BiSeNet icon indicating copy to clipboard operation
BiSeNet copied to clipboard

Add bisenetv2. My implementation of BiSeNet

Results 61 BiSeNet issues
Sort by recently updated
recently updated
newest added

![训练结束条件](https://github.com/CoinCheung/BiSeNet/assets/92571806/bede2d67-6679-4e27-8741-2d542549ec2d)

Using my own custom dataset, the following error occurred. May I ask how to set the label ![1F QVF8$LKI55FH{8X2T5 R](https://github.com/CoinCheung/BiSeNet/assets/121050088/c2eeb133-9ff7-4c6c-803b-11e3c4780288)

![Uploading 785eb16a91c1dfc9cbaef404b75203e.png…]() 有好心人知道为什么训练老是显示No module named 'lib'吗,是因为lib下边的__init__.py是空的吗

Greetings, First of all, really appreciate the work you have done. This has been really useful for my experiments. Now, to my question, I have seen that generally, for fine-tuning,...

你好,之前跑tensorrt官网上提供的https://github.com/NVIDIA/TensorRT/tree/release/8.5/quickstart/SemanticSegmentation 例子中,也使用了argmax,pth模型转onnx转tensorrt,可以直接可以使用;想问下,程序中有使用argmax plugin,为什么会使用argmax plugin,或者说使用argmax plugin的目的是什么?希望能帮解答一下,多谢

你好,同一个模型,使用不同版本的tensorRT进行量化,在TRT8上推理速度为6ms,在TRT10上推理速度为9ms,使用的是仓库提供的预训练模型,这是什么原因呢?

python -m torch.distributed.run --nproc_per_node=2 tools/train_amp.py WARNING:__main__:***************************************** Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for...

/home/c24702/code/C++_project/bisenet1/BiSeNet/tensorrt/trt_dep.cpp: In member function 'void SemanticSegmentTrt::parse_to_engine(std::string, std::string, std::string, std::string)': /home/c24702/code/C++_project/bisenet1/BiSeNet/tensorrt/trt_dep.cpp:106:48: error: 'kBF16' is not a member of 'nvinfer1::BuilderFlag' 106 | config->setFlag(nvinfer1::BuilderFlag::kBF16); | ^~~~~ /home/c24702/code/C++_project/bisenet1/BiSeNet/tensorrt/trt_dep.cpp: In member function 'std::vector SemanticSegmentTrt::inference(std::vector&)': /home/c24702/code/C++_project/bisenet1/BiSeNet/tensorrt/trt_dep.cpp:229:14:...

请问一下为什么在cityscapes的分割结果中,收集数据的车辆本身分割不出来呢?而在原文中是可以分割出来的。 ![原图](https://github.com/user-attachments/assets/1df6ad7e-d61e-4372-af01-47f77f574f84) ![分割后](https://github.com/user-attachments/assets/5e08f457-b091-40a7-9ed0-1415f300633f)