AssertionError: Have you provided bias as a kwarg? If so, please remove bias=.

[Mustansar Saeed]

Hi, I am using PySyft 0.2.x and trying to build training_plan for CNN model. My Model is as follows:

Model

import syft as sy
from syft.frameworks.torch.nn import Conv2d, max_pool2d

class FemnistNet(nn.Module):
    def __init__(self):
        super(FemnistNet, self).__init__()
        self.conv1 = Conv2d(1, 32, kernel_size=5, stride=1, padding=0, bias=True)
        self.pool1 = nn.MaxPool2d(2, stride=2, )
        
        self.conv2 = Conv2d(32, 64, kernel_size=5, stride=1, padding=0, bias=True)
        self.pool2 = nn.MaxPool2d(2, stride=2)

        self.fc1 = nn.Linear(1024, 512)        
        self.fc2 = nn.Linear(512 ,62)

    def forward(self, x):
        x = x.view(-1, 1, 28, 28)
        x = self.conv1(x)
        x = th.nn.functional.relu(x)
        x = self.pool1(x)
        
        x=self.conv2(x)
        x = th.nn.functional.relu(x)
        x = self.pool2(x)
        x = x.flatten(start_dim=1)
        
        x = self.fc1(x)
        l1_activations = th.nn.functional.relu(x)
        
        x = self.fc2(l1_activations)

        x = x.softmax()

        return x, l1_activations

Training Plan

@sy.func2plan()
def training_plan(X, y, batch_size, lr, model_params):
    model.train()
    
    # inject params into model
    set_model_params(model, model_params)
    
    logits, activations = model.forward(X)
    loss = cross_entropy_with_logits(logits, y, batch_size)
    # backprop
    loss.backward()
    
    updated_params = [
        naive_sgd(param, lr=lr)
        for param in model_params
    ]

    # accuracy
    pred = th.argmax(logits, dim=1)
    target = th.argmax(y, dim=1)
    acc = pred.eq(target).sum()/ batch_size
    
    return (
        loss,
        acc,
        logits,
        None,
        *updated_params,        
    )

Building training plan

num = 50
X = th.tensor((dataX), dtype=th.float) ##784 sized 1D array. would be reshaped to 1, 28, 28
y = nn.functional.one_hot(th.tensor(dataY), 62) ##62-length sized one hot vectors

lr = th.tensor([0.0003]) ##0.0003 learning rate
batch_size = th.tensor([float(num)]) ##20 is our batch size

loss, acc, logits, target, *updated_params = training_plan.build(X[0:num], y[0:num], batch_size, lr, model_params, trace_autograd=True)

Error:

---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
<ipython-input-196-b1aa83563586> in <module>
      6 batch_size = th.tensor([float(num)]) ##20 is our batch size
      7 
----> 8 loss, acc, logits, target, *updated_params = training_plan.build(X[0:num], y[0:num], batch_size, lr, model_params, trace_autograd=True)
      9 # updated_params
     10 loss

/usr/local/lib/python3.7/dist-packages/syft-0.2.9-py3.7.egg/syft/execution/plan.py in build(self, trace_autograd, *args)
    273                 framework_kwargs[f_name] = wrap_framework_func(self.role)
    274 
--> 275         results = self.forward(*args, **framework_kwargs)
    276 
    277         # Register inputs in role

<ipython-input-131-e49ab938819a> in training_plan(X, y, batch_size, lr, model_params)
     12 
     13 
---> 14     logits, activations = model.forward(X)
     15 #     print("Logits", logits[:1])
     16     loss = cross_entropy_with_logits(logits, y, batch_size)

<ipython-input-193-80ed0341fc1a> in forward(self, x)
     45 #         print(x.shape)
     46 
---> 47         x = self.conv1(x)
     48         x = th.nn.functional.relu(x)
     49 #         print("conv1 shape", x.shape)

/usr/local/lib/python3.7/dist-packages/torch-1.4.0-py3.7-linux-x86_64.egg/torch/nn/modules/module.py in __call__(self, *input, **kwargs)
    530             result = self._slow_forward(*input, **kwargs)
    531         else:
--> 532             result = self.forward(*input, **kwargs)
    533         for hook in self._forward_hooks.values():
    534             hook_result = hook(self, input, result)

/usr/local/lib/python3.7/dist-packages/syft-0.2.9-py3.7.egg/syft/frameworks/torch/nn/conv.py in forward(self, input)
     70 
     71         return conv2d(
---> 72             input, self.weight, self.bias, self.stride, self.padding, self.dilation, self.groups
     73         )
     74 

/usr/local/lib/python3.7/dist-packages/syft-0.2.9-py3.7.egg/syft/frameworks/torch/nn/functional.py in conv2d(input, weight, bias, stride, padding, dilation, groups)
    245         assert isinstance(
    246             bias, sy.AdditiveSharingTensor
--> 247         ), "Have you provided bias as a kwarg? If so, please remove `bias=`."
    248 
    249     locations = input.locations

AssertionError: Have you provided bias as a kwarg? If so, please remove `bias=`.

Can anyone please tell what is the issue?