[pnnx] onnx的unsqueeze参数支持多个维度，torch的只能单个维度，pnnx转换onnx模型存在问题

[robvoid]

error log | 日志或报错信息 | ログ

运行 python xxx_pnnx.py错误:
TypeError: unsqueeze(): argument 'dim' must be int, not tuple

运行 python xxx_ncnn.py不报错，但是输出tensor的shape错误

model | 模型 | モデル

1. original model
   unsqueeze_error.sim.onnx.zip

import torch
import torch.nn as nn

class Model(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(
            3,4,kernel_size=3,stride=4,
        )
        self.conv2 = nn.Conv2d(
            4,4,kernel_size=7,stride=4
        )
        self.bn = nn.BatchNorm2d(4)
        self.act = nn.ReLU()
        self.linear = nn.Linear(144,7)

    def forward(self,x):
        x = self.act(self.bn(self.conv1(x)))
        x = self.act(self.bn(self.conv2(x)))
        x = x.flatten(1)
        x = self.linear(x)
        x = torch.unsqueeze(x,-1)
        x = torch.unsqueeze(x,-1)
        return x

def main():
    model = Model()
    model.eval()
    x = torch.rand(1,3,112,112)
    
    torch.onnx.export(model, x, 'unsqueeze_error.onnx', verbose=True, opset_version=11,
            training=torch.onnx.TrainingMode.EVAL,
            do_constant_folding=True,
            input_names=["data"],
            output_names=["ouput"]
    )
    

if __name__ == '__main__':
    main()

onnxsim unsqueeze_error.onnx unsqueeze_error.sim.onnx

how to reproduce | 复现步骤 | 再現方法

1. pnnx unsqueeze_error.sim.onnx inputshape=[1,3,112,112]
2. 查看ncnn.param最后的ExpandDims，运行_pnnx.py
   预期输出shape [1, 7, 1, 1]，ncnn输出shape [1, 1, 7]