Awesome backbones 训练的模型如何部署到移动端?

[yao-th]

我用torch.jit.trace() export 模型的时候报错, 之前也有人遇到同样的问题. 不知道如何解决? 或者有其他的方法让AB输出的模型在移动端上使用么?

(torch) E:\python\Awesome-Backbones>python tools/pth2pt.py models\repvgg\repvgg_A0.py
Loading Train_Epoch042-Loss0.012.pth

Traceback (most recent call last):
File "tools/pth2pt.py", line 125, in
main()
File "tools/pth2pt.py", line 121, in main
traced_script_module = torch.jit.trace(model, example)
File "D:\anaconda3\envs\torch\lib\site-packages\torch\jit_trace.py", line 742, in trace
_module_class,
File "D:\anaconda3\envs\torch\lib\site-packages\torch\jit_trace.py", line 940, in trace_module
_force_outplace,
File "D:\anaconda3\envs\torch\lib\site-packages\torch\nn\modules\module.py", line 725, in _call_impl
result = self._slow_forward(*input, **kwargs)
File "D:\anaconda3\envs\torch\lib\site-packages\torch\nn\modules\module.py", line 709, in _slow_forward
result = self.forward(*input, **kwargs)
File "E:\python\Awesome-Backbones\models\build.py", line 125, in forward
return self.forward_train(x,**kwargs)
TypeError: forward_train() missing 1 required positional argument: 'targets'

[Fafa-DL]

这里应该不能调用forward_train

[PurpleSky-NS]

给模型传参数的时候，多传两个False就会走forward_test了，参考导出onnx
#68