search

voldemortX / pytorch-auto-drive

PytorchAutoDrive: Segmentation models (ERFNet, ENet, DeepLab, FCN...) and Lane detection models (SCNN, RESA, LSTR, LaneATT, BézierLaneNet...) based on PyTorch with fast training, visualization, benchmarking & deployment help
BSD 3-Clause "New" or "Revised" License
837 stars 137 forks source link

TensorRT #92

Open liuYYYYYYYYy opened 2 years ago

liuYYYYYYYYy commented 2 years ago

What should I do after generating the TensorRT model?

voldemortX commented 2 years ago

@liuYYYYYYYYy You can use it in NVIDIA's related packages. I believe there is pip inference toolkits like tensorrt. See https://developer.nvidia.com/tensorrt for introductions.

liuYYYYYYYYy commented 2 years ago

Is there no code in this project that can run TensorRT model？

voldemortX commented 2 years ago

@liuYYYYYYYYy Currently we only provide conversion codes. However, we do have some verification codes that does inference here. You could use it similarly like here.

In the future, we might want to provide onnx/tensorrt execution for full dataset/custom input inference or visualization processes.

Rsweater commented 1 year ago

Hi @voldemortX Seniors! Is the FPS of the model in the repository, using TensorRT?

voldemortX commented 1 year ago

@Rsweater Not yet, lots of them have conversion issues and I feel like it should be done on a jetson, which I don't have.

Rsweater commented 1 year ago

@Rsweater Not yet, lots of them have conversion issues and I feel like it should be done on a jetson, which I don't have.

Oh, I see. Thank you very much, senior.