使用Pytorch2Keras将Pytorch模型转换为Keras时出错

我正在使用Pytorch2Keras转换签名验证 （）使用以下代码在Pytorch至Keras中提供

from pytorch2keras.converter import pytorch_to_keras
state_dict, _, _ = torch.load('models/signet_f_lambda_0.95.pth')
base_model = SigNet().to(device).eval()
input_np = np.random.uniform(0, 1, (1,1,150,220))
input_var = Variable(torch.FloatTensor(input_np))
k_model = pytorch_to_keras(base_model, input_var, [(1, 150, 220)], verbose=True)  

我得到以下错误

TypeError                                 Traceback (most recent call last)
----> 6 k_model = pytorch_to_keras(base_model, input_var, [(1, 150, 220)], verbose=True)

/pytorch2keras/converter.py in pytorch_to_keras(model, args, input_shapes, change_ordering, verbose, name_policy)
 71     k_model = onnx_to_keras(onnx_model=onnx_model, input_names=input_names,
 72                             input_shapes=input_shapes, name_policy=name_policy,
---> 73                             verbose=verbose, change_ordering=change_ordering)
 74 
 75     return k_model

/onnx2keras/converter.py in onnx_to_keras(onnx_model, input_names, input_shapes, name_policy, verbose, change_ordering)
144             node_params,
145             layers,
--> 146             node_name
147         )
148 

/onnx2keras/reshape_layers.py in convert_unsqueeze(node, params, layers, node_name)
151         layers[node_name] = np.expand_dims(layers[node.input[0]], params['axes'][0])
152     else:
--> 153         if len(params['axes'][0]) != 0:
154             raise AttributeError('Axes is not 0. Cannot unsqueeze')
155 

 TypeError: object of type 'int' has no len()

帮我解决这个问题