torch2trt

torch2trt copied to clipboard

I getting this error on torch.save(trt.state_dict,"name.pth")

arunmack789

There are questions about the parameters needed for forward propagation in model transformation

2

Hello john, once again I have to ask for your advice sincerely. In a previous demo, I was able to successfully convert a .pth to a .engine model with the...

huangshilong911

Segmentation fault on nn.Linear

1

Hi, trying to use `torch2trt` on a simple example. But i receive a strange sefault on nn.Linear. Here are details. Libs: ```tensorrt==10.0.1 tensorrt-cu12==10.0.1 tensorrt-cu12-bindings==10.0.1 tensorrt-cu12-libs==10.0.1 torch==2.0.0 torch2trt==0.5.0 ``` Code: ```...

brazhenko

AttributeError: 'NoneType' object has no attribute 'serialize'

2

root@51a03800c3eb:/data/track# python3 tools/trt.py -f exps/example/mot/yolox_s_mix_det.py -c bytetrack_s_mot17.pth.tar 2021-12-22 08:51:09.491 | INFO | __main__:main:49 - loaded checkpoint done. [TensorRT] WARNING: Tensor DataType is determined at build time for tensors not marked...

uniyushu

AttributeError: 'tensorrt_bindings.tensorrt.INetworkDefinition' object has no attribute 'add_convolution'. Did you mean: 'add_convolution_nd'?

5

I'm converting a PyTorch model with MobileNetV2 backbone, However I get this error: `AttributeError: 'tensorrt_bindings.tensorrt.INetworkDefinition' object has no attribute 'add_convolution'. Did you mean: 'add_convolution_nd'?` Versions: - tensorrt==10.0.0b6 - torch2trt==0.4.0 -...

AWallyAllah

Encountered known unsupported method torch.Tensor.add

5

Hi, Facing this issue when use tensorrt 7.1.3 + pytorch 1.7.0 + torchvision 0.8.1 The network is a very simple MLP. ``` Warning: Encountered known unsupported method torch.Tensor.add Warning: Encountered...

yuchen2580

I use [poetry](https://python-poetry.org) to run most of my projects. running `poetry add torch2trt` fails due to the project being non-PEP517 compliant. Please update the project to be self contained and...

elisa-aleman

I am encountering an issue when trying to convert a model based on the hybridnets architecture using the torch2trt package. The error seems to be related to the conversion of...

moscaAle29

Unsupported method!

1

Warning: Encountered known unsupported method torch.nn.functional.has_torch_function_variadic Warning: Encountered known unsupported method torch.embedding Warning: Encountered known unsupported method torch.nn.functional.embedding Warning: Encountered known unsupported method torch.arange Warning: Encountered known unsupported method torch.Tensor.to

boge2023

Hi, refer to https://github.com/NVIDIA-AI-IOT/torch2trt/blob/master/torch2trt/converters/native_converters.py?plain=1#L439, `out_channels = int(weight.shape[0])` should be modified as `out_channels = int(weight.shape[1])`. ? For transpose3d, weight was defined `(in_channels,outchannels​, ...)` which was exactly opposite of nn.conv3d.

Yuxiang1990