qnq 1.0.2

qnq's not quantization

QNQ -- QNQ's not quantization

version 1.0.2 2021.1.29

Description

The toolkit is for Techart algorithm team to quantize their custom neural network's pretrained model. The toolkit is beta now, you can contact me with email([e-mail]) for adding ops and fixing bugs.

How to install

pip install qnq

How to use

This README.MD is in very early stages, and will be updated soon. you can visit https://git.zwdong.com/zhiwei.dong/qnq_tutorial for more examples for QNQ.

1. Prepare your model.

   1. Check if your model contains non-class operator, like torch.matmul.
   2. If True, add from qnq.operators.torchfunc_ops import * to your code.
   3. Then use class replace non-class operator, you can refer fellow #! add by dongz

   class BasicBlock(nn.Module):
       expansion = 1
   
       def __init__(self, inplanes, planes, stride=1, downsample=None):
           super(BasicBlock, self).__init__()
           self.conv1 = conv3x3(inplanes, planes, stride)
           self.bn1 = nn.BatchNorm2d(planes)
           self.relu1 = nn.ReLU(inplace=True)
           self.relu2 = nn.ReLU(inplace=True)
           self.conv2 = conv3x3(planes, planes)
           self.bn2 = nn.BatchNorm2d(planes)
           self.downsample = downsample
           self.stride = stride
   
           #! add by dongz
           self.torch_add = TorchAdd()
   
       def forward(self, x):
           identity = x
   
           out = self.conv1(x)
           out = self.bn1(out)
           out = self.relu1(out)
   
           out = self.conv2(out)
           out = self.bn2(out)
   
           if self.downsample is not None:
               identity = self.downsample(x)
   
           #! add by dongz
           out = self.torch_add(out, identity)
           # out += identity
           out = self.relu2(out)
   
           return out
   

2. Prepare 'metrics', 'metrics_light'(optional) and 'steper'.

   1. Choose at least 1k data to calibration your quantized model.
   2. 'metrics' inference without input params, return metrics value(a float number).
   3. 'metrics_light' inference without input params, return metrics value(a float number), you can choose 1/10 testsets to test.
   4. 'steper' done inference and without input params too, but add quant.step(), and no return.
   5. Check qnq_tutorial for details.

3. Prepare pretrained checkpoints.

   1. Train your model and use torch.save() to save your checkpoints.
   2. Use checkpoints = torch.load(checkpoints_path) and model.load_state_dict(checkpoints) to load your checkpoints.

4. Quantize

   1. For code
      1. Add from qnq import QNQ
      2. Add quant = QNQ(model, save_path, config_path, metrics, metrics_light, steper).
      3. Add quant.search()
   2. First run the program will exit, but the config_path will show a yaml file.
   3. Edit config.yaml and rerun for quantization.

Operators supported

• Convolution Layers
  • Conv
  • ConvTranspose
• Pooling Layers
  • MaxPool
  • AveragePool
  • AdaptiveAvgPool
• Activation
  • Relu、Relu6
  • PRelu、LeakyRelu
  • LogSoftmax
• Normalization Layers
  • BatchNorm
  • LayerNorm
• Recurrent
  • LSTM
• Linear Layers
  • Linear
• Vision Layers
  • Upsample
  • Embedding
• Torch Function
  • Add, Sum, Minus, DotMul, MatMul, Div,
  • Sqrt, Exp
  • Sin, Cos
  • SoftMax, Sigmoid, Tanh
  • TorchTemplate, TorchDummy