https://lpcv.ai/2023LPCVC/introduction
https://lpcv.ai/scoreboard/Segmentation23
- 10/31/2023 IEEE Computer Society published a blog about the LPCV 2023. Our team ModelTC was announced as the winner of this year’s competition.
refer to official sample: 23LPCVC_Segmentation_Track-Sample_Solution
- file
submit_pyz/submit_0803.pyz
- SHA256 = f1db90947eebedc3229bd5dd70ce5af586893ed8e370acce4f1b9ce33c62c315
- Submitted at 2023-08-03 09:16:28 EST
- Perfomance Score 75.608
- train a segmentation model on your server
- export a '.onnx' file from your framework on your server
- Use
trtexec
on Jetson Nano, convert onnx model to tensorrt model:
trtexec --workspace=4096 --onnx=xxx.onnx --saveEngine=xxx.trt --best --useSpinWait --outputIOFormats=fp16:chw --inputIOFormats=fp16:chw --verbose
- put the tensorrt model file into
inference_code
folder - pack
inference_code
folder into pyz and submit(this step is same as the sample solution)
Goto Train readme for training details, log, codebase, trained weights and reproduction