Aten

@yjh0410 @developer0hye cls_loss = torch.sum(cls_loss_function(pred_cls, gt_cls) * gt_mask) / batch_size Why divide by batch_size？？ Instead of all samples. Could you explain that！ Hope to reply，Thanks!

Deploy with TensorRT: https://github.com/Broad-sky/feature-detection-matching-algorithm

> Hi. > > I would like to use the code you provide, but I can't sign up Baidu cloud. Baidu doesn't allow creating account for who is outside of...

> hi！ Have you tried tensorrt to transform the model into TRT file and then predict https://github.com/Broad-sky/SOLOv2-TensorRT

> 您好，我在自己的电脑运行./convertModel ./solov2_r50_fpn_8gpu_3x.onnx -g ./seg_coco_permute.bin时出现以下错误 > > -->Generate model start > > Input filename: solov2_r50_fpn_8gpu_3x.onnx ONNX IR version: 0.0.6 Opset version: 11 Producer name: pytorch Producer version: 1.7 Domain: Model...

> Hi developer, thanks for you great repo, how did you convert from torch to onnx and tensorRT models, > > best Thank you very much for paying attention to...

> Hi @Broad-sky , I figered out how to convert to trt model, but I met a error in step 3 `make clean`, could you provide the environment setting? >...

> 请问这个需要solov2的相关环境吗，我在build里make报错如下： CMake Error：The source directory ".../SOLOv2-TensorRT" does not appear to contain CMakeLists.txt 按照readme中的步骤编译一下

> > > 请问这个需要solov2的相关环境吗，我在build里make报错如下： CMake Error：The source directory ".../SOLOv2-TensorRT" does not appear to contain CMakeLists.txt > > > > > > 按照readme中的步骤编译一下 > > 您能帮忙解决一下吗？ 我看了MakeFile的内容，make的话也不会生成convertModel这个文件，我不知道该如何进行了 环境配置好了吗

> 我没有配置SOLOv2相关的环境，但是tensorrt的环境配置好了，这个不是脱离原环境的加速吗 检查一下TensorRT版本