Open fiorano10 opened 5 years ago
Maybe no warn/error mentioned, you can check the converted prototxt file using netscope or netron, whether match your darknet cfg defined.
There are a bunch of errors in the generated protoxt after checking it on netscope. I used darknet_alexey to train a model to detect custom class. Here are the weights, config, and generated prototxt I can detect objects using the weights so the training went okay, can't figure what's wrong. Thoughts?
Hi @fiorano10 I too have the same problem.. Is there any solution ?
@shiva13425 Couldn't fix it
I have a custom model trained on Yolov2 for 1 class, when I try converting it caffe I get the protoxt file but no caffemodel. The script finishes with no other outputs/errors. what could be wrong?