch.Tensor | None = None):
|
||||
"""
|
||||
Initialize TINT4Linear from raw torchao INT4 tensors.
|
||||
def __init__(self, in_features: int, out_features: int,
|
||||
qdata: torch.Tensor, scale: torch.Tensor,
|
||||
zp: torch.Tensor, block_size: tuple,
|
||||
bias: torch.Tensor | None = None):
|
||||
"""
|
||||
Initialize TINT4Linear from raw torchao INT4 tensors.
|
||||
|
||||
Args:
|
||||
in_features: Input feature dimension.
|
||||
out_features: Output feature dimension.
|
||||
qdata: Packed int32 quantized weight data (out_f, in_f // 8).
|
||||
scale: Per-block fp16 scales (n_blocks, out_f).
|
||||
zp: Per-block int8 zero-points (n_blocks, out_f).
|
||||
block_size: (b0, b1) tuple defining quantization block shape.
|
||||
bias: Optional bias tensor (out_f,).
|
||||
"""
|
||||
super().__init__()
|
||||
self.in_features = in_features
|
||||
self.out_features = out_features
|
||||
Args:
|
||||
in_features: Input feature dimension.
|
||||
out_features: Output feature dimension.
|
||||
qdata: Packed int32 quantized weight data (out_f, in_f // 8).
|
||||
scale: Per-block fp16 scales (n_blocks, out_f).
|
||||
zp: Per-block int8 zero-points (n_blocks, out_f).
|
||||
block_size: (b0, b1) tuple defining quantization block shape.
|
||||
bias: Optional bias tensor (out_f,).
|
||||
"""
|
||||
super().__init__()
|
||||
self.in_features = in_features
|
||||
self.out_features = out_features
|
||||
|
||||
if bias is not None:
|
||||
self.bias = nn.Parameter(bias)
|
||||
else:
|
||||
self.register_parameter('bias', None)
|
||||
if bias is not None:
|
||||
self.bias = nn.Parameter(bias)
|
||||
else:
|
||||
self.register_parameter('bias', None)
|
||||
|
||||
# ★ FIX 1: 使用 register_buffer 而不是直接用 = 赋值
|
||||
# 好处：model.to(device) / state_dict() / load_state_dict() 能自动管理
|
||||
# 旧代码：self._qdata = qdata → PyTorch 不知道这是持久化数据
|
||||
self.register_buffer('_qdata', qdata)
|
||||
self.register_buffer('_scale', scale)
|
||||
self.register_buffer('_zp', zp)
|
||||
self._block_size = block_size
|
||||
self._qt: Int4PlainInt32Tensor | None = None
|
||||
# ★ FIX 1: 使用 register_buffer 而不是直接用 = 赋值
|
||||
# 好处：model.to(device) / state_dict() / load_state_dict() 能自动管理
|
||||
# 旧代码：self._qdata = qdata → PyTorch 不知道这是持久化数据
|
||||
self.register_buffer('_qdata', qdata)
|
||||
self.register_buffer('_scale', scale)
|
||||
self.register_buffer('_zp', zp)
|
||||
self._block_size = block_size
|
||||
self._qt: Int4PlainInt32Tensor | None = None
|
||||
|
||||
# QuaRot (optional)
|
||||
self._use_quarot: bool = False
|
||||
self._group_size: int = 128
|
||||
self._hadamard_H: torch.Tensor | None = None
|
||||
# QuaRot (optional)
|
||||
self._use_quarot: bool = False
|
||||
self._group_size: int = 128
|
||||
self._hadamard_H: torch.Tensor | None = None
|
||||
|
||||
# LoRA entries slot
|
||||
self._tint4_lora_entries: dict | None = None
|
||||
# LoRA entries slot
|
||||
self._tint4_lora_entries: dict | None = None
|
||||
|
||||
# Pre-rotation flag
|
||||
self._lora_prerotated: bool = False
|
||||
# Pre-rotation flag
|
||||
self._lora_prerotated: bool = False
|
||||
|
||||
@property
|
||||
def weight(self) -> Int4PlainInt32Tensor:
|
||||
"""Lazily construct and return the Int4PlainInt32Tensor view."""
|
||||
if self._qt is None:
|
||||
self._qt = Int4PlainInt32Tensor(
|
||||
self._qdata, self._scale, self._zp,
|
||||
self._block_size, [self.out_features, self.in_features],
|
||||
)
|
||||
return self._qt
|
||||
@property
|
||||
def weight(self) -> Int4PlainInt32Tensor:
|
||||
"""Lazily construct and return the Int4PlainInt32Tensor view."""
|
||||
if self._qt is None:
|
||||
self._qt = Int4PlainInt32Tensor(
|
||||
self._qdata, self._scale, self._zp,
|
||||
self._block_size, [self.out_features, self.in_features],
|
||||
)
|
||||
return self._qt
|
||||
|
||||
@weight.setter
|
||||
def weight(self, value: Int4PlainInt32Tensor) -> None:
|
||||
"""Replace the underlying Int4PlainInt32Tensor."""
|
||||
if isinstance(value, Int4PlainInt32Tensor):
|
||||
self._qt = value
|
||||
@weight.setter
|
||||
def weight(self, value: Int4PlainInt32Tensor) -> None:
|
||||
"""Replace the underlying Int4PlainInt32Tensor."""
|
||||
if isinstance(value, Int4PlainInt32Tensor):
|
||||
self._qt = value
|
||||
|
||||
def _rotate_into_quarot(self, tensor: torch.Tensor) -> torch.Tensor:
|
||||
"""
|
||||
Rotate a LoRA A or w2 matrix into the QuaRot Hadamard basis.
|
||||
def _rotate_into_quarot(self, tensor: torch.Tensor) -> torch.Tensor:
|
||||
"""
|
||||
Rotate a LoRA A or w2 matrix into the QuaRot Hadamard basis.
|
||||
|
||||
Skipped when _lora_prerotated is True (caller already rotated).
|
||||
Returns unchanged tensor if QuaRot is disabled or dimensions
|
||||
are not divisible by _group_size.
|
||||
"""
|
||||
if self._lora_prerotated:
|
||||
return tensor
|
||||
if not self._use_quarot or self._hadamard_H is None:
|
||||
return tensor
|
||||
if tensor.shape[1] % self._group_size != 0:
|
||||
return tensor
|
||||
Ht = self._hadamard_H.T.to(device=tensor.device, dtype=tensor.dtype)
|
||||
ng = tensor.shape[1] // self._group_size
|
||||
return (tensor.reshape(tensor.shape[0], ng, self._group_size)
|
||||
@ Ht).reshape(tensor.shape[0], -1)
|
||||
Skipped when _lora_prerotated is True (caller already rotated).
|
||||
Returns unchanged tensor if QuaRot is disabled or dimensions
|
||||
are not divisible by _group_size.
|
||||
"""
|
||||
if self._lora_prerotated:
|
||||
return tensor
|
||||
if not self._use_quarot or self._hadamard_H is None:
|
||||
return tensor
|
||||
if tensor.shape[1] % self._group_size != 0:
|
||||
return tensor
|
||||
Ht = self._hadamard_H.T.to(device=tensor.device, dtype=tensor.dtype)
|
||||
ng = tensor.shape[1] // self._group_size
|
||||
return (tensor.reshape(tensor.shape[0], ng, self._group_size)
|
||||
@ Ht).reshape(tensor.shape[0], -1)
|
||||
|
||||
def forward(self, x: torch.Tensor) -> torch.Tensor:
|
||||
"""
|
||||
Forward pass with optional QuaRot activation rotation and LoRA injection.
|
||||
def forward(self, x: torch.Tensor) -> torch.Tensor:
|
||||
"""
|
||||
Forward pass with optional QuaRot activation rotation and LoRA injection.
|
||||
|
||||
Args:
|
||||
x: Input tensor, shape (..., in_features).
|
||||
Args:
|
||||
x: Input tensor, shape (..., in_features).
|
||||
|
||||
Returns:
|
||||
Output tensor, shape (..., out_features).
|
||||
Returns:
|
||||
Output tensor, shape (..., out_features).
|
||||
|
||||
Raises:
|
||||
ValueError: If activation dim is not divisible by QuaRot group_size,
|
||||
or if LoRA output shapes are incompatible.
|
||||
"""
|
||||
x_flat = x.reshape(-1, x.shape[-1])
|
||||
Raises:
|
||||
ValueError: If activation dim is not divisible by QuaRot group_size,
|
||||
or if LoRA output shapes are incompatible.
|
||||
"""
|
||||
x_flat = x.reshape(-1, x.shape[-1])
|
||||
|
||||
# QuaRot activation rotation (online)
|
||||
if self._use_quarot and self._hadamard_H is not None:
|
||||
if x_flat.shape[-1] % self._group_size != 0:
|
||||
raise ValueError(
|
||||
f"activation dim {x_flat.shape[-1]} not divisible "
|
||||
f"by QuaRot group_size {self._group_size}"
|
||||
)
|
||||
H_dev = self._hadamard_H.to(device=x.device, dtype=x.dtype)
|
||||
n_groups = x_flat.shape[-1] // self._group_size
|
||||
x_g = x_flat.view(-1, n_groups, self._group_size)
|
||||
x_flat = torch.matmul(x_g, H_dev).view(-1, n_groups * self._group_size)
|
||||
# QuaRot activation rotation (online)
|
||||
if self._use_quarot and self._hadamard_H is not None:
|
||||
if x_flat.shape[-1] % self._group_size != 0:
|
||||
raise ValueError(
|
||||
f"activation dim {x_flat.shape[-1]} not divisible "
|
||||
f"by QuaRot group_size {self._group_size}"
|
||||
)
|
||||
H_dev = self._hadamard_H.to(device=x.device, dtype=x.dtype)
|
||||
n_groups = x_flat.shape[-1] // self._group_size
|
||||
x_g = x_flat.view(-1, n_groups, self._group_size)
|
||||
x_flat = torch.matmul(x_g, H_dev).view(-1, n_groups * self._group_size)
|
||||
|
||||
dev = x.device
|
||||
dev = x.device
|
||||
|
||||
# Rebuild Int4PlainInt32Tensor on device change
|
||||
if self._qt is None or self._qt.device != dev:
|
||||
self._qt = Int4PlainInt32Tensor(
|
||||
self._qdata.to(dev), self._scale.to(dev), self._zp.to(dev),
|
||||
self._block_size, [self.out_features, self.in_features],
|
||||
)
|
||||
# Rebuild Int4PlainInt32Tensor on device change
|
||||
if self._qt is None or self._qt.device != dev:
|
||||
self._qt = Int4PlainInt32Tensor(
|
||||
self._qdata.to(dev), self._scale.to(dev), self._zp.to(dev),
|
||||
self._block_size, [self.out_features, self.in_features],
|
||||
)
|
||||
|
||||
out = F.linear(x_flat, self._qt, None)
|
||||
out = F.linear(x_flat, self._qt, None)
|
||||
|
||||
# LoRA forward injection
|
||||
entries = self._tint4_lora_entries
|
||||
if entries is not None and entries:
|
||||
cd = (x.dtype if x.dtype in (torch.float16, torch.bfloat16)
|
||||
else torch.float16)
|
||||
for lora_entries in entries.values():
|
||||
for e in lora_entries:
|
||||
# ── LoKr path ──
|
||||
# ★ FIX 2: 用 torch.kron 替代 repeat_interleave
|
||||
# 旧方案要求 w2 维度能被 factor 整除（例如 64÷60=不整除→报错）
|
||||
# 新方案 kron 展开 + pad/trim 到目标维度，兼容任意分解
|
||||
if isinstance(e[0], str) and e[0] == "lokr":
|
||||
_, w1, w2, mult, factor = e[:5]
|
||||
sl = e[5] if len(e) > 5 else None
|
||||
se = e[6] if len(e) > 6 else None
|
||||
# LoRA forward injection
|
||||
entries = self._tint4_lora_entries
|
||||
if entries is not None and entries:
|
||||
cd = (x.dtype if x.dtype in (torch.float16, torch.bfloat16)
|
||||
else torch.float16)
|
||||
for lora_entries in entries.values():
|
||||
for e in lora_entries:
|
||||
# ── LoKr path ──
|
||||
# ★ FIX 2: 用 torch.kron 替代 repeat_interleave
|
||||
# 旧方案要求 w2 维度能被 factor 整除（例如 64÷60=不整除→报错）
|
||||
# 新方案 kron 展开 + pad/trim 到目标维度，兼容任意分解
|
||||
if isinstance(e[0], str) and e[0] == "lokr":
|
||||
_, w1, w2, mult, factor = e[:5]
|
||||
sl = e[5] if len(e) > 5 else None
|
||||
se = e[6] if len(e) > 6 else None
|
||||
|
||||
# QuaRot: 先旋转 w2（小矩阵，转发负担小）
|
||||
# 数学上等价于旋转完整 kron 展开（Hadamard 是分块对角的）
|
||||
w2 = self._rotate_into_quarot(w2)
|
||||
w1d = w1.to(device=dev, dtype=cd)
|
||||
w2d = w2.to(device=dev, dtype=cd)
|
||||
# QuaRot: 先旋转 w2（小矩阵，转发负担小）
|
||||
# 数学上等价于旋转完整 kron 展开（Hadamard 是分块对角的）
|
||||
w2 = self._rotate_into_quarot(w2)
|
||||
w1d = w1.to(device=dev, dtype=cd)
|
||||
w2d = w2.to(device=dev, dtype=cd)
|
||||
|
||||
# kron 展开为完整 delta（代替 repeat_interleave）
|
||||
dw = torch.kron(w1d, w2d)
|
||||
# kron 展开为完整 delta（代替 repeat_interleave）
|
||||
dw = torch.kron(w1d, w2d)
|
||||
|
||||
# pad/trim 到层实际维度
|
||||
# QKV 融合层：kron 覆盖 1 个 head，模块持有 3 个
|
||||
if dw.shape[0] < self.out_features:
|
||||
dw = dw.repeat(
|
||||
self.out_features // dw.shape[0], 1)
|
||||
elif dw.shape[0] > self.out_features:
|
||||
dw = dw[:self.out_features, :]
|
||||
if dw.shape[1] < self.in_features:
|
||||
dw = dw.repeat(
|
||||
1, self.in_features // dw.shape[1])
|
||||
elif dw.shape[1] > self.in_features:
|
||||
dw = dw[:, :self.in_features]
|
||||
# Pad/trim to layer (or slice) dimensions
|
||||
tgt_out = (se - sl) if sl is not None else self.out_features
|
||||
if dw.shape[0] < tgt_out:
|
||||
dw = dw.repeat(
|
||||
(tgt_out + dw.shape[0] - 1) // dw.shape[0], 1)
|
||||
if dw.shape[0] > tgt_out:
|
||||
dw = dw[:tgt_out, :]
|
||||
if dw.shape[1] < self.in_features:
|
||||
dw = dw.repeat(
|
||||
1, (self.in_features + dw.shape[1] - 1) // dw.shape[1])
|
||||
if dw.shape[1] > self.in_features:
|
||||
dw = dw[:, :self.in_features]
|
||||
|
||||
dw = dw.contiguous().mul_(mult)
|
||||
lo = x_flat.to(cd) @ dw.T
|
||||
if sl is not None:
|
||||
if lo.shape[1] != (se - sl):
|
||||
raise ValueError(
|
||||
f"LoKr output width {lo.shape[1]} "
|
||||
f"doesn't match target slice width "
|
||||
f"{se - sl}"
|
||||
)
|
||||
out[:, sl:se] += lo
|
||||
elif lo.shape == out.shape:
|
||||
out += lo
|
||||
else:
|
||||
raise ValueError(
|
||||
f"LoKr output shape {lo.shape} "
|
||||
f"incompatible with {out.shape}"
|
||||
)
|
||||
continue
|
||||
dw = dw.contiguous().mul_(mult)
|
||||
lo = x_flat.to(cd) @ dw.T
|
||||
if sl is not None:
|
||||
if lo.shape[1] != (se - sl):
|
||||
raise ValueError(
|
||||
f"LoKr output width {lo.shape[1]} "
|
||||
f"doesn't match target slice width "
|
||||
f"{se - sl}"
|
||||
)
|
||||
out[:, sl:se] += lo
|
||||
elif lo.shape == out.shape:
|
||||
out += lo
|
||||
else:
|
||||
raise ValueError(
|
||||
f"LoKr output shape {lo.shape} "
|
||||
f"incompatible with {out.shape}"
|
||||
)
|
||||
continue
|
||||
|
||||
# Standard LoRA path
|
||||
A, B, mult = e[:3]
|
||||
sl = e[3] if len(e) > 3 else None
|
||||
se = e[4] if len(e) > 4 else None
|
||||
A = self._rotate_into_quarot(A)
|
||||
Ad = A.to(device=dev, dtype=cd)
|
||||
Bd = B.to(device=dev, dtype=cd)
|
||||
lo = (x_flat.to(cd) @ Ad.T) @ Bd.T * mult
|
||||
if sl is not None:
|
||||
if lo.shape[1] != (se - sl):
|
||||
raise ValueError(
|
||||
f"LoRA output width {lo.shape[1]} "
|
||||
f"doesn't match target slice width "
|
||||
f"{se - sl}"
|
||||
)
|
||||
out[:, sl:se] += lo
|
||||
elif lo.shape[1] == out.shape[1]:
|
||||
out += lo
|
||||
else:
|
||||
raise ValueError(
|
||||
f"LoRA output shape {lo.shape} "
|
||||
f"incompatible with {out.shape}"
|
||||
)
|
||||
# Standard LoRA path
|
||||
A, B, mult = e[:3]
|
||||
sl = e[3] if len(e) > 3 else None
|
||||
se = e[4] if len(e) > 4 else None
|
||||
A = self._rotate_into_quarot(A)
|
||||
Ad = A.to(device=dev, dtype=cd)
|
||||
Bd = B.to(device=dev, dtype=cd)
|
||||
lo = (x_flat.to(cd) @ Ad.T) @ Bd.T * mult
|
||||
if sl is not None:
|
||||
if lo.shape[1] != (se - sl):
|
||||
raise ValueError(
|
||||
f"LoRA output width {lo.shape[1]} "
|
||||
f"doesn't match target slice width "
|
||||
f"{se - sl}"
|
||||
)
|
||||
out[:, sl:se] += lo
|
||||
elif lo.shape[1] == out.shape[1]:
|
||||
out += lo
|
||||
else:
|
||||
raise ValueError(
|
||||
f"LoRA output shape {lo.shape} "
|
||||
f"incompatible with {out.shape}"
|
||||
)
|
||||
|
||||
if self.bias is not None:
|
||||
out += self.bias.to(device=dev, dtype=out.dtype)
|
||||
if self.bias is not None:
|
||||
out += self.bias.to(device=dev, dtype=out.dtype)
|
||||
|
||||
return out.reshape(*x.shape[:-1], out.shape[-1])
|
||||
return out.reshape(*x.shape[:-1], out.shape[-1])
|
||||
|
||||
def release(self) -> None:
|
||||
"""Drop cached Int4PlainInt32Tensor to free VRAM."""
|
||||
self._qt = None
|
||||
def release(self) -> None:
|
||||
"""Drop cached Int4PlainInt32Tensor to free VRAM."""
|
||||
self._qt = None
|
||||
|
||||
def __del__(self):
|
||||
"""Cleanup raw tensor references on deletion."""
|
||||
self._qdata = None
|
||||
self._scale = None
|
||||
self._zp = None
|
||||
self._qt = None
|
||||
self._tint4_lora_entries = None
|
||||
def __del__(self):
|
||||
"""Cleanup raw tensor references on deletion."""
|
||||
self._qdata = None
|
||||
self._scale = None
|
||||
self._zp = None
|
||||
self._qt = None
|
||||
self._tint4_lora_entries = None
|
||||
|
||||
Loading…
Reference in New Issue
Block a user