How can I convert my model?

[ruyijidan]

model = alexnet(pretrained=True).eval().cuda()
it;s regular pytorch model
how can I convert my alexnet model ?

[jaybdub]

Hi ruyijidan,

Thanks for reaching out!

The main README has some documentation on how to convert alexnet.

Would you mind clarifying what issue you're running into?

Thanks,
John

[ruyijidan]

Thank for reply！
I have know I just load_state_dict my model and torch2trt
but I got this:

File "/home/oeasy/Desktop/arcface_RT/test1/test5.py", line 20, in
model_trt = torch2trt(model, [x])
File "/home/oeasy/Desktop/arcface_RT/test1/torch2trt/torch2trt.py", line 252, in torch2trt
outputs = module(*inputs)
File "/home/oeasy/anaconda3/envs/tensorrt37/lib/python3.7/site-packages/torch/nn/modules/module.py", line 547, in call
result = self.forward(*input, **kwargs)
File "/home/oeasy/Desktop/arcface_RT/test1/resnet50/arcface_model.py", line 138, in forward
x = self.body(x)
File "/home/oeasy/anaconda3/envs/tensorrt37/lib/python3.7/site-packages/torch/nn/modules/module.py", line 547, in call
result = self.forward(*input, **kwargs)
File "/home/oeasy/anaconda3/envs/tensorrt37/lib/python3.7/site-packages/torch/nn/modules/container.py", line 92, in forward
input = module(input)
File "/home/oeasy/anaconda3/envs/tensorrt37/lib/python3.7/site-packages/torch/nn/modules/module.py", line 547, in call
result = self.forward(*input, **kwargs)
File "/home/oeasy/Desktop/arcface_RT/test1/resnet50/arcface_model.py", line 75, in forward
shortcut = self.shortcut_layer(x)
File "/home/oeasy/anaconda3/envs/tensorrt37/lib/python3.7/site-packages/torch/nn/modules/module.py", line 547, in call
result = self.forward(*input, **kwargs)
File "/home/oeasy/Desktop/arcface_RT/test1/torch2trt/torch2trt.py", line 97, in wrapper
converter(ctx)
File "/home/oeasy/Desktop/arcface_RT/test1/torch2trt/converters/MaxPool2d.py", line 23, in convert_MaxPool2d
input=input._trt, type=trt.PoolingType.MAX, window_size=kernel_size)
AttributeError: 'Tensor' object has no attribute '_trt'