search

ilhamsyahids / ncnn-android-yolox

Waste Object Detection in Android with YOLOX
5 stars 0 forks source link

Why is there no reshape layer in the .param file after I convert my yolox-tiny model to the ncnn model? #2

Open MyXym opened 1 year ago

MyXym commented 1 year ago

I followed the steps to convert my PyTorch-trained model to onnx-simplifier, then converted that model to ncnn and modified the focus layer. However, in the output section at the bottom, there is no reshape layer. What could be causing this issue?

ilhamsyahids commented 1 year ago

I am not aware of your PyTorch model, have you tried another pretrained model? How do you convert the model? I personally convert using this

MyXym commented 1 year ago

I am not aware of your PyTorch model, have you tried another pretrained model? How do you convert the model? I personally convert using this I've used the website conversion tool you mentioned, but it doesn't work either. I seem to know what's wrong, it seems that the trained model doesn't have a reshape layer inside the file after converting it to ncnn. Maybe because the YOLO Head was not loaded after I improved the model.