KeyError: 'MatMul'

Hi, after I generate the onnx file using a Lenet 5 pytorch model, running the code
dtypes = {} (outputs, placeholders, variables, constants, operators) = ng.from_onnx(onnx_filename, value_dtypes=dtypes, default_placeholder_dtype=act_dtype, default_variable_dtype=weight_dtype, default_constant_dtype=weight_dtype, default_operator_dtype=act_dtype, default_scale_dtype=scale_dtype, default_bias_dtype=bias_dtype, disable_fusion=disable_fusion)

I got this error.

---

KeyError Traceback (most recent call last)
in
1 dtypes = {}
2 (outputs, placeholders, variables,
----> 3 constants, operators) = ng.from_onnx(onnx_filename,
4 value_dtypes=dtypes,
5 default_placeholder_dtype=act_dtype,

~/.linuxbrew/opt/python@3.8/lib/python3.8/site-packages/nngen/onnx/init.py in from_onnx(filename, value_dtypes, value_shapes, default_placeholder_dtype, default_variable_dtype, default_constant_dtype, default_operator_dtype, default_scale_dtype, default_bias_dtype, onnx_input_layout, onnx_filter_layout, disable_fusion, verbose)
327
328 for name, output_node in output_nodes.items():
--> 329 visitor.visit(name)
330
331 # outputs

~/.linuxbrew/opt/python@3.8/lib/python3.8/site-packages/nngen/onnx/init.py in visit(self, name)
144 node = util.search_node_from_model(self.model, name)
145
--> 146 node_func = _get_func(node.op_type)
147 node_op = node_func(self, node)
148

~/.linuxbrew/opt/python@3.8/lib/python3.8/site-packages/nngen/onnx/init.py in _get_func(op_type)
73
74 def _get_func(op_type):
---> 75 return func_map[op_type]
76
77

KeyError: 'MatMul'

My Lenet 5 model:

`import torch
import torchvision
import torch.nn as nn
import torch.nn.functional as F

class Lenet5(nn.Module):
# Constructer
#def init(self, out_1 = 20, out_2 = 50, hidden_1 = 500):
def init(self, out_1 = 6, out_2 = 16, hidden_1 = 120, hidden_2 = 84):
super(Lenet5, self).init()
# CNN1 out channel num and CNN2 out channel num
self.out1 = out_1
self.out2 = out_2

    # CNN1
    self.conv1 = nn.Conv2d(in_channels = 1, out_channels= out_1, kernel_size=5, stride = 1, bias = False)
    self.maxpool1 = nn.MaxPool2d(kernel_size = 2, padding = 2)
    # CNN2
    self.conv2 = nn.Conv2d(in_channels = out_1, out_channels = out_2, kernel_size=5, stride = 1, bias = False)
    self.maxpool2 = nn.MaxPool2d(kernel_size = 2, padding = 2)
    
    # Linear layer
    self.fc1 = nn.Linear(out_2 * 4 * 4, hidden_1, bias = False)
    self.fc2 = nn.Linear(hidden_1, hidden_2, bias = False)
    self.fc3 = nn.Linear(hidden_2, 10, bias = False)
def forward(self, x):
    x = F.relu(self.conv1(x))
    x = F.max_pool2d(x, 2, 2)
    x = F.relu(self.conv2(x))
    x = F.max_pool2d(x, 2, 2)
    x = x.reshape(-1, 4*4*self.out2)
    x = F.relu(self.fc1(x))
    x = F.relu(self.fc2(x))
    x = self.fc3(x)
    return x
    #return F.log_softmax(x, dim=1)

# Output in each steps
def activation(self, x):
    z1 = self.conv1(x)
    a1 = torch.relu(z1)
    out = self.maxpool1(a1)
    
    z2 = self.conv2(out)
    a2 = torch.relu(z2)
    out1 = self.maxpool2(a2)
    out = out.view(out.size(0), -1)
    return z1, a1, z2, a2, out1, out`

[neushaj]

Hello. I'm getting the same error while using an onnx file as input. How did you fix it?