File size: 1,328 Bytes
8798b9e |
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 |
import torch
from torch import nn
import torch.nn.functional as F
def modified_weight_quant(w):
""" Per−tensor quantization to 1.58 bits. No grouping is needed for quantization.
Args:
w: a weight tensor with shape [d, k]
Returns:
u: a quantized weight with shape [d, k]
"""
u = w.clamp(-1, 1).round()
return u
def normalize(w):
w = w / torch.norm(w, dim=1, keepdim=True)
return w
class QLinear(nn.Linear):
def __init__(self,
*kargs,
**kwargs
):
super(QLinear, self).__init__(*kargs, **kwargs)
"""
This is only for training, and kernel optimization is needed for efficiency.
"""
self.scales = nn.Parameter(torch.ones(self.out_features))
self.quantizer = modified_weight_quant
def forward(self, x):
"""i
Args:
x: an input tensor with shape [n, d]
Returns:
y: an output tensor with shape [n, d]
"""
w_quant = self.weight
x = x.to(w_quant.device)
# STE weight quantization
w_quant = w_quant + (self.quantizer(w_quant) - w_quant).detach()
y = F.linear(x, w_quant)
# apply scales post matmul
y = y * self.scales
if self.bias is not None:
y = y + self.bias
return y |